From 9a08c352d05305ca7651540c3b107da1e4e1f40b Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Wed, 17 Feb 2016 16:49:38 -0800 Subject: fs: add filp_clone_open API I need an API that allows me to obtain a clone of the current file pointer to pass in to an exec handler. I've labelled this as an internal API because I can't see how it would be useful outside of the fs subsystem. The use case will be a persistent binfmt_misc handler. Signed-off-by: James Bottomley Acked-by: Serge Hallyn Acked-by: Jan Kara --- fs/internal.h | 1 + fs/open.c | 20 ++++++++++++++++++++ 2 files changed, 21 insertions(+) (limited to 'fs') diff --git a/fs/internal.h b/fs/internal.h index b71deeecea17..c8ca0c957743 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -108,6 +108,7 @@ extern long do_handle_open(int mountdirfd, struct file_handle __user *ufh, int open_flag); extern int open_check_o_direct(struct file *f); extern int vfs_open(const struct path *, struct file *, const struct cred *); +extern struct file *filp_clone_open(struct file *); /* * inode.c diff --git a/fs/open.c b/fs/open.c index 17cb6b1dab75..bfe6f2b8345f 100644 --- a/fs/open.c +++ b/fs/open.c @@ -1002,6 +1002,26 @@ struct file *file_open_root(struct dentry *dentry, struct vfsmount *mnt, } EXPORT_SYMBOL(file_open_root); +struct file *filp_clone_open(struct file *oldfile) +{ + struct file *file; + int retval; + + file = get_empty_filp(); + if (IS_ERR(file)) + return file; + + file->f_flags = oldfile->f_flags; + retval = vfs_open(&oldfile->f_path, file, oldfile->f_cred); + if (retval) { + put_filp(file); + return ERR_PTR(retval); + } + + return file; +} +EXPORT_SYMBOL(filp_clone_open); + long do_sys_open(int dfd, const char __user *filename, int flags, umode_t mode) { struct open_flags op; -- cgit v1.2.3 From 948b701a607f123df92ed29084413e5dd8cda2ed Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Wed, 17 Feb 2016 16:51:16 -0800 Subject: binfmt_misc: add persistent opened binary handler for containers This patch adds a new flag 'F' to the binfmt handlers. If you pass in 'F' the binary that runs the emulation will be opened immediately and in future, will be cloned from the open file. The net effect is that the handler survives both changeroots and mount namespace changes, making it easy to work with foreign architecture containers without contaminating the container image with the emulator. Signed-off-by: James Bottomley Acked-by: Serge Hallyn --- fs/binfmt_misc.c | 41 +++++++++++++++++++++++++++++++++++++++-- 1 file changed, 39 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index 3a3ced779fc7..8a108c435bc6 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -26,6 +26,8 @@ #include #include +#include "internal.h" + #ifdef DEBUG # define USE_DEBUG 1 #else @@ -43,6 +45,7 @@ enum {Enabled, Magic}; #define MISC_FMT_PRESERVE_ARGV0 (1 << 31) #define MISC_FMT_OPEN_BINARY (1 << 30) #define MISC_FMT_CREDENTIALS (1 << 29) +#define MISC_FMT_OPEN_FILE (1 << 28) typedef struct { struct list_head list; @@ -54,6 +57,7 @@ typedef struct { char *interpreter; /* filename of interpreter */ char *name; struct dentry *dentry; + struct file *interp_file; } Node; static DEFINE_RWLOCK(entries_lock); @@ -201,7 +205,13 @@ static int load_misc_binary(struct linux_binprm *bprm) if (retval < 0) goto error; - interp_file = open_exec(iname); + if (fmt->flags & MISC_FMT_OPEN_FILE && fmt->interp_file) { + interp_file = filp_clone_open(fmt->interp_file); + if (!IS_ERR(interp_file)) + deny_write_access(interp_file); + } else { + interp_file = open_exec(iname); + } retval = PTR_ERR(interp_file); if (IS_ERR(interp_file)) goto error; @@ -285,6 +295,11 @@ static char *check_special_flags(char *sfs, Node *e) e->flags |= (MISC_FMT_CREDENTIALS | MISC_FMT_OPEN_BINARY); break; + case 'F': + pr_debug("register: flag: F: open interpreter file now\n"); + p++; + e->flags |= MISC_FMT_OPEN_FILE; + break; default: cont = 0; } @@ -543,6 +558,8 @@ static void entry_status(Node *e, char *page) *dp++ = 'O'; if (e->flags & MISC_FMT_CREDENTIALS) *dp++ = 'C'; + if (e->flags & MISC_FMT_OPEN_FILE) + *dp++ = 'F'; *dp++ = '\n'; if (!test_bit(Magic, &e->flags)) { @@ -590,6 +607,11 @@ static void kill_node(Node *e) } write_unlock(&entries_lock); + if ((e->flags & MISC_FMT_OPEN_FILE) && e->interp_file) { + filp_close(e->interp_file, NULL); + e->interp_file = NULL; + } + if (dentry) { drop_nlink(d_inode(dentry)); d_drop(dentry); @@ -698,6 +720,21 @@ static ssize_t bm_register_write(struct file *file, const char __user *buffer, goto out2; } + if (e->flags & MISC_FMT_OPEN_FILE) { + struct file *f; + + f = open_exec(e->interpreter); + if (IS_ERR(f)) { + err = PTR_ERR(f); + pr_notice("register: failed to install interpreter file %s\n", e->interpreter); + simple_release_fs(&bm_mnt, &entry_count); + iput(inode); + inode = NULL; + goto out2; + } + e->interp_file = f; + } + e->dentry = dget(dentry); inode->i_private = e; inode->i_fop = &bm_entry_operations; @@ -716,7 +753,7 @@ out: if (err) { kfree(e); - return -EINVAL; + return err; } return count; } -- cgit v1.2.3 From 38512aa98a3feb6acd7da8f0ed5dade5b592b426 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 7 Jun 2016 21:44:08 -0400 Subject: NFS: Don't flush caches for a getattr that races with writeback If there were outstanding writes then chalk up the unexpected change attribute on the server to them. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 52e7d6869e3b..60051e62d3f1 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1729,12 +1729,15 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) if (inode->i_version != fattr->change_attr) { dprintk("NFS: change_attr change on server for file %s/%ld\n", inode->i_sb->s_id, inode->i_ino); - invalid |= NFS_INO_INVALID_ATTR - | NFS_INO_INVALID_DATA - | NFS_INO_INVALID_ACCESS - | NFS_INO_INVALID_ACL; - if (S_ISDIR(inode->i_mode)) - nfs_force_lookup_revalidate(inode); + /* Could it be a race with writeback? */ + if (nfsi->nrequests == 0) { + invalid |= NFS_INO_INVALID_ATTR + | NFS_INO_INVALID_DATA + | NFS_INO_INVALID_ACCESS + | NFS_INO_INVALID_ACL; + if (S_ISDIR(inode->i_mode)) + nfs_force_lookup_revalidate(inode); + } inode->i_version = fattr->change_attr; } } else { -- cgit v1.2.3 From 57b691819ee2b095da505b34abdcd3193d0af75c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 3 Jun 2016 17:07:19 -0400 Subject: NFS: Cache access checks more aggressively If an attribute revalidation fails, then we already know that we'll zap the access cache. If, OTOH, the inode isn't changing, there should be no need to eject access calls just because they are old. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 52 +++++++++++++++++++++++++++++++--------------------- 1 file changed, 31 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index aaf7bd0cbae2..210b33636fe4 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -2228,21 +2228,37 @@ static struct nfs_access_entry *nfs_access_search_rbtree(struct inode *inode, st return NULL; } -static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res) +static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, struct nfs_access_entry *res, bool may_block) { struct nfs_inode *nfsi = NFS_I(inode); struct nfs_access_entry *cache; - int err = -ENOENT; + bool retry = true; + int err; spin_lock(&inode->i_lock); - if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS) - goto out_zap; - cache = nfs_access_search_rbtree(inode, cred); - if (cache == NULL) - goto out; - if (!nfs_have_delegated_attributes(inode) && - !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) - goto out_stale; + for(;;) { + if (nfsi->cache_validity & NFS_INO_INVALID_ACCESS) + goto out_zap; + cache = nfs_access_search_rbtree(inode, cred); + err = -ENOENT; + if (cache == NULL) + goto out; + /* Found an entry, is our attribute cache valid? */ + if (!nfs_attribute_cache_expired(inode) && + !(nfsi->cache_validity & NFS_INO_INVALID_ATTR)) + break; + err = -ECHILD; + if (!may_block) + goto out; + if (!retry) + goto out_zap; + spin_unlock(&inode->i_lock); + err = __nfs_revalidate_inode(NFS_SERVER(inode), inode); + if (err) + return err; + spin_lock(&inode->i_lock); + retry = false; + } res->jiffies = cache->jiffies; res->cred = cache->cred; res->mask = cache->mask; @@ -2251,12 +2267,6 @@ static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, str out: spin_unlock(&inode->i_lock); return err; -out_stale: - rb_erase(&cache->rb_node, &nfsi->access_cache); - list_del(&cache->lru); - spin_unlock(&inode->i_lock); - nfs_access_free_entry(cache); - return -ENOENT; out_zap: spin_unlock(&inode->i_lock); nfs_access_zap_cache(inode); @@ -2283,13 +2293,12 @@ static int nfs_access_get_cached_rcu(struct inode *inode, struct rpc_cred *cred, cache = NULL; if (cache == NULL) goto out; - if (!nfs_have_delegated_attributes(inode) && - !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo)) + err = nfs_revalidate_inode_rcu(NFS_SERVER(inode), inode); + if (err) goto out; res->jiffies = cache->jiffies; res->cred = cache->cred; res->mask = cache->mask; - err = 0; out: rcu_read_unlock(); return err; @@ -2378,18 +2387,19 @@ EXPORT_SYMBOL_GPL(nfs_access_set_mask); static int nfs_do_access(struct inode *inode, struct rpc_cred *cred, int mask) { struct nfs_access_entry cache; + bool may_block = (mask & MAY_NOT_BLOCK) == 0; int status; trace_nfs_access_enter(inode); status = nfs_access_get_cached_rcu(inode, cred, &cache); if (status != 0) - status = nfs_access_get_cached(inode, cred, &cache); + status = nfs_access_get_cached(inode, cred, &cache, may_block); if (status == 0) goto out_cached; status = -ECHILD; - if (mask & MAY_NOT_BLOCK) + if (!may_block) goto out; /* Be clever: ask server to check for all possible rights */ -- cgit v1.2.3 From ca0daa277acac1029f74d9fea838c9e507398226 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 8 Jun 2016 17:08:28 -0400 Subject: NFS: Cache aggressively when file is open for writing Unless the user is using file locking, we must assume close-to-open cache consistency when the file is open for writing. Adjust the caching algorithm so that it does not clear the cache on out-of-order writes and/or attribute revalidations. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 13 ++---------- fs/nfs/inode.c | 62 +++++++++++++++++++++++++++++++++++++++++----------------- 2 files changed, 46 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 717a8d6af52d..2d39d9f9da7d 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -779,11 +779,6 @@ do_unlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) return status; } -static int -is_time_granular(struct timespec *ts) { - return ((ts->tv_sec == 0) && (ts->tv_nsec <= 1000)); -} - static int do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) { @@ -817,12 +812,8 @@ do_setlk(struct file *filp, int cmd, struct file_lock *fl, int is_local) * This makes locking act as a cache coherency point. */ nfs_sync_mapping(filp->f_mapping); - if (!NFS_PROTO(inode)->have_delegation(inode, FMODE_READ)) { - if (is_time_granular(&NFS_SERVER(inode)->time_delta)) - __nfs_revalidate_inode(NFS_SERVER(inode), inode); - else - nfs_zap_caches(inode); - } + if (!NFS_PROTO(inode)->have_delegation(inode, FMODE_READ)) + nfs_zap_mapping(inode, filp->f_mapping); out: return status; } diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 60051e62d3f1..4e65a5a8a01b 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -878,7 +878,10 @@ void nfs_inode_attach_open_context(struct nfs_open_context *ctx) struct nfs_inode *nfsi = NFS_I(inode); spin_lock(&inode->i_lock); - list_add(&ctx->list, &nfsi->open_files); + if (ctx->mode & FMODE_WRITE) + list_add(&ctx->list, &nfsi->open_files); + else + list_add_tail(&ctx->list, &nfsi->open_files); spin_unlock(&inode->i_lock); } EXPORT_SYMBOL_GPL(nfs_inode_attach_open_context); @@ -1215,6 +1218,25 @@ int nfs_revalidate_mapping_protected(struct inode *inode, struct address_space * return __nfs_revalidate_mapping(inode, mapping, true); } +static bool nfs_file_has_writers(struct nfs_inode *nfsi) +{ + struct inode *inode = &nfsi->vfs_inode; + + assert_spin_locked(&inode->i_lock); + + if (!S_ISREG(inode->i_mode)) + return false; + if (list_empty(&nfsi->open_files)) + return false; + /* Note: This relies on nfsi->open_files being ordered with writers + * being placed at the head of the list. + * See nfs_inode_attach_open_context() + */ + return (list_first_entry(&nfsi->open_files, + struct nfs_open_context, + list)->mode & FMODE_WRITE) == FMODE_WRITE; +} + static unsigned long nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr) { struct nfs_inode *nfsi = NFS_I(inode); @@ -1279,22 +1301,24 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat if ((fattr->valid & NFS_ATTR_FATTR_TYPE) && (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) return -EIO; - if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && - inode->i_version != fattr->change_attr) - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + if (!nfs_file_has_writers(nfsi)) { + /* Verify a few of the more important attributes */ + if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && inode->i_version != fattr->change_attr) + invalid |= NFS_INO_INVALID_ATTR | NFS_INO_REVAL_PAGECACHE; - /* Verify a few of the more important attributes */ - if ((fattr->valid & NFS_ATTR_FATTR_MTIME) && !timespec_equal(&inode->i_mtime, &fattr->mtime)) - invalid |= NFS_INO_INVALID_ATTR; + if ((fattr->valid & NFS_ATTR_FATTR_MTIME) && !timespec_equal(&inode->i_mtime, &fattr->mtime)) + invalid |= NFS_INO_INVALID_ATTR; - if (fattr->valid & NFS_ATTR_FATTR_SIZE) { - cur_size = i_size_read(inode); - new_isize = nfs_size_to_loff_t(fattr->size); - if (cur_size != new_isize) - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + if ((fattr->valid & NFS_ATTR_FATTR_CTIME) && !timespec_equal(&inode->i_ctime, &fattr->ctime)) + invalid |= NFS_INO_INVALID_ATTR; + + if (fattr->valid & NFS_ATTR_FATTR_SIZE) { + cur_size = i_size_read(inode); + new_isize = nfs_size_to_loff_t(fattr->size); + if (cur_size != new_isize) + invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + } } - if (nfsi->nrequests != 0) - invalid &= ~NFS_INO_REVAL_PAGECACHE; /* Have any file permissions changed? */ if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)) @@ -1526,7 +1550,7 @@ EXPORT_SYMBOL_GPL(nfs_refresh_inode); static int nfs_post_op_update_inode_locked(struct inode *inode, struct nfs_fattr *fattr) { - unsigned long invalid = NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE; + unsigned long invalid = NFS_INO_INVALID_ATTR; /* * Don't revalidate the pagecache if we hold a delegation, but do @@ -1675,6 +1699,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long invalid = 0; unsigned long now = jiffies; unsigned long save_cache_validity; + bool have_writers = nfs_file_has_writers(nfsi); bool cache_revalidated = true; dfprintk(VFS, "NFS: %s(%s/%lu fh_crc=0x%08x ct=%d info=0x%x)\n", @@ -1730,7 +1755,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) dprintk("NFS: change_attr change on server for file %s/%ld\n", inode->i_sb->s_id, inode->i_ino); /* Could it be a race with writeback? */ - if (nfsi->nrequests == 0) { + if (!have_writers) { invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_DATA | NFS_INO_INVALID_ACCESS @@ -1770,9 +1795,10 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) if (new_isize != cur_isize) { /* Do we perhaps have any outstanding writes, or has * the file grown beyond our last write? */ - if ((nfsi->nrequests == 0) || new_isize > cur_isize) { + if (nfsi->nrequests == 0 || new_isize > cur_isize) { i_size_write(inode, new_isize); - invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA; + if (!have_writers) + invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA; } dprintk("NFS: isize change on server for file %s/%ld " "(%Ld to %Ld)\n", -- cgit v1.2.3 From 6b56a89833fa7903595c8d138bb4927187315cba Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Jun 2016 18:23:01 -0400 Subject: NFS: Kill NFS_INO_NFS_INO_FLUSHING: it is a performance killer filemap_datawrite() and friends already deal just fine with livelock. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 8 -------- fs/nfs/nfstrace.h | 1 - fs/nfs/write.c | 11 ----------- include/linux/nfs_fs.h | 1 - 4 files changed, 21 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 2d39d9f9da7d..29d7477a62e8 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -359,14 +359,6 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping, file, mapping->host->i_ino, len, (long long) pos); start: - /* - * Prevent starvation issues if someone is doing a consistency - * sync-to-disk - */ - ret = wait_on_bit_action(&NFS_I(mapping->host)->flags, NFS_INO_FLUSHING, - nfs_wait_bit_killable, TASK_KILLABLE); - if (ret) - return ret; /* * Wait for O_DIRECT to complete */ diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h index 0b9e5cc9a747..fe80a1c26340 100644 --- a/fs/nfs/nfstrace.h +++ b/fs/nfs/nfstrace.h @@ -37,7 +37,6 @@ { 1 << NFS_INO_ADVISE_RDPLUS, "ADVISE_RDPLUS" }, \ { 1 << NFS_INO_STALE, "STALE" }, \ { 1 << NFS_INO_INVALIDATING, "INVALIDATING" }, \ - { 1 << NFS_INO_FLUSHING, "FLUSHING" }, \ { 1 << NFS_INO_FSCACHE, "FSCACHE" }, \ { 1 << NFS_INO_LAYOUTCOMMIT, "NEED_LAYOUTCOMMIT" }, \ { 1 << NFS_INO_LAYOUTCOMMITTING, "LAYOUTCOMMIT" }) diff --git a/fs/nfs/write.c b/fs/nfs/write.c index e1c74d3db64d..980d44f3a84c 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -657,16 +657,9 @@ static int nfs_writepages_callback(struct page *page, struct writeback_control * int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) { struct inode *inode = mapping->host; - unsigned long *bitlock = &NFS_I(inode)->flags; struct nfs_pageio_descriptor pgio; int err; - /* Stop dirtying of new pages while we sync */ - err = wait_on_bit_lock_action(bitlock, NFS_INO_FLUSHING, - nfs_wait_bit_killable, TASK_KILLABLE); - if (err) - goto out_err; - nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGES); nfs_pageio_init_write(&pgio, inode, wb_priority(wbc), false, @@ -674,10 +667,6 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) err = write_cache_pages(mapping, wbc, nfs_writepages_callback, &pgio); nfs_pageio_complete(&pgio); - clear_bit_unlock(NFS_INO_FLUSHING, bitlock); - smp_mb__after_atomic(); - wake_up_bit(bitlock, NFS_INO_FLUSHING); - if (err < 0) goto out_err; err = pgio.pg_error; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index d71278c3c5bd..120dd04b553c 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -205,7 +205,6 @@ struct nfs_inode { #define NFS_INO_STALE (1) /* possible stale inode */ #define NFS_INO_ACL_LRU_SET (2) /* Inode is on the LRU list */ #define NFS_INO_INVALIDATING (3) /* inode is being invalidated */ -#define NFS_INO_FLUSHING (4) /* inode is flushing out data */ #define NFS_INO_FSCACHE (5) /* inode can be cached by FS-Cache */ #define NFS_INO_FSCACHE_LOCK (6) /* FS-Cache cookie management lock */ #define NFS_INO_LAYOUTCOMMIT (9) /* layoutcommit required */ -- cgit v1.2.3 From 811ed92ecc9f47eee90beabcf5c2133f2a6d2440 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Jun 2016 18:25:56 -0400 Subject: NFS: writepage of a single page should not be synchronous It is almost always better to wait for more so that we can issue a bulk commit. Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 980d44f3a84c..b13d48881d3a 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -625,7 +625,7 @@ static int nfs_writepage_locked(struct page *page, int err; nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE); - nfs_pageio_init_write(&pgio, inode, wb_priority(wbc), + nfs_pageio_init_write(&pgio, inode, 0, false, &nfs_async_write_completion_ops); err = nfs_do_writepage(page, wbc, &pgio, launder); nfs_pageio_complete(&pgio); -- cgit v1.2.3 From 93761d9863c332d1099d80629f89cf48eb745e48 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 2 Jun 2016 11:03:00 -0400 Subject: NFS: Don't hold the inode lock across fsync() Commits are no longer required to be serialised. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 29d7477a62e8..249262b6bcbe 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -277,11 +277,9 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret != 0) break; - inode_lock(inode); ret = nfs_file_fsync_commit(file, start, end, datasync); if (!ret) ret = pnfs_sync_inode(inode, !!datasync); - inode_unlock(inode); /* * If nfs_file_fsync_commit detected a server reboot, then * resend all dirty pages that might have been covered by -- cgit v1.2.3 From 4f52b6bb8c57b9accafad526a429d6c0851cc62f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 2 Jun 2016 18:10:33 -0400 Subject: NFS: Don't call COMMIT in ->releasepage() While COMMIT has the potential to free up a lot of memory that is being taken by unstable writes, it isn't guaranteed to free up this particular page. Also, calling fsync() on the server is expensive and so we want to do it in a more controlled fashion, rather than have it triggered at random by the VM. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 23 ----------------------- 1 file changed, 23 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 249262b6bcbe..df4dd8e7e62e 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -460,31 +460,8 @@ static void nfs_invalidate_page(struct page *page, unsigned int offset, */ static int nfs_release_page(struct page *page, gfp_t gfp) { - struct address_space *mapping = page->mapping; - dfprintk(PAGECACHE, "NFS: release_page(%p)\n", page); - /* Always try to initiate a 'commit' if relevant, but only - * wait for it if the caller allows blocking. Even then, - * only wait 1 second and only if the 'bdi' is not congested. - * Waiting indefinitely can cause deadlocks when the NFS - * server is on this machine, when a new TCP connection is - * needed and in other rare cases. There is no particular - * need to wait extensively here. A short wait has the - * benefit that someone else can worry about the freezer. - */ - if (mapping) { - struct nfs_server *nfss = NFS_SERVER(mapping->host); - nfs_commit_inode(mapping->host, 0); - if (gfpflags_allow_blocking(gfp) && - !bdi_write_congested(&nfss->backing_dev_info)) { - wait_on_page_bit_killable_timeout(page, PG_private, - HZ); - if (PagePrivate(page)) - set_bdi_congested(&nfss->backing_dev_info, - BLK_RW_ASYNC); - } - } /* If PagePrivate() is set, then the page is not freeable */ if (PagePrivate(page)) return 0; -- cgit v1.2.3 From 5c6e5b60aae4347223f176966455010a5715b863 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 22 Jun 2016 14:13:12 -0400 Subject: NFS: Fix an Oops in the pNFS files and flexfiles connection setup to the DS Chris Worley reports: RIP: 0010:[] [] rpc_new_client+0x2a0/0x2e0 [sunrpc] RSP: 0018:ffff880158f6f548 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff880234f8bc00 RCX: 000000000000ea60 RDX: 0000000000074cc0 RSI: 000000000000ea60 RDI: ffff880234f8bcf0 RBP: ffff880158f6f588 R08: 000000000001ac80 R09: ffff880237003300 R10: ffff880201171000 R11: ffffea0000d75200 R12: ffffffffa03afc60 R13: ffff880230c18800 R14: 0000000000000000 R15: ffff880158f6f680 FS: 00007f0e32673740(0000) GS:ffff88023fc40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000008 CR3: 0000000234886000 CR4: 00000000001406e0 Stack: ffffffffa047a680 0000000000000000 ffff880158f6f598 ffff880158f6f680 ffff880158f6f680 ffff880234d11d00 ffff88023357f800 ffff880158f6f7d0 ffff880158f6f5b8 ffffffffa024660a ffff880158f6f5b8 ffffffffa02492ec Call Trace: [] rpc_create_xprt+0x1a/0xb0 [sunrpc] [] ? xprt_create_transport+0x13c/0x240 [sunrpc] [] rpc_create+0xc6/0x1a0 [sunrpc] [] nfs_create_rpc_client+0xf5/0x140 [nfs] [] nfs_init_client+0x3a/0xd0 [nfs] [] nfs_get_client+0x25f/0x310 [nfs] [] ? rpc_ntop+0xe8/0x100 [sunrpc] [] nfs3_set_ds_client+0xcc/0x100 [nfsv3] [] nfs4_pnfs_ds_connect+0x120/0x400 [nfsv4] [] nfs4_ff_layout_prepare_ds+0xe7/0x330 [nfs_layout_flexfiles] [] ff_layout_pg_init_write+0xcb/0x280 [nfs_layout_flexfiles] [] __nfs_pageio_add_request+0x12c/0x490 [nfs] [] nfs_pageio_add_request+0xc2/0x2a0 [nfs] [] ? nfs_pageio_init+0x75/0x120 [nfs] [] nfs_do_writepage+0x120/0x270 [nfs] [] nfs_writepage_locked+0x61/0xc0 [nfs] [] ? __percpu_counter_add+0x55/0x70 [] nfs_wb_single_page+0xef/0x1c0 [nfs] [] ? __dec_zone_page_state+0x33/0x40 [] nfs_launder_page+0x41/0x90 [nfs] [] invalidate_inode_pages2_range+0x340/0x3a0 [] invalidate_inode_pages2+0x17/0x20 [] nfs_release+0x9e/0xb0 [nfs] [] ? nfs_open+0x60/0x60 [nfs] [] nfs_file_release+0x3d/0x60 [nfs] [] __fput+0xdc/0x1e0 [] ____fput+0xe/0x10 [] task_work_run+0xc4/0xe0 [] do_exit+0x2e8/0xb30 [] ? do_audit_syscall_entry+0x6c/0x70 [] ? __audit_syscall_exit+0x1e6/0x280 [] do_group_exit+0x3f/0xa0 [] SyS_exit_group+0x14/0x20 [] system_call_fastpath+0x12/0x71 Which seems to be due to a call to utsname() when in a task exit context in order to determine the hostname to set in rpc_new_client(). In reality, what we want here is not the hostname of the current task, but the hostname that was used to set up the metadata server. Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 22 ++++++++++------------ fs/nfs/internal.h | 16 ++++++++-------- fs/nfs/nfs3client.c | 8 +++++--- fs/nfs/nfs4client.c | 20 ++++++++++++-------- include/linux/nfs_xdr.h | 5 ++--- 5 files changed, 37 insertions(+), 34 deletions(-) (limited to 'fs') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 0c96528db94a..4849d0f778dc 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -367,8 +367,6 @@ nfs_found_client(const struct nfs_client_initdata *cl_init, */ struct nfs_client * nfs_get_client(const struct nfs_client_initdata *cl_init, - const struct rpc_timeout *timeparms, - const char *ip_addr, rpc_authflavor_t authflavour) { struct nfs_client *clp, *new = NULL; @@ -399,7 +397,7 @@ nfs_get_client(const struct nfs_client_initdata *cl_init, &nn->nfs_client_list); spin_unlock(&nn->nfs_client_lock); new->cl_flags = cl_init->init_flags; - return rpc_ops->init_client(new, timeparms, ip_addr); + return rpc_ops->init_client(new, cl_init); } spin_unlock(&nn->nfs_client_lock); @@ -470,7 +468,7 @@ EXPORT_SYMBOL_GPL(nfs_init_timeout_values); * Create an RPC client handle */ int nfs_create_rpc_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, + const struct nfs_client_initdata *cl_init, rpc_authflavor_t flavor) { struct rpc_clnt *clnt = NULL; @@ -479,8 +477,9 @@ int nfs_create_rpc_client(struct nfs_client *clp, .protocol = clp->cl_proto, .address = (struct sockaddr *)&clp->cl_addr, .addrsize = clp->cl_addrlen, - .timeout = timeparms, + .timeout = cl_init->timeparms, .servername = clp->cl_hostname, + .nodename = cl_init->nodename, .program = &nfs_program, .version = clp->rpc_ops->version, .authflavor = flavor, @@ -591,14 +590,12 @@ EXPORT_SYMBOL_GPL(nfs_init_server_rpcclient); * nfs_init_client - Initialise an NFS2 or NFS3 client * * @clp: nfs_client to initialise - * @timeparms: timeout parameters for underlying RPC transport - * @ip_addr: IP presentation address (not used) + * @cl_init: Initialisation parameters * * Returns pointer to an NFS client, or an ERR_PTR value. */ struct nfs_client *nfs_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr) + const struct nfs_client_initdata *cl_init) { int error; @@ -612,7 +609,7 @@ struct nfs_client *nfs_init_client(struct nfs_client *clp, * Create a client RPC handle for doing FSSTAT with UNIX auth only * - RFC 2623, sec 2.3.2 */ - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_UNIX); if (error < 0) goto error; nfs_mark_client_ready(clp, NFS_CS_READY); @@ -633,6 +630,7 @@ static int nfs_init_server(struct nfs_server *server, const struct nfs_parsed_mount_data *data, struct nfs_subversion *nfs_mod) { + struct rpc_timeout timeparms; struct nfs_client_initdata cl_init = { .hostname = data->nfs_server.hostname, .addr = (const struct sockaddr *)&data->nfs_server.address, @@ -640,8 +638,8 @@ static int nfs_init_server(struct nfs_server *server, .nfs_mod = nfs_mod, .proto = data->nfs_server.protocol, .net = data->net, + .timeparms = &timeparms, }; - struct rpc_timeout timeparms; struct nfs_client *clp; int error; @@ -653,7 +651,7 @@ static int nfs_init_server(struct nfs_server *server, set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); /* Allocate or find a client reference we can use */ - clp = nfs_get_client(&cl_init, &timeparms, NULL, RPC_AUTH_UNIX); + clp = nfs_get_client(&cl_init, RPC_AUTH_UNIX); if (IS_ERR(clp)) { dprintk("<-- nfs_init_server() = error %ld\n", PTR_ERR(clp)); return PTR_ERR(clp); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 5154fa65a2f2..fa88609f85e3 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -66,13 +66,16 @@ struct nfs_clone_mount { struct nfs_client_initdata { unsigned long init_flags; - const char *hostname; - const struct sockaddr *addr; + const char *hostname; /* Hostname of the server */ + const struct sockaddr *addr; /* Address of the server */ + const char *nodename; /* Hostname of the client */ + const char *ip_addr; /* IP address of the client */ size_t addrlen; struct nfs_subversion *nfs_mod; int proto; u32 minorversion; struct net *net; + const struct rpc_timeout *timeparms; }; /* @@ -147,9 +150,8 @@ extern void nfs_umount(const struct nfs_mount_request *info); extern const struct rpc_program nfs_program; extern void nfs_clients_init(struct net *net); extern struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *); -int nfs_create_rpc_client(struct nfs_client *, const struct rpc_timeout *, rpc_authflavor_t); +int nfs_create_rpc_client(struct nfs_client *, const struct nfs_client_initdata *, rpc_authflavor_t); struct nfs_client *nfs_get_client(const struct nfs_client_initdata *, - const struct rpc_timeout *, const char *, rpc_authflavor_t); int nfs_probe_fsinfo(struct nfs_server *server, struct nfs_fh *, struct nfs_fattr *); void nfs_server_insert_lists(struct nfs_server *); @@ -338,8 +340,7 @@ nfs4_label_copy(struct nfs4_label *dst, struct nfs4_label *src) /* proc.c */ void nfs_close_context(struct nfs_open_context *ctx, int is_sync); extern struct nfs_client *nfs_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr); + const struct nfs_client_initdata *); /* dir.c */ extern void nfs_force_use_readdirplus(struct inode *dir); @@ -521,8 +522,7 @@ extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq); /* nfs4proc.c */ extern void __nfs4_read_done_cb(struct nfs_pgio_header *); extern struct nfs_client *nfs4_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr); + const struct nfs_client_initdata *); extern int nfs40_walk_client_list(struct nfs_client *clp, struct nfs_client **result, struct rpc_cred *cred); diff --git a/fs/nfs/nfs3client.c b/fs/nfs/nfs3client.c index 9e9fa347a948..0457b4129421 100644 --- a/fs/nfs/nfs3client.c +++ b/fs/nfs/nfs3client.c @@ -81,14 +81,17 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, rpc_authflavor_t au_flavor) { + struct rpc_timeout ds_timeout; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, + .nodename = mds_clp->cl_rpcclient->cl_nodename, + .ip_addr = mds_clp->cl_ipaddr, .nfs_mod = &nfs_v3, .proto = ds_proto, .net = mds_clp->cl_net, + .timeparms = &ds_timeout, }; - struct rpc_timeout ds_timeout; struct nfs_client *clp; char buf[INET6_ADDRSTRLEN + 1]; @@ -99,8 +102,7 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, /* Use the MDS nfs_client cl_ipaddr. */ nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans); - clp = nfs_get_client(&cl_init, &ds_timeout, mds_clp->cl_ipaddr, - au_flavor); + clp = nfs_get_client(&cl_init, au_flavor); return clp; } diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 10410e8b5853..5fc7fbbfdcef 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -349,10 +349,10 @@ static int nfs4_init_client_minor_version(struct nfs_client *clp) * Returns pointer to an NFS client, or an ERR_PTR value. */ struct nfs_client *nfs4_init_client(struct nfs_client *clp, - const struct rpc_timeout *timeparms, - const char *ip_addr) + const struct nfs_client_initdata *cl_init) { char buf[INET6_ADDRSTRLEN + 1]; + const char *ip_addr = cl_init->ip_addr; struct nfs_client *old; int error; @@ -370,9 +370,9 @@ struct nfs_client *nfs4_init_client(struct nfs_client *clp, __set_bit(NFS_CS_DISCRTRY, &clp->cl_flags); __set_bit(NFS_CS_NO_RETRANS_TIMEOUT, &clp->cl_flags); - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_GSS_KRB5I); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_GSS_KRB5I); if (error == -EINVAL) - error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX); + error = nfs_create_rpc_client(clp, cl_init, RPC_AUTH_UNIX); if (error < 0) goto error; @@ -793,10 +793,12 @@ static int nfs4_set_client(struct nfs_server *server, .hostname = hostname, .addr = addr, .addrlen = addrlen, + .ip_addr = ip_addr, .nfs_mod = &nfs_v4, .proto = proto, .minorversion = minorversion, .net = net, + .timeparms = timeparms, }; struct nfs_client *clp; int error; @@ -809,7 +811,7 @@ static int nfs4_set_client(struct nfs_server *server, set_bit(NFS_CS_MIGRATION, &cl_init.init_flags); /* Allocate or find a client reference we can use */ - clp = nfs_get_client(&cl_init, timeparms, ip_addr, authflavour); + clp = nfs_get_client(&cl_init, authflavour); if (IS_ERR(clp)) { error = PTR_ERR(clp); goto error; @@ -847,15 +849,18 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, u32 minor_version, rpc_authflavor_t au_flavor) { + struct rpc_timeout ds_timeout; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, + .nodename = mds_clp->cl_rpcclient->cl_nodename, + .ip_addr = mds_clp->cl_ipaddr, .nfs_mod = &nfs_v4, .proto = ds_proto, .minorversion = minor_version, .net = mds_clp->cl_net, + .timeparms = &ds_timeout, }; - struct rpc_timeout ds_timeout; struct nfs_client *clp; char buf[INET6_ADDRSTRLEN + 1]; @@ -869,8 +874,7 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, * (section 13.1 RFC 5661). */ nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans); - clp = nfs_get_client(&cl_init, &ds_timeout, mds_clp->cl_ipaddr, - au_flavor); + clp = nfs_get_client(&cl_init, au_flavor); dprintk("<-- %s %p\n", __func__, clp); return clp; diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index c304a11b5b1a..82b81a1c2438 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1596,9 +1596,8 @@ struct nfs_rpc_ops { int (*have_delegation)(struct inode *, fmode_t); int (*return_delegation)(struct inode *); struct nfs_client *(*alloc_client) (const struct nfs_client_initdata *); - struct nfs_client * - (*init_client) (struct nfs_client *, const struct rpc_timeout *, - const char *); + struct nfs_client *(*init_client) (struct nfs_client *, + const struct nfs_client_initdata *); void (*free_client) (struct nfs_client *); struct nfs_server *(*create_server)(struct nfs_mount_info *, struct nfs_subversion *); struct nfs_server *(*clone_server)(struct nfs_server *, struct nfs_fh *, -- cgit v1.2.3 From ca857cc1d4cf17aba4bbb3b95d35454ad96924b3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 28 Jun 2016 13:54:09 -0400 Subject: NFS/pnfs: Do not clobber existing pgio_done_cb in nfs4_proc_read_setup If a pNFS client sets hdr->pgio_done_cb, then we should not overwrite that in nfs4_proc_read_setup() Fixes: 75bf47ebf6b5 ("pNFS/flexfile: Fix erroneous fall back to...") Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ff416d0e24bc..6191b7e46913 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -4392,7 +4392,8 @@ static void nfs4_proc_read_setup(struct nfs_pgio_header *hdr, struct rpc_message *msg) { hdr->timestamp = jiffies; - hdr->pgio_done_cb = nfs4_read_done_cb; + if (!hdr->pgio_done_cb) + hdr->pgio_done_cb = nfs4_read_done_cb; msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ]; nfs4_init_sequence(&hdr->args.seq_args, &hdr->res.seq_res, 0); } -- cgit v1.2.3 From 8487c479e2668dd1231e9c3c77a203d744aec081 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 26 Jun 2016 08:44:35 -0400 Subject: NFSv4: Allow retry of operations that used a returned delegation stateid Fix up nfs4_do_handle_exception() so that it can check if the operation that received the NFS4ERR_BAD_STATEID was using a defunct delegation. Apply that to the case of SETATTR, which will currently return EIO in some cases where this happens. Reported-by: Olga Kornievskaia Signed-off-by: Trond Myklebust --- fs/nfs/nfs4_fs.h | 1 + fs/nfs/nfs4proc.c | 79 ++++++++++++++++++++++++++++++++----------------------- 2 files changed, 47 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 768456fa1b17..4be567a54958 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -185,6 +185,7 @@ struct nfs4_state { struct nfs4_exception { struct nfs4_state *state; struct inode *inode; + nfs4_stateid *stateid; long timeout; unsigned char delay : 1, recovering : 1, diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 6191b7e46913..519368b98762 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -363,6 +363,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server, { struct nfs_client *clp = server->nfs_client; struct nfs4_state *state = exception->state; + const nfs4_stateid *stateid = exception->stateid; struct inode *inode = exception->inode; int ret = errorcode; @@ -376,9 +377,18 @@ static int nfs4_do_handle_exception(struct nfs_server *server, case -NFS4ERR_DELEG_REVOKED: case -NFS4ERR_ADMIN_REVOKED: case -NFS4ERR_BAD_STATEID: - if (inode && nfs_async_inode_return_delegation(inode, - NULL) == 0) - goto wait_on_recovery; + if (inode) { + int err; + + err = nfs_async_inode_return_delegation(inode, + stateid); + if (err == 0) + goto wait_on_recovery; + if (stateid != NULL && stateid->type == NFS4_DELEGATION_STATEID_TYPE) { + exception->retry = 1; + break; + } + } if (state == NULL) break; ret = nfs4_schedule_stateid_recovery(server, state); @@ -2669,28 +2679,17 @@ static struct nfs4_state *nfs4_do_open(struct inode *dir, return res; } -static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, - struct nfs_fattr *fattr, struct iattr *sattr, - struct nfs4_state *state, struct nfs4_label *ilabel, - struct nfs4_label *olabel) +static int _nfs4_do_setattr(struct inode *inode, + struct nfs_setattrargs *arg, + struct nfs_setattrres *res, + struct rpc_cred *cred, + struct nfs4_state *state) { struct nfs_server *server = NFS_SERVER(inode); - struct nfs_setattrargs arg = { - .fh = NFS_FH(inode), - .iap = sattr, - .server = server, - .bitmask = server->attr_bitmask, - .label = ilabel, - }; - struct nfs_setattrres res = { - .fattr = fattr, - .label = olabel, - .server = server, - }; struct rpc_message msg = { .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETATTR], - .rpc_argp = &arg, - .rpc_resp = &res, + .rpc_argp = arg, + .rpc_resp = res, .rpc_cred = cred, }; struct rpc_cred *delegation_cred = NULL; @@ -2699,17 +2698,13 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, bool truncate; int status; - arg.bitmask = nfs4_bitmask(server, ilabel); - if (ilabel) - arg.bitmask = nfs4_bitmask(server, olabel); - - nfs_fattr_init(fattr); + nfs_fattr_init(res->fattr); /* Servers should only apply open mode checks for file size changes */ - truncate = (sattr->ia_valid & ATTR_SIZE) ? true : false; + truncate = (arg->iap->ia_valid & ATTR_SIZE) ? true : false; fmode = truncate ? FMODE_WRITE : FMODE_READ; - if (nfs4_copy_delegation_stateid(inode, fmode, &arg.stateid, &delegation_cred)) { + if (nfs4_copy_delegation_stateid(inode, fmode, &arg->stateid, &delegation_cred)) { /* Use that stateid */ } else if (truncate && state != NULL) { struct nfs_lockowner lockowner = { @@ -2719,19 +2714,19 @@ static int _nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, if (!nfs4_valid_open_stateid(state)) return -EBADF; if (nfs4_select_rw_stateid(state, FMODE_WRITE, &lockowner, - &arg.stateid, &delegation_cred) == -EIO) + &arg->stateid, &delegation_cred) == -EIO) return -EBADF; } else - nfs4_stateid_copy(&arg.stateid, &zero_stateid); + nfs4_stateid_copy(&arg->stateid, &zero_stateid); if (delegation_cred) msg.rpc_cred = delegation_cred; - status = nfs4_call_sync(server->client, server, &msg, &arg.seq_args, &res.seq_res, 1); + status = nfs4_call_sync(server->client, server, &msg, &arg->seq_args, &res->seq_res, 1); put_rpccred(delegation_cred); if (status == 0 && state != NULL) renew_lease(server, timestamp); - trace_nfs4_setattr(inode, &arg.stateid, status); + trace_nfs4_setattr(inode, &arg->stateid, status); return status; } @@ -2741,13 +2736,31 @@ static int nfs4_do_setattr(struct inode *inode, struct rpc_cred *cred, struct nfs4_label *olabel) { struct nfs_server *server = NFS_SERVER(inode); + struct nfs_setattrargs arg = { + .fh = NFS_FH(inode), + .iap = sattr, + .server = server, + .bitmask = server->attr_bitmask, + .label = ilabel, + }; + struct nfs_setattrres res = { + .fattr = fattr, + .label = olabel, + .server = server, + }; struct nfs4_exception exception = { .state = state, .inode = inode, + .stateid = &arg.stateid, }; int err; + + arg.bitmask = nfs4_bitmask(server, ilabel); + if (ilabel) + arg.bitmask = nfs4_bitmask(server, olabel); + do { - err = _nfs4_do_setattr(inode, cred, fattr, sattr, state, ilabel, olabel); + err = _nfs4_do_setattr(inode, &arg, &res, cred, state); switch (err) { case -NFS4ERR_OPENMODE: if (!(sattr->ia_valid & ATTR_SIZE)) { -- cgit v1.2.3 From 73e6c5d854d3f7f75e8b46d3e54aeb5d83fe6b1f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 26 Jun 2016 12:27:25 -0400 Subject: pNFS/files: Fix layoutcommit after a commit to DS According to the errata https://www.rfc-editor.org/errata_search.php?rfc=5661&eid=2751 we should always send layout commit after a commit to DS. Fixes: bc7d4b8fd091 ("nfs/filelayout: set layoutcommit...") Signed-off-by: Trond Myklebust --- fs/nfs/filelayout/filelayout.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index aa59757389dc..b4c1407e8fe4 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -375,8 +375,7 @@ static int filelayout_commit_done_cb(struct rpc_task *task, return -EAGAIN; } - if (data->verf.committed == NFS_UNSTABLE) - pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); + pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); return 0; } -- cgit v1.2.3 From c001c87a63aa2f35358e33eb05e45e4cbcb34f54 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 26 Jun 2016 12:39:49 -0400 Subject: pNFS/flexfiles: Fix layoutcommit after a commit to DS We should always do a layoutcommit after commit to DS, except if the layout segment we're using has set FF_FLAGS_NO_LAYOUTCOMMIT. Fixes: d67ae825a59d ("pnfs/flexfiles: Add the FlexFile Layout Driver") Signed-off-by: Trond Myklebust --- fs/nfs/flexfilelayout/flexfilelayout.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 0e8018bc9880..2689c9e9dc3c 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1530,8 +1530,7 @@ static int ff_layout_commit_done_cb(struct rpc_task *task, return -EAGAIN; } - if (data->verf.committed == NFS_UNSTABLE - && ff_layout_need_layoutcommit(data->lseg)) + if (ff_layout_need_layoutcommit(data->lseg)) pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); return 0; -- cgit v1.2.3 From bc28e1c2e3c8a4c5198ebfd8bbae0afd73dfafd5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 26 Jun 2016 16:14:40 -0400 Subject: pNFS/flexfiles: Clean up calls to pnfs_set_layoutcommit() Let's just have one place where we check ff_layout_need_layoutcommit(). Signed-off-by: Trond Myklebust --- fs/nfs/flexfilelayout/flexfilelayout.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 2689c9e9dc3c..14f2ed3f1a5b 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1325,15 +1325,16 @@ ff_layout_need_layoutcommit(struct pnfs_layout_segment *lseg) * we always send layoutcommit after DS writes. */ static void -ff_layout_set_layoutcommit(struct nfs_pgio_header *hdr) +ff_layout_set_layoutcommit(struct inode *inode, + struct pnfs_layout_segment *lseg, + loff_t end_offset) { - if (!ff_layout_need_layoutcommit(hdr->lseg)) + if (!ff_layout_need_layoutcommit(lseg)) return; - pnfs_set_layoutcommit(hdr->inode, hdr->lseg, - hdr->mds_offset + hdr->res.count); - dprintk("%s inode %lu pls_end_pos %lu\n", __func__, hdr->inode->i_ino, - (unsigned long) NFS_I(hdr->inode)->layout->plh_lwb); + pnfs_set_layoutcommit(inode, lseg, end_offset); + dprintk("%s inode %lu pls_end_pos %llu\n", __func__, inode->i_ino, + (unsigned long long) NFS_I(inode)->layout->plh_lwb); } static bool @@ -1494,7 +1495,8 @@ static int ff_layout_write_done_cb(struct rpc_task *task, if (hdr->res.verf->committed == NFS_FILE_SYNC || hdr->res.verf->committed == NFS_DATA_SYNC) - ff_layout_set_layoutcommit(hdr); + ff_layout_set_layoutcommit(hdr->inode, hdr->lseg, + hdr->mds_offset + (loff_t)hdr->res.count); /* zero out fattr since we don't care DS attr at all */ hdr->fattr.valid = 0; @@ -1530,8 +1532,7 @@ static int ff_layout_commit_done_cb(struct rpc_task *task, return -EAGAIN; } - if (ff_layout_need_layoutcommit(data->lseg)) - pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb); + ff_layout_set_layoutcommit(data->inode, data->lseg, data->lwb); return 0; } -- cgit v1.2.3 From 2e18d4d822ea9cc811ea26a880cf2ed47cbf8889 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 26 Jun 2016 18:54:58 -0400 Subject: pNFS: Files and flexfiles always need to commit before layoutcommit So ensure that we mark the layout for commit once the write is done, and then ensure that the commit to ds is finished before sending layoutcommit. Note that by doing this, we're able to optimise away the commit for the case of servers that don't need layoutcommit in order to return updated attributes. Signed-off-by: Trond Myklebust --- fs/nfs/filelayout/filelayout.c | 9 ++++++--- fs/nfs/flexfilelayout/flexfilelayout.c | 7 +++++-- fs/nfs/nfs4xdr.c | 11 ++++++++--- fs/nfs/pnfs.c | 5 ++++- fs/nfs/pnfs_nfs.c | 7 +++++++ 5 files changed, 30 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index b4c1407e8fe4..25bd91a6e088 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -255,13 +255,16 @@ static int filelayout_read_done_cb(struct rpc_task *task, static void filelayout_set_layoutcommit(struct nfs_pgio_header *hdr) { + loff_t end_offs = 0; if (FILELAYOUT_LSEG(hdr->lseg)->commit_through_mds || - hdr->res.verf->committed != NFS_DATA_SYNC) + hdr->res.verf->committed == NFS_FILE_SYNC) return; + if (hdr->res.verf->committed == NFS_DATA_SYNC) + end_offs = hdr->mds_offset + (loff_t)hdr->res.count; - pnfs_set_layoutcommit(hdr->inode, hdr->lseg, - hdr->mds_offset + hdr->res.count); + /* Note: if the write is unstable, don't set end_offs until commit */ + pnfs_set_layoutcommit(hdr->inode, hdr->lseg, end_offs); dprintk("%s inode %lu pls_end_pos %lu\n", __func__, hdr->inode->i_ino, (unsigned long) NFS_I(hdr->inode)->layout->plh_lwb); } diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index 14f2ed3f1a5b..e6206eaf2bdf 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1470,6 +1470,7 @@ static void ff_layout_read_release(void *data) static int ff_layout_write_done_cb(struct rpc_task *task, struct nfs_pgio_header *hdr) { + loff_t end_offs = 0; int err; trace_nfs4_pnfs_write(hdr, task->tk_status); @@ -1495,8 +1496,10 @@ static int ff_layout_write_done_cb(struct rpc_task *task, if (hdr->res.verf->committed == NFS_FILE_SYNC || hdr->res.verf->committed == NFS_DATA_SYNC) - ff_layout_set_layoutcommit(hdr->inode, hdr->lseg, - hdr->mds_offset + (loff_t)hdr->res.count); + end_offs = hdr->mds_offset + (loff_t)hdr->res.count; + + /* Note: if the write is unstable, don't set end_offs until commit */ + ff_layout_set_layoutcommit(hdr->inode, hdr->lseg, end_offs); /* zero out fattr since we don't care DS attr at all */ hdr->fattr.valid = 0; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 661e753fe1c9..7bd3a5c09d31 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1985,9 +1985,14 @@ encode_layoutcommit(struct xdr_stream *xdr, p = xdr_encode_hyper(p, args->lastbytewritten + 1); /* length */ *p = cpu_to_be32(0); /* reclaim */ encode_nfs4_stateid(xdr, &args->stateid); - p = reserve_space(xdr, 20); - *p++ = cpu_to_be32(1); /* newoffset = TRUE */ - p = xdr_encode_hyper(p, args->lastbytewritten); + if (args->lastbytewritten != U64_MAX) { + p = reserve_space(xdr, 20); + *p++ = cpu_to_be32(1); /* newoffset = TRUE */ + p = xdr_encode_hyper(p, args->lastbytewritten); + } else { + p = reserve_space(xdr, 12); + *p++ = cpu_to_be32(0); /* newoffset = FALSE */ + } *p++ = cpu_to_be32(0); /* Never send time_modify_changed */ *p++ = cpu_to_be32(NFS_SERVER(args->inode)->pnfs_curr_ld->id);/* type */ diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 0c7e0d45a4de..62553182514e 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -2378,7 +2378,10 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync) nfs_fattr_init(&data->fattr); data->args.bitmask = NFS_SERVER(inode)->cache_consistency_bitmask; data->res.fattr = &data->fattr; - data->args.lastbytewritten = end_pos - 1; + if (end_pos != 0) + data->args.lastbytewritten = end_pos - 1; + else + data->args.lastbytewritten = U64_MAX; data->res.server = NFS_SERVER(inode); if (ld->prepare_layoutcommit) { diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index 0dfc476da3e1..0d10cc280a23 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -932,6 +932,13 @@ EXPORT_SYMBOL_GPL(pnfs_layout_mark_request_commit); int pnfs_nfs_generic_sync(struct inode *inode, bool datasync) { + int ret; + + if (!pnfs_layoutcommit_outstanding(inode)) + return 0; + ret = nfs_commit_inode(inode, FLUSH_SYNC); + if (ret < 0) + return ret; if (datasync) return 0; return pnfs_layoutcommit_inode(inode, true); -- cgit v1.2.3 From ac46bd374c9a838874c450c528e2e922ee748ff9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Jul 2016 13:46:53 -0400 Subject: pNFS: Ensure we layoutcommit before revalidating attributes If we need to update the cached attributes, then we'd better make sure that we also layoutcommit first. Otherwise, the server may have stale attributes. Prior to this patch, the revalidation code tried to "fix" this problem by simply disabling attributes that would be affected by the layoutcommit. That approach breaks nfs_writeback_check_extend(), leading to a file size corruption. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 23 +++++++---------------- 1 file changed, 7 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 4e65a5a8a01b..6c0618eb5d57 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -974,6 +974,13 @@ __nfs_revalidate_inode(struct nfs_server *server, struct inode *inode) if (NFS_STALE(inode)) goto out; + /* pNFS: Attributes aren't updated until we layoutcommit */ + if (S_ISREG(inode->i_mode)) { + status = pnfs_sync_inode(inode, false); + if (status) + goto out; + } + status = -ENOMEM; fattr = nfs_alloc_fattr(); if (fattr == NULL) @@ -1493,28 +1500,12 @@ static int nfs_inode_attrs_need_update(const struct inode *inode, const struct n ((long)nfsi->attr_gencount - (long)nfs_read_attr_generation_counter() > 0); } -/* - * Don't trust the change_attribute, mtime, ctime or size if - * a pnfs LAYOUTCOMMIT is outstanding - */ -static void nfs_inode_attrs_handle_layoutcommit(struct inode *inode, - struct nfs_fattr *fattr) -{ - if (pnfs_layoutcommit_outstanding(inode)) - fattr->valid &= ~(NFS_ATTR_FATTR_CHANGE | - NFS_ATTR_FATTR_MTIME | - NFS_ATTR_FATTR_CTIME | - NFS_ATTR_FATTR_SIZE); -} - static int nfs_refresh_inode_locked(struct inode *inode, struct nfs_fattr *fattr) { int ret; trace_nfs_refresh_inode_enter(inode); - nfs_inode_attrs_handle_layoutcommit(inode, fattr); - if (nfs_inode_attrs_need_update(inode, fattr)) ret = nfs_update_inode(inode, fattr); else -- cgit v1.2.3 From 6712007734cbd64ff924af16fc236751d47ff80b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 5 Jul 2016 19:08:58 -0400 Subject: pNFS: pnfs_layoutcommit_outstanding() is no longer used when !CONFIG_NFS_V4_1 Cleanup... Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.h | 7 ------- 1 file changed, 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index b21bd0bee784..d6be5299a55a 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -716,13 +716,6 @@ pnfs_use_threshold(struct nfs4_threshold **dst, struct nfs4_threshold *src, return false; } -static inline bool -pnfs_layoutcommit_outstanding(struct inode *inode) -{ - return false; -} - - static inline struct nfs4_threshold *pnfs_mdsthreshold_alloc(void) { return NULL; -- cgit v1.2.3 From 8fc3c3862728373e0d0f5abccc6afc56c69e0c63 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Jun 2016 21:32:24 -0400 Subject: NFS: Fix O_DIRECT verifier problems We should not be interested in looking at the value of the stable field, since that could take any value. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 10 ++++++++-- fs/nfs/internal.h | 7 +++++++ fs/nfs/write.c | 2 +- 3 files changed, 16 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 979b3c4dee6a..d6d43b5eafb3 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -196,6 +196,12 @@ static void nfs_direct_set_hdr_verf(struct nfs_direct_req *dreq, WARN_ON_ONCE(verfp->committed < 0); } +static int nfs_direct_cmp_verf(const struct nfs_writeverf *v1, + const struct nfs_writeverf *v2) +{ + return nfs_write_verifier_cmp(&v1->verifier, &v2->verifier); +} + /* * nfs_direct_cmp_hdr_verf - compare verifier for pgio header * @dreq - direct request possibly spanning multiple servers @@ -215,7 +221,7 @@ static int nfs_direct_set_or_cmp_hdr_verf(struct nfs_direct_req *dreq, nfs_direct_set_hdr_verf(dreq, hdr); return 0; } - return memcmp(verfp, &hdr->verf, sizeof(struct nfs_writeverf)); + return nfs_direct_cmp_verf(verfp, &hdr->verf); } /* @@ -238,7 +244,7 @@ static int nfs_direct_cmp_commit_data_verf(struct nfs_direct_req *dreq, if (verfp->committed < 0) return 1; - return memcmp(verfp, &data->verf, sizeof(struct nfs_writeverf)); + return nfs_direct_cmp_verf(verfp, &data->verf); } /** diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 5154fa65a2f2..150a8eb0f323 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -506,6 +506,13 @@ extern int nfs_migrate_page(struct address_space *, #define nfs_migrate_page NULL #endif +static inline int +nfs_write_verifier_cmp(const struct nfs_write_verifier *v1, + const struct nfs_write_verifier *v2) +{ + return memcmp(v1->data, v2->data, sizeof(v1->data)); +} + /* unlink.c */ extern struct rpc_task * nfs_async_rename(struct inode *old_dir, struct inode *new_dir, diff --git a/fs/nfs/write.c b/fs/nfs/write.c index b13d48881d3a..3087fb6f1983 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1789,7 +1789,7 @@ static void nfs_commit_release_pages(struct nfs_commit_data *data) /* Okay, COMMIT succeeded, apparently. Check the verifier * returned by the server against all stored verfs. */ - if (!memcmp(&req->wb_verf, &data->verf.verifier, sizeof(req->wb_verf))) { + if (!nfs_write_verifier_cmp(&req->wb_verf, &data->verf.verifier)) { /* We have a match */ nfs_inode_remove_request(req); dprintk(" OK\n"); -- cgit v1.2.3 From a5314a74928fa6dbc4503a8c64f43bb5c1c12ac1 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 1 Jun 2016 21:42:32 -0400 Subject: NFS: Ensure we reset the write verifier 'committed' value on resend. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 2 ++ fs/nfs/internal.h | 17 +++++++++++++++++ 2 files changed, 19 insertions(+) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index d6d43b5eafb3..fb659bb50678 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -661,6 +661,8 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) nfs_direct_write_scan_commit_list(dreq->inode, &reqs, &cinfo); dreq->count = 0; + dreq->verf.committed = NFS_INVALID_STABLE_HOW; + nfs_clear_pnfs_ds_commit_verifiers(&dreq->ds_cinfo); for (i = 0; i < dreq->mirror_count; i++) dreq->mirrors[i].count = 0; get_dreq(dreq); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 150a8eb0f323..0eb5c924886d 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -499,6 +499,23 @@ int nfs_key_timeout_notify(struct file *filp, struct inode *inode); bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx); void nfs_pageio_stop_mirroring(struct nfs_pageio_descriptor *pgio); +#ifdef CONFIG_NFS_V4_1 +static inline +void nfs_clear_pnfs_ds_commit_verifiers(struct pnfs_ds_commit_info *cinfo) +{ + int i; + + for (i = 0; i < cinfo->nbuckets; i++) + cinfo->buckets[i].direct_verf.committed = NFS_INVALID_STABLE_HOW; +} +#else +static inline +void nfs_clear_pnfs_ds_commit_verifiers(struct pnfs_ds_commit_info *cinfo) +{ +} +#endif + + #ifdef CONFIG_MIGRATION extern int nfs_migrate_page(struct address_space *, struct page *, struct page *, enum migrate_mode); -- cgit v1.2.3 From 2f3c7d87a347b12f725f6128b3097727b91b230e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 22 Jun 2016 14:38:06 -0400 Subject: NFS: Remove racy size manipulations in O_DIRECT On success, the RPC callbacks will ensure that we make the appropriate calls to nfs_writeback_update_inode() Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 16 ---------------- 1 file changed, 16 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index fb659bb50678..826d4dace0e5 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -376,15 +376,6 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq, bool write) { struct inode *inode = dreq->inode; - if (dreq->iocb && write) { - loff_t pos = dreq->iocb->ki_pos + dreq->count; - - spin_lock(&inode->i_lock); - if (i_size_read(inode) < pos) - i_size_write(inode, pos); - spin_unlock(&inode->i_lock); - } - if (write) nfs_zap_mapping(inode, inode->i_mapping); @@ -1058,14 +1049,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) if (!result) { result = nfs_direct_wait(dreq); if (result > 0) { - struct inode *inode = mapping->host; - iocb->ki_pos = pos + result; - spin_lock(&inode->i_lock); - if (i_size_read(inode) < iocb->ki_pos) - i_size_write(inode, iocb->ki_pos); - spin_unlock(&inode->i_lock); - /* XXX: should check the generic_write_sync retval */ generic_write_sync(iocb, result); } -- cgit v1.2.3 From 89698b24d24f9c8b470a73351b0b7199c17e0153 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Jun 2016 10:35:48 -0400 Subject: NFS Cleanup: move call to generic_write_checks() into fs/nfs/direct.c Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 12 ++++++++---- fs/nfs/file.c | 6 +----- 2 files changed, 9 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 826d4dace0e5..0169eca8eb42 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -988,6 +988,7 @@ static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq, ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) { ssize_t result = -EINVAL; + size_t count; struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; struct inode *inode = mapping->host; @@ -998,8 +999,11 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) dfprintk(FILE, "NFS: direct write(%pD2, %zd@%Ld)\n", file, iov_iter_count(iter), (long long) iocb->ki_pos); - nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, - iov_iter_count(iter)); + result = generic_write_checks(iocb, iter); + if (result <= 0) + return result; + count = result; + nfs_add_stats(mapping->host, NFSIOS_DIRECTWRITTENBYTES, count); pos = iocb->ki_pos; end = (pos + iov_iter_count(iter) - 1) >> PAGE_SHIFT; @@ -1017,7 +1021,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) goto out_unlock; } - task_io_account_write(iov_iter_count(iter)); + task_io_account_write(count); result = -ENOMEM; dreq = nfs_direct_req_alloc(); @@ -1025,7 +1029,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) goto out_unlock; dreq->inode = inode; - dreq->bytes_left = dreq->max_count = iov_iter_count(iter); + dreq->bytes_left = dreq->max_count = count; dreq->io_start = pos; dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp)); l_ctx = nfs_get_lock_context(dreq->ctx); diff --git a/fs/nfs/file.c b/fs/nfs/file.c index df4dd8e7e62e..c26847c84d00 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -629,12 +629,8 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) if (result) return result; - if (iocb->ki_flags & IOCB_DIRECT) { - result = generic_write_checks(iocb, from); - if (result <= 0) - return result; + if (iocb->ki_flags & IOCB_DIRECT) return nfs_file_direct_write(iocb, from); - } dprintk("NFS: write(%pD2, %zu@%Ld)\n", file, count, (long long) iocb->ki_pos); -- cgit v1.2.3 From 18290650b1c8655cfe6e0d63dd34942a037a130b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Jun 2016 15:00:42 -0400 Subject: NFS: Move buffered I/O locking into nfs_file_write() Preparation for the patch that de-serialises O_DIRECT reads and writes. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 27 +++++++++++++++------------ 1 file changed, 15 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index c26847c84d00..46cf0afe3c0f 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -623,7 +623,6 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) struct inode *inode = file_inode(file); unsigned long written = 0; ssize_t result; - size_t count = iov_iter_count(from); result = nfs_key_timeout_notify(file, inode); if (result) @@ -633,9 +632,8 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) return nfs_file_direct_write(iocb, from); dprintk("NFS: write(%pD2, %zu@%Ld)\n", - file, count, (long long) iocb->ki_pos); + file, iov_iter_count(from), (long long) iocb->ki_pos); - result = -EBUSY; if (IS_SWAPFILE(inode)) goto out_swapfile; /* @@ -647,28 +645,33 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) goto out; } - result = count; - if (!count) + inode_lock(inode); + result = generic_write_checks(iocb, from); + if (result > 0) { + current->backing_dev_info = inode_to_bdi(inode); + result = generic_perform_write(file, from, iocb->ki_pos); + current->backing_dev_info = NULL; + } + inode_unlock(inode); + if (result <= 0) goto out; - result = generic_file_write_iter(iocb, from); - if (result > 0) - written = result; + written = generic_write_sync(iocb, result); + iocb->ki_pos += written; /* Return error values */ - if (result >= 0 && nfs_need_check_write(file, inode)) { + if (nfs_need_check_write(file, inode)) { int err = vfs_fsync(file, 0); if (err < 0) result = err; } - if (result > 0) - nfs_add_stats(inode, NFSIOS_NORMALWRITTENBYTES, written); + nfs_add_stats(inode, NFSIOS_NORMALWRITTENBYTES, written); out: return result; out_swapfile: printk(KERN_INFO "NFS: attempt to write to active swap file!\n"); - goto out; + return -EBUSY; } EXPORT_SYMBOL_GPL(nfs_file_write); -- cgit v1.2.3 From a5864c999de6703f7ce908f72337568520c6cad3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 3 Jun 2016 17:07:19 -0400 Subject: NFS: Do not serialise O_DIRECT reads and writes Allow dio requests to be scheduled in parallel, but ensuring that they do not conflict with buffered I/O. Signed-off-by: Trond Myklebust --- fs/nfs/Makefile | 2 +- fs/nfs/direct.c | 41 +++----------- fs/nfs/file.c | 12 ++-- fs/nfs/internal.h | 8 +++ fs/nfs/io.c | 147 +++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/nfs_fs.h | 1 + 6 files changed, 174 insertions(+), 37 deletions(-) create mode 100644 fs/nfs/io.c (limited to 'fs') diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 8664417955a2..6abdda209642 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -6,7 +6,7 @@ obj-$(CONFIG_NFS_FS) += nfs.o CFLAGS_nfstrace.o += -I$(src) nfs-y := client.o dir.o file.o getroot.o inode.o super.o \ - direct.o pagelist.o read.o symlink.o unlink.o \ + io.o direct.o pagelist.o read.o symlink.o unlink.o \ write.o namespace.o mount_clnt.o nfstrace.o nfs-$(CONFIG_ROOT_NFS) += nfsroot.o nfs-$(CONFIG_SYSCTL) += sysctl.o diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 0169eca8eb42..6d0e88096440 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -578,17 +578,12 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) if (!count) goto out; - inode_lock(inode); - result = nfs_sync_mapping(mapping); - if (result) - goto out_unlock; - task_io_account_read(count); result = -ENOMEM; dreq = nfs_direct_req_alloc(); if (dreq == NULL) - goto out_unlock; + goto out; dreq->inode = inode; dreq->bytes_left = dreq->max_count = count; @@ -603,10 +598,12 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) if (!is_sync_kiocb(iocb)) dreq->iocb = iocb; + nfs_start_io_direct(inode); + NFS_I(inode)->read_io += count; result = nfs_direct_read_schedule_iovec(dreq, iter, iocb->ki_pos); - inode_unlock(inode); + nfs_end_io_direct(inode); if (!result) { result = nfs_direct_wait(dreq); @@ -614,13 +611,8 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter) iocb->ki_pos += result; } - nfs_direct_req_release(dreq); - return result; - out_release: nfs_direct_req_release(dreq); -out_unlock: - inode_unlock(inode); out: return result; } @@ -1008,25 +1000,12 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) pos = iocb->ki_pos; end = (pos + iov_iter_count(iter) - 1) >> PAGE_SHIFT; - inode_lock(inode); - - result = nfs_sync_mapping(mapping); - if (result) - goto out_unlock; - - if (mapping->nrpages) { - result = invalidate_inode_pages2_range(mapping, - pos >> PAGE_SHIFT, end); - if (result) - goto out_unlock; - } - task_io_account_write(count); result = -ENOMEM; dreq = nfs_direct_req_alloc(); if (!dreq) - goto out_unlock; + goto out; dreq->inode = inode; dreq->bytes_left = dreq->max_count = count; @@ -1041,6 +1020,8 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) if (!is_sync_kiocb(iocb)) dreq->iocb = iocb; + nfs_start_io_direct(inode); + result = nfs_direct_write_schedule_iovec(dreq, iter, pos); if (mapping->nrpages) { @@ -1048,7 +1029,7 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) pos >> PAGE_SHIFT, end); } - inode_unlock(inode); + nfs_end_io_direct(inode); if (!result) { result = nfs_direct_wait(dreq); @@ -1058,13 +1039,9 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, struct iov_iter *iter) generic_write_sync(iocb, result); } } - nfs_direct_req_release(dreq); - return result; - out_release: nfs_direct_req_release(dreq); -out_unlock: - inode_unlock(inode); +out: return result; } diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 46cf0afe3c0f..9f8da9e1b23f 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -170,12 +170,14 @@ nfs_file_read(struct kiocb *iocb, struct iov_iter *to) iocb->ki_filp, iov_iter_count(to), (unsigned long) iocb->ki_pos); - result = nfs_revalidate_mapping_protected(inode, iocb->ki_filp->f_mapping); + nfs_start_io_read(inode); + result = nfs_revalidate_mapping(inode, iocb->ki_filp->f_mapping); if (!result) { result = generic_file_read_iter(iocb, to); if (result > 0) nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, result); } + nfs_end_io_read(inode); return result; } EXPORT_SYMBOL_GPL(nfs_file_read); @@ -191,12 +193,14 @@ nfs_file_splice_read(struct file *filp, loff_t *ppos, dprintk("NFS: splice_read(%pD2, %lu@%Lu)\n", filp, (unsigned long) count, (unsigned long long) *ppos); - res = nfs_revalidate_mapping_protected(inode, filp->f_mapping); + nfs_start_io_read(inode); + res = nfs_revalidate_mapping(inode, filp->f_mapping); if (!res) { res = generic_file_splice_read(filp, ppos, pipe, count, flags); if (res > 0) nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, res); } + nfs_end_io_read(inode); return res; } EXPORT_SYMBOL_GPL(nfs_file_splice_read); @@ -645,14 +649,14 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from) goto out; } - inode_lock(inode); + nfs_start_io_write(inode); result = generic_write_checks(iocb, from); if (result > 0) { current->backing_dev_info = inode_to_bdi(inode); result = generic_perform_write(file, from, iocb->ki_pos); current->backing_dev_info = NULL; } - inode_unlock(inode); + nfs_end_io_write(inode); if (result <= 0) goto out; diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 0eb5c924886d..159b64ede82a 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -411,6 +411,14 @@ extern void __exit unregister_nfs_fs(void); extern bool nfs_sb_active(struct super_block *sb); extern void nfs_sb_deactive(struct super_block *sb); +/* io.c */ +extern void nfs_start_io_read(struct inode *inode); +extern void nfs_end_io_read(struct inode *inode); +extern void nfs_start_io_write(struct inode *inode); +extern void nfs_end_io_write(struct inode *inode); +extern void nfs_start_io_direct(struct inode *inode); +extern void nfs_end_io_direct(struct inode *inode); + /* namespace.c */ #define NFS_PATH_CANONICAL 1 extern char *nfs_path(char **p, struct dentry *dentry, diff --git a/fs/nfs/io.c b/fs/nfs/io.c new file mode 100644 index 000000000000..1fc5d1ce327e --- /dev/null +++ b/fs/nfs/io.c @@ -0,0 +1,147 @@ +/* + * Copyright (c) 2016 Trond Myklebust + * + * I/O and data path helper functionality. + */ + +#include +#include +#include +#include +#include +#include + +#include "internal.h" + +/* Call with exclusively locked inode->i_rwsem */ +static void nfs_block_o_direct(struct nfs_inode *nfsi, struct inode *inode) +{ + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags)) { + clear_bit(NFS_INO_ODIRECT, &nfsi->flags); + inode_dio_wait(inode); + } +} + +/** + * nfs_start_io_read - declare the file is being used for buffered reads + * @inode - file inode + * + * Declare that a buffered read operation is about to start, and ensure + * that we block all direct I/O. + * On exit, the function ensures that the NFS_INO_ODIRECT flag is unset, + * and holds a shared lock on inode->i_rwsem to ensure that the flag + * cannot be changed. + * In practice, this means that buffered read operations are allowed to + * execute in parallel, thanks to the shared lock, whereas direct I/O + * operations need to wait to grab an exclusive lock in order to set + * NFS_INO_ODIRECT. + * Note that buffered writes and truncates both take a write lock on + * inode->i_rwsem, meaning that those are serialised w.r.t. the reads. + */ +void +nfs_start_io_read(struct inode *inode) +{ + struct nfs_inode *nfsi = NFS_I(inode); + /* Be an optimist! */ + down_read(&inode->i_rwsem); + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags) == 0) + return; + up_read(&inode->i_rwsem); + /* Slow path.... */ + down_write(&inode->i_rwsem); + nfs_block_o_direct(nfsi, inode); + downgrade_write(&inode->i_rwsem); +} + +/** + * nfs_end_io_read - declare that the buffered read operation is done + * @inode - file inode + * + * Declare that a buffered read operation is done, and release the shared + * lock on inode->i_rwsem. + */ +void +nfs_end_io_read(struct inode *inode) +{ + up_read(&inode->i_rwsem); +} + +/** + * nfs_start_io_write - declare the file is being used for buffered writes + * @inode - file inode + * + * Declare that a buffered read operation is about to start, and ensure + * that we block all direct I/O. + */ +void +nfs_start_io_write(struct inode *inode) +{ + down_write(&inode->i_rwsem); + nfs_block_o_direct(NFS_I(inode), inode); +} + +/** + * nfs_end_io_write - declare that the buffered write operation is done + * @inode - file inode + * + * Declare that a buffered write operation is done, and release the + * lock on inode->i_rwsem. + */ +void +nfs_end_io_write(struct inode *inode) +{ + up_write(&inode->i_rwsem); +} + +/* Call with exclusively locked inode->i_rwsem */ +static void nfs_block_buffered(struct nfs_inode *nfsi, struct inode *inode) +{ + if (!test_bit(NFS_INO_ODIRECT, &nfsi->flags)) { + set_bit(NFS_INO_ODIRECT, &nfsi->flags); + nfs_wb_all(inode); + } +} + +/** + * nfs_end_io_direct - declare the file is being used for direct i/o + * @inode - file inode + * + * Declare that a direct I/O operation is about to start, and ensure + * that we block all buffered I/O. + * On exit, the function ensures that the NFS_INO_ODIRECT flag is set, + * and holds a shared lock on inode->i_rwsem to ensure that the flag + * cannot be changed. + * In practice, this means that direct I/O operations are allowed to + * execute in parallel, thanks to the shared lock, whereas buffered I/O + * operations need to wait to grab an exclusive lock in order to clear + * NFS_INO_ODIRECT. + * Note that buffered writes and truncates both take a write lock on + * inode->i_rwsem, meaning that those are serialised w.r.t. O_DIRECT. + */ +void +nfs_start_io_direct(struct inode *inode) +{ + struct nfs_inode *nfsi = NFS_I(inode); + /* Be an optimist! */ + down_read(&inode->i_rwsem); + if (test_bit(NFS_INO_ODIRECT, &nfsi->flags) != 0) + return; + up_read(&inode->i_rwsem); + /* Slow path.... */ + down_write(&inode->i_rwsem); + nfs_block_buffered(nfsi, inode); + downgrade_write(&inode->i_rwsem); +} + +/** + * nfs_end_io_direct - declare that the direct i/o operation is done + * @inode - file inode + * + * Declare that a direct I/O operation is done, and release the shared + * lock on inode->i_rwsem. + */ +void +nfs_end_io_direct(struct inode *inode) +{ + up_read(&inode->i_rwsem); +} diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 120dd04b553c..225d17d35277 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -210,6 +210,7 @@ struct nfs_inode { #define NFS_INO_LAYOUTCOMMIT (9) /* layoutcommit required */ #define NFS_INO_LAYOUTCOMMITTING (10) /* layoutcommit inflight */ #define NFS_INO_LAYOUTSTATS (11) /* layoutstats inflight */ +#define NFS_INO_ODIRECT (12) /* I/O setting is O_DIRECT */ static inline struct nfs_inode *NFS_I(const struct inode *inode) { -- cgit v1.2.3 From f7b5c340aca87d736a6b15aa40bf135f1baab011 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Jun 2016 09:29:47 -0400 Subject: NFS: Cleanup nfs_direct_complete() There is only one caller that sets the "write" argument to true, so just move the call to nfs_zap_mapping() and get rid of the now redundant argument. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 6d0e88096440..c16d33eb1ddf 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -372,13 +372,10 @@ out: * Synchronous I/O uses a stack-allocated iocb. Thus we can't trust * the iocb is still valid here if this is a synchronous request. */ -static void nfs_direct_complete(struct nfs_direct_req *dreq, bool write) +static void nfs_direct_complete(struct nfs_direct_req *dreq) { struct inode *inode = dreq->inode; - if (write) - nfs_zap_mapping(inode, inode->i_mapping); - inode_dio_end(inode); if (dreq->iocb) { @@ -431,7 +428,7 @@ static void nfs_direct_read_completion(struct nfs_pgio_header *hdr) } out_put: if (put_dreq(dreq)) - nfs_direct_complete(dreq, false); + nfs_direct_complete(dreq); hdr->release(hdr); } @@ -537,7 +534,7 @@ static ssize_t nfs_direct_read_schedule_iovec(struct nfs_direct_req *dreq, } if (put_dreq(dreq)) - nfs_direct_complete(dreq, false); + nfs_direct_complete(dreq); return 0; } @@ -764,7 +761,8 @@ static void nfs_direct_write_schedule_work(struct work_struct *work) nfs_direct_write_reschedule(dreq); break; default: - nfs_direct_complete(dreq, true); + nfs_zap_mapping(dreq->inode, dreq->inode->i_mapping); + nfs_direct_complete(dreq); } } -- cgit v1.2.3 From f508d46ae41a796036aef566637685dbf83b554f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Jun 2016 09:55:48 -0400 Subject: NFS: Remove redundant waits for O_DIRECT in fsync() and write_begin() We're now waiting immediately after taking the locks, so waiting in fsync() and write_begin() is either redundant or potentially subject to livelock (if not holding the lock). Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 6 ------ 1 file changed, 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 9f8da9e1b23f..0e9b4a068f13 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -276,7 +276,6 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync) trace_nfs_fsync_enter(inode); - inode_dio_wait(inode); do { ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret != 0) @@ -361,11 +360,6 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping, file, mapping->host->i_ino, len, (long long) pos); start: - /* - * Wait for O_DIRECT to complete - */ - inode_dio_wait(mapping->host); - page = grab_cache_page_write_begin(mapping, index, flags); if (!page) return -ENOMEM; -- cgit v1.2.3 From be527494e02b89e03485955b30de6c1e976a07eb Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 22 Jun 2016 08:19:36 -0400 Subject: NFS: Remove unused function nfs_revalidate_mapping_protected() Clean up... Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 38 ++++---------------------------------- include/linux/nfs_fs.h | 1 - 2 files changed, 4 insertions(+), 35 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 6c0618eb5d57..0e0500f2bb6b 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1131,14 +1131,12 @@ out: } /** - * __nfs_revalidate_mapping - Revalidate the pagecache + * nfs_revalidate_mapping - Revalidate the pagecache * @inode - pointer to host inode * @mapping - pointer to mapping - * @may_lock - take inode->i_mutex? */ -static int __nfs_revalidate_mapping(struct inode *inode, - struct address_space *mapping, - bool may_lock) +int nfs_revalidate_mapping(struct inode *inode, + struct address_space *mapping) { struct nfs_inode *nfsi = NFS_I(inode); unsigned long *bitlock = &nfsi->flags; @@ -1187,12 +1185,7 @@ static int __nfs_revalidate_mapping(struct inode *inode, nfsi->cache_validity &= ~NFS_INO_INVALID_DATA; spin_unlock(&inode->i_lock); trace_nfs_invalidate_mapping_enter(inode); - if (may_lock) { - inode_lock(inode); - ret = nfs_invalidate_mapping(inode, mapping); - inode_unlock(inode); - } else - ret = nfs_invalidate_mapping(inode, mapping); + ret = nfs_invalidate_mapping(inode, mapping); trace_nfs_invalidate_mapping_exit(inode, ret); clear_bit_unlock(NFS_INO_INVALIDATING, bitlock); @@ -1202,29 +1195,6 @@ out: return ret; } -/** - * nfs_revalidate_mapping - Revalidate the pagecache - * @inode - pointer to host inode - * @mapping - pointer to mapping - */ -int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping) -{ - return __nfs_revalidate_mapping(inode, mapping, false); -} - -/** - * nfs_revalidate_mapping_protected - Revalidate the pagecache - * @inode - pointer to host inode - * @mapping - pointer to mapping - * - * Differs from nfs_revalidate_mapping() in that it grabs the inode->i_mutex - * while invalidating the mapping. - */ -int nfs_revalidate_mapping_protected(struct inode *inode, struct address_space *mapping) -{ - return __nfs_revalidate_mapping(inode, mapping, true); -} - static bool nfs_file_has_writers(struct nfs_inode *nfsi) { struct inode *inode = &nfsi->vfs_inode; diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 225d17d35277..810124b33327 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -351,7 +351,6 @@ extern int nfs_revalidate_inode_rcu(struct nfs_server *server, struct inode *ino extern int __nfs_revalidate_inode(struct nfs_server *, struct inode *); extern int nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping); extern int nfs_revalidate_mapping_rcu(struct inode *inode); -extern int nfs_revalidate_mapping_protected(struct inode *inode, struct address_space *mapping); extern int nfs_setattr(struct dentry *, struct iattr *); extern void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr, struct nfs_fattr *); extern void nfs_setsecurity(struct inode *inode, struct nfs_fattr *fattr, -- cgit v1.2.3 From 651b0e702981304f77091b82870a01480705f4fe Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 25 Jun 2016 17:24:46 -0400 Subject: NFS: Do not aggressively cache file attributes in the case of O_DIRECT A file that is open for O_DIRECT is by definition not obeying close-to-open cache consistency semantics, so let's not cache the attributes too aggressively either. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 9 +++++++-- fs/nfs/internal.h | 5 +++++ 2 files changed, 12 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 0e0500f2bb6b..7688436b19ba 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1214,6 +1214,11 @@ static bool nfs_file_has_writers(struct nfs_inode *nfsi) list)->mode & FMODE_WRITE) == FMODE_WRITE; } +static bool nfs_file_has_buffered_writers(struct nfs_inode *nfsi) +{ + return nfs_file_has_writers(nfsi) && nfs_file_io_is_buffered(nfsi); +} + static unsigned long nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr) { struct nfs_inode *nfsi = NFS_I(inode); @@ -1278,7 +1283,7 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat if ((fattr->valid & NFS_ATTR_FATTR_TYPE) && (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) return -EIO; - if (!nfs_file_has_writers(nfsi)) { + if (!nfs_file_has_buffered_writers(nfsi)) { /* Verify a few of the more important attributes */ if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && inode->i_version != fattr->change_attr) invalid |= NFS_INO_INVALID_ATTR | NFS_INO_REVAL_PAGECACHE; @@ -1660,7 +1665,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long invalid = 0; unsigned long now = jiffies; unsigned long save_cache_validity; - bool have_writers = nfs_file_has_writers(nfsi); + bool have_writers = nfs_file_has_buffered_writers(nfsi); bool cache_revalidated = true; dfprintk(VFS, "NFS: %s(%s/%lu fh_crc=0x%08x ct=%d info=0x%x)\n", diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 159b64ede82a..01dccf18da0a 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -419,6 +419,11 @@ extern void nfs_end_io_write(struct inode *inode); extern void nfs_start_io_direct(struct inode *inode); extern void nfs_end_io_direct(struct inode *inode); +static inline bool nfs_file_io_is_buffered(struct nfs_inode *nfsi) +{ + return test_bit(NFS_INO_ODIRECT, &nfsi->flags) == 0; +} + /* namespace.c */ #define NFS_PATH_CANONICAL 1 extern char *nfs_path(char **p, struct dentry *dentry, -- cgit v1.2.3 From 79566ef018f53a181f067afdf7bef9cc53f9d34b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 25 Jun 2016 17:45:40 -0400 Subject: NFS: Getattr doesn't require data sync semantics When retrieving stat() information, NFS unfortunately does require us to sync writes to disk in order to ensure that mtime and ctime are up to date. However we shouldn't have to ensure that those writes are persisted. Relaxing that requirement does mean that we may see an mtime/ctime change if the server reboots and forces us to replay all writes. The exception to this rule are pNFS clients that are required to send layoutcommit, however that is dealt with by the call to pnfs_sync_inode() in _nfs_revalidate_inode(). Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 7688436b19ba..35fda08dc4f6 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -661,9 +661,7 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) trace_nfs_getattr_enter(inode); /* Flush out writes to the server in order to update c/mtime. */ if (S_ISREG(inode->i_mode)) { - inode_lock(inode); - err = nfs_sync_inode(inode); - inode_unlock(inode); + err = filemap_write_and_wait(inode->i_mapping); if (err) goto out; } -- cgit v1.2.3 From 1e564d3dbd684a105582471cb9ff2aada64a9052 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 25 Jun 2016 17:50:53 -0400 Subject: NFSv4.2: Fix a race in nfs42_proc_deallocate() When punching holes in a file, we want to ensure the operation is serialised w.r.t. other writes, meaning that we want to call nfs_sync_inode() while holding the inode lock. Signed-off-by: Trond Myklebust --- fs/nfs/nfs42proc.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index aa03ed09ba06..0f9f536e647b 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -113,15 +113,17 @@ int nfs42_proc_deallocate(struct file *filep, loff_t offset, loff_t len) if (!nfs_server_capable(inode, NFS_CAP_DEALLOCATE)) return -EOPNOTSUPP; - nfs_wb_all(inode); inode_lock(inode); + err = nfs_sync_inode(inode); + if (err) + goto out_unlock; err = nfs42_proc_fallocate(&msg, filep, offset, len); if (err == 0) truncate_pagecache_range(inode, offset, (offset + len) -1); if (err == -EOPNOTSUPP) NFS_SERVER(inode)->caps &= ~NFS_CAP_DEALLOCATE; - +out_unlock: inode_unlock(inode); return err; } -- cgit v1.2.3 From 837bb1d752d92ea4d870877ffbd6ec5cf76624b3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 25 Jun 2016 18:12:03 -0400 Subject: NFSv4.2: Fix writeback races in nfs4_copy_file_range We need to ensure that any writes to the destination file are serialised with the copy, meaning that the writeback has to occur under the inode lock. Also relax the writeback requirement on the source, and rely on the stateid checking to tell us if the source rebooted. Add the helper nfs_filemap_write_and_wait_range() to call pnfs_sync_inode() as is appropriate for pNFS servers that may need a layoutcommit. Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 3 +++ fs/nfs/nfs42proc.c | 9 +++++++++ fs/nfs/nfs4file.c | 14 +------------- fs/nfs/write.c | 18 ++++++++++++++++++ 4 files changed, 31 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 01dccf18da0a..3b01c9146e15 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -512,6 +512,9 @@ int nfs_key_timeout_notify(struct file *filp, struct inode *inode); bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx); void nfs_pageio_stop_mirroring(struct nfs_pageio_descriptor *pgio); +int nfs_filemap_write_and_wait_range(struct address_space *mapping, + loff_t lstart, loff_t lend); + #ifdef CONFIG_NFS_V4_1 static inline void nfs_clear_pnfs_ds_commit_verifiers(struct pnfs_ds_commit_info *cinfo) diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index 0f9f536e647b..b7d457cea03f 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -156,11 +156,20 @@ static ssize_t _nfs42_proc_copy(struct file *src, loff_t pos_src, if (status) return status; + status = nfs_filemap_write_and_wait_range(file_inode(src)->i_mapping, + pos_src, pos_src + (loff_t)count - 1); + if (status) + return status; + status = nfs4_set_rw_stateid(&args.dst_stateid, dst_lock->open_context, dst_lock, FMODE_WRITE); if (status) return status; + status = nfs_sync_inode(dst_inode); + if (status) + return status; + status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0); if (status == -ENOTSUPP) diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index 014b0e41ace5..7cdc0ab9e6f5 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -133,21 +133,9 @@ static ssize_t nfs4_copy_file_range(struct file *file_in, loff_t pos_in, struct file *file_out, loff_t pos_out, size_t count, unsigned int flags) { - struct inode *in_inode = file_inode(file_in); - struct inode *out_inode = file_inode(file_out); - int ret; - - if (in_inode == out_inode) + if (file_inode(file_in) == file_inode(file_out)) return -EINVAL; - /* flush any pending writes */ - ret = nfs_sync_inode(in_inode); - if (ret) - return ret; - ret = nfs_sync_inode(out_inode); - if (ret) - return ret; - return nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count); } diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 3087fb6f1983..538a473b324b 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1912,6 +1912,24 @@ out_mark_dirty: } EXPORT_SYMBOL_GPL(nfs_write_inode); +/* + * Wrapper for filemap_write_and_wait_range() + * + * Needed for pNFS in order to ensure data becomes visible to the + * client. + */ +int nfs_filemap_write_and_wait_range(struct address_space *mapping, + loff_t lstart, loff_t lend) +{ + int ret; + + ret = filemap_write_and_wait_range(mapping, lstart, lend); + if (ret == 0) + ret = pnfs_sync_inode(mapping->host, true); + return ret; +} +EXPORT_SYMBOL_GPL(nfs_filemap_write_and_wait_range); + /* * flush the inode to disk. */ -- cgit v1.2.3 From e95fc4a06983c14273a39d26aad9cc5a8a09ff64 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 25 Jun 2016 17:57:39 -0400 Subject: NFSv4.2: llseek(SEEK_HOLE) and llseek(SEEK_DATA) don't require data sync We want to ensure that we write the cached data to the server, but don't require it be synced to disk. If the server reboots, we will get a stateid error, which will cause us to retry anyway. Signed-off-by: Trond Myklebust --- fs/nfs/nfs42proc.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index b7d457cea03f..616dc254b38b 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -269,7 +269,11 @@ static loff_t _nfs42_proc_llseek(struct file *filep, if (status) return status; - nfs_wb_all(inode); + status = nfs_filemap_write_and_wait_range(inode->i_mapping, + offset, LLONG_MAX); + if (status) + return status; + status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0); if (status == -ENOTSUPP) -- cgit v1.2.3 From 9a773e7c8de2a34ae682624624e95a96b121b6d1 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 23 Jun 2016 11:09:04 -0400 Subject: NFS nfs_vm_page_mkwrite: Don't freeze me, Bro... Prevent filesystem freezes while handling the write page fault. Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 0e9b4a068f13..039d58790629 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -569,6 +569,8 @@ static int nfs_vm_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) filp, filp->f_mapping->host->i_ino, (long long)page_offset(page)); + sb_start_pagefault(inode->i_sb); + /* make sure the cache has finished storing the page */ nfs_fscache_wait_on_page_write(NFS_I(inode), page); @@ -595,6 +597,7 @@ static int nfs_vm_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) out_unlock: unlock_page(page); out: + sb_end_pagefault(inode->i_sb); return ret; } -- cgit v1.2.3 From e40edf2da4fc6ab10c707b0522372b08b38171e6 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:47 -0400 Subject: Btrfs: add bytes_readonly to the spaceinfo at once For some reason we're adding bytes_readonly to the space info after we update the space info with the block group info. This creates a tiny race where we could over-reserve space because we haven't yet taken out the bytes_readonly bit. Since we already know this information at the time we call update_space_info, just pass it along so it can be updated all at once. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 29 +++++++++++------------------ 1 file changed, 11 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 82b912a293ab..777d9a5aa83f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3913,6 +3913,7 @@ static const char *alloc_name(u64 flags) static int update_space_info(struct btrfs_fs_info *info, u64 flags, u64 total_bytes, u64 bytes_used, + u64 bytes_readonly, struct btrfs_space_info **space_info) { struct btrfs_space_info *found; @@ -3933,6 +3934,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->disk_total += total_bytes * factor; found->bytes_used += bytes_used; found->disk_used += bytes_used * factor; + found->bytes_readonly += bytes_readonly; if (total_bytes > 0) found->full = 0; spin_unlock(&found->lock); @@ -3960,7 +3962,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->disk_used = bytes_used * factor; found->bytes_pinned = 0; found->bytes_reserved = 0; - found->bytes_readonly = 0; + found->bytes_readonly = bytes_readonly; found->bytes_may_use = 0; found->full = 0; found->max_extent_size = 0; @@ -4470,7 +4472,7 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, space_info = __find_space_info(extent_root->fs_info, flags); if (!space_info) { ret = update_space_info(extent_root->fs_info, flags, - 0, 0, &space_info); + 0, 0, 0, &space_info); BUG_ON(ret); /* -ENOMEM */ } BUG_ON(!space_info); /* Logic error */ @@ -10007,7 +10009,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) ret = update_space_info(info, cache->flags, found_key.offset, btrfs_block_group_used(&cache->item), - &space_info); + cache->bytes_super, &space_info); if (ret) { btrfs_remove_free_space_cache(cache); spin_lock(&info->block_group_cache_lock); @@ -10020,9 +10022,6 @@ int btrfs_read_block_groups(struct btrfs_root *root) } cache->space_info = space_info; - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_readonly += cache->bytes_super; - spin_unlock(&cache->space_info->lock); __link_block_group(space_info, cache); @@ -10114,7 +10113,6 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, int ret; struct btrfs_root *extent_root; struct btrfs_block_group_cache *cache; - extent_root = root->fs_info->extent_root; btrfs_set_log_full_commit(root->fs_info, trans); @@ -10160,7 +10158,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * assigned to our block group, but don't update its counters just yet. * We want our bg to be added to the rbtree with its ->space_info set. */ - ret = update_space_info(root->fs_info, cache->flags, 0, 0, + ret = update_space_info(root->fs_info, cache->flags, 0, 0, 0, &cache->space_info); if (ret) { btrfs_remove_free_space_cache(cache); @@ -10180,7 +10178,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * the rbtree, update the space info's counters. */ ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, - &cache->space_info); + cache->bytes_super, &cache->space_info); if (ret) { btrfs_remove_free_space_cache(cache); spin_lock(&root->fs_info->block_group_cache_lock); @@ -10193,16 +10191,11 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, } update_global_block_rsv(root->fs_info); - spin_lock(&cache->space_info->lock); - cache->space_info->bytes_readonly += cache->bytes_super; - spin_unlock(&cache->space_info->lock); - __link_block_group(cache->space_info, cache); list_add_tail(&cache->bg_list, &trans->new_bgs); set_avail_alloc_bits(extent_root->fs_info, type); - return 0; } @@ -10747,21 +10740,21 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) mixed = 1; flags = BTRFS_BLOCK_GROUP_SYSTEM; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); if (ret) goto out; if (mixed) { flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); } else { flags = BTRFS_BLOCK_GROUP_METADATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); if (ret) goto out; flags = BTRFS_BLOCK_GROUP_DATA; - ret = update_space_info(fs_info, flags, 0, 0, &space_info); + ret = update_space_info(fs_info, flags, 0, 0, 0, &space_info); } out: return ret; -- cgit v1.2.3 From 25d609f86d6808eb1f0e8a6cafc3edb4a2b5ae35 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:48 -0400 Subject: Btrfs: fix callers of btrfs_block_rsv_migrate So btrfs_block_rsv_migrate just unconditionally calls block_rsv_migrate_bytes. Not only this but it unconditionally changes the size of the block_rsv. This isn't a bug strictly speaking, but it makes truncate block rsv's look funny because every time we migrate bytes over its size grows, even though we only want it to be a specific size. So collapse this into one function that takes an update_size argument and make truncate and evict not update the size for consistency sake. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 4 ++-- fs/btrfs/delayed-inode.c | 8 ++++---- fs/btrfs/extent-tree.c | 18 ++++++------------ fs/btrfs/file.c | 4 ++-- fs/btrfs/inode.c | 7 +++---- fs/btrfs/relocation.c | 2 +- 6 files changed, 18 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 4274a7bfdaed..fe16474fabf3 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2661,8 +2661,8 @@ int btrfs_block_rsv_refill(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 min_reserved, enum btrfs_reserve_flush_enum flush); int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, - struct btrfs_block_rsv *dst_rsv, - u64 num_bytes); + struct btrfs_block_rsv *dst_rsv, u64 num_bytes, + int update_size); int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, struct btrfs_block_rsv *dest, u64 num_bytes, int min_factor); diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index d3aaabbfada0..f749a5447b2a 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -553,7 +553,7 @@ static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, dst_rsv = &root->fs_info->delayed_block_rsv; num_bytes = btrfs_calc_trans_metadata_size(root, 1); - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); if (!ret) { trace_btrfs_space_reservation(root->fs_info, "delayed_item", item->key.objectid, @@ -649,7 +649,7 @@ static int btrfs_delayed_inode_reserve_metadata( if (!ret) goto out; - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); if (!ret) goto out; @@ -663,12 +663,12 @@ static int btrfs_delayed_inode_reserve_metadata( * since this really shouldn't happen that often. */ ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, - dst_rsv, num_bytes); + dst_rsv, num_bytes, 1); goto out; } migrate: - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); out: /* diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 777d9a5aa83f..1ceabb7a4479 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5265,8 +5265,9 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, } } -static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, - struct btrfs_block_rsv *dst, u64 num_bytes) +int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src, + struct btrfs_block_rsv *dst, u64 num_bytes, + int update_size) { int ret; @@ -5274,7 +5275,7 @@ static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, if (ret) return ret; - block_rsv_add_bytes(dst, num_bytes, 1); + block_rsv_add_bytes(dst, num_bytes, update_size); return 0; } @@ -5381,13 +5382,6 @@ int btrfs_block_rsv_refill(struct btrfs_root *root, return ret; } -int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, - struct btrfs_block_rsv *dst_rsv, - u64 num_bytes) -{ - return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); -} - void btrfs_block_rsv_release(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes) @@ -5567,7 +5561,7 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); trace_btrfs_space_reservation(root->fs_info, "orphan", btrfs_ino(inode), num_bytes, 1); - return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); + return btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); } void btrfs_orphan_release_metadata(struct inode *inode) @@ -5622,7 +5616,7 @@ int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, BTRFS_RESERVE_FLUSH_ALL); if (ret == -ENOSPC && use_global_rsv) - ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes); + ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes, 1); if (ret && *qgroup_reserved) btrfs_qgroup_free_meta(root, *qgroup_reserved); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 2234e88cf674..f3f61d1ad18a 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -2479,7 +2479,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, - min_size); + min_size, 0); BUG_ON(ret); trans->block_rsv = rsv; @@ -2522,7 +2522,7 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, - rsv, min_size); + rsv, min_size, 0); BUG_ON(ret); /* shouldn't happen */ trans->block_rsv = rsv; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4421954720b8..3e61bd1f3f65 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5264,7 +5264,7 @@ void btrfs_evict_inode(struct inode *inode) if (steal_from_global) { if (!btrfs_check_space_for_delayed_refs(trans, root)) ret = btrfs_block_rsv_migrate(global_rsv, rsv, - min_size); + min_size, 0); else ret = -ENOSPC; } @@ -9116,7 +9116,7 @@ static int btrfs_truncate(struct inode *inode) /* Migrate the slack space for the truncate to our reserve */ ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, - min_size); + min_size, 0); BUG_ON(ret); /* @@ -9156,7 +9156,7 @@ static int btrfs_truncate(struct inode *inode) } ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, - rsv, min_size); + rsv, min_size, 0); BUG_ON(ret); /* shouldn't happen */ trans->block_rsv = rsv; } @@ -9177,7 +9177,6 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root); } - out: btrfs_free_block_rsv(root, rsv); diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 0477dca154ed..0bf3bc879503 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -4643,7 +4643,7 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, if (rc->merge_reloc_tree) { ret = btrfs_block_rsv_migrate(&pending->block_rsv, rc->block_rsv, - rc->nodes_relocated); + rc->nodes_relocated, 1); if (ret) return ret; } -- cgit v1.2.3 From 48c3d480e4f746edb0bd3b84b4bd987ef8728560 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:49 -0400 Subject: Btrfs: always reserve metadata for delalloc extents There are a few races in the metadata reservation stuff. First we add the bytes to the block_rsv well after we've set the bit on the inode saying that we have space for it and after we've reserved the bytes. So use the normal btrfs_block_rsv_add helper for this case. Secondly we can flush delalloc extents when we try to reserve space for our write, which means that we could have used up the space for the inode and we wouldn't know because we only check before the reservation. So instead make sure we are always reserving space for the inode update, and then if we don't need it release those bytes afterward. Thanks, Signed-off-by: Josef Bacik Reviewed-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 35 +++++++++++++---------------------- 1 file changed, 13 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 1ceabb7a4479..ac80fcce25f7 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5726,12 +5726,12 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) u64 to_reserve = 0; u64 csum_bytes; unsigned nr_extents = 0; - int extra_reserve = 0; enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_FLUSH_ALL; int ret = 0; bool delalloc_lock = true; u64 to_free = 0; unsigned dropped; + bool release_extra = false; /* If we are a free space inode we need to not flush since we will be in * the middle of a transaction commit. We also don't need the delalloc @@ -5757,24 +5757,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) BTRFS_MAX_EXTENT_SIZE - 1, BTRFS_MAX_EXTENT_SIZE); BTRFS_I(inode)->outstanding_extents += nr_extents; - nr_extents = 0; + nr_extents = 0; if (BTRFS_I(inode)->outstanding_extents > BTRFS_I(inode)->reserved_extents) - nr_extents = BTRFS_I(inode)->outstanding_extents - + nr_extents += BTRFS_I(inode)->outstanding_extents - BTRFS_I(inode)->reserved_extents; - /* - * Add an item to reserve for updating the inode when we complete the - * delalloc io. - */ - if (!test_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) { - nr_extents++; - extra_reserve = 1; - } - - to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); + /* We always want to reserve a slot for updating the inode. */ + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents + 1); to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); csum_bytes = BTRFS_I(inode)->csum_bytes; spin_unlock(&BTRFS_I(inode)->lock); @@ -5786,18 +5777,16 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) goto out_fail; } - ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); + ret = btrfs_block_rsv_add(root, block_rsv, to_reserve, flush); if (unlikely(ret)) { btrfs_qgroup_free_meta(root, nr_extents * root->nodesize); goto out_fail; } spin_lock(&BTRFS_I(inode)->lock); - if (extra_reserve) { - set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags); - nr_extents--; - } + if (test_and_set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) + release_extra = true; BTRFS_I(inode)->reserved_extents += nr_extents; spin_unlock(&BTRFS_I(inode)->lock); @@ -5807,8 +5796,10 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) if (to_reserve) trace_btrfs_space_reservation(root->fs_info, "delalloc", btrfs_ino(inode), to_reserve, 1); - block_rsv_add_bytes(block_rsv, to_reserve, 1); - + if (release_extra) + btrfs_block_rsv_release(root, block_rsv, + btrfs_calc_trans_metadata_size(root, + 1)); return 0; out_fail: -- cgit v1.2.3 From c48f49d63dfea12c0173fb2e99db1308f39c23ae Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:50 -0400 Subject: Btrfs: change delayed reservation fallback behavior We reserve space for the inode update when we first reserve space for writing to a file. However there are lots of ways that we can use this reservation and not have it for subsequent ordered extents. Previously we'd fall through and try to reserve metadata bytes for this, then we'd just steal the full reservation from the delalloc_block_rsv, and if that didn't have enough space we'd steal the full reservation from the global reserve. The problem with this is we can easily just return ENOSPC and fallback to updating the inode item directly. In the worst case (assuming 4k nodesize) we'd steal 64kib from the global reserve if we fall all the way through, however if we just fallback and update the inode directly we'd only steal 4k * BTRFS_PATH_MAX in the worst case which is 32kib. We would have also just added the extent item for the inode so we likely will have already cow'ed down most of the way to the leaf containing the inode item, so we are more often than not only need one or two nodesize's worth of reservations. Given the reservation for the extent itself is also a worst case we will likely already have space to cover the inode update. This change will make us behave better in the theoretical worst case, and much better in the case that we don't have our reservation and cannot reserve more metadata. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/delayed-inode.c | 64 +++++++++++++++++------------------------------- 1 file changed, 23 insertions(+), 41 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index f749a5447b2a..dd3c040139a2 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -597,6 +597,29 @@ static int btrfs_delayed_inode_reserve_metadata( num_bytes = btrfs_calc_trans_metadata_size(root, 1); + /* + * If our block_rsv is the delalloc block reserve then check and see if + * we have our extra reservation for updating the inode. If not fall + * through and try to reserve space quickly. + * + * We used to try and steal from the delalloc block rsv or the global + * reserve, but we'd steal a full reservation, which isn't kind. We are + * here through delalloc which means we've likely just cowed down close + * to the leaf that contains the inode, so we would steal less just + * doing the fallback inode update, so if we do end up having to steal + * from the global block rsv we hopefully only steal one or two blocks + * worth which is less likely to hurt us. + */ + if (src_rsv && src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { + spin_lock(&BTRFS_I(inode)->lock); + if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, + &BTRFS_I(inode)->runtime_flags)) + release = true; + else + src_rsv = NULL; + spin_unlock(&BTRFS_I(inode)->lock); + } + /* * btrfs_dirty_inode will update the inode under btrfs_join_transaction * which doesn't reserve space for speed. This is a problem since we @@ -626,51 +649,10 @@ static int btrfs_delayed_inode_reserve_metadata( num_bytes, 1); } return ret; - } else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { - spin_lock(&BTRFS_I(inode)->lock); - if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) { - spin_unlock(&BTRFS_I(inode)->lock); - release = true; - goto migrate; - } - spin_unlock(&BTRFS_I(inode)->lock); - - /* Ok we didn't have space pre-reserved. This shouldn't happen - * too often but it can happen if we do delalloc to an existing - * inode which gets dirtied because of the time update, and then - * isn't touched again until after the transaction commits and - * then we try to write out the data. First try to be nice and - * reserve something strictly for us. If not be a pain and try - * to steal from the delalloc block rsv. - */ - ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, - BTRFS_RESERVE_NO_FLUSH); - if (!ret) - goto out; - - ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); - if (!ret) - goto out; - - if (btrfs_test_opt(root, ENOSPC_DEBUG)) { - btrfs_debug(root->fs_info, - "block rsv migrate returned %d", ret); - WARN_ON(1); - } - /* - * Ok this is a problem, let's just steal from the global rsv - * since this really shouldn't happen that often. - */ - ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, - dst_rsv, num_bytes, 1); - goto out; } -migrate: ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes, 1); -out: /* * Migrate only takes a reservation, it doesn't touch the size of the * block_rsv. This is to simplify people who don't normally have things -- cgit v1.2.3 From d555b6c380c644af63dbdaa7cc14bba041a4e4dd Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:51 -0400 Subject: Btrfs: warn_on for unaccounted spaces These were hidden behind enospc_debug, which isn't helpful as they indicate actual bugs, unlike the rest of the enospc_debug stuff which is really debug information. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index ac80fcce25f7..a1355a33e7e9 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -9778,13 +9778,15 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info) space_info = list_entry(info->space_info.next, struct btrfs_space_info, list); - if (btrfs_test_opt(info->tree_root, ENOSPC_DEBUG)) { - if (WARN_ON(space_info->bytes_pinned > 0 || + + /* + * Do not hide this behind enospc_debug, this is actually + * important and indicates a real bug if this happens. + */ + if (WARN_ON(space_info->bytes_pinned > 0 || space_info->bytes_reserved > 0 || - space_info->bytes_may_use > 0)) { - dump_space_info(space_info, 0, 0); - } - } + space_info->bytes_may_use > 0)) + dump_space_info(space_info, 0, 0); list_del(&space_info->list); for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) { struct kobject *kobj; -- cgit v1.2.3 From c83f8effefa46c15f2fd43de598d9839d0056096 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:52 -0400 Subject: Btrfs: add tracepoint for adding block groups I'm writing a tool to visualize the enospc system inside btrfs, I need this tracepoint in order to keep track of the block groups in the system. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 2 ++ include/trace/events/btrfs.h | 40 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 42 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a1355a33e7e9..aae7b04afa9f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -9994,6 +9994,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) goto error; } + trace_btrfs_add_block_group(root->fs_info, cache, 0); ret = update_space_info(info, cache->flags, found_key.offset, btrfs_block_group_used(&cache->item), cache->bytes_super, &space_info); @@ -10164,6 +10165,7 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, * Now that our block group has its ->space_info set and is inserted in * the rbtree, update the space info's counters. */ + trace_btrfs_add_block_group(root->fs_info, cache, 1); ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, cache->bytes_super, &cache->space_info); if (ret) { diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h index e90e82ad6875..985e01b6c849 100644 --- a/include/trace/events/btrfs.h +++ b/include/trace/events/btrfs.h @@ -440,6 +440,46 @@ TRACE_EVENT(btrfs_sync_fs, TP_printk("wait = %d", __entry->wait) ); +TRACE_EVENT(btrfs_add_block_group, + + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_block_group_cache *block_group, int create), + + TP_ARGS(fs_info, block_group, create), + + TP_STRUCT__entry( + __array( u8, fsid, BTRFS_UUID_SIZE ) + __field( u64, offset ) + __field( u64, size ) + __field( u64, flags ) + __field( u64, bytes_used ) + __field( u64, bytes_super ) + __field( int, create ) + ), + + TP_fast_assign( + memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE); + __entry->offset = block_group->key.objectid; + __entry->size = block_group->key.offset; + __entry->flags = block_group->flags; + __entry->bytes_used = + btrfs_block_group_used(&block_group->item); + __entry->bytes_super = block_group->bytes_super; + __entry->create = create; + ), + + TP_printk("%pU: block_group offset = %llu, size = %llu, " + "flags = %llu(%s), bytes_used = %llu, bytes_super = %llu, " + "create = %d", __entry->fsid, + (unsigned long long)__entry->offset, + (unsigned long long)__entry->size, + (unsigned long long)__entry->flags, + __print_flags((unsigned long)__entry->flags, "|", + BTRFS_GROUP_FLAGS), + (unsigned long long)__entry->bytes_used, + (unsigned long long)__entry->bytes_super, __entry->create) +); + #define show_ref_action(action) \ __print_symbolic(action, \ { BTRFS_ADD_DELAYED_REF, "ADD_DELAYED_REF" }, \ -- cgit v1.2.3 From 957780eb2788d8c218d539e19a85653f51a96dc1 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 17 May 2016 13:30:55 -0400 Subject: Btrfs: introduce ticketed enospc infrastructure Our enospc flushing sucks. It is born from a time where we were early enospc'ing constantly because multiple threads would race in for the same reservation and randomly starve other ones out. So I came up with this solution to block any other reservations from happening while one guy tried to flush stuff to satisfy his reservation. This gives us pretty good correctness, but completely crap latency. The solution I've come up with is ticketed reservations. Basically we try to make our reservation, and if we can't we put a ticket on a list in order and kick off an async flusher thread. This async flusher thread does the same old flushing we always did, just asynchronously. As space is freed and added back to the space_info it checks and sees if we have any tickets that need satisfying, and adds space to the tickets and wakes up anything we've satisfied. Once the flusher thread stops making progress it wakes up all the current tickets and tells them to take a hike. There is a priority list for things that can't flush, since the async flusher could do anything we need to avoid deadlocks. These guys get priority for having their reservation made, and will still do manual flushing themselves in case the async flusher isn't running. This patch gives us significantly better latencies. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 2 + fs/btrfs/extent-tree.c | 529 +++++++++++++++++++++++++++++++++++-------------- 2 files changed, 380 insertions(+), 151 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index fe16474fabf3..2e04c9d6f21d 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -439,6 +439,8 @@ struct btrfs_space_info { struct list_head list; /* Protected by the spinlock 'lock'. */ struct list_head ro_bgs; + struct list_head priority_tickets; + struct list_head tickets; struct rw_semaphore groups_sem; /* for block groups in our same type */ diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index aae7b04afa9f..2c17b621a661 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -111,6 +111,16 @@ static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, u64 num_bytes); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num_bytes, int reserved); +static int __reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush); +static void space_info_add_new_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes); +static void space_info_add_old_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes); static noinline int block_group_cache_done(struct btrfs_block_group_cache *cache) @@ -3937,6 +3947,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->bytes_readonly += bytes_readonly; if (total_bytes > 0) found->full = 0; + space_info_add_new_bytes(info, found, total_bytes - + bytes_used - bytes_readonly); spin_unlock(&found->lock); *space_info = found; return 0; @@ -3971,6 +3983,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags, found->flush = 0; init_waitqueue_head(&found->wait); INIT_LIST_HEAD(&found->ro_bgs); + INIT_LIST_HEAD(&found->tickets); + INIT_LIST_HEAD(&found->priority_tickets); ret = kobject_init_and_add(&found->kobj, &space_info_ktype, info->space_info_kobj, "%s", @@ -4584,12 +4598,19 @@ static int can_overcommit(struct btrfs_root *root, struct btrfs_space_info *space_info, u64 bytes, enum btrfs_reserve_flush_enum flush) { - struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; - u64 profile = btrfs_get_alloc_profile(root, 0); + struct btrfs_block_rsv *global_rsv; + u64 profile; u64 space_size; u64 avail; u64 used; + /* Don't overcommit when in mixed mode. */ + if (space_info->flags & BTRFS_BLOCK_GROUP_DATA) + return 0; + + BUG_ON(root->fs_info == NULL); + global_rsv = &root->fs_info->global_block_rsv; + profile = btrfs_get_alloc_profile(root, 0); used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly; @@ -4741,6 +4762,11 @@ skip_async: spin_unlock(&space_info->lock); break; } + if (list_empty(&space_info->tickets) && + list_empty(&space_info->priority_tickets)) { + spin_unlock(&space_info->lock); + break; + } spin_unlock(&space_info->lock); loops++; @@ -4818,6 +4844,13 @@ enum flush_state { COMMIT_TRANS = 6, }; +struct reserve_ticket { + u64 bytes; + int error; + struct list_head list; + wait_queue_head_t wait; +}; + static int flush_space(struct btrfs_root *root, struct btrfs_space_info *space_info, u64 num_bytes, u64 orig_bytes, int state) @@ -4875,17 +4908,22 @@ static inline u64 btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, struct btrfs_space_info *space_info) { + struct reserve_ticket *ticket; u64 used; u64 expected; - u64 to_reclaim; + u64 to_reclaim = 0; to_reclaim = min_t(u64, num_online_cpus() * SZ_1M, SZ_16M); - spin_lock(&space_info->lock); if (can_overcommit(root, space_info, to_reclaim, - BTRFS_RESERVE_FLUSH_ALL)) { - to_reclaim = 0; - goto out; - } + BTRFS_RESERVE_FLUSH_ALL)) + return 0; + + list_for_each_entry(ticket, &space_info->tickets, list) + to_reclaim += ticket->bytes; + list_for_each_entry(ticket, &space_info->priority_tickets, list) + to_reclaim += ticket->bytes; + if (to_reclaim) + return to_reclaim; used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly + @@ -4901,9 +4939,6 @@ btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, to_reclaim = 0; to_reclaim = min(to_reclaim, space_info->bytes_may_use + space_info->bytes_reserved); -out: - spin_unlock(&space_info->lock); - return to_reclaim; } @@ -4920,69 +4955,169 @@ static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); } -static int btrfs_need_do_async_reclaim(struct btrfs_space_info *space_info, - struct btrfs_fs_info *fs_info, - int flush_state) +static void wake_all_tickets(struct list_head *head) { - u64 used; - - spin_lock(&space_info->lock); - /* - * We run out of space and have not got any free space via flush_space, - * so don't bother doing async reclaim. - */ - if (flush_state > COMMIT_TRANS && space_info->full) { - spin_unlock(&space_info->lock); - return 0; - } + struct reserve_ticket *ticket; - used = space_info->bytes_used + space_info->bytes_reserved + - space_info->bytes_pinned + space_info->bytes_readonly + - space_info->bytes_may_use; - if (need_do_async_reclaim(space_info, fs_info, used)) { - spin_unlock(&space_info->lock); - return 1; + while (!list_empty(head)) { + ticket = list_first_entry(head, struct reserve_ticket, list); + list_del_init(&ticket->list); + ticket->error = -ENOSPC; + wake_up(&ticket->wait); } - spin_unlock(&space_info->lock); - - return 0; } +/* + * This is for normal flushers, we can wait all goddamned day if we want to. We + * will loop and continuously try to flush as long as we are making progress. + * We count progress as clearing off tickets each time we have to loop. + */ static void btrfs_async_reclaim_metadata_space(struct work_struct *work) { + struct reserve_ticket *last_ticket = NULL; struct btrfs_fs_info *fs_info; struct btrfs_space_info *space_info; u64 to_reclaim; int flush_state; + int commit_cycles = 0; fs_info = container_of(work, struct btrfs_fs_info, async_reclaim_work); space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); + spin_lock(&space_info->lock); to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, space_info); - if (!to_reclaim) + if (!to_reclaim) { + space_info->flush = 0; + spin_unlock(&space_info->lock); return; + } + last_ticket = list_first_entry(&space_info->tickets, + struct reserve_ticket, list); + spin_unlock(&space_info->lock); flush_state = FLUSH_DELAYED_ITEMS_NR; + do { + struct reserve_ticket *ticket; + int ret; + + ret = flush_space(fs_info->fs_root, space_info, to_reclaim, + to_reclaim, flush_state); + spin_lock(&space_info->lock); + if (list_empty(&space_info->tickets)) { + space_info->flush = 0; + spin_unlock(&space_info->lock); + return; + } + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, + space_info); + ticket = list_first_entry(&space_info->tickets, + struct reserve_ticket, list); + if (last_ticket == ticket) { + flush_state++; + } else { + last_ticket = ticket; + flush_state = FLUSH_DELAYED_ITEMS_NR; + if (commit_cycles) + commit_cycles--; + } + + if (flush_state > COMMIT_TRANS) { + commit_cycles++; + if (commit_cycles > 2) { + wake_all_tickets(&space_info->tickets); + space_info->flush = 0; + } else { + flush_state = FLUSH_DELAYED_ITEMS_NR; + } + } + spin_unlock(&space_info->lock); + } while (flush_state <= COMMIT_TRANS); +} + +void btrfs_init_async_reclaim_work(struct work_struct *work) +{ + INIT_WORK(work, btrfs_async_reclaim_metadata_space); +} + +static void priority_reclaim_metadata_space(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + struct reserve_ticket *ticket) +{ + u64 to_reclaim; + int flush_state = FLUSH_DELAYED_ITEMS_NR; + + spin_lock(&space_info->lock); + to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, + space_info); + if (!to_reclaim) { + spin_unlock(&space_info->lock); + return; + } + spin_unlock(&space_info->lock); + do { flush_space(fs_info->fs_root, space_info, to_reclaim, to_reclaim, flush_state); flush_state++; - if (!btrfs_need_do_async_reclaim(space_info, fs_info, - flush_state)) + spin_lock(&space_info->lock); + if (ticket->bytes == 0) { + spin_unlock(&space_info->lock); return; + } + spin_unlock(&space_info->lock); + + /* + * Priority flushers can't wait on delalloc without + * deadlocking. + */ + if (flush_state == FLUSH_DELALLOC || + flush_state == FLUSH_DELALLOC_WAIT) + flush_state = ALLOC_CHUNK; } while (flush_state < COMMIT_TRANS); } -void btrfs_init_async_reclaim_work(struct work_struct *work) +static int wait_reserve_ticket(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + struct reserve_ticket *ticket, u64 orig_bytes) + { - INIT_WORK(work, btrfs_async_reclaim_metadata_space); + DEFINE_WAIT(wait); + int ret = 0; + + spin_lock(&space_info->lock); + while (ticket->bytes > 0 && ticket->error == 0) { + ret = prepare_to_wait_event(&ticket->wait, &wait, TASK_KILLABLE); + if (ret) { + ret = -EINTR; + break; + } + spin_unlock(&space_info->lock); + + schedule(); + + finish_wait(&ticket->wait, &wait); + spin_lock(&space_info->lock); + } + if (!ret) + ret = ticket->error; + if (!list_empty(&ticket->list)) + list_del_init(&ticket->list); + if (ticket->bytes && ticket->bytes < orig_bytes) { + u64 num_bytes = orig_bytes - ticket->bytes; + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, num_bytes, 0); + } + spin_unlock(&space_info->lock); + + return ret; } /** * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space * @root - the root we're allocating for - * @block_rsv - the block_rsv we're allocating for + * @space_info - the space info we want to allocate from * @orig_bytes - the number of bytes we want * @flush - whether or not we can flush to make our reservation * @@ -4993,81 +5128,34 @@ void btrfs_init_async_reclaim_work(struct work_struct *work) * regain reservations will be made and this will fail if there is not enough * space already. */ -static int reserve_metadata_bytes(struct btrfs_root *root, - struct btrfs_block_rsv *block_rsv, - u64 orig_bytes, - enum btrfs_reserve_flush_enum flush) +static int __reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_space_info *space_info, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush) { - struct btrfs_space_info *space_info = block_rsv->space_info; + struct reserve_ticket ticket; u64 used; - u64 num_bytes = orig_bytes; - int flush_state = FLUSH_DELAYED_ITEMS_NR; int ret = 0; - bool flushing = false; -again: - ret = 0; + ASSERT(orig_bytes); spin_lock(&space_info->lock); - /* - * We only want to wait if somebody other than us is flushing and we - * are actually allowed to flush all things. - */ - while (flush == BTRFS_RESERVE_FLUSH_ALL && !flushing && - space_info->flush) { - spin_unlock(&space_info->lock); - /* - * If we have a trans handle we can't wait because the flusher - * may have to commit the transaction, which would mean we would - * deadlock since we are waiting for the flusher to finish, but - * hold the current transaction open. - */ - if (current->journal_info) - return -EAGAIN; - ret = wait_event_killable(space_info->wait, !space_info->flush); - /* Must have been killed, return */ - if (ret) - return -EINTR; - - spin_lock(&space_info->lock); - } - ret = -ENOSPC; used = space_info->bytes_used + space_info->bytes_reserved + space_info->bytes_pinned + space_info->bytes_readonly + space_info->bytes_may_use; /* - * The idea here is that we've not already over-reserved the block group - * then we can go ahead and save our reservation first and then start - * flushing if we need to. Otherwise if we've already overcommitted - * lets start flushing stuff first and then come back and try to make - * our reservation. + * If we have enough space then hooray, make our reservation and carry + * on. If not see if we can overcommit, and if we can, hooray carry on. + * If not things get more complicated. */ - if (used <= space_info->total_bytes) { - if (used + orig_bytes <= space_info->total_bytes) { - space_info->bytes_may_use += orig_bytes; - trace_btrfs_space_reservation(root->fs_info, - "space_info", space_info->flags, orig_bytes, 1); - ret = 0; - } else { - /* - * Ok set num_bytes to orig_bytes since we aren't - * overocmmitted, this way we only try and reclaim what - * we need. - */ - num_bytes = orig_bytes; - } - } else { - /* - * Ok we're over committed, set num_bytes to the overcommitted - * amount plus the amount of bytes that we need for this - * reservation. - */ - num_bytes = used - space_info->total_bytes + - (orig_bytes * 2); - } - - if (ret && can_overcommit(root, space_info, orig_bytes, flush)) { + if (used + orig_bytes <= space_info->total_bytes) { + space_info->bytes_may_use += orig_bytes; + trace_btrfs_space_reservation(root->fs_info, "space_info", + space_info->flags, orig_bytes, + 1); + ret = 0; + } else if (can_overcommit(root, space_info, orig_bytes, flush)) { space_info->bytes_may_use += orig_bytes; trace_btrfs_space_reservation(root->fs_info, "space_info", space_info->flags, orig_bytes, @@ -5076,16 +5164,27 @@ again: } /* - * Couldn't make our reservation, save our place so while we're trying - * to reclaim space we can actually use it instead of somebody else - * stealing it from us. + * If we couldn't make a reservation then setup our reservation ticket + * and kick the async worker if it's not already running. * - * We make the other tasks wait for the flush only when we can flush - * all things. + * If we are a priority flusher then we just need to add our ticket to + * the list and we will do our own flushing further down. */ if (ret && flush != BTRFS_RESERVE_NO_FLUSH) { - flushing = true; - space_info->flush = 1; + ticket.bytes = orig_bytes; + ticket.error = 0; + init_waitqueue_head(&ticket.wait); + if (flush == BTRFS_RESERVE_FLUSH_ALL) { + list_add_tail(&ticket.list, &space_info->tickets); + if (!space_info->flush) { + space_info->flush = 1; + queue_work(system_unbound_wq, + &root->fs_info->async_reclaim_work); + } + } else { + list_add_tail(&ticket.list, + &space_info->priority_tickets); + } } else if (!ret && space_info->flags & BTRFS_BLOCK_GROUP_METADATA) { used += orig_bytes; /* @@ -5100,33 +5199,56 @@ again: &root->fs_info->async_reclaim_work); } spin_unlock(&space_info->lock); - if (!ret || flush == BTRFS_RESERVE_NO_FLUSH) - goto out; + return ret; - ret = flush_space(root, space_info, num_bytes, orig_bytes, - flush_state); - flush_state++; + if (flush == BTRFS_RESERVE_FLUSH_ALL) + return wait_reserve_ticket(root->fs_info, space_info, &ticket, + orig_bytes); - /* - * If we are FLUSH_LIMIT, we can not flush delalloc, or the deadlock - * would happen. So skip delalloc flush. - */ - if (flush == BTRFS_RESERVE_FLUSH_LIMIT && - (flush_state == FLUSH_DELALLOC || - flush_state == FLUSH_DELALLOC_WAIT)) - flush_state = ALLOC_CHUNK; + ret = 0; + priority_reclaim_metadata_space(root->fs_info, space_info, &ticket); + spin_lock(&space_info->lock); + if (ticket.bytes) { + if (ticket.bytes < orig_bytes) { + u64 num_bytes = orig_bytes - ticket.bytes; + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(root->fs_info, + "space_info", space_info->flags, + num_bytes, 0); - if (!ret) - goto again; - else if (flush == BTRFS_RESERVE_FLUSH_LIMIT && - flush_state < COMMIT_TRANS) - goto again; - else if (flush == BTRFS_RESERVE_FLUSH_ALL && - flush_state <= COMMIT_TRANS) - goto again; + } + list_del_init(&ticket.list); + ret = -ENOSPC; + } + spin_unlock(&space_info->lock); + ASSERT(list_empty(&ticket.list)); + return ret; +} -out: +/** + * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space + * @root - the root we're allocating for + * @block_rsv - the block_rsv we're allocating for + * @orig_bytes - the number of bytes we want + * @flush - whether or not we can flush to make our reservation + * + * This will reserve orgi_bytes number of bytes from the space info associated + * with the block_rsv. If there is not enough space it will make an attempt to + * flush out space to make room. It will do this by flushing delalloc if + * possible or committing the transaction. If flush is 0 then no attempts to + * regain reservations will be made and this will fail if there is not enough + * space already. + */ +static int reserve_metadata_bytes(struct btrfs_root *root, + struct btrfs_block_rsv *block_rsv, + u64 orig_bytes, + enum btrfs_reserve_flush_enum flush) +{ + int ret; + + ret = __reserve_metadata_bytes(root, block_rsv->space_info, orig_bytes, + flush); if (ret == -ENOSPC && unlikely(root->orphan_cleanup_state == ORPHAN_CLEANUP_STARTED)) { struct btrfs_block_rsv *global_rsv = @@ -5139,13 +5261,8 @@ out: if (ret == -ENOSPC) trace_btrfs_space_reservation(root->fs_info, "space_info:enospc", - space_info->flags, orig_bytes, 1); - if (flushing) { - spin_lock(&space_info->lock); - space_info->flush = 0; - wake_up_all(&space_info->wait); - spin_unlock(&space_info->lock); - } + block_rsv->space_info->flags, + orig_bytes, 1); return ret; } @@ -5221,6 +5338,108 @@ int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, return 0; } +/* + * This is for space we already have accounted in space_info->bytes_may_use, so + * basically when we're returning space from block_rsv's. + */ +static void space_info_add_old_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes) +{ + struct reserve_ticket *ticket; + struct list_head *head; + u64 used; + enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_NO_FLUSH; + bool check_overcommit = false; + + spin_lock(&space_info->lock); + head = &space_info->priority_tickets; + + /* + * If we are over our limit then we need to check and see if we can + * overcommit, and if we can't then we just need to free up our space + * and not satisfy any requests. + */ + used = space_info->bytes_used + space_info->bytes_reserved + + space_info->bytes_pinned + space_info->bytes_readonly + + space_info->bytes_may_use; + if (used - num_bytes >= space_info->total_bytes) + check_overcommit = true; +again: + while (!list_empty(head) && num_bytes) { + ticket = list_first_entry(head, struct reserve_ticket, + list); + /* + * We use 0 bytes because this space is already reserved, so + * adding the ticket space would be a double count. + */ + if (check_overcommit && + !can_overcommit(fs_info->extent_root, space_info, 0, + flush)) + break; + if (num_bytes >= ticket->bytes) { + list_del_init(&ticket->list); + num_bytes -= ticket->bytes; + ticket->bytes = 0; + wake_up(&ticket->wait); + } else { + ticket->bytes -= num_bytes; + num_bytes = 0; + } + } + + if (num_bytes && head == &space_info->priority_tickets) { + head = &space_info->tickets; + flush = BTRFS_RESERVE_FLUSH_ALL; + goto again; + } + space_info->bytes_may_use -= num_bytes; + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, num_bytes, 0); + spin_unlock(&space_info->lock); +} + +/* + * This is for newly allocated space that isn't accounted in + * space_info->bytes_may_use yet. So if we allocate a chunk or unpin an extent + * we use this helper. + */ +static void space_info_add_new_bytes(struct btrfs_fs_info *fs_info, + struct btrfs_space_info *space_info, + u64 num_bytes) +{ + struct reserve_ticket *ticket; + struct list_head *head = &space_info->priority_tickets; + +again: + while (!list_empty(head) && num_bytes) { + ticket = list_first_entry(head, struct reserve_ticket, + list); + if (num_bytes >= ticket->bytes) { + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, + ticket->bytes, 1); + list_del_init(&ticket->list); + num_bytes -= ticket->bytes; + space_info->bytes_may_use += ticket->bytes; + ticket->bytes = 0; + wake_up(&ticket->wait); + } else { + trace_btrfs_space_reservation(fs_info, "space_info", + space_info->flags, + num_bytes, 1); + space_info->bytes_may_use += num_bytes; + ticket->bytes -= num_bytes; + num_bytes = 0; + } + } + + if (num_bytes && head == &space_info->priority_tickets) { + head = &space_info->tickets; + goto again; + } +} + static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, struct btrfs_block_rsv *block_rsv, struct btrfs_block_rsv *dest, u64 num_bytes) @@ -5255,13 +5474,9 @@ static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, } spin_unlock(&dest->lock); } - if (num_bytes) { - spin_lock(&space_info->lock); - space_info->bytes_may_use -= num_bytes; - trace_btrfs_space_reservation(fs_info, "space_info", - space_info->flags, num_bytes, 0); - spin_unlock(&space_info->lock); - } + if (num_bytes) + space_info_add_old_bytes(fs_info, space_info, + num_bytes); } } @@ -6470,17 +6685,29 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, readonly = true; } spin_unlock(&cache->lock); - if (!readonly && global_rsv->space_info == space_info) { + if (!readonly && return_free_space && + global_rsv->space_info == space_info) { + u64 to_add = len; + WARN_ON(!return_free_space); spin_lock(&global_rsv->lock); if (!global_rsv->full) { - len = min(len, global_rsv->size - - global_rsv->reserved); - global_rsv->reserved += len; - space_info->bytes_may_use += len; + to_add = min(len, global_rsv->size - + global_rsv->reserved); + global_rsv->reserved += to_add; + space_info->bytes_may_use += to_add; if (global_rsv->reserved >= global_rsv->size) global_rsv->full = 1; + trace_btrfs_space_reservation(fs_info, + "space_info", + space_info->flags, + to_add, 1); + len -= to_add; } spin_unlock(&global_rsv->lock); + /* Add to any tickets we may have */ + if (len) + space_info_add_new_bytes(fs_info, space_info, + len); } spin_unlock(&space_info->lock); } -- cgit v1.2.3 From c51e7bb1845079fda515a7acf0514e289aa6a07c Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:54 -0400 Subject: Btrfs: trace pinned extents Pinned extents are an important metric to keep track of for enospc. Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 2c17b621a661..9aa64f1aa413 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6246,6 +6246,9 @@ static int update_block_group(struct btrfs_trans_handle *trans, spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); + trace_btrfs_space_reservation(root->fs_info, "pinned", + cache->space_info->flags, + num_bytes, 1); set_extent_dirty(info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); @@ -6320,6 +6323,8 @@ static int pin_down_extent(struct btrfs_root *root, spin_unlock(&cache->lock); spin_unlock(&cache->space_info->lock); + trace_btrfs_space_reservation(root->fs_info, "pinned", + cache->space_info->flags, num_bytes, 1); set_extent_dirty(root->fs_info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); if (reserved) @@ -6678,6 +6683,9 @@ static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, spin_lock(&cache->lock); cache->pinned -= len; space_info->bytes_pinned -= len; + + trace_btrfs_space_reservation(fs_info, "pinned", + space_info->flags, len, 0); space_info->max_extent_size = 0; percpu_counter_add(&space_info->total_bytes_pinned, -len); if (cache->ro) { -- cgit v1.2.3 From f485c9ee32531794e39631437fcedc82d5b8e650 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:55 -0400 Subject: Btrfs: fix delalloc reservation amount tracepoint We can sometimes drop the reservation we had for our inode, so we need to remove that amount from to_reserve so that our tracepoint reports a valid amount of space. Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 9aa64f1aa413..5d24ec44d99b 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6000,8 +6000,10 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) spin_lock(&BTRFS_I(inode)->lock); if (test_and_set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, - &BTRFS_I(inode)->runtime_flags)) + &BTRFS_I(inode)->runtime_flags)) { + to_reserve -= btrfs_calc_trans_metadata_size(root, 1); release_extra = true; + } BTRFS_I(inode)->reserved_extents += nr_extents; spin_unlock(&BTRFS_I(inode)->lock); -- cgit v1.2.3 From f376df2b7da3a40f62f861a65efdd8c29fa1b877 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:56 -0400 Subject: Btrfs: add tracepoints for flush events We want to track when we're triggering flushing from our reservation code and what flushing is being done when we start flushing. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 9 +++++ fs/btrfs/extent-tree.c | 22 ++++++------ include/trace/events/btrfs.h | 82 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 103 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 2e04c9d6f21d..83a6a931af09 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2626,6 +2626,15 @@ enum btrfs_reserve_flush_enum { BTRFS_RESERVE_FLUSH_ALL, }; +enum btrfs_flush_state { + FLUSH_DELAYED_ITEMS_NR = 1, + FLUSH_DELAYED_ITEMS = 2, + FLUSH_DELALLOC = 3, + FLUSH_DELALLOC_WAIT = 4, + ALLOC_CHUNK = 5, + COMMIT_TRANS = 6, +}; + int btrfs_check_data_free_space(struct inode *inode, u64 start, u64 len); int btrfs_alloc_data_chunk_ondemand(struct inode *inode, u64 bytes); void btrfs_free_reserved_data_space(struct inode *inode, u64 start, u64 len); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5d24ec44d99b..31ded6aae1f0 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4835,15 +4835,6 @@ commit: return btrfs_commit_transaction(trans, root); } -enum flush_state { - FLUSH_DELAYED_ITEMS_NR = 1, - FLUSH_DELAYED_ITEMS = 2, - FLUSH_DELALLOC = 3, - FLUSH_DELALLOC_WAIT = 4, - ALLOC_CHUNK = 5, - COMMIT_TRANS = 6, -}; - struct reserve_ticket { u64 bytes; int error; @@ -4901,6 +4892,8 @@ static int flush_space(struct btrfs_root *root, break; } + trace_btrfs_flush_space(root->fs_info, space_info->flags, num_bytes, + orig_bytes, state, ret); return ret; } @@ -5178,6 +5171,10 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, list_add_tail(&ticket.list, &space_info->tickets); if (!space_info->flush) { space_info->flush = 1; + trace_btrfs_trigger_flush(root->fs_info, + space_info->flags, + orig_bytes, flush, + "enospc"); queue_work(system_unbound_wq, &root->fs_info->async_reclaim_work); } @@ -5194,9 +5191,14 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, */ if (!root->fs_info->log_root_recovering && need_do_async_reclaim(space_info, root->fs_info, used) && - !work_busy(&root->fs_info->async_reclaim_work)) + !work_busy(&root->fs_info->async_reclaim_work)) { + trace_btrfs_trigger_flush(root->fs_info, + space_info->flags, + orig_bytes, flush, + "preempt"); queue_work(system_unbound_wq, &root->fs_info->async_reclaim_work); + } } spin_unlock(&space_info->lock); if (!ret || flush == BTRFS_RESERVE_NO_FLUSH) diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h index 985e01b6c849..226c7f283247 100644 --- a/include/trace/events/btrfs.h +++ b/include/trace/events/btrfs.h @@ -784,6 +784,88 @@ TRACE_EVENT(btrfs_space_reservation, __entry->bytes) ); +#define show_flush_action(action) \ + __print_symbolic(action, \ + { BTRFS_RESERVE_NO_FLUSH, "BTRFS_RESERVE_NO_FLUSH"}, \ + { BTRFS_RESERVE_FLUSH_LIMIT, "BTRFS_RESERVE_FLUSH_LIMIT"}, \ + { BTRFS_RESERVE_FLUSH_ALL, "BTRFS_RESERVE_FLUSH_ALL"}) + +TRACE_EVENT(btrfs_trigger_flush, + + TP_PROTO(struct btrfs_fs_info *fs_info, u64 flags, u64 bytes, + int flush, char *reason), + + TP_ARGS(fs_info, flags, bytes, flush, reason), + + TP_STRUCT__entry( + __array( u8, fsid, BTRFS_UUID_SIZE ) + __field( u64, flags ) + __field( u64, bytes ) + __field( int, flush ) + __string( reason, reason ) + ), + + TP_fast_assign( + memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE); + __entry->flags = flags; + __entry->bytes = bytes; + __entry->flush = flush; + __assign_str(reason, reason) + ), + + TP_printk("%pU: %s: flush = %d(%s), flags = %llu(%s), bytes = %llu", + __entry->fsid, __get_str(reason), __entry->flush, + show_flush_action(__entry->flush), + (unsigned long long)__entry->flags, + __print_flags((unsigned long)__entry->flags, "|", + BTRFS_GROUP_FLAGS), + (unsigned long long)__entry->bytes) +); + +#define show_flush_state(state) \ + __print_symbolic(state, \ + { FLUSH_DELAYED_ITEMS_NR, "FLUSH_DELAYED_ITEMS_NR"}, \ + { FLUSH_DELAYED_ITEMS, "FLUSH_DELAYED_ITEMS"}, \ + { FLUSH_DELALLOC, "FLUSH_DELALLOC"}, \ + { FLUSH_DELALLOC_WAIT, "FLUSH_DELALLOC_WAIT"}, \ + { ALLOC_CHUNK, "ALLOC_CHUNK"}, \ + { COMMIT_TRANS, "COMMIT_TRANS"}) + +TRACE_EVENT(btrfs_flush_space, + + TP_PROTO(struct btrfs_fs_info *fs_info, u64 flags, u64 num_bytes, + u64 orig_bytes, int state, int ret), + + TP_ARGS(fs_info, flags, num_bytes, orig_bytes, state, ret), + + TP_STRUCT__entry( + __array( u8, fsid, BTRFS_UUID_SIZE ) + __field( u64, flags ) + __field( u64, num_bytes ) + __field( u64, orig_bytes ) + __field( int, state ) + __field( int, ret ) + ), + + TP_fast_assign( + memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE); + __entry->flags = flags; + __entry->num_bytes = num_bytes; + __entry->orig_bytes = orig_bytes; + __entry->state = state; + __entry->ret = ret; + ), + + TP_printk("%pU: state = %d(%s), flags = %llu(%s), num_bytes = %llu, " + "orig_bytes = %llu, ret = %d", __entry->fsid, __entry->state, + show_flush_state(__entry->state), + (unsigned long long)__entry->flags, + __print_flags((unsigned long)__entry->flags, "|", + BTRFS_GROUP_FLAGS), + (unsigned long long)__entry->num_bytes, + (unsigned long long)__entry->orig_bytes, __entry->ret) +); + DECLARE_EVENT_CLASS(btrfs__reserved_extent, TP_PROTO(struct btrfs_root *root, u64 start, u64 len), -- cgit v1.2.3 From 31bada7c4e5e80b9cc031e89247175bed58ed9e9 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:58 -0400 Subject: Btrfs: fix release reserved extents trace points We were doing trace_btrfs_release_reserved_extent() in pin_down_extent which isn't quite right because we will go through and free that extent later when we unpin, so it messes up apps that are accounting for the reservation space. We were also unconditionally doing it in __btrfs_free_reserved_extent(), when we only actually free the reservation instead of pinning the extent. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 31ded6aae1f0..a012fd49f214 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6331,8 +6331,6 @@ static int pin_down_extent(struct btrfs_root *root, cache->space_info->flags, num_bytes, 1); set_extent_dirty(root->fs_info->pinned_extents, bytenr, bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); - if (reserved) - trace_btrfs_reserved_extent_free(root, bytenr, num_bytes); return 0; } @@ -8008,12 +8006,10 @@ static int __btrfs_free_reserved_extent(struct btrfs_root *root, ret = btrfs_discard_extent(root, start, len, NULL); btrfs_add_free_space(cache, start, len); btrfs_update_reserved_bytes(cache, len, RESERVE_FREE, delalloc); + trace_btrfs_reserved_extent_free(root, start, len); } btrfs_put_block_group(cache); - - trace_btrfs_reserved_extent_free(root, start, len); - return ret; } -- cgit v1.2.3 From d38b349c39a9feb1aca82c3280b9b1234497d582 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 25 Mar 2016 13:25:59 -0400 Subject: Btrfs: don't bother kicking async if there's nothing to reclaim We do this check when we start the async reclaimer thread, might as well check before we kick it off to save us some cycles. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a012fd49f214..a0ca2b5914d0 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4944,6 +4944,9 @@ static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, if ((space_info->bytes_used + space_info->bytes_reserved) >= thresh) return 0; + if (!btrfs_calc_reclaim_metadata_size(fs_info->fs_root, space_info)) + return 0; + return (used >= thresh && !btrfs_fs_closing(fs_info) && !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); } -- cgit v1.2.3 From 87241c2e68451c4d50b60af84f9a3ab119001b4c Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 25 Apr 2016 09:58:18 -0400 Subject: Btrfs: use root when checking need_async_flush Instead of doing fs_info->fs_root in need_async_flush, which may not be set during recovery when mounting, just pass the root itself in, which makes more sense as thats what btrfs_calc_reclaim_metadata_size takes. Signed-off-by: Josef Bacik Reported-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index a0ca2b5914d0..9ef8f99d5a2a 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4936,7 +4936,7 @@ btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, } static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, - struct btrfs_fs_info *fs_info, u64 used) + struct btrfs_root *root, u64 used) { u64 thresh = div_factor_fine(space_info->total_bytes, 98); @@ -4944,11 +4944,12 @@ static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, if ((space_info->bytes_used + space_info->bytes_reserved) >= thresh) return 0; - if (!btrfs_calc_reclaim_metadata_size(fs_info->fs_root, space_info)) + if (!btrfs_calc_reclaim_metadata_size(root, space_info)) return 0; - return (used >= thresh && !btrfs_fs_closing(fs_info) && - !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); + return (used >= thresh && !btrfs_fs_closing(root->fs_info) && + !test_bit(BTRFS_FS_STATE_REMOUNTING, + &root->fs_info->fs_state)); } static void wake_all_tickets(struct list_head *head) @@ -5193,7 +5194,7 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, * the async reclaim as we will panic. */ if (!root->fs_info->log_root_recovering && - need_do_async_reclaim(space_info, root->fs_info, used) && + need_do_async_reclaim(space_info, root, used) && !work_busy(&root->fs_info->async_reclaim_work)) { trace_btrfs_trigger_flush(root->fs_info, space_info->flags, -- cgit v1.2.3 From ae2e4728816510ae4e8ed2b732b34ea0aba05a34 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 12:58:35 -0400 Subject: Btrfs: change how we calculate the global block rsv Traditionally we've calculated the global block rsv by guessing how much of the metadata used amount was the extent tree, and then taking the data size and figuring out how large the csum tree would have to be to hold that much data. This is imprecise and falls down on MIXED file systems as we can't trust the data used amount. This resulted in failures for xfstests generic/333 because it creates lots of clones, which explodes out the extent tree. Our global reserve calculations were woefully inaccurate in this case which meant we got into a situation where we did not have enough reserved to do our work. We know we only use the global block rsv for the extent, csum, and root trees, so just get the bytes used for these trees and use that as the basis of our global reserve. Since these are not reference counted trees the bytes_used value will be accurate. This fixed the transaction aborts seen with generic/333. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 45 +++++++++------------------------------------ 1 file changed, 9 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 9ef8f99d5a2a..7c644f810816 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5615,48 +5615,21 @@ void btrfs_block_rsv_release(struct btrfs_root *root, num_bytes); } -/* - * helper to calculate size of global block reservation. - * the desired value is sum of space used by extent tree, - * checksum tree and root tree - */ -static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) -{ - struct btrfs_space_info *sinfo; - u64 num_bytes; - u64 meta_used; - u64 data_used; - int csum_size = btrfs_super_csum_size(fs_info->super_copy); - - sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); - spin_lock(&sinfo->lock); - data_used = sinfo->bytes_used; - spin_unlock(&sinfo->lock); - - sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); - spin_lock(&sinfo->lock); - if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA) - data_used = 0; - meta_used = sinfo->bytes_used; - spin_unlock(&sinfo->lock); - - num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) * - csum_size * 2; - num_bytes += div_u64(data_used + meta_used, 50); - - if (num_bytes * 3 > meta_used) - num_bytes = div_u64(meta_used, 3); - - return ALIGN(num_bytes, fs_info->extent_root->nodesize << 10); -} - static void update_global_block_rsv(struct btrfs_fs_info *fs_info) { struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; struct btrfs_space_info *sinfo = block_rsv->space_info; u64 num_bytes; - num_bytes = calc_global_metadata_size(fs_info); + /* + * The global block rsv is based on the size of the extent tree, the + * checksum tree and the root tree. If the fs is empty we want to set + * it to a minimal amount for safety. + */ + num_bytes = btrfs_root_used(&fs_info->extent_root->root_item) + + btrfs_root_used(&fs_info->csum_root->root_item) + + btrfs_root_used(&fs_info->tree_root->root_item); + num_bytes = max_t(u64, num_bytes, SZ_16M); spin_lock(&sinfo->lock); spin_lock(&block_rsv->lock); -- cgit v1.2.3 From 40acc3eededbe16f6104a32776e541f223d00b5e Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 13:01:08 -0400 Subject: Btrfs: always use trans->block_rsv for orphans This is the case all the time anyway except for relocation which could be doing a reloc root for a non ref counted root, in which case we'd end up with some random block rsv rather than the one we have our reservation in. If there isn't enough space in the block rsv we are trying to steal from we'll BUG() because we expect there to be space for the orphan to make its reservation. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 7c644f810816..f3de211173ad 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5744,7 +5744,13 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root); + /* + * We always use trans->block_rsv here as we will have reserved space + * for our orphan when starting the transaction, using get_block_rsv() + * here will sometimes make us choose the wrong block rsv as we could be + * doing a reloc inode for a non refcounted root. + */ + struct btrfs_block_rsv *src_rsv = trans->block_rsv; struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv; /* -- cgit v1.2.3 From ac2fabac4211431b607b326c2233e73b81e86af2 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 13:08:26 -0400 Subject: Btrfs: fill relocation block rsv after allocation Since we set the reloc control before we've reserved our space for relocation we could race with a root being dirtied and not actually have space to do our init reloc root. So once we've allocated it and set it up go ahead and make our reservation before setting the relocate control, that way anybody who tries to do the reloc root init has space to use. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/relocation.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 0bf3bc879503..23e16de52e23 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -3871,6 +3871,7 @@ static noinline_for_stack int prepare_to_relocate(struct reloc_control *rc) { struct btrfs_trans_handle *trans; + int ret; rc->block_rsv = btrfs_alloc_block_rsv(rc->extent_root, BTRFS_BLOCK_RSV_TEMP); @@ -3885,6 +3886,11 @@ int prepare_to_relocate(struct reloc_control *rc) rc->reserved_bytes = 0; rc->block_rsv->size = rc->extent_root->nodesize * RELOCATION_RESERVED_NODES; + ret = btrfs_block_rsv_refill(rc->extent_root, + rc->block_rsv, rc->block_rsv->size, + BTRFS_RESERVE_FLUSH_ALL); + if (ret) + return ret; rc->create_reloc_tree = 1; set_reloc_control(rc); -- cgit v1.2.3 From 8ca17f0f59529df9e2132ca594ff00c52a9bc556 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 27 May 2016 13:24:13 -0400 Subject: Btrfs: use FLUSH_LIMIT for relocation in reserve_metadata_bytes We used to allow you to set FLUSH_ALL and then just wouldn't do things like commit transactions or wait on ordered extents if we noticed you were in a transaction. However now that all the flushing for FLUSH_ALL is asynchronous we've lost the ability to tell, and we could end up deadlocking. So instead use FLUSH_LIMIT in reserve_metadata_bytes in relocation and then return -EAGAIN if we error out to preserve the previous behavior. I've also added an ASSERT() to catch anybody else who tries to do this. Thanks, Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 2 ++ fs/btrfs/relocation.c | 37 ++++++++++++++++++++----------------- 2 files changed, 22 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index f3de211173ad..3a129c42658e 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5135,6 +5135,8 @@ static int __reserve_metadata_bytes(struct btrfs_root *root, int ret = 0; ASSERT(orig_bytes); + ASSERT(!current->journal_info || flush != BTRFS_RESERVE_FLUSH_ALL); + spin_lock(&space_info->lock); ret = -ENOSPC; used = space_info->bytes_used + space_info->bytes_reserved + diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 23e16de52e23..fc067b07e31f 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2604,25 +2604,28 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans, trans->block_rsv = rc->block_rsv; rc->reserved_bytes += num_bytes; + + /* + * We are under a transaction here so we can only do limited flushing. + * If we get an enospc just kick back -EAGAIN so we know to drop the + * transaction and try to refill when we can flush all the things. + */ ret = btrfs_block_rsv_refill(root, rc->block_rsv, num_bytes, - BTRFS_RESERVE_FLUSH_ALL); + BTRFS_RESERVE_FLUSH_LIMIT); if (ret) { - if (ret == -EAGAIN) { - tmp = rc->extent_root->nodesize * - RELOCATION_RESERVED_NODES; - while (tmp <= rc->reserved_bytes) - tmp <<= 1; - /* - * only one thread can access block_rsv at this point, - * so we don't need hold lock to protect block_rsv. - * we expand more reservation size here to allow enough - * space for relocation and we will return earlier in - * enospc case. - */ - rc->block_rsv->size = tmp + rc->extent_root->nodesize * - RELOCATION_RESERVED_NODES; - } - return ret; + tmp = rc->extent_root->nodesize * RELOCATION_RESERVED_NODES; + while (tmp <= rc->reserved_bytes) + tmp <<= 1; + /* + * only one thread can access block_rsv at this point, + * so we don't need hold lock to protect block_rsv. + * we expand more reservation size here to allow enough + * space for relocation and we will return eailer in + * enospc case. + */ + rc->block_rsv->size = tmp + rc->extent_root->nodesize * + RELOCATION_RESERVED_NODES; + return -EAGAIN; } return 0; -- cgit v1.2.3 From a4e187d83d88eeaba6252aac0a2ffe5eaa73a818 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 29 Jun 2016 13:55:22 -0400 Subject: NFS: Don't drop CB requests with invalid principals Before commit 778be232a207 ("NFS do not find client in NFSv4 pg_authenticate"), the Linux callback server replied with RPC_AUTH_ERROR / RPC_AUTH_BADCRED, instead of dropping the CB request. Let's restore that behavior so the server has a chance to do something useful about it, and provide a warning that helps admins correct the problem. Fixes: 778be232a207 ("NFS do not find client in NFSv4 ...") Signed-off-by: Chuck Lever Tested-by: Steve Wise Signed-off-by: Anna Schumaker --- fs/nfs/callback_xdr.c | 6 +++++- net/sunrpc/svc.c | 5 +++++ 2 files changed, 10 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index d81f96aacd51..656f68f7fe53 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -925,7 +925,7 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r if (hdr_arg.minorversion == 0) { cps.clp = nfs4_find_client_ident(SVC_NET(rqstp), hdr_arg.cb_ident); if (!cps.clp || !check_gss_callback_principal(cps.clp, rqstp)) - return rpc_drop_reply; + goto out_invalidcred; } cps.minorversion = hdr_arg.minorversion; @@ -953,6 +953,10 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r nfs_put_client(cps.clp); dprintk("%s: done, status = %u\n", __func__, ntohl(status)); return rpc_success; + +out_invalidcred: + pr_warn_ratelimited("NFS: NFSv4 callback contains invalid cred\n"); + return rpc_autherr_badcred; } /* diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index 87290a5a9ac7..c5b0cb4f4056 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -1194,6 +1194,11 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) procp->pc_release(rqstp, NULL, rqstp->rq_resp); goto dropit; } + if (*statp == rpc_autherr_badcred) { + if (procp->pc_release) + procp->pc_release(rqstp, NULL, rqstp->rq_resp); + goto err_bad_auth; + } if (*statp == rpc_success && (xdr = procp->pc_encode) && !xdr(rqstp, resv->iov_base+resv->iov_len, rqstp->rq_resp)) { -- cgit v1.2.3 From dedeb13f9efb4439a37cf56317c8f25860dd667b Mon Sep 17 00:00:00 2001 From: Andrew Elble Date: Wed, 15 Jun 2016 12:52:08 -0400 Subject: nfsd: allow mach_creds_match to be used more broadly Rename mach_creds_match() to nfsd4_mach_creds_match() and un-staticify Signed-off-by: Andrew Elble Reviewed-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 14 +++++++------- fs/nfsd/xdr4.h | 2 ++ 2 files changed, 9 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 70d0b9b33031..ef583507d276 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1972,7 +1972,7 @@ static bool svc_rqst_integrity_protected(struct svc_rqst *rqstp) service == RPC_GSS_SVC_PRIVACY; } -static bool mach_creds_match(struct nfs4_client *cl, struct svc_rqst *rqstp) +bool nfsd4_mach_creds_match(struct nfs4_client *cl, struct svc_rqst *rqstp) { struct svc_cred *cr = &rqstp->rq_cred; @@ -2424,7 +2424,7 @@ nfsd4_exchange_id(struct svc_rqst *rqstp, status = nfserr_inval; goto out; } - if (!mach_creds_match(conf, rqstp)) { + if (!nfsd4_mach_creds_match(conf, rqstp)) { status = nfserr_wrong_cred; goto out; } @@ -2676,7 +2676,7 @@ nfsd4_create_session(struct svc_rqst *rqstp, if (conf) { status = nfserr_wrong_cred; - if (!mach_creds_match(conf, rqstp)) + if (!nfsd4_mach_creds_match(conf, rqstp)) goto out_free_conn; cs_slot = &conf->cl_cs_slot; status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0); @@ -2692,7 +2692,7 @@ nfsd4_create_session(struct svc_rqst *rqstp, goto out_free_conn; } status = nfserr_wrong_cred; - if (!mach_creds_match(unconf, rqstp)) + if (!nfsd4_mach_creds_match(unconf, rqstp)) goto out_free_conn; cs_slot = &unconf->cl_cs_slot; status = check_slot_seqid(cr_ses->seqid, cs_slot->sl_seqid, 0); @@ -2801,7 +2801,7 @@ __be32 nfsd4_bind_conn_to_session(struct svc_rqst *rqstp, if (!session) goto out_no_session; status = nfserr_wrong_cred; - if (!mach_creds_match(session->se_client, rqstp)) + if (!nfsd4_mach_creds_match(session->se_client, rqstp)) goto out; status = nfsd4_map_bcts_dir(&bcts->dir); if (status) @@ -2848,7 +2848,7 @@ nfsd4_destroy_session(struct svc_rqst *r, if (!ses) goto out_client_lock; status = nfserr_wrong_cred; - if (!mach_creds_match(ses->se_client, r)) + if (!nfsd4_mach_creds_match(ses->se_client, r)) goto out_put_session; status = mark_session_dead_locked(ses, 1 + ref_held_by_me); if (status) @@ -3087,7 +3087,7 @@ nfsd4_destroy_clientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *csta status = nfserr_stale_clientid; goto out; } - if (!mach_creds_match(clp, rqstp)) { + if (!nfsd4_mach_creds_match(clp, rqstp)) { clp = NULL; status = nfserr_wrong_cred; goto out; diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h index d9554813e58a..74342a7c208a 100644 --- a/fs/nfsd/xdr4.h +++ b/fs/nfsd/xdr4.h @@ -654,6 +654,8 @@ set_change_info(struct nfsd4_change_info *cinfo, struct svc_fh *fhp) } + +bool nfsd4_mach_creds_match(struct nfs4_client *cl, struct svc_rqst *rqstp); int nfs4svc_encode_voidres(struct svc_rqst *, __be32 *, void *); int nfs4svc_decode_compoundargs(struct svc_rqst *, __be32 *, struct nfsd4_compoundargs *); -- cgit v1.2.3 From ed94164398c935a42be7b129a478eb19c598b68a Mon Sep 17 00:00:00 2001 From: Andrew Elble Date: Wed, 15 Jun 2016 12:52:09 -0400 Subject: nfsd: implement machine credential support for some operations This addresses the conundrum referenced in RFC5661 18.35.3, and will allow clients to return state to the server using the machine credentials. The biggest part of the problem is that we need to allow the client to send a compound op with integrity/privacy on mounts that don't have it enabled. Add server support for properly decoding and using spo_must_enforce and spo_must_allow bits. Add support for machine credentials to be used for CLOSE, OPEN_DOWNGRADE, LOCKU, DELEGRETURN, and TEST/FREE STATEID. Implement a check so as to not throw WRONGSEC errors when these operations are used if integrity/privacy isn't turned on. Without this, Linux clients with credentials that expired while holding delegations were getting stuck in an endless loop. Signed-off-by: Andrew Elble Reviewed-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/export.c | 10 ++++++++++ fs/nfsd/nfs4proc.c | 39 +++++++++++++++++++++++++++++++++++++++ fs/nfsd/nfs4state.c | 18 ++++++++++++++++++ fs/nfsd/nfs4xdr.c | 51 +++++++++++++++++++++++---------------------------- fs/nfsd/nfsd.h | 5 +++++ fs/nfsd/state.h | 1 + fs/nfsd/xdr4.h | 3 +++ 7 files changed, 99 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index b4d84b579f20..79de2f38dd63 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -954,6 +954,16 @@ __be32 check_nfsd_access(struct svc_export *exp, struct svc_rqst *rqstp) rqstp->rq_cred.cr_flavor == RPC_AUTH_UNIX) return 0; } + + /* If the compound op contains a spo_must_allowed op, + * it will be sent with integrity/protection which + * will have to be expressly allowed on mounts that + * don't support it + */ + + if (nfsd4_spo_must_allow(rqstp)) + return 0; + return nfserr_wrongsec; } diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index de1ff1d98bb1..b1159b3e9816 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -2335,6 +2335,45 @@ static struct nfsd4_operation nfsd4_ops[] = { }, }; +/** + * nfsd4_spo_must_allow - Determine if the compound op contains an + * operation that is allowed to be sent with machine credentials + * + * @rqstp: a pointer to the struct svc_rqst + * + * Checks to see if the compound contains a spo_must_allow op + * and confirms that it was sent with the proper machine creds. + */ + +bool nfsd4_spo_must_allow(struct svc_rqst *rqstp) +{ + struct nfsd4_compoundres *resp = rqstp->rq_resp; + struct nfsd4_compoundargs *argp = rqstp->rq_argp; + struct nfsd4_op *this = &argp->ops[resp->opcnt - 1]; + struct nfsd4_compound_state *cstate = &resp->cstate; + struct nfs4_op_map *allow = &cstate->clp->cl_spo_must_allow; + u32 opiter; + + if (!cstate->minorversion) + return false; + + if (cstate->spo_must_allowed == true) + return true; + + opiter = resp->opcnt; + while (opiter < argp->opcnt) { + this = &argp->ops[opiter++]; + if (test_bit(this->opnum, allow->u.longs) && + cstate->clp->cl_mach_cred && + nfsd4_mach_creds_match(cstate->clp, rqstp)) { + cstate->spo_must_allowed = true; + return true; + } + } + cstate->spo_must_allowed = false; + return false; +} + int nfsd4_max_reply(struct svc_rqst *rqstp, struct nfsd4_op *op) { struct nfsd4_operation *opdesc; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index ef583507d276..ebfcebd5eab1 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2388,6 +2388,22 @@ nfsd4_exchange_id(struct svc_rqst *rqstp, switch (exid->spa_how) { case SP4_MACH_CRED: + exid->spo_must_enforce[0] = 0; + exid->spo_must_enforce[1] = ( + 1 << (OP_BIND_CONN_TO_SESSION - 32) | + 1 << (OP_EXCHANGE_ID - 32) | + 1 << (OP_CREATE_SESSION - 32) | + 1 << (OP_DESTROY_SESSION - 32) | + 1 << (OP_DESTROY_CLIENTID - 32)); + + exid->spo_must_allow[0] &= (1 << (OP_CLOSE) | + 1 << (OP_OPEN_DOWNGRADE) | + 1 << (OP_LOCKU) | + 1 << (OP_DELEGRETURN)); + + exid->spo_must_allow[1] &= ( + 1 << (OP_TEST_STATEID - 32) | + 1 << (OP_FREE_STATEID - 32)); if (!svc_rqst_integrity_protected(rqstp)) { status = nfserr_inval; goto out_nolock; @@ -2473,6 +2489,8 @@ out_new: goto out; } new->cl_minorversion = cstate->minorversion; + new->cl_spo_must_allow.u.words[0] = exid->spo_must_allow[0]; + new->cl_spo_must_allow.u.words[1] = exid->spo_must_allow[1]; gen_clid(new, nn); add_to_unconfirmed(new); diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 9df898ba648f..84ef94794496 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -1299,16 +1299,14 @@ nfsd4_decode_exchange_id(struct nfsd4_compoundargs *argp, break; case SP4_MACH_CRED: /* spo_must_enforce */ - READ_BUF(4); - dummy = be32_to_cpup(p++); - READ_BUF(dummy * 4); - p += dummy; - + status = nfsd4_decode_bitmap(argp, + exid->spo_must_enforce); + if (status) + goto out; /* spo_must_allow */ - READ_BUF(4); - dummy = be32_to_cpup(p++); - READ_BUF(dummy * 4); - p += dummy; + status = nfsd4_decode_bitmap(argp, exid->spo_must_allow); + if (status) + goto out; break; case SP4_SSV: /* ssp_ops */ @@ -3867,14 +3865,6 @@ nfsd4_encode_write(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_w return nfserr; } -static const u32 nfs4_minimal_spo_must_enforce[2] = { - [1] = 1 << (OP_BIND_CONN_TO_SESSION - 32) | - 1 << (OP_EXCHANGE_ID - 32) | - 1 << (OP_CREATE_SESSION - 32) | - 1 << (OP_DESTROY_SESSION - 32) | - 1 << (OP_DESTROY_CLIENTID - 32) -}; - static __be32 nfsd4_encode_exchange_id(struct nfsd4_compoundres *resp, __be32 nfserr, struct nfsd4_exchange_id *exid) @@ -3885,6 +3875,7 @@ nfsd4_encode_exchange_id(struct nfsd4_compoundres *resp, __be32 nfserr, char *server_scope; int major_id_sz; int server_scope_sz; + int status = 0; uint64_t minor_id = 0; if (nfserr) @@ -3913,18 +3904,20 @@ nfsd4_encode_exchange_id(struct nfsd4_compoundres *resp, __be32 nfserr, case SP4_NONE: break; case SP4_MACH_CRED: - /* spo_must_enforce, spo_must_allow */ - p = xdr_reserve_space(xdr, 16); - if (!p) - return nfserr_resource; - /* spo_must_enforce bitmap: */ - *p++ = cpu_to_be32(2); - *p++ = cpu_to_be32(nfs4_minimal_spo_must_enforce[0]); - *p++ = cpu_to_be32(nfs4_minimal_spo_must_enforce[1]); - /* empty spo_must_allow bitmap: */ - *p++ = cpu_to_be32(0); - + status = nfsd4_encode_bitmap(xdr, + exid->spo_must_enforce[0], + exid->spo_must_enforce[1], + exid->spo_must_enforce[2]); + if (status) + goto out; + /* spo_must_allow bitmap: */ + status = nfsd4_encode_bitmap(xdr, + exid->spo_must_allow[0], + exid->spo_must_allow[1], + exid->spo_must_allow[2]); + if (status) + goto out; break; default: WARN_ON_ONCE(1); @@ -3951,6 +3944,8 @@ nfsd4_encode_exchange_id(struct nfsd4_compoundres *resp, __be32 nfserr, /* Implementation id */ *p++ = cpu_to_be32(0); /* zero length nfs_impl_id4 array */ return 0; +out: + return status; } static __be32 diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h index cf980523898b..9446849888d5 100644 --- a/fs/nfsd/nfsd.h +++ b/fs/nfsd/nfsd.h @@ -124,6 +124,7 @@ void nfs4_state_shutdown_net(struct net *net); void nfs4_reset_lease(time_t leasetime); int nfs4_reset_recoverydir(char *recdir); char * nfs4_recoverydir(void); +bool nfsd4_spo_must_allow(struct svc_rqst *rqstp); #else static inline int nfsd4_init_slabs(void) { return 0; } static inline void nfsd4_free_slabs(void) { } @@ -134,6 +135,10 @@ static inline void nfs4_state_shutdown_net(struct net *net) { } static inline void nfs4_reset_lease(time_t leasetime) { } static inline int nfs4_reset_recoverydir(char *recdir) { return 0; } static inline char * nfs4_recoverydir(void) {return NULL; } +static inline bool nfsd4_spo_must_allow(struct svc_rqst *rqstp) +{ + return false; +} #endif /* diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index 64053eadeb81..b95adf9a1595 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -345,6 +345,7 @@ struct nfs4_client { u32 cl_exchange_flags; /* number of rpc's in progress over an associated session: */ atomic_t cl_refcount; + struct nfs4_op_map cl_spo_must_allow; /* for nfs41 callbacks */ /* We currently support a single back channel with a single slot */ diff --git a/fs/nfsd/xdr4.h b/fs/nfsd/xdr4.h index 74342a7c208a..beea0c5edc51 100644 --- a/fs/nfsd/xdr4.h +++ b/fs/nfsd/xdr4.h @@ -59,6 +59,7 @@ struct nfsd4_compound_state { struct nfsd4_session *session; struct nfsd4_slot *slot; int data_offset; + bool spo_must_allowed; size_t iovlen; u32 minorversion; __be32 status; @@ -403,6 +404,8 @@ struct nfsd4_exchange_id { clientid_t clientid; u32 seqid; int spa_how; + u32 spo_must_enforce[3]; + u32 spo_must_allow[3]; }; struct nfsd4_sequence { -- cgit v1.2.3 From d7c920d1349ad78b9181d8933404e1a378da6feb Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Tue, 14 Jun 2016 13:41:27 -0700 Subject: nfsd: flex file device id encoding will need the server address Signed-off-by: Tom Haynes Reviewed-by: Christoph Hellwig Reviewed-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/blocklayout.c | 2 ++ fs/nfsd/nfs4proc.c | 2 +- fs/nfsd/pnfs.h | 1 + 3 files changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/blocklayout.c b/fs/nfsd/blocklayout.c index 31f3df193bdb..c98563d35822 100644 --- a/fs/nfsd/blocklayout.c +++ b/fs/nfsd/blocklayout.c @@ -162,6 +162,7 @@ nfsd4_block_get_device_info_simple(struct super_block *sb, static __be32 nfsd4_block_proc_getdeviceinfo(struct super_block *sb, + struct svc_rqst *rqstp, struct nfs4_client *clp, struct nfsd4_getdeviceinfo *gdp) { @@ -354,6 +355,7 @@ nfsd4_block_get_device_info_scsi(struct super_block *sb, static __be32 nfsd4_scsi_proc_getdeviceinfo(struct super_block *sb, + struct svc_rqst *rqstp, struct nfs4_client *clp, struct nfsd4_getdeviceinfo *gdp) { diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index b1159b3e9816..89fad8aeffe2 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -1270,7 +1270,7 @@ nfsd4_getdeviceinfo(struct svc_rqst *rqstp, nfserr = nfs_ok; if (gdp->gd_maxcount != 0) { nfserr = ops->proc_getdeviceinfo(exp->ex_path.mnt->mnt_sb, - cstate->session->se_client, gdp); + rqstp, cstate->session->se_client, gdp); } gdp->gd_notify_types &= ops->notify_types; diff --git a/fs/nfsd/pnfs.h b/fs/nfsd/pnfs.h index 7d073b9b1553..e855677fb132 100644 --- a/fs/nfsd/pnfs.h +++ b/fs/nfsd/pnfs.h @@ -21,6 +21,7 @@ struct nfsd4_layout_ops { u32 notify_types; __be32 (*proc_getdeviceinfo)(struct super_block *sb, + struct svc_rqst *rqstp, struct nfs4_client *clp, struct nfsd4_getdeviceinfo *gdevp); __be32 (*encode_getdeviceinfo)(struct xdr_stream *xdr, -- cgit v1.2.3 From 9b9960a0ca4773e21c4b153ed355583946346b25 Mon Sep 17 00:00:00 2001 From: Tom Haynes Date: Tue, 14 Jun 2016 13:41:28 -0700 Subject: nfsd: Add a super simple flex file server Have a simple flex file server where the mds (NFSv4.1 or NFSv4.2) is also the ds (NFSv3). I.e., the metadata and the data file are the exact same file. This will allow testing of the flex file client. Simply add the "pnfs" export option to your export in /etc/exports and mount from a client that supports flex files. Signed-off-by: Tom Haynes Reviewed-by: Christoph Hellwig Signed-off-by: J. Bruce Fields --- fs/nfsd/Kconfig | 17 ++++++ fs/nfsd/Makefile | 1 + fs/nfsd/flexfilelayout.c | 133 ++++++++++++++++++++++++++++++++++++++++++++ fs/nfsd/flexfilelayoutxdr.c | 115 ++++++++++++++++++++++++++++++++++++++ fs/nfsd/flexfilelayoutxdr.h | 49 ++++++++++++++++ fs/nfsd/nfs4layouts.c | 12 +++- fs/nfsd/pnfs.h | 3 + 7 files changed, 329 insertions(+), 1 deletion(-) create mode 100644 fs/nfsd/flexfilelayout.c create mode 100644 fs/nfsd/flexfilelayoutxdr.c create mode 100644 fs/nfsd/flexfilelayoutxdr.h (limited to 'fs') diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig index c9f583d7bac8..e64d1a3ec770 100644 --- a/fs/nfsd/Kconfig +++ b/fs/nfsd/Kconfig @@ -111,6 +111,23 @@ config NFSD_SCSILAYOUT If unsure, say N. +config NFSD_FLEXFILELAYOUT + bool "NFSv4.1 server support for pNFS Flex File layouts" + depends on NFSD_V4 + select NFSD_PNFS + help + This option enables support for the exporting pNFS Flex File + layouts in the kernel's NFS server. The pNFS Flex File layout + enables NFS clients to directly perform I/O to NFSv3 devices + accesible to both the server and the clients. See + draft-ietf-nfsv4-flex-files for more details. + + Warning, this server implements the bare minimum functionality + to be a flex file server - it is for testing the client, + not for use in production. + + If unsure, say N. + config NFSD_V4_SECURITY_LABEL bool "Provide Security Label support for NFSv4 server" depends on NFSD_V4 && SECURITY diff --git a/fs/nfsd/Makefile b/fs/nfsd/Makefile index 3ae5f3c77e28..5f5d3a76980c 100644 --- a/fs/nfsd/Makefile +++ b/fs/nfsd/Makefile @@ -20,3 +20,4 @@ nfsd-$(CONFIG_NFSD_V4) += nfs4proc.o nfs4xdr.o nfs4state.o nfs4idmap.o \ nfsd-$(CONFIG_NFSD_PNFS) += nfs4layouts.o nfsd-$(CONFIG_NFSD_BLOCKLAYOUT) += blocklayout.o blocklayoutxdr.o nfsd-$(CONFIG_NFSD_SCSILAYOUT) += blocklayout.o blocklayoutxdr.o +nfsd-$(CONFIG_NFSD_FLEXFILELAYOUT) += flexfilelayout.o flexfilelayoutxdr.o diff --git a/fs/nfsd/flexfilelayout.c b/fs/nfsd/flexfilelayout.c new file mode 100644 index 000000000000..df880e9fa71f --- /dev/null +++ b/fs/nfsd/flexfilelayout.c @@ -0,0 +1,133 @@ +/* + * Copyright (c) 2016 Tom Haynes + * + * The following implements a super-simple flex-file server + * where the NFSv4.1 mds is also the ds. And the storage is + * the same. I.e., writing to the mds via a NFSv4.1 WRITE + * goes to the same location as the NFSv3 WRITE. + */ +#include + +#include + +#include + +#include "flexfilelayoutxdr.h" +#include "pnfs.h" + +#define NFSDDBG_FACILITY NFSDDBG_PNFS + +static __be32 +nfsd4_ff_proc_layoutget(struct inode *inode, const struct svc_fh *fhp, + struct nfsd4_layoutget *args) +{ + struct nfsd4_layout_seg *seg = &args->lg_seg; + u32 device_generation = 0; + int error; + uid_t u; + + struct pnfs_ff_layout *fl; + + /* + * The super simple flex file server has 1 mirror, 1 data server, + * and 1 file handle. So instead of 4 allocs, do 1 for now. + * Zero it out for the stateid - don't want junk in there! + */ + error = -ENOMEM; + fl = kzalloc(sizeof(*fl), GFP_KERNEL); + if (!fl) + goto out_error; + args->lg_content = fl; + + /* + * Avoid layout commit, try to force the I/O to the DS, + * and for fun, cause all IOMODE_RW layout segments to + * effectively be WRITE only. + */ + fl->flags = FF_FLAGS_NO_LAYOUTCOMMIT | FF_FLAGS_NO_IO_THRU_MDS | + FF_FLAGS_NO_READ_IO; + + /* Do not allow a IOMODE_READ segment to have write pemissions */ + if (seg->iomode == IOMODE_READ) { + u = from_kuid(&init_user_ns, inode->i_uid) + 1; + fl->uid = make_kuid(&init_user_ns, u); + } else + fl->uid = inode->i_uid; + fl->gid = inode->i_gid; + + error = nfsd4_set_deviceid(&fl->deviceid, fhp, device_generation); + if (error) + goto out_error; + + fl->fh.size = fhp->fh_handle.fh_size; + memcpy(fl->fh.data, &fhp->fh_handle.fh_base, fl->fh.size); + + /* Give whole file layout segments */ + seg->offset = 0; + seg->length = NFS4_MAX_UINT64; + + dprintk("GET: 0x%llx:0x%llx %d\n", seg->offset, seg->length, + seg->iomode); + return 0; + +out_error: + seg->length = 0; + return nfserrno(error); +} + +static __be32 +nfsd4_ff_proc_getdeviceinfo(struct super_block *sb, struct svc_rqst *rqstp, + struct nfs4_client *clp, struct nfsd4_getdeviceinfo *gdp) +{ + struct pnfs_ff_device_addr *da; + + u16 port; + char addr[INET6_ADDRSTRLEN]; + + da = kzalloc(sizeof(struct pnfs_ff_device_addr), GFP_KERNEL); + if (!da) + return nfserrno(-ENOMEM); + + gdp->gd_device = da; + + da->version = 3; + da->minor_version = 0; + + da->rsize = svc_max_payload(rqstp); + da->wsize = da->rsize; + + rpc_ntop((struct sockaddr *)&rqstp->rq_daddr, + addr, INET6_ADDRSTRLEN); + if (rqstp->rq_daddr.ss_family == AF_INET) { + struct sockaddr_in *sin; + + sin = (struct sockaddr_in *)&rqstp->rq_daddr; + port = ntohs(sin->sin_port); + snprintf(da->netaddr.netid, FF_NETID_LEN + 1, "tcp"); + da->netaddr.netid_len = 3; + } else { + struct sockaddr_in6 *sin6; + + sin6 = (struct sockaddr_in6 *)&rqstp->rq_daddr; + port = ntohs(sin6->sin6_port); + snprintf(da->netaddr.netid, FF_NETID_LEN + 1, "tcp6"); + da->netaddr.netid_len = 4; + } + + da->netaddr.addr_len = + snprintf(da->netaddr.addr, FF_ADDR_LEN + 1, + "%s.%hhu.%hhu", addr, port >> 8, port & 0xff); + + da->tightly_coupled = false; + + return 0; +} + +const struct nfsd4_layout_ops ff_layout_ops = { + .notify_types = + NOTIFY_DEVICEID4_DELETE | NOTIFY_DEVICEID4_CHANGE, + .proc_getdeviceinfo = nfsd4_ff_proc_getdeviceinfo, + .encode_getdeviceinfo = nfsd4_ff_encode_getdeviceinfo, + .proc_layoutget = nfsd4_ff_proc_layoutget, + .encode_layoutget = nfsd4_ff_encode_layoutget, +}; diff --git a/fs/nfsd/flexfilelayoutxdr.c b/fs/nfsd/flexfilelayoutxdr.c new file mode 100644 index 000000000000..5e3fd7fc1a9f --- /dev/null +++ b/fs/nfsd/flexfilelayoutxdr.c @@ -0,0 +1,115 @@ +/* + * Copyright (c) 2016 Tom Haynes + */ +#include +#include + +#include "nfsd.h" +#include "flexfilelayoutxdr.h" + +#define NFSDDBG_FACILITY NFSDDBG_PNFS + +struct ff_idmap { + char buf[11]; + int len; +}; + +__be32 +nfsd4_ff_encode_layoutget(struct xdr_stream *xdr, + struct nfsd4_layoutget *lgp) +{ + struct pnfs_ff_layout *fl = lgp->lg_content; + int len, mirror_len, ds_len, fh_len; + __be32 *p; + + /* + * Unlike nfsd4_encode_user, we know these will + * always be stringified. + */ + struct ff_idmap uid; + struct ff_idmap gid; + + fh_len = 4 + fl->fh.size; + + uid.len = sprintf(uid.buf, "%u", from_kuid(&init_user_ns, fl->uid)); + gid.len = sprintf(gid.buf, "%u", from_kgid(&init_user_ns, fl->gid)); + + /* 8 + len for recording the length, name, and padding */ + ds_len = 20 + sizeof(stateid_opaque_t) + 4 + fh_len + + 8 + uid.len + 8 + gid.len; + + mirror_len = 4 + ds_len; + + /* The layout segment */ + len = 20 + mirror_len; + + p = xdr_reserve_space(xdr, sizeof(__be32) + len); + if (!p) + return nfserr_toosmall; + + *p++ = cpu_to_be32(len); + p = xdr_encode_hyper(p, 0); /* stripe unit of 1 */ + + *p++ = cpu_to_be32(1); /* single mirror */ + *p++ = cpu_to_be32(1); /* single data server */ + + p = xdr_encode_opaque_fixed(p, &fl->deviceid, + sizeof(struct nfsd4_deviceid)); + + *p++ = cpu_to_be32(1); /* efficiency */ + + *p++ = cpu_to_be32(fl->stateid.si_generation); + p = xdr_encode_opaque_fixed(p, &fl->stateid.si_opaque, + sizeof(stateid_opaque_t)); + + *p++ = cpu_to_be32(1); /* single file handle */ + p = xdr_encode_opaque(p, fl->fh.data, fl->fh.size); + + p = xdr_encode_opaque(p, uid.buf, uid.len); + p = xdr_encode_opaque(p, gid.buf, gid.len); + + *p++ = cpu_to_be32(fl->flags); + *p++ = cpu_to_be32(0); /* No stats collect hint */ + + return 0; +} + +__be32 +nfsd4_ff_encode_getdeviceinfo(struct xdr_stream *xdr, + struct nfsd4_getdeviceinfo *gdp) +{ + struct pnfs_ff_device_addr *da = gdp->gd_device; + int len; + int ver_len; + int addr_len; + __be32 *p; + + /* len + padding for two strings */ + addr_len = 16 + da->netaddr.netid_len + da->netaddr.addr_len; + ver_len = 20; + + len = 4 + ver_len + 4 + addr_len; + + p = xdr_reserve_space(xdr, len + sizeof(__be32)); + if (!p) + return nfserr_resource; + + /* + * Fill in the overall length and number of volumes at the beginning + * of the layout. + */ + *p++ = cpu_to_be32(len); + *p++ = cpu_to_be32(1); /* 1 netaddr */ + p = xdr_encode_opaque(p, da->netaddr.netid, da->netaddr.netid_len); + p = xdr_encode_opaque(p, da->netaddr.addr, da->netaddr.addr_len); + + *p++ = cpu_to_be32(1); /* 1 versions */ + + *p++ = cpu_to_be32(da->version); + *p++ = cpu_to_be32(da->minor_version); + *p++ = cpu_to_be32(da->rsize); + *p++ = cpu_to_be32(da->wsize); + *p++ = cpu_to_be32(da->tightly_coupled); + + return 0; +} diff --git a/fs/nfsd/flexfilelayoutxdr.h b/fs/nfsd/flexfilelayoutxdr.h new file mode 100644 index 000000000000..467defd4e563 --- /dev/null +++ b/fs/nfsd/flexfilelayoutxdr.h @@ -0,0 +1,49 @@ +/* + * Copyright (c) 2016 Tom Haynes + */ +#ifndef _NFSD_FLEXFILELAYOUTXDR_H +#define _NFSD_FLEXFILELAYOUTXDR_H 1 + +#include +#include "xdr4.h" + +#define FF_FLAGS_NO_LAYOUTCOMMIT 1 +#define FF_FLAGS_NO_IO_THRU_MDS 2 +#define FF_FLAGS_NO_READ_IO 4 + +struct xdr_stream; + +#define FF_NETID_LEN (4) +#define FF_ADDR_LEN (INET6_ADDRSTRLEN + 8) +struct pnfs_ff_netaddr { + char netid[FF_NETID_LEN + 1]; + char addr[FF_ADDR_LEN + 1]; + u32 netid_len; + u32 addr_len; +}; + +struct pnfs_ff_device_addr { + struct pnfs_ff_netaddr netaddr; + u32 version; + u32 minor_version; + u32 rsize; + u32 wsize; + bool tightly_coupled; +}; + +struct pnfs_ff_layout { + u32 flags; + u32 stats_collect_hint; + kuid_t uid; + kgid_t gid; + struct nfsd4_deviceid deviceid; + stateid_t stateid; + struct nfs_fh fh; +}; + +__be32 nfsd4_ff_encode_getdeviceinfo(struct xdr_stream *xdr, + struct nfsd4_getdeviceinfo *gdp); +__be32 nfsd4_ff_encode_layoutget(struct xdr_stream *xdr, + struct nfsd4_layoutget *lgp); + +#endif /* _NFSD_FLEXFILELAYOUTXDR_H */ diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c index 953c0755cb37..6d98d16b3354 100644 --- a/fs/nfsd/nfs4layouts.c +++ b/fs/nfsd/nfs4layouts.c @@ -27,6 +27,9 @@ static const struct nfsd4_callback_ops nfsd4_cb_layout_ops; static const struct lock_manager_operations nfsd4_layouts_lm_ops; const struct nfsd4_layout_ops *nfsd4_layout_ops[LAYOUT_TYPE_MAX] = { +#ifdef CONFIG_NFSD_FLEXFILELAYOUT + [LAYOUT_FLEX_FILES] = &ff_layout_ops, +#endif #ifdef CONFIG_NFSD_BLOCKLAYOUT [LAYOUT_BLOCK_VOLUME] = &bl_layout_ops, #endif @@ -122,17 +125,24 @@ nfsd4_set_deviceid(struct nfsd4_deviceid *id, const struct svc_fh *fhp, void nfsd4_setup_layout_type(struct svc_export *exp) { +#if defined(CONFIG_NFSD_BLOCKLAYOUT) || defined(CONFIG_NFSD_SCSILAYOUT) struct super_block *sb = exp->ex_path.mnt->mnt_sb; +#endif if (!(exp->ex_flags & NFSEXP_PNFS)) return; /* - * Check if the file system supports exporting a block-like layout. + * If flex file is configured, use it by default. Otherwise + * check if the file system supports exporting a block-like layout. * If the block device supports reservations prefer the SCSI layout, * otherwise advertise the block layout. */ +#ifdef CONFIG_NFSD_FLEXFILELAYOUT + exp->ex_layout_type = LAYOUT_FLEX_FILES; +#endif #ifdef CONFIG_NFSD_BLOCKLAYOUT + /* overwrite flex file layout selection if needed */ if (sb->s_export_op->get_uuid && sb->s_export_op->map_blocks && sb->s_export_op->commit_blocks) diff --git a/fs/nfsd/pnfs.h b/fs/nfsd/pnfs.h index e855677fb132..0c2a716e8741 100644 --- a/fs/nfsd/pnfs.h +++ b/fs/nfsd/pnfs.h @@ -45,6 +45,9 @@ extern const struct nfsd4_layout_ops bl_layout_ops; #ifdef CONFIG_NFSD_SCSILAYOUT extern const struct nfsd4_layout_ops scsi_layout_ops; #endif +#ifdef CONFIG_NFSD_FLEXFILELAYOUT +extern const struct nfsd4_layout_ops ff_layout_ops; +#endif __be32 nfsd4_preprocess_layout_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, stateid_t *stateid, -- cgit v1.2.3 From 93f580a9a2413def2fcc5fd07085efcd95943f1b Mon Sep 17 00:00:00 2001 From: Oleg Drokin Date: Thu, 7 Jul 2016 21:49:38 -0400 Subject: nfsd: Correct a comment for NFSD_MAY_ defines location Those are now defined in fs/nfsd/vfs.h Signed-off-by: Oleg Drokin Reviewed-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/nfsfh.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index a8919444c460..7e49dcbbf452 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -298,7 +298,7 @@ out: * that it expects something not of the given type. * * @access is formed from the NFSD_MAY_* constants defined in - * include/linux/nfsd/nfsd.h. + * fs/nfsd/vfs.h. */ __be32 fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, int access) -- cgit v1.2.3 From d28c442f5bd5471adc4d1add5fa78109f811c0a3 Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Sat, 2 Jul 2016 08:24:32 +0200 Subject: nfsd: Fix some indent inconsistancy Silent a few smatch warnings about indentation Signed-off-by: Christophe JAILLET Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 2 +- fs/nfsd/nfsfh.c | 2 +- fs/nfsd/nfsxdr.c | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index ebfcebd5eab1..bc2173a31b5d 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -3130,7 +3130,7 @@ nfsd4_reclaim_complete(struct svc_rqst *rqstp, struct nfsd4_compound_state *csta * We don't take advantage of the rca_one_fs case. * That's OK, it's optional, we can safely ignore it. */ - return nfs_ok; + return nfs_ok; } status = nfserr_complete_already; diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 7e49dcbbf452..27250e279c37 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -533,7 +533,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry, * the reference filehandle (if it is in the same export) * or the export options. */ - set_version_and_fsid_type(fhp, exp, ref_fh); + set_version_and_fsid_type(fhp, exp, ref_fh); if (ref_fh == fhp) fh_put(ref_fh); diff --git a/fs/nfsd/nfsxdr.c b/fs/nfsd/nfsxdr.c index 79d964aa8079..41b468a6a90f 100644 --- a/fs/nfsd/nfsxdr.c +++ b/fs/nfsd/nfsxdr.c @@ -240,7 +240,7 @@ nfssvc_decode_diropargs(struct svc_rqst *rqstp, __be32 *p, || !(p = decode_filename(p, &args->name, &args->len))) return 0; - return xdr_argsize_check(rqstp, p); + return xdr_argsize_check(rqstp, p); } int -- cgit v1.2.3 From 8b7d9d09b24f4ef16f7ae34b6d9e59857fda0870 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Jul 2016 12:42:40 -0400 Subject: NFSv4: Revert "Truncating file opens should also sync O_DIRECT writes" We're not holding any locks, so both nfs_wb_all() and inode_dio_wait() are unenforcible and have livelock potential. Just limit ourselves to flushing out the data. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c index 7cdc0ab9e6f5..d085ad794884 100644 --- a/fs/nfs/nfs4file.c +++ b/fs/nfs/nfs4file.c @@ -66,7 +66,7 @@ nfs4_file_open(struct inode *inode, struct file *filp) if (openflags & O_TRUNC) { attr.ia_valid |= ATTR_SIZE; attr.ia_size = 0; - nfs_sync_inode(inode); + filemap_write_and_wait(inode->i_mapping); } inode = NFS_PROTO(dir)->open_context(dir, ctx, openflags, &attr, NULL); -- cgit v1.2.3 From 15d66ac20927e00ac0e65af6707eff60b92e61b8 Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Fri, 8 Jul 2016 09:53:20 -0400 Subject: xfs: abstract block export operations from nfsd layouts Instead of creeping pnfs layout configuration into filesystems, move the definition of block-based export operations under a more abstract configuration. Signed-off-by: Benjamin Coddington Reviewed-by: Christoph Hellwig Acked-by: Dave Chinner Signed-off-by: J. Bruce Fields --- fs/Kconfig | 6 ++++++ fs/nfsd/Kconfig | 2 ++ fs/xfs/Makefile | 3 +-- fs/xfs/xfs_export.c | 2 +- fs/xfs/xfs_pnfs.h | 4 ++-- 5 files changed, 12 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index b8fcb416be72..4d2f9210e748 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -67,6 +67,12 @@ config FS_POSIX_ACL config EXPORTFS tristate +config EXPORTFS_BLOCK_OPS + bool "Enable filesystem export operations for block IO" + help + This option enables the export operations for a filesystem to support + external block IO. + config FILE_LOCKING bool "Enable POSIX file locking API" if EXPERT default y diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig index e64d1a3ec770..47febcf99185 100644 --- a/fs/nfsd/Kconfig +++ b/fs/nfsd/Kconfig @@ -90,6 +90,7 @@ config NFSD_BLOCKLAYOUT bool "NFSv4.1 server support for pNFS block layouts" depends on NFSD_V4 && BLOCK select NFSD_PNFS + select EXPORTFS_BLOCK_OPS help This option enables support for the exporting pNFS block layouts in the kernel's NFS server. The pNFS block layout enables NFS @@ -102,6 +103,7 @@ config NFSD_SCSILAYOUT bool "NFSv4.1 server support for pNFS SCSI layouts" depends on NFSD_V4 && BLOCK select NFSD_PNFS + select EXPORTFS_BLOCK_OPS help This option enables support for the exporting pNFS SCSI layouts in the kernel's NFS server. The pNFS SCSI layout enables NFS diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 3542d94fddce..52c288514be1 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -121,5 +121,4 @@ xfs-$(CONFIG_XFS_RT) += xfs_rtalloc.o xfs-$(CONFIG_XFS_POSIX_ACL) += xfs_acl.o xfs-$(CONFIG_SYSCTL) += xfs_sysctl.o xfs-$(CONFIG_COMPAT) += xfs_ioctl32.o -xfs-$(CONFIG_NFSD_BLOCKLAYOUT) += xfs_pnfs.o -xfs-$(CONFIG_NFSD_SCSILAYOUT) += xfs_pnfs.o +xfs-$(CONFIG_EXPORTFS_BLOCK_OPS) += xfs_pnfs.o diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c index a1b2dd828b9d..fe1bfee35898 100644 --- a/fs/xfs/xfs_export.c +++ b/fs/xfs/xfs_export.c @@ -246,7 +246,7 @@ const struct export_operations xfs_export_operations = { .fh_to_parent = xfs_fs_fh_to_parent, .get_parent = xfs_fs_get_parent, .commit_metadata = xfs_fs_nfs_commit_metadata, -#ifdef CONFIG_NFSD_BLOCKLAYOUT +#ifdef CONFIG_EXPORTFS_BLOCK_OPS .get_uuid = xfs_fs_get_uuid, .map_blocks = xfs_fs_map_blocks, .commit_blocks = xfs_fs_commit_blocks, diff --git a/fs/xfs/xfs_pnfs.h b/fs/xfs/xfs_pnfs.h index 93f74853961b..e8339f74966b 100644 --- a/fs/xfs/xfs_pnfs.h +++ b/fs/xfs/xfs_pnfs.h @@ -1,7 +1,7 @@ #ifndef _XFS_PNFS_H #define _XFS_PNFS_H 1 -#if defined(CONFIG_NFSD_BLOCKLAYOUT) || defined(CONFIG_NFSD_SCSILAYOUT) +#ifdef CONFIG_EXPORTFS_BLOCK_OPS int xfs_fs_get_uuid(struct super_block *sb, u8 *buf, u32 *len, u64 *offset); int xfs_fs_map_blocks(struct inode *inode, loff_t offset, u64 length, struct iomap *iomap, bool write, u32 *device_generation); @@ -15,5 +15,5 @@ xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex) { return 0; } -#endif /* CONFIG_NFSD_PNFS */ +#endif /* CONFIG_EXPORTFS_BLOCK_OPS */ #endif /* _XFS_PNFS_H */ -- cgit v1.2.3 From dd51db188600c8c54c46cff84e83358ece8ccb4d Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Thu, 14 Jul 2016 12:01:17 +0800 Subject: nfsd/blocklayout: Make sure calculate signature/designator length aligned These values are all multiples of 4 already, so there's no change in behavior from this patch. But perhaps this will prevent mistakes in the future. Signed-off-by: Kinglong Mee Signed-off-by: J. Bruce Fields --- fs/nfsd/blocklayoutxdr.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/blocklayoutxdr.c b/fs/nfsd/blocklayoutxdr.c index 6c3b316f932e..d72467e5eb6e 100644 --- a/fs/nfsd/blocklayoutxdr.c +++ b/fs/nfsd/blocklayoutxdr.c @@ -43,7 +43,7 @@ nfsd4_block_encode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) switch (b->type) { case PNFS_BLOCK_VOLUME_SIMPLE: - len = 4 + 4 + 8 + 4 + b->simple.sig_len; + len = 4 + 4 + 8 + 4 + (XDR_QUADLEN(b->simple.sig_len) << 2); p = xdr_reserve_space(xdr, len); if (!p) return -ETOOSMALL; @@ -54,7 +54,7 @@ nfsd4_block_encode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) p = xdr_encode_opaque(p, b->simple.sig, b->simple.sig_len); break; case PNFS_BLOCK_VOLUME_SCSI: - len = 4 + 4 + 4 + 4 + b->scsi.designator_len + 8; + len = 4 + 4 + 4 + 4 + (XDR_QUADLEN(b->scsi.designator_len) << 2) + 8; p = xdr_reserve_space(xdr, len); if (!p) return -ETOOSMALL; -- cgit v1.2.3 From 885848186fbc2d1d8fb6d2fdc2156638ae289a46 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Wed, 13 Jul 2016 16:40:14 -0400 Subject: nfsd: Close race between nfsd4_release_lockowner and nfsd4_lock nfsd4_release_lockowner finds a lock owner that has no lock state, and drops cl_lock. Then release_lockowner picks up cl_lock and unhashes the lock owner. During the window where cl_lock is dropped, I don't see anything preventing a concurrent nfsd4_lock from finding that same lock owner and adding lock state to it. Move release_lockowner() into nfsd4_release_lockowner and hang onto the cl_lock until after the lock owner's state cannot be found again. Found by inspection, we don't currently have a reproducer. Fixes: 2c41beb0e5cf ("nfsd: reduce cl_lock thrashing in ... ") Reviewed-by: Jeff Layton Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 40 +++++++++++++++++----------------------- 1 file changed, 17 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index bc2173a31b5d..8410ca275db1 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1200,27 +1200,6 @@ free_ol_stateid_reaplist(struct list_head *reaplist) } } -static void release_lockowner(struct nfs4_lockowner *lo) -{ - struct nfs4_client *clp = lo->lo_owner.so_client; - struct nfs4_ol_stateid *stp; - struct list_head reaplist; - - INIT_LIST_HEAD(&reaplist); - - spin_lock(&clp->cl_lock); - unhash_lockowner_locked(lo); - while (!list_empty(&lo->lo_owner.so_stateids)) { - stp = list_first_entry(&lo->lo_owner.so_stateids, - struct nfs4_ol_stateid, st_perstateowner); - WARN_ON(!unhash_lock_stateid(stp)); - put_ol_stateid_locked(stp, &reaplist); - } - spin_unlock(&clp->cl_lock); - free_ol_stateid_reaplist(&reaplist); - nfs4_put_stateowner(&lo->lo_owner); -} - static void release_open_stateid_locks(struct nfs4_ol_stateid *open_stp, struct list_head *reaplist) { @@ -5963,6 +5942,7 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp, __be32 status; struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id); struct nfs4_client *clp; + LIST_HEAD (reaplist); dprintk("nfsd4_release_lockowner clientid: (%08x/%08x):\n", clid->cl_boot, clid->cl_id); @@ -5993,9 +5973,23 @@ nfsd4_release_lockowner(struct svc_rqst *rqstp, nfs4_get_stateowner(sop); break; } + if (!lo) { + spin_unlock(&clp->cl_lock); + return status; + } + + unhash_lockowner_locked(lo); + while (!list_empty(&lo->lo_owner.so_stateids)) { + stp = list_first_entry(&lo->lo_owner.so_stateids, + struct nfs4_ol_stateid, + st_perstateowner); + WARN_ON(!unhash_lock_stateid(stp)); + put_ol_stateid_locked(stp, &reaplist); + } spin_unlock(&clp->cl_lock); - if (lo) - release_lockowner(lo); + free_ol_stateid_reaplist(&reaplist); + nfs4_put_stateowner(&lo->lo_owner); + return status; } -- cgit v1.2.3 From 8a4c3926889e7bf226e9f0254e7eface1f85f312 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Sun, 10 Jul 2016 15:55:58 -0400 Subject: nfsd: allow nfsd to advertise multiple layout types If the underlying filesystem supports multiple layout types, then there is little reason not to advertise that fact to clients and let them choose what type to use. Turn the ex_layout_type field into a bitfield. For each supported layout type, we set a bit in that field. When the client requests a layout, ensure that the bit for that layout type is set. When the client requests attributes, send back a list of supported types. Signed-off-by: Jeff Layton Reviewed-by: Weston Andros Adamson Signed-off-by: J. Bruce Fields --- fs/nfsd/export.c | 4 ++-- fs/nfsd/export.h | 2 +- fs/nfsd/nfs4layouts.c | 6 +++--- fs/nfsd/nfs4proc.c | 4 ++-- fs/nfsd/nfs4xdr.c | 30 ++++++++++++++---------------- 5 files changed, 22 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 79de2f38dd63..43e109cc0ccc 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -706,7 +706,7 @@ static void svc_export_init(struct cache_head *cnew, struct cache_head *citem) new->ex_fslocs.locations = NULL; new->ex_fslocs.locations_count = 0; new->ex_fslocs.migrated = 0; - new->ex_layout_type = 0; + new->ex_layout_types = 0; new->ex_uuid = NULL; new->cd = item->cd; } @@ -731,7 +731,7 @@ static void export_update(struct cache_head *cnew, struct cache_head *citem) item->ex_fslocs.locations_count = 0; new->ex_fslocs.migrated = item->ex_fslocs.migrated; item->ex_fslocs.migrated = 0; - new->ex_layout_type = item->ex_layout_type; + new->ex_layout_types = item->ex_layout_types; new->ex_nflavors = item->ex_nflavors; for (i = 0; i < MAX_SECINFO_LIST; i++) { new->ex_flavors[i] = item->ex_flavors[i]; diff --git a/fs/nfsd/export.h b/fs/nfsd/export.h index 2e315072bf3f..730f15eeb7ed 100644 --- a/fs/nfsd/export.h +++ b/fs/nfsd/export.h @@ -57,7 +57,7 @@ struct svc_export { struct nfsd4_fs_locations ex_fslocs; uint32_t ex_nflavors; struct exp_flavor_info ex_flavors[MAX_SECINFO_LIST]; - enum pnfs_layouttype ex_layout_type; + u32 ex_layout_types; struct nfsd4_deviceid_map *ex_devid_map; struct cache_detail *cd; }; diff --git a/fs/nfsd/nfs4layouts.c b/fs/nfsd/nfs4layouts.c index 6d98d16b3354..2be9602b0221 100644 --- a/fs/nfsd/nfs4layouts.c +++ b/fs/nfsd/nfs4layouts.c @@ -139,21 +139,21 @@ void nfsd4_setup_layout_type(struct svc_export *exp) * otherwise advertise the block layout. */ #ifdef CONFIG_NFSD_FLEXFILELAYOUT - exp->ex_layout_type = LAYOUT_FLEX_FILES; + exp->ex_layout_types |= 1 << LAYOUT_FLEX_FILES; #endif #ifdef CONFIG_NFSD_BLOCKLAYOUT /* overwrite flex file layout selection if needed */ if (sb->s_export_op->get_uuid && sb->s_export_op->map_blocks && sb->s_export_op->commit_blocks) - exp->ex_layout_type = LAYOUT_BLOCK_VOLUME; + exp->ex_layout_types |= 1 << LAYOUT_BLOCK_VOLUME; #endif #ifdef CONFIG_NFSD_SCSILAYOUT /* overwrite block layout selection if needed */ if (sb->s_export_op->map_blocks && sb->s_export_op->commit_blocks && sb->s_bdev && sb->s_bdev->bd_disk->fops->pr_ops) - exp->ex_layout_type = LAYOUT_SCSI; + exp->ex_layout_types |= 1 << LAYOUT_SCSI; #endif } diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 89fad8aeffe2..e0c15f879d89 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -1219,12 +1219,12 @@ nfsd4_verify(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, static const struct nfsd4_layout_ops * nfsd4_layout_verify(struct svc_export *exp, unsigned int layout_type) { - if (!exp->ex_layout_type) { + if (!exp->ex_layout_types) { dprintk("%s: export does not support pNFS\n", __func__); return NULL; } - if (exp->ex_layout_type != layout_type) { + if (!(exp->ex_layout_types & (1 << layout_type))) { dprintk("%s: layout type %d not supported\n", __func__, layout_type); return NULL; diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 84ef94794496..0aa0236a1429 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -2162,22 +2162,20 @@ nfsd4_encode_aclname(struct xdr_stream *xdr, struct svc_rqst *rqstp, } static inline __be32 -nfsd4_encode_layout_type(struct xdr_stream *xdr, enum pnfs_layouttype layout_type) +nfsd4_encode_layout_types(struct xdr_stream *xdr, u32 layout_types) { - __be32 *p; + __be32 *p; + unsigned long i = hweight_long(layout_types); - if (layout_type) { - p = xdr_reserve_space(xdr, 8); - if (!p) - return nfserr_resource; - *p++ = cpu_to_be32(1); - *p++ = cpu_to_be32(layout_type); - } else { - p = xdr_reserve_space(xdr, 4); - if (!p) - return nfserr_resource; - *p++ = cpu_to_be32(0); - } + p = xdr_reserve_space(xdr, 4 + 4 * i); + if (!p) + return nfserr_resource; + + *p++ = cpu_to_be32(i); + + for (i = LAYOUT_NFSV4_1_FILES; i < LAYOUT_TYPE_MAX; ++i) + if (layout_types & (1 << i)) + *p++ = cpu_to_be32(i); return 0; } @@ -2752,13 +2750,13 @@ out_acl: } #ifdef CONFIG_NFSD_PNFS if (bmval1 & FATTR4_WORD1_FS_LAYOUT_TYPES) { - status = nfsd4_encode_layout_type(xdr, exp->ex_layout_type); + status = nfsd4_encode_layout_types(xdr, exp->ex_layout_types); if (status) goto out; } if (bmval2 & FATTR4_WORD2_LAYOUT_TYPES) { - status = nfsd4_encode_layout_type(xdr, exp->ex_layout_type); + status = nfsd4_encode_layout_types(xdr, exp->ex_layout_types); if (status) goto out; } -- cgit v1.2.3 From 0173ca0544b682b7b313269dc0600d4774098a14 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 8 Jul 2016 18:41:28 +0900 Subject: nfs/blocklayout: use proper fmode for opening block devices This was fixed for the original block layout code a while ago, but also needs to be fixed for the SCSI layout path. Signed-off-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index e5b89675263e..7fb9c07c078c 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -316,7 +316,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, return -EINVAL; } - d->bdev = blkdev_get_by_path(devname, FMODE_READ, NULL); + d->bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); if (IS_ERR(d->bdev)) { pr_warn("pNFS: failed to open device %s (%ld)\n", devname, PTR_ERR(d->bdev)); @@ -352,7 +352,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, return 0; out_blkdev_put: - blkdev_put(d->bdev, FMODE_READ); + blkdev_put(d->bdev, FMODE_READ | FMODE_WRITE); return error; } -- cgit v1.2.3 From d702d41ed41328487bd3b270467721222f8036e4 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 8 Jul 2016 18:41:29 +0900 Subject: nfs/blocklayout: refactor open-by-wwn The current code works with the standard udev/systemd names, but we'll have to add another method in the next patch. Refactor it into a separate helper to make room for the new variant. Signed-off-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 53 ++++++++++++++++++++++++------------------------ 1 file changed, 27 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index 7fb9c07c078c..ea70883a174a 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -287,44 +287,45 @@ bl_validate_designator(struct pnfs_block_volume *v) } } +/* + * Try to open the udev path for the WWN. At least on Debian the udev + * by-id path will always point to the dm-multipath device if one exists. + */ +static struct block_device * +bl_open_udev_path(struct pnfs_block_volume *v) +{ + struct block_device *bdev; + const char *devname; + + devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%*phN", + v->scsi.designator_len, v->scsi.designator); + if (!devname) + return ERR_PTR(-ENOMEM); + + bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); + if (IS_ERR(bdev)) { + pr_warn("pNFS: failed to open device %s (%ld)\n", + devname, PTR_ERR(bdev)); + } + + kfree(devname); + return bdev; +} + static int bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) { struct pnfs_block_volume *v = &volumes[idx]; const struct pr_ops *ops; - const char *devname; int error; if (!bl_validate_designator(v)) return -EINVAL; - switch (v->scsi.designator_len) { - case 8: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%8phN", - v->scsi.designator); - break; - case 12: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%12phN", - v->scsi.designator); - break; - case 16: - devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%16phN", - v->scsi.designator); - break; - default: - return -EINVAL; - } - - d->bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); - if (IS_ERR(d->bdev)) { - pr_warn("pNFS: failed to open device %s (%ld)\n", - devname, PTR_ERR(d->bdev)); - kfree(devname); + d->bdev = bl_open_udev_path(v); + if (IS_ERR(d->bdev)) return PTR_ERR(d->bdev); - } - - kfree(devname); d->len = i_size_read(d->bdev->bd_inode); d->map = bl_map_simple; -- cgit v1.2.3 From 11487ddbdb12c36e094bedcc5c906ff219905a03 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 8 Jul 2016 18:41:30 +0900 Subject: nfs/blocklayout: support RH/Fedora dm-mpath device nodes Instead of reusing the wwn-* names for multipath devices nodes RHEL and Fedora introduce new dm-mpath-uuid-* nodes with a slightly different naming scheme. Try these names first to ensure we always get a multipath-capable device if it exists. Signed-off-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 26 +++++++++++++++++++++++++- 1 file changed, 25 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index ea70883a174a..436bb303d856 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -312,6 +312,28 @@ bl_open_udev_path(struct pnfs_block_volume *v) return bdev; } +/* + * Try to open the RH/Fedora specific dm-mpath udev path for this WWN, as the + * wwn- links will only point to the first discovered SCSI device there. + */ +static struct block_device * +bl_open_dm_mpath_udev_path(struct pnfs_block_volume *v) +{ + struct block_device *bdev; + const char *devname; + + devname = kasprintf(GFP_KERNEL, + "/dev/disk/by-id/dm-uuid-mpath-%d%*phN", + v->scsi.designator_type, + v->scsi.designator_len, v->scsi.designator); + if (!devname) + return ERR_PTR(-ENOMEM); + + bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL); + kfree(devname); + return bdev; +} + static int bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) @@ -323,7 +345,9 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, if (!bl_validate_designator(v)) return -EINVAL; - d->bdev = bl_open_udev_path(v); + d->bdev = bl_open_dm_mpath_udev_path(v); + if (IS_ERR(d->bdev)) + d->bdev = bl_open_udev_path(v); if (IS_ERR(d->bdev)) return PTR_ERR(d->bdev); -- cgit v1.2.3 From ecc2b88c4a48cbabb3ed7617f55e92c123fb634b Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Thu, 14 Jul 2016 12:01:28 +0800 Subject: nfs/blocklayout: Make sure calculate signature length aligned Avoid a bad nfs server return an unaligned length of signature. Signed-off-by: Kinglong Mee Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index 436bb303d856..7d9f570e1b52 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -89,7 +89,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) memcpy(&b->simple.sigs[i].sig, p, b->simple.sigs[i].sig_len); - b->simple.len += 8 + 4 + b->simple.sigs[i].sig_len; + b->simple.len += 8 + 4 + \ + (XDR_QUADLEN(b->simple.sigs[i].sig_len) << 2); } break; case PNFS_BLOCK_VOLUME_SLICE: -- cgit v1.2.3 From c77efc1e7884c818ba67ec36b08e220202d9428c Mon Sep 17 00:00:00 2001 From: Kinglong Mee Date: Thu, 14 Jul 2016 12:02:01 +0800 Subject: nfs/blocklayout: Check max uuids and devices before decoding Avoid nfs return uuids/devices larger than maximum. Signed-off-by: Kinglong Mee Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index 7d9f570e1b52..118252fd1d64 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -65,8 +65,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) if (!p) return -EIO; b->simple.nr_sigs = be32_to_cpup(p++); - if (!b->simple.nr_sigs) { - dprintk("no signature\n"); + if (!b->simple.nr_sigs || b->simple.nr_sigs > PNFS_BLOCK_MAX_UUIDS) { + dprintk("Bad signature count: %d\n", b->simple.nr_sigs); return -EIO; } @@ -105,7 +105,12 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) p = xdr_inline_decode(xdr, 4); if (!p) return -EIO; + b->concat.volumes_count = be32_to_cpup(p++); + if (b->concat.volumes_count > PNFS_BLOCK_MAX_DEVICES) { + dprintk("Too many volumes: %d\n", b->concat.volumes_count); + return -EIO; + } p = xdr_inline_decode(xdr, b->concat.volumes_count * 4); if (!p) @@ -117,8 +122,13 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b) p = xdr_inline_decode(xdr, 8 + 4); if (!p) return -EIO; + p = xdr_decode_hyper(p, &b->stripe.chunk_size); b->stripe.volumes_count = be32_to_cpup(p++); + if (b->stripe.volumes_count > PNFS_BLOCK_MAX_DEVICES) { + dprintk("Too many volumes: %d\n", b->stripe.volumes_count); + return -EIO; + } p = xdr_inline_decode(xdr, b->stripe.volumes_count * 4); if (!p) -- cgit v1.2.3 From 10b7e9ad44881fcd46ac24eb7374377c6e8962ed Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 18 Jul 2016 00:51:01 -0400 Subject: pNFS: Don't mark the inode as revalidated if a LAYOUTCOMMIT is outstanding We know that the attributes will need updating if there is still a LAYOUTCOMMIT outstanding. Reported-by: Christoph Hellwig Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 5 ++++- fs/nfs/pnfs.h | 7 +++++++ 2 files changed, 11 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 35fda08dc4f6..9df45832e28b 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1664,7 +1664,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long now = jiffies; unsigned long save_cache_validity; bool have_writers = nfs_file_has_buffered_writers(nfsi); - bool cache_revalidated = true; + bool cache_revalidated; dfprintk(VFS, "NFS: %s(%s/%lu fh_crc=0x%08x ct=%d info=0x%x)\n", __func__, inode->i_sb->s_id, inode->i_ino, @@ -1713,6 +1713,9 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) /* Do atomic weak cache consistency updates */ invalid |= nfs_wcc_update_inode(inode, fattr); + + cache_revalidated = !pnfs_layoutcommit_outstanding(inode); + /* More cache consistency checks */ if (fattr->valid & NFS_ATTR_FATTR_CHANGE) { if (inode->i_version != fattr->change_attr) { diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index d6be5299a55a..181283c4ebc3 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -628,6 +628,13 @@ pnfs_sync_inode(struct inode *inode, bool datasync) return 0; } +static inline bool +pnfs_layoutcommit_outstanding(struct inode *inode) +{ + return false; +} + + static inline bool pnfs_roc(struct inode *ino) { -- cgit v1.2.3 From 56b38a1f7c781519eef09c1668a3c97ea911f86b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Jul 2016 18:34:12 -0400 Subject: pNFS: Fix post-layoutget error handling in pnfs_update_layout() The non-retry error path is currently broken and ends up releasing the reference to the layout twice. It also can end up clearing the NFS_LAYOUT_FIRST_LAYOUTGET flag twice, causing a race. In addition, the retry path will fail to decrement the plh_outstanding counter. Fixes: 183d9e7b112aa ("pnfs: rework LAYOUTGET retry handling") Cc: stable@vger.kernel.org # 4.7 Signed-off-by: Trond Myklebust Reviewed-by: Jeff Layton --- fs/nfs/pnfs.c | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 0fbe734cc38c..563f131c9abe 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1645,6 +1645,7 @@ lookup_again: lseg = send_layoutget(lo, ctx, &stateid, &arg, &timeout, gfp_flags); trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg, PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET); + atomic_dec(&lo->plh_outstanding); if (IS_ERR(lseg)) { switch(PTR_ERR(lseg)) { case -ERECALLCONFLICT: @@ -1652,26 +1653,26 @@ lookup_again: lseg = NULL; /* Fallthrough */ case -EAGAIN: - pnfs_put_layout_hdr(lo); - if (first) - pnfs_clear_first_layoutget(lo); - if (lseg) { - trace_pnfs_update_layout(ino, pos, count, - iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETRY); - goto lookup_again; - } - /* Fallthrough */ + break; default: if (!nfs_error_is_fatal(PTR_ERR(lseg))) { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); lseg = NULL; } + goto out_put_layout_hdr; + } + if (lseg) { + if (first) + pnfs_clear_first_layoutget(lo); + trace_pnfs_update_layout(ino, pos, count, + iomode, lo, lseg, PNFS_UPDATE_LAYOUT_RETRY); + pnfs_put_layout_hdr(lo); + goto lookup_again; } } else { pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode)); } - atomic_dec(&lo->plh_outstanding); out_put_layout_hdr: if (first) pnfs_clear_first_layoutget(lo); -- cgit v1.2.3 From e85d7ee42003314652ab3ae2c60e3b8cd793b65f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Jul 2016 18:46:24 -0400 Subject: pNFS: Separate handling of NFS4ERR_LAYOUTTRYLATER and RECALLCONFLICT They are not the same error, and need to be handled differently. Fixes: 183d9e7b112aa ("pnfs: rework LAYOUTGET retry handling") Cc: stable@vger.kernel.org # 4.7 Signed-off-by: Trond Myklebust Reviewed-by: Jeff Layton --- fs/nfs/nfs4proc.c | 23 ++++++++++++++--------- fs/nfs/pnfs.c | 1 + 2 files changed, 15 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 519368b98762..ee8efe0a5202 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -437,6 +437,7 @@ static int nfs4_do_handle_exception(struct nfs_server *server, case -NFS4ERR_DELAY: nfs_inc_server_stats(server, NFSIOS_DELAY); case -NFS4ERR_GRACE: + case -NFS4ERR_LAYOUTTRYLATER: case -NFS4ERR_RECALLCONFLICT: exception->delay = 1; return 0; @@ -7883,11 +7884,12 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, struct inode *inode = lgp->args.inode; struct nfs_server *server = NFS_SERVER(inode); struct pnfs_layout_hdr *lo; - int status = task->tk_status; + int nfs4err = task->tk_status; + int err, status = 0; dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status); - switch (status) { + switch (nfs4err) { case 0: goto out; @@ -7919,12 +7921,11 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, status = -EOVERFLOW; goto out; } - /* Fallthrough */ + status = -EBUSY; + break; case -NFS4ERR_RECALLCONFLICT: - nfs4_handle_exception(server, -NFS4ERR_RECALLCONFLICT, - exception); status = -ERECALLCONFLICT; - goto out; + break; case -NFS4ERR_EXPIRED: case -NFS4ERR_BAD_STATEID: exception->timeout = 0; @@ -7955,9 +7956,13 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, spin_unlock(&inode->i_lock); } - status = nfs4_handle_exception(server, status, exception); - if (exception->retry) - status = -EAGAIN; + err = nfs4_handle_exception(server, nfs4err, exception); + if (!status) { + if (exception->retry) + status = -EAGAIN; + else + status = err; + } out: dprintk("<-- %s\n", __func__); return status; diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 563f131c9abe..c50d4ebab5c5 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1648,6 +1648,7 @@ lookup_again: atomic_dec(&lo->plh_outstanding); if (IS_ERR(lseg)) { switch(PTR_ERR(lseg)) { + case -EBUSY: case -ERECALLCONFLICT: if (time_after(jiffies, giveup)) lseg = NULL; -- cgit v1.2.3 From 66b53f325876703b7ab815c482cd104609f8772c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Jul 2016 14:28:31 -0400 Subject: pNFS: Handle NFS4ERR_RECALLCONFLICT correctly in LAYOUTGET Instead of giving up altogether and falling back to doing I/O through the MDS, which may make the situation worse, wait for 2 lease periods for the callback to resolve itself, and then try destroying the existing layout. Only if this was an attempt at getting a first layout, do we give up altogether, as the server is clearly crazy. Fixes: 183d9e7b112aa ("pnfs: rework LAYOUTGET retry handling") Cc: stable@vger.kernel.org # 4.7 Signed-off-by: Trond Myklebust Reviewed-by: Jeff Layton --- fs/nfs/pnfs.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index c50d4ebab5c5..7d992362ff04 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1505,7 +1505,7 @@ pnfs_update_layout(struct inode *ino, struct pnfs_layout_segment *lseg = NULL; nfs4_stateid stateid; long timeout = 0; - unsigned long giveup = jiffies + rpc_get_timeout(server->client); + unsigned long giveup = jiffies + (clp->cl_lease_time << 1); bool first; if (!pnfs_enabled_sb(NFS_SERVER(ino))) { @@ -1649,9 +1649,18 @@ lookup_again: if (IS_ERR(lseg)) { switch(PTR_ERR(lseg)) { case -EBUSY: - case -ERECALLCONFLICT: if (time_after(jiffies, giveup)) lseg = NULL; + break; + case -ERECALLCONFLICT: + /* Huh? We hold no layouts, how is there a recall? */ + if (first) { + lseg = NULL; + break; + } + /* Destroy the existing layout and start over */ + if (time_after(jiffies, giveup)) + pnfs_destroy_layout(NFS_I(ino)); /* Fallthrough */ case -EAGAIN: break; -- cgit v1.2.3 From f7db0b283868411dc6bc8a223fd032b211d2d91f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 14 Jul 2016 15:14:02 -0400 Subject: pNFS: Fix LAYOUTGET handling of NFS4ERR_BAD_STATEID and NFS4ERR_EXPIRED We want to recover the open stateid if there is no layout stateid and/or the stateid argument matches an open stateid. Otherwise throw out the existing layout and recover from scratch, as the layout stateid is bad. Fixes: 183d9e7b112aa ("pnfs: rework LAYOUTGET retry handling") Cc: stable@vger.kernel.org # 4.7 Signed-off-by: Trond Myklebust Reviewed-by: Jeff Layton --- fs/nfs/nfs4proc.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ee8efe0a5202..a1a3b4c9a563 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7886,6 +7886,7 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, struct pnfs_layout_hdr *lo; int nfs4err = task->tk_status; int err, status = 0; + LIST_HEAD(head); dprintk("--> %s tk_status => %d\n", __func__, -task->tk_status); @@ -7930,30 +7931,25 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, case -NFS4ERR_BAD_STATEID: exception->timeout = 0; spin_lock(&inode->i_lock); - if (nfs4_stateid_match(&lgp->args.stateid, + lo = NFS_I(inode)->layout; + /* If the open stateid was bad, then recover it. */ + if (!lo || test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) || + nfs4_stateid_match_other(&lgp->args.stateid, &lgp->args.ctx->state->stateid)) { spin_unlock(&inode->i_lock); - /* If the open stateid was bad, then recover it. */ exception->state = lgp->args.ctx->state; break; } - lo = NFS_I(inode)->layout; - if (lo && !test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) && - nfs4_stateid_match_other(&lgp->args.stateid, &lo->plh_stateid)) { - LIST_HEAD(head); - - /* - * Mark the bad layout state as invalid, then retry - * with the current stateid. - */ - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); - spin_unlock(&inode->i_lock); - pnfs_free_lseg_list(&head); - status = -EAGAIN; - goto out; - } else - spin_unlock(&inode->i_lock); + + /* + * Mark the bad layout state as invalid, then retry + */ + set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); + pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); + spin_unlock(&inode->i_lock); + pnfs_free_lseg_list(&head); + status = -EAGAIN; + goto out; } err = nfs4_handle_exception(server, nfs4err, exception); -- cgit v1.2.3 From e68fd7c8071d541d3f2f7eed5814b63e865dd277 Mon Sep 17 00:00:00 2001 From: Steve Dickson Date: Wed, 25 May 2016 10:36:50 -0400 Subject: mount: use sec= that was specified on the command line When older servers return RPC_AUTH_NULL, it means the rpc creds will be ignored. In that case use the sec= that was specified instead of setting sec=null Fixes Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1112983 Signed-off-by: Steve Dickson Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 2137e0202f25..18d446e1a82b 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1684,6 +1684,7 @@ static int nfs_verify_authflavors(struct nfs_parsed_mount_data *args, { rpc_authflavor_t flavor = RPC_AUTH_MAXFLAVOR; unsigned int i; + int use_auth_null = false; /* * If the sec= mount option is used, the specified flavor or AUTH_NULL @@ -1691,14 +1692,21 @@ static int nfs_verify_authflavors(struct nfs_parsed_mount_data *args, * * AUTH_NULL has a special meaning when it's in the server list - it * means that the server will ignore the rpc creds, so any flavor - * can be used. + * can be used but still use the sec= that was specified. */ for (i = 0; i < count; i++) { flavor = server_authlist[i]; - if (nfs_auth_info_match(&args->auth_info, flavor) || - flavor == RPC_AUTH_NULL) + if (nfs_auth_info_match(&args->auth_info, flavor)) goto out; + + if (flavor == RPC_AUTH_NULL) + use_auth_null = true; + } + + if (use_auth_null) { + flavor = RPC_AUTH_NULL; + goto out; } dfprintk(MOUNT, -- cgit v1.2.3 From ce52914eb76efd62aa48d738cf845b37852bf920 Mon Sep 17 00:00:00 2001 From: Scott Mayhew Date: Tue, 7 Jun 2016 15:14:48 -0400 Subject: sunrpc: move NO_CRKEY_TIMEOUT to the auth->au_flags A generic_cred can be used to look up a unx_cred or a gss_cred, so it's not really safe to use the the generic_cred->acred->ac_flags to store the NO_CRKEY_TIMEOUT flag. A lookup for a unx_cred triggered while the KEY_EXPIRE_SOON flag is already set will cause both NO_CRKEY_TIMEOUT and KEY_EXPIRE_SOON to be set in the ac_flags, leaving the user associated with the auth_cred to be in a state where they're perpetually doing 4K NFS_FILE_SYNC writes. This can be reproduced as follows: 1. Mount two NFS filesystems, one with sec=krb5 and one with sec=sys. They do not need to be the same export, nor do they even need to be from the same NFS server. Also, v3 is fine. $ sudo mount -o v3,sec=krb5 server1:/export /mnt/krb5 $ sudo mount -o v3,sec=sys server2:/export /mnt/sys 2. As the normal user, before accessing the kerberized mount, kinit with a short lifetime (but not so short that renewing the ticket would leave you within the 4-minute window again by the time the original ticket expires), e.g. $ kinit -l 10m -r 60m 3. Do some I/O to the kerberized mount and verify that the writes are wsize, UNSTABLE: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 4. Wait until you're within 4 minutes of key expiry, then do some more I/O to the kerberized mount to ensure that RPC_CRED_KEY_EXPIRE_SOON gets set. Verify that the writes are 4K, FILE_SYNC: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 5. Now do some I/O to the sec=sys mount. This will cause RPC_CRED_NO_CRKEY_TIMEOUT to be set: $ dd if=/dev/zero of=/mnt/sys/file bs=1M count=1 6. Writes for that user will now be permanently 4K, FILE_SYNC for that user, regardless of which mount is being written to, until you reboot the client. Renewing the kerberos ticket (assuming it hasn't already expired) will have no effect. Grabbing a new kerberos ticket at this point will have no effect either. Move the flag to the auth->au_flags field (which is currently unused) and rename it slightly to reflect that it's no longer associated with the auth_cred->ac_flags. Add the rpc_auth to the arg list of rpcauth_cred_key_to_expire and check the au_flags there too. Finally, add the inode to the arg list of nfs_ctx_key_to_expire so we can determine the rpc_auth to pass to rpcauth_cred_key_to_expire. Signed-off-by: Scott Mayhew Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 4 ++-- fs/nfs/internal.h | 2 +- fs/nfs/write.c | 6 ++++-- include/linux/sunrpc/auth.h | 6 ++++-- net/sunrpc/auth.c | 4 +++- net/sunrpc/auth_generic.c | 9 +-------- net/sunrpc/auth_gss/auth_gss.c | 1 + net/sunrpc/auth_null.c | 1 + net/sunrpc/auth_unix.c | 1 + 9 files changed, 18 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 717a8d6af52d..6bcd8913e8a9 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -432,7 +432,7 @@ static int nfs_write_end(struct file *file, struct address_space *mapping, return status; NFS_I(mapping->host)->write_io += copied; - if (nfs_ctx_key_to_expire(ctx)) { + if (nfs_ctx_key_to_expire(ctx, mapping->host)) { status = nfs_wb_all(mapping->host); if (status < 0) return status; @@ -645,7 +645,7 @@ static int nfs_need_check_write(struct file *filp, struct inode *inode) ctx = nfs_file_open_context(filp); if (test_bit(NFS_CONTEXT_ERROR_WRITE, &ctx->flags) || - nfs_ctx_key_to_expire(ctx)) + nfs_ctx_key_to_expire(ctx, inode)) return 1; return 0; } diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index fa88609f85e3..d2260e67334f 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -497,7 +497,7 @@ void nfs_init_cinfo(struct nfs_commit_info *cinfo, struct inode *inode, struct nfs_direct_req *dreq); int nfs_key_timeout_notify(struct file *filp, struct inode *inode); -bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx); +bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode); void nfs_pageio_stop_mirroring(struct nfs_pageio_descriptor *pgio); #ifdef CONFIG_MIGRATION diff --git a/fs/nfs/write.c b/fs/nfs/write.c index e1c74d3db64d..0b949a06b297 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1195,9 +1195,11 @@ nfs_key_timeout_notify(struct file *filp, struct inode *inode) /* * Test if the open context credential key is marked to expire soon. */ -bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx) +bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode) { - return rpcauth_cred_key_to_expire(ctx->cred); + struct rpc_auth *auth = NFS_SERVER(inode)->client->cl_auth; + + return rpcauth_cred_key_to_expire(auth, ctx->cred); } /* diff --git a/include/linux/sunrpc/auth.h b/include/linux/sunrpc/auth.h index 899791573a40..f890a295a7ff 100644 --- a/include/linux/sunrpc/auth.h +++ b/include/linux/sunrpc/auth.h @@ -37,7 +37,6 @@ struct rpcsec_gss_info; /* auth_cred ac_flags bits */ enum { - RPC_CRED_NO_CRKEY_TIMEOUT = 0, /* underlying cred has no key timeout */ RPC_CRED_KEY_EXPIRE_SOON = 1, /* underlying cred key will expire soon */ RPC_CRED_NOTIFY_TIMEOUT = 2, /* nofity generic cred when underlying key will expire soon */ @@ -82,6 +81,9 @@ struct rpc_cred { #define RPCAUTH_CRED_MAGIC 0x0f4aa4f0 +/* rpc_auth au_flags */ +#define RPCAUTH_AUTH_NO_CRKEY_TIMEOUT 0x0001 /* underlying cred has no key timeout */ + /* * Client authentication handle */ @@ -196,7 +198,7 @@ void rpcauth_destroy_credcache(struct rpc_auth *); void rpcauth_clear_credcache(struct rpc_cred_cache *); int rpcauth_key_timeout_notify(struct rpc_auth *, struct rpc_cred *); -bool rpcauth_cred_key_to_expire(struct rpc_cred *); +bool rpcauth_cred_key_to_expire(struct rpc_auth *, struct rpc_cred *); char * rpcauth_stringify_acceptor(struct rpc_cred *); static inline diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c index 040ff627c18a..696eb39fc1cb 100644 --- a/net/sunrpc/auth.c +++ b/net/sunrpc/auth.c @@ -359,8 +359,10 @@ rpcauth_key_timeout_notify(struct rpc_auth *auth, struct rpc_cred *cred) EXPORT_SYMBOL_GPL(rpcauth_key_timeout_notify); bool -rpcauth_cred_key_to_expire(struct rpc_cred *cred) +rpcauth_cred_key_to_expire(struct rpc_auth *auth, struct rpc_cred *cred) { + if (auth->au_flags & RPCAUTH_AUTH_NO_CRKEY_TIMEOUT) + return false; if (!cred->cr_ops->crkey_to_expire) return false; return cred->cr_ops->crkey_to_expire(cred); diff --git a/net/sunrpc/auth_generic.c b/net/sunrpc/auth_generic.c index 54dd3fdead54..168219535a34 100644 --- a/net/sunrpc/auth_generic.c +++ b/net/sunrpc/auth_generic.c @@ -224,7 +224,7 @@ generic_key_timeout(struct rpc_auth *auth, struct rpc_cred *cred) /* Fast track for non crkey_timeout (no key) underlying credentials */ - if (test_bit(RPC_CRED_NO_CRKEY_TIMEOUT, &acred->ac_flags)) + if (auth->au_flags & RPCAUTH_AUTH_NO_CRKEY_TIMEOUT) return 0; /* Fast track for the normal case */ @@ -236,12 +236,6 @@ generic_key_timeout(struct rpc_auth *auth, struct rpc_cred *cred) if (IS_ERR(tcred)) return -EACCES; - if (!tcred->cr_ops->crkey_timeout) { - set_bit(RPC_CRED_NO_CRKEY_TIMEOUT, &acred->ac_flags); - ret = 0; - goto out_put; - } - /* Test for the almost error case */ ret = tcred->cr_ops->crkey_timeout(tcred); if (ret != 0) { @@ -257,7 +251,6 @@ generic_key_timeout(struct rpc_auth *auth, struct rpc_cred *cred) set_bit(RPC_CRED_NOTIFY_TIMEOUT, &acred->ac_flags); } -out_put: put_rpccred(tcred); return ret; } diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index e64ae93d5b4f..813a3cdfb573 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -1015,6 +1015,7 @@ gss_create_new(struct rpc_auth_create_args *args, struct rpc_clnt *clnt) auth = &gss_auth->rpc_auth; auth->au_cslack = GSS_CRED_SLACK >> 2; auth->au_rslack = GSS_VERF_SLACK >> 2; + auth->au_flags = 0; auth->au_ops = &authgss_ops; auth->au_flavor = flavor; atomic_set(&auth->au_count, 1); diff --git a/net/sunrpc/auth_null.c b/net/sunrpc/auth_null.c index 8d9eb4d5ddd8..4d17376b2acb 100644 --- a/net/sunrpc/auth_null.c +++ b/net/sunrpc/auth_null.c @@ -115,6 +115,7 @@ static struct rpc_auth null_auth = { .au_cslack = NUL_CALLSLACK, .au_rslack = NUL_REPLYSLACK, + .au_flags = RPCAUTH_AUTH_NO_CRKEY_TIMEOUT, .au_ops = &authnull_ops, .au_flavor = RPC_AUTH_NULL, .au_count = ATOMIC_INIT(0), diff --git a/net/sunrpc/auth_unix.c b/net/sunrpc/auth_unix.c index 9f65452b7cbc..a99278c984e8 100644 --- a/net/sunrpc/auth_unix.c +++ b/net/sunrpc/auth_unix.c @@ -228,6 +228,7 @@ static struct rpc_auth unix_auth = { .au_cslack = UNX_CALLSLACK, .au_rslack = NUL_REPLYSLACK, + .au_flags = RPCAUTH_AUTH_NO_CRKEY_TIMEOUT, .au_ops = &authunix_ops, .au_flavor = RPC_AUTH_UNIX, .au_count = ATOMIC_INIT(0), -- cgit v1.2.3 From d9c0ce0e45723a4924d22ac2e5e13c2dede76cad Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Fri, 10 Jun 2016 16:37:35 -0400 Subject: pnfs/blocklayout: put deviceid node after releasing bl_ext_lock The last put of deviceid nodes for SCSI layouts may sleep, so we shouldn't hold any spinlocks. Make sure we put them outside the bl_ext_lock. Signed-off-by: Benjamin Coddington Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/extent_tree.c | 27 +++++++++++++++++++++------ 1 file changed, 21 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/extent_tree.c b/fs/nfs/blocklayout/extent_tree.c index 720b3ff55fa9..992bcb19c11e 100644 --- a/fs/nfs/blocklayout/extent_tree.c +++ b/fs/nfs/blocklayout/extent_tree.c @@ -121,6 +121,16 @@ ext_try_to_merge_right(struct rb_root *root, struct pnfs_block_extent *be) return be; } +static void __ext_put_deviceids(struct list_head *head) +{ + struct pnfs_block_extent *be, *tmp; + + list_for_each_entry_safe(be, tmp, head, be_list) { + nfs4_put_deviceid_node(be->be_device); + kfree(be); + } +} + static void __ext_tree_insert(struct rb_root *root, struct pnfs_block_extent *new, bool merge_ok) @@ -163,7 +173,8 @@ free_new: } static int -__ext_tree_remove(struct rb_root *root, sector_t start, sector_t end) +__ext_tree_remove(struct rb_root *root, + sector_t start, sector_t end, struct list_head *tmp) { struct pnfs_block_extent *be; sector_t len1 = 0, len2 = 0; @@ -223,8 +234,7 @@ __ext_tree_remove(struct rb_root *root, sector_t start, sector_t end) struct pnfs_block_extent *next = ext_tree_next(be); rb_erase(&be->be_node, root); - nfs4_put_deviceid_node(be->be_device); - kfree(be); + list_add_tail(&be->be_list, tmp); be = next; } @@ -350,16 +360,18 @@ int ext_tree_remove(struct pnfs_block_layout *bl, bool rw, sector_t start, sector_t end) { int err, err2; + LIST_HEAD(tmp); spin_lock(&bl->bl_ext_lock); - err = __ext_tree_remove(&bl->bl_ext_ro, start, end); + err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp); if (rw) { - err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end); + err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end, &tmp); if (!err) err = err2; } spin_unlock(&bl->bl_ext_lock); + __ext_put_deviceids(&tmp); return err; } @@ -396,12 +408,13 @@ ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start, sector_t end = start + len; struct pnfs_block_extent *be; int err = 0; + LIST_HEAD(tmp); spin_lock(&bl->bl_ext_lock); /* * First remove all COW extents or holes from written to range. */ - err = __ext_tree_remove(&bl->bl_ext_ro, start, end); + err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp); if (err) goto out; @@ -459,6 +472,8 @@ ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start, } out: spin_unlock(&bl->bl_ext_lock); + + __ext_put_deviceids(&tmp); return err; } -- cgit v1.2.3 From 3fc75f12089eab6bf3f5350a5f760f241060bd5d Mon Sep 17 00:00:00 2001 From: Tigran Mkrtchyan Date: Mon, 13 Jun 2016 19:57:35 +0200 Subject: nfs4: clnt: respect noresvport when establishing connections to DSes result: $ mount -o vers=4.1 dcache-lab007:/ /pnfs $ cp /etc/profile /pnfs tcp 0 0 131.169.185.68:1005 131.169.191.141:32049 ESTABLISHED tcp 0 0 131.169.185.68:751 131.169.191.144:2049 ESTABLISHED $ $ mount -o vers=4.1,noresvport dcache-lab007:/ /pnfs $ cp /etc/profile /pnfs tcp 0 0 131.169.185.68:34894 131.169.191.141:32049 ESTABLISHED tcp 0 0 131.169.185.68:35722 131.169.191.144:2049 ESTABLISHED $ Signed-off-by: Tigran Mkrtchyan Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 2 +- fs/nfs/nfs4client.c | 6 +++++- fs/nfs/pnfs_nfs.c | 2 +- 3 files changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index d2260e67334f..b257a2eaf92c 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -186,7 +186,7 @@ extern struct nfs_server *nfs_clone_server(struct nfs_server *, rpc_authflavor_t); extern int nfs_wait_client_init_complete(const struct nfs_client *clp); extern void nfs_mark_client_ready(struct nfs_client *clp, int state); -extern struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, +extern struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 5fc7fbbfdcef..8d7d08d4f95f 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -844,12 +844,13 @@ error: * low timeout interval so that if a connection is lost, we retry through * the MDS. */ -struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, +struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, u32 minor_version, rpc_authflavor_t au_flavor) { struct rpc_timeout ds_timeout; + struct nfs_client *mds_clp = mds_srv->nfs_client; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, @@ -868,6 +869,9 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp, return ERR_PTR(-EINVAL); cl_init.hostname = buf; + if (mds_srv->flags & NFS_MOUNT_NORESVPORT) + __set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); + /* * Set an authflavor equual to the MDS value. Use the MDS nfs_client * cl_ipaddr so as to use the same EXCHANGE_ID co_ownerid as the MDS diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index b38e3c0dc790..5856b2c66234 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -690,7 +690,7 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, dprintk("%s: DS %s: trying address %s\n", __func__, ds->ds_remotestr, da->da_remotestr); - clp = nfs4_set_ds_client(mds_srv->nfs_client, + clp = nfs4_set_ds_client(mds_srv, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, minor_version, -- cgit v1.2.3 From b224f7cb635f0a1a0a80c1dae93699a2a1161604 Mon Sep 17 00:00:00 2001 From: Tigran Mkrtchyan Date: Mon, 13 Jun 2016 20:52:00 +0200 Subject: nfs4: flexfiles: respect noresvport when establishing connections to DSes Signed-off-by: Tigran Mkrtchyan Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 2 +- fs/nfs/nfs3client.c | 6 +++++- fs/nfs/pnfs_nfs.c | 4 ++-- 3 files changed, 8 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index b257a2eaf92c..f7e33a5984d8 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -195,7 +195,7 @@ extern struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, rpc_authflavor_t au_flavor); extern struct rpc_clnt *nfs4_find_or_create_ds_client(struct nfs_client *, struct inode *); -extern struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, +extern struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, rpc_authflavor_t au_flavor); diff --git a/fs/nfs/nfs3client.c b/fs/nfs/nfs3client.c index 0457b4129421..ee753547fb0a 100644 --- a/fs/nfs/nfs3client.c +++ b/fs/nfs/nfs3client.c @@ -76,12 +76,13 @@ struct nfs_server *nfs3_clone_server(struct nfs_server *source, * low timeout interval so that if a connection is lost, we retry through * the MDS. */ -struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, +struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, unsigned int ds_timeo, unsigned int ds_retrans, rpc_authflavor_t au_flavor) { struct rpc_timeout ds_timeout; + struct nfs_client *mds_clp = mds_srv->nfs_client; struct nfs_client_initdata cl_init = { .addr = ds_addr, .addrlen = ds_addrlen, @@ -100,6 +101,9 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_client *mds_clp, return ERR_PTR(-EINVAL); cl_init.hostname = buf; + if (mds_srv->flags & NFS_MOUNT_NORESVPORT) + set_bit(NFS_CS_NORESVPORT, &cl_init.init_flags); + /* Use the MDS nfs_client cl_ipaddr. */ nfs_init_timeout_values(&ds_timeout, ds_proto, ds_timeo, ds_retrans); clp = nfs_get_client(&cl_init, au_flavor); diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index 5856b2c66234..fe183fbc4b90 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -595,7 +595,7 @@ static void nfs4_clear_ds_conn_bit(struct nfs4_pnfs_ds *ds) } static struct nfs_client *(*get_v3_ds_connect)( - struct nfs_client *mds_clp, + struct nfs_server *mds_srv, const struct sockaddr *ds_addr, int ds_addrlen, int ds_proto, @@ -654,7 +654,7 @@ static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv, rpc_clnt_add_xprt(clp->cl_rpcclient, &xprt_args, rpc_clnt_test_and_add_xprt, NULL); } else - clp = get_v3_ds_connect(mds_srv->nfs_client, + clp = get_v3_ds_connect(mds_srv, (struct sockaddr *)&da->da_addr, da->da_addrlen, IPPROTO_TCP, timeo, retrans, au_flavor); -- cgit v1.2.3 From bac357dcec2956f01df9da5365be257741b534dc Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 20 Jul 2016 16:48:45 -0700 Subject: Btrfs: avoid deadlocks during reservations in btrfs_truncate_block The new enospc code makes it possible to deadlock if we don't use FLUSH_LIMIT during reservations inside a transaction. This enforces the correct flush type to avoid both deadlocks and assertions Signed-off-by: Chris Mason Signed-off-by: Josef Bacik --- fs/btrfs/extent-tree.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 3a129c42658e..9fcb8c97083b 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -5939,10 +5939,15 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) * the middle of a transaction commit. We also don't need the delalloc * mutex since we won't race with anybody. We need this mostly to make * lockdep shut its filthy mouth. + * + * If we have a transaction open (can happen if we call truncate_block + * from truncate), then we need FLUSH_LIMIT so we don't deadlock. */ if (btrfs_is_free_space_inode(inode)) { flush = BTRFS_RESERVE_NO_FLUSH; delalloc_lock = false; + } else if (current->journal_info) { + flush = BTRFS_RESERVE_FLUSH_LIMIT; } if (flush != BTRFS_RESERVE_NO_FLUSH && -- cgit v1.2.3 From 7f5458ec5c13b4489eff220b6429e571e2140aa6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:11:45 -0400 Subject: qstr: constify instances in logfs Signed-off-by: Al Viro --- fs/logfs/dir.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/logfs/dir.c b/fs/logfs/dir.c index 2d5336bd4efd..438b7c516fd6 100644 --- a/fs/logfs/dir.c +++ b/fs/logfs/dir.c @@ -156,7 +156,7 @@ static pgoff_t hash_index(u32 hash, int round) static struct page *logfs_get_dd_page(struct inode *dir, struct dentry *dentry) { - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; struct page *page; struct logfs_disk_dentry *dd; u32 hash = hash_32(name->name, name->len, 0); @@ -323,7 +323,7 @@ static int logfs_readdir(struct file *file, struct dir_context *ctx) return 0; } -static void logfs_set_name(struct logfs_disk_dentry *dd, struct qstr *name) +static void logfs_set_name(struct logfs_disk_dentry *dd, const struct qstr *name) { dd->namelen = cpu_to_be16(name->len); memcpy(dd->name, name->name, name->len); -- cgit v1.2.3 From b5cce521e849c998e169820c734731a468fcf2b4 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:17:26 -0400 Subject: qstr: constify instances in hfsplus Signed-off-by: Al Viro --- fs/hfsplus/catalog.c | 12 ++++++------ fs/hfsplus/hfsplus_fs.h | 10 +++++----- 2 files changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/hfsplus/catalog.c b/fs/hfsplus/catalog.c index fb707e8f423a..142534d3c2d5 100644 --- a/fs/hfsplus/catalog.c +++ b/fs/hfsplus/catalog.c @@ -40,7 +40,7 @@ int hfsplus_cat_bin_cmp_key(const hfsplus_btree_key *k1, /* Generates key for catalog file/folders record. */ int hfsplus_cat_build_key(struct super_block *sb, - hfsplus_btree_key *key, u32 parent, struct qstr *str) + hfsplus_btree_key *key, u32 parent, const struct qstr *str) { int len, err; @@ -174,7 +174,7 @@ static int hfsplus_cat_build_record(hfsplus_cat_entry *entry, static int hfsplus_fill_cat_thread(struct super_block *sb, hfsplus_cat_entry *entry, int type, - u32 parentid, struct qstr *str) + u32 parentid, const struct qstr *str) { int err; @@ -250,7 +250,7 @@ static void hfsplus_subfolders_dec(struct inode *dir) } int hfsplus_create_cat(u32 cnid, struct inode *dir, - struct qstr *str, struct inode *inode) + const struct qstr *str, struct inode *inode) { struct super_block *sb = dir->i_sb; struct hfs_find_data fd; @@ -318,7 +318,7 @@ err2: return err; } -int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str) +int hfsplus_delete_cat(u32 cnid, struct inode *dir, const struct qstr *str) { struct super_block *sb = dir->i_sb; struct hfs_find_data fd; @@ -415,8 +415,8 @@ out: } int hfsplus_rename_cat(u32 cnid, - struct inode *src_dir, struct qstr *src_name, - struct inode *dst_dir, struct qstr *dst_name) + struct inode *src_dir, const struct qstr *src_name, + struct inode *dst_dir, const struct qstr *dst_name) { struct super_block *sb = src_dir->i_sb; struct hfs_find_data src_fd, dst_fd; diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index fdc3446d934a..9ed1d9b0a543 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -445,17 +445,17 @@ int hfsplus_cat_case_cmp_key(const hfsplus_btree_key *k1, int hfsplus_cat_bin_cmp_key(const hfsplus_btree_key *k1, const hfsplus_btree_key *k2); int hfsplus_cat_build_key(struct super_block *sb, hfsplus_btree_key *key, - u32 parent, struct qstr *str); + u32 parent, const struct qstr *str); void hfsplus_cat_build_key_with_cnid(struct super_block *sb, hfsplus_btree_key *key, u32 parent); void hfsplus_cat_set_perms(struct inode *inode, struct hfsplus_perm *perms); int hfsplus_find_cat(struct super_block *sb, u32 cnid, struct hfs_find_data *fd); -int hfsplus_create_cat(u32 cnid, struct inode *dir, struct qstr *str, +int hfsplus_create_cat(u32 cnid, struct inode *dir, const struct qstr *str, struct inode *inode); -int hfsplus_delete_cat(u32 cnid, struct inode *dir, struct qstr *str); -int hfsplus_rename_cat(u32 cnid, struct inode *src_dir, struct qstr *src_name, - struct inode *dst_dir, struct qstr *dst_name); +int hfsplus_delete_cat(u32 cnid, struct inode *dir, const struct qstr *str); +int hfsplus_rename_cat(u32 cnid, struct inode *src_dir, const struct qstr *src_name, + struct inode *dst_dir, const struct qstr *dst_name); /* dir.c */ extern const struct inode_operations hfsplus_dir_inode_operations; -- cgit v1.2.3 From 71e939634de7ed6aad2978b55d131ec6f1f9aaa6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:22:29 -0400 Subject: qstr: constify instances in hfs Signed-off-by: Al Viro --- fs/hfs/catalog.c | 12 ++++++------ fs/hfs/hfs_fs.h | 14 +++++++------- fs/hfs/inode.c | 2 +- fs/hfs/trans.c | 2 +- 4 files changed, 15 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/hfs/catalog.c b/fs/hfs/catalog.c index 98cde8ba5dc2..8f4afd3f5108 100644 --- a/fs/hfs/catalog.c +++ b/fs/hfs/catalog.c @@ -20,7 +20,7 @@ * * Given the ID of the parent and the name build a search key. */ -void hfs_cat_build_key(struct super_block *sb, btree_key *key, u32 parent, struct qstr *name) +void hfs_cat_build_key(struct super_block *sb, btree_key *key, u32 parent, const struct qstr *name) { key->cat.reserved = 0; key->cat.ParID = cpu_to_be32(parent); @@ -64,7 +64,7 @@ static int hfs_cat_build_record(hfs_cat_rec *rec, u32 cnid, struct inode *inode) static int hfs_cat_build_thread(struct super_block *sb, hfs_cat_rec *rec, int type, - u32 parentid, struct qstr *name) + u32 parentid, const struct qstr *name) { rec->type = type; memset(rec->thread.reserved, 0, sizeof(rec->thread.reserved)); @@ -79,7 +79,7 @@ static int hfs_cat_build_thread(struct super_block *sb, * Add a new file or directory to the catalog B-tree and * return a (struct hfs_cat_entry) for it in '*result'. */ -int hfs_cat_create(u32 cnid, struct inode *dir, struct qstr *str, struct inode *inode) +int hfs_cat_create(u32 cnid, struct inode *dir, const struct qstr *str, struct inode *inode) { struct hfs_find_data fd; struct super_block *sb; @@ -210,7 +210,7 @@ int hfs_cat_find_brec(struct super_block *sb, u32 cnid, * Delete the indicated file or directory. * The associated thread is also removed unless ('with_thread'==0). */ -int hfs_cat_delete(u32 cnid, struct inode *dir, struct qstr *str) +int hfs_cat_delete(u32 cnid, struct inode *dir, const struct qstr *str) { struct super_block *sb; struct hfs_find_data fd; @@ -277,8 +277,8 @@ out: * If the destination exists it is removed and a * (struct hfs_cat_entry) for it is returned in '*result'. */ -int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name, - struct inode *dst_dir, struct qstr *dst_name) +int hfs_cat_move(u32 cnid, struct inode *src_dir, const struct qstr *src_name, + struct inode *dst_dir, const struct qstr *dst_name) { struct super_block *sb; struct hfs_find_data src_fd, dst_fd; diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index ee2f385811c8..e799ebe71b51 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -178,11 +178,11 @@ extern int hfs_clear_vbm_bits(struct super_block *, u16, u16); extern int hfs_cat_keycmp(const btree_key *, const btree_key *); struct hfs_find_data; extern int hfs_cat_find_brec(struct super_block *, u32, struct hfs_find_data *); -extern int hfs_cat_create(u32, struct inode *, struct qstr *, struct inode *); -extern int hfs_cat_delete(u32, struct inode *, struct qstr *); -extern int hfs_cat_move(u32, struct inode *, struct qstr *, - struct inode *, struct qstr *); -extern void hfs_cat_build_key(struct super_block *, btree_key *, u32, struct qstr *); +extern int hfs_cat_create(u32, struct inode *, const struct qstr *, struct inode *); +extern int hfs_cat_delete(u32, struct inode *, const struct qstr *); +extern int hfs_cat_move(u32, struct inode *, const struct qstr *, + struct inode *, const struct qstr *); +extern void hfs_cat_build_key(struct super_block *, btree_key *, u32, const struct qstr *); /* dir.c */ extern const struct file_operations hfs_dir_operations; @@ -201,7 +201,7 @@ extern int hfs_get_block(struct inode *, sector_t, struct buffer_head *, int); extern const struct address_space_operations hfs_aops; extern const struct address_space_operations hfs_btree_aops; -extern struct inode *hfs_new_inode(struct inode *, struct qstr *, umode_t); +extern struct inode *hfs_new_inode(struct inode *, const struct qstr *, umode_t); extern void hfs_inode_write_fork(struct inode *, struct hfs_extent *, __be32 *, __be32 *); extern int hfs_write_inode(struct inode *, struct writeback_control *); extern int hfs_inode_setattr(struct dentry *, struct iattr *); @@ -237,7 +237,7 @@ extern int hfs_compare_dentry(const struct dentry *parent, const struct dentry * unsigned int len, const char *str, const struct qstr *name); /* trans.c */ -extern void hfs_asc2mac(struct super_block *, struct hfs_name *, struct qstr *); +extern void hfs_asc2mac(struct super_block *, struct hfs_name *, const struct qstr *); extern int hfs_mac2asc(struct super_block *, char *, const struct hfs_name *); /* super.c */ diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c index 8eed66af5b82..6d6affd3ba5b 100644 --- a/fs/hfs/inode.c +++ b/fs/hfs/inode.c @@ -177,7 +177,7 @@ const struct address_space_operations hfs_aops = { /* * hfs_new_inode */ -struct inode *hfs_new_inode(struct inode *dir, struct qstr *name, umode_t mode) +struct inode *hfs_new_inode(struct inode *dir, const struct qstr *name, umode_t mode) { struct super_block *sb = dir->i_sb; struct inode *inode = new_inode(sb); diff --git a/fs/hfs/trans.c b/fs/hfs/trans.c index b1ce4c7ad3fb..39f5e343bf4d 100644 --- a/fs/hfs/trans.c +++ b/fs/hfs/trans.c @@ -94,7 +94,7 @@ out: * This routine is a inverse to hfs_mac2triv(). * A ':' is replaced by a '/'. */ -void hfs_asc2mac(struct super_block *sb, struct hfs_name *out, struct qstr *in) +void hfs_asc2mac(struct super_block *sb, struct hfs_name *out, const struct qstr *in) { struct nls_table *nls_disk = HFS_SB(sb)->nls_disk; struct nls_table *nls_io = HFS_SB(sb)->nls_io; -- cgit v1.2.3 From 8ac790f312c47bb378fb129d3832d401945a0ca3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:25:07 -0400 Subject: qstr: constify instances in autofs4 Signed-off-by: Al Viro --- fs/autofs4/root.c | 8 ++++---- fs/autofs4/waitq.c | 4 ++-- 2 files changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 3767f6641af1..fa84bb8832e0 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -159,7 +159,7 @@ static struct dentry *autofs4_lookup_active(struct dentry *dentry) { struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); struct dentry *parent = dentry->d_parent; - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; unsigned int len = name->len; unsigned int hash = name->hash; const unsigned char *str = name->name; @@ -172,7 +172,7 @@ static struct dentry *autofs4_lookup_active(struct dentry *dentry) list_for_each(p, head) { struct autofs_info *ino; struct dentry *active; - struct qstr *qstr; + const struct qstr *qstr; ino = list_entry(p, struct autofs_info, active); active = ino->dentry; @@ -214,7 +214,7 @@ static struct dentry *autofs4_lookup_expiring(struct dentry *dentry, { struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb); struct dentry *parent = dentry->d_parent; - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; unsigned int len = name->len; unsigned int hash = name->hash; const unsigned char *str = name->name; @@ -227,7 +227,7 @@ static struct dentry *autofs4_lookup_expiring(struct dentry *dentry, list_for_each(p, head) { struct autofs_info *ino; struct dentry *expiring; - struct qstr *qstr; + const struct qstr *qstr; if (rcu_walk) { spin_unlock(&sbi->lookup_lock); diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c index 631f1554c87b..432b0d5a3c75 100644 --- a/fs/autofs4/waitq.c +++ b/fs/autofs4/waitq.c @@ -225,7 +225,7 @@ rename_retry: } static struct autofs_wait_queue * -autofs4_find_wait(struct autofs_sb_info *sbi, struct qstr *qstr) +autofs4_find_wait(struct autofs_sb_info *sbi, const struct qstr *qstr) { struct autofs_wait_queue *wq; @@ -249,7 +249,7 @@ autofs4_find_wait(struct autofs_sb_info *sbi, struct qstr *qstr) */ static int validate_request(struct autofs_wait_queue **wait, struct autofs_sb_info *sbi, - struct qstr *qstr, + const struct qstr *qstr, struct dentry *dentry, enum autofs_notify notify) { struct autofs_wait_queue *wq; -- cgit v1.2.3 From 612645f7cfe1221f8066b87d62089719bcf43f68 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:27:42 -0400 Subject: qstr: constify instances in ocfs2 Signed-off-by: Al Viro --- fs/ocfs2/dlm/dlmdomain.c | 4 +--- fs/ocfs2/dlmfs/dlmfs.c | 4 ++-- fs/ocfs2/dlmfs/userdlm.c | 2 +- fs/ocfs2/dlmfs/userdlm.h | 2 +- 4 files changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 12e064b8be9a..533bd524e41e 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -172,12 +172,10 @@ void __dlm_unhash_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) void __dlm_insert_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { struct hlist_head *bucket; - struct qstr *q; assert_spin_locked(&dlm->spinlock); - q = &res->lockname; - bucket = dlm_lockres_hash(dlm, q->hash); + bucket = dlm_lockres_hash(dlm, res->lockname.hash); /* get a reference for our hashtable */ dlm_lockres_get(res); diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c index 47b3b2d4e775..ef474cdd6404 100644 --- a/fs/ocfs2/dlmfs/dlmfs.c +++ b/fs/ocfs2/dlmfs/dlmfs.c @@ -469,7 +469,7 @@ static int dlmfs_mkdir(struct inode * dir, { int status; struct inode *inode = NULL; - struct qstr *domain = &dentry->d_name; + const struct qstr *domain = &dentry->d_name; struct dlmfs_inode_private *ip; struct ocfs2_cluster_connection *conn; @@ -518,7 +518,7 @@ static int dlmfs_create(struct inode *dir, { int status = 0; struct inode *inode; - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; mlog(0, "create %.*s\n", name->len, name->name); diff --git a/fs/ocfs2/dlmfs/userdlm.c b/fs/ocfs2/dlmfs/userdlm.c index 0499e3fb7bdb..f70cda2f090d 100644 --- a/fs/ocfs2/dlmfs/userdlm.c +++ b/fs/ocfs2/dlmfs/userdlm.c @@ -667,7 +667,7 @@ void user_dlm_set_locking_protocol(void) ocfs2_stack_glue_set_max_proto_version(&user_dlm_lproto.lp_max_version); } -struct ocfs2_cluster_connection *user_dlm_register(struct qstr *name) +struct ocfs2_cluster_connection *user_dlm_register(const struct qstr *name) { int rc; struct ocfs2_cluster_connection *conn; diff --git a/fs/ocfs2/dlmfs/userdlm.h b/fs/ocfs2/dlmfs/userdlm.h index 3b42d79531d7..ede94a6e7fd3 100644 --- a/fs/ocfs2/dlmfs/userdlm.h +++ b/fs/ocfs2/dlmfs/userdlm.h @@ -83,7 +83,7 @@ void user_dlm_write_lvb(struct inode *inode, ssize_t user_dlm_read_lvb(struct inode *inode, char *val, unsigned int len); -struct ocfs2_cluster_connection *user_dlm_register(struct qstr *name); +struct ocfs2_cluster_connection *user_dlm_register(const struct qstr *name); void user_dlm_unregister(struct ocfs2_cluster_connection *conn); void user_dlm_set_locking_protocol(void); -- cgit v1.2.3 From beffb8feb6e26851165ee927e0eb963d6db334b0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 16:34:42 -0400 Subject: qstr: constify instances in nfs Signed-off-by: Al Viro --- fs/nfs/nfs3proc.c | 8 ++++---- fs/nfs/nfs4_fs.h | 5 +++-- fs/nfs/nfs4namespace.c | 4 ++-- fs/nfs/nfs4proc.c | 16 ++++++++-------- fs/nfs/proc.c | 8 ++++---- fs/nfs/unlink.c | 2 +- include/linux/nfs_xdr.h | 8 ++++---- 7 files changed, 26 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index cb28cceefebe..698be9361280 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -144,7 +144,7 @@ nfs3_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, } static int -nfs3_proc_lookup(struct inode *dir, struct qstr *name, +nfs3_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label) { @@ -404,7 +404,7 @@ out: } static int -nfs3_proc_remove(struct inode *dir, struct qstr *name) +nfs3_proc_remove(struct inode *dir, const struct qstr *name) { struct nfs_removeargs arg = { .fh = NFS_FH(dir), @@ -480,7 +480,7 @@ nfs3_proc_rename_done(struct rpc_task *task, struct inode *old_dir, } static int -nfs3_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) +nfs3_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name) { struct nfs3_linkargs arg = { .fromfh = NFS_FH(inode), @@ -582,7 +582,7 @@ out: } static int -nfs3_proc_rmdir(struct inode *dir, struct qstr *name) +nfs3_proc_rmdir(struct inode *dir, const struct qstr *name) { struct nfs_fattr *dir_attr; struct nfs3_diropargs arg = { diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 768456fa1b17..81e938554f10 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -224,7 +224,8 @@ int nfs_atomic_open(struct inode *, struct dentry *, struct file *, extern struct file_system_type nfs4_fs_type; /* nfs4namespace.c */ -struct rpc_clnt *nfs4_negotiate_security(struct rpc_clnt *, struct inode *, struct qstr *); +struct rpc_clnt *nfs4_negotiate_security(struct rpc_clnt *, struct inode *, + const struct qstr *); struct vfsmount *nfs4_submount(struct nfs_server *, struct dentry *, struct nfs_fh *, struct nfs_fattr *); int nfs4_replace_transport(struct nfs_server *server, @@ -251,7 +252,7 @@ extern int nfs4_proc_fs_locations(struct rpc_clnt *, struct inode *, const struc extern int nfs4_proc_get_locations(struct inode *, struct nfs4_fs_locations *, struct page *page, struct rpc_cred *); extern int nfs4_proc_fsid_present(struct inode *, struct rpc_cred *); -extern struct rpc_clnt *nfs4_proc_lookup_mountpoint(struct inode *, struct qstr *, +extern struct rpc_clnt *nfs4_proc_lookup_mountpoint(struct inode *, const struct qstr *, struct nfs_fh *, struct nfs_fattr *); extern int nfs4_proc_secinfo(struct inode *, const struct qstr *, struct nfs4_secinfo_flavors *); extern const struct xattr_handler *nfs4_xattr_handlers[]; diff --git a/fs/nfs/nfs4namespace.c b/fs/nfs/nfs4namespace.c index f592672373cb..d21104912676 100644 --- a/fs/nfs/nfs4namespace.c +++ b/fs/nfs/nfs4namespace.c @@ -208,7 +208,7 @@ static struct rpc_clnt *nfs_find_best_sec(struct rpc_clnt *clnt, */ struct rpc_clnt * nfs4_negotiate_security(struct rpc_clnt *clnt, struct inode *inode, - struct qstr *name) + const struct qstr *name) { struct page *page; struct nfs4_secinfo_flavors *flavors; @@ -397,7 +397,7 @@ struct vfsmount *nfs4_submount(struct nfs_server *server, struct dentry *dentry, rpc_authflavor_t flavor = server->client->cl_auth->au_flavor; struct dentry *parent = dget_parent(dentry); struct inode *dir = d_inode(parent); - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; struct rpc_clnt *client; struct vfsmount *mnt; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ff416d0e24bc..e75815c69d17 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3530,7 +3530,7 @@ static void nfs_fixup_secinfo_attributes(struct nfs_fattr *fattr) } static int nfs4_proc_lookup_common(struct rpc_clnt **clnt, struct inode *dir, - struct qstr *name, struct nfs_fh *fhandle, + const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label) { struct nfs4_exception exception = { }; @@ -3572,7 +3572,7 @@ out: return err; } -static int nfs4_proc_lookup(struct inode *dir, struct qstr *name, +static int nfs4_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label) { @@ -3588,7 +3588,7 @@ static int nfs4_proc_lookup(struct inode *dir, struct qstr *name, } struct rpc_clnt * -nfs4_proc_lookup_mountpoint(struct inode *dir, struct qstr *name, +nfs4_proc_lookup_mountpoint(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr) { struct rpc_clnt *client = NFS_CLIENT(dir); @@ -3747,7 +3747,7 @@ out: return status; } -static int _nfs4_proc_remove(struct inode *dir, struct qstr *name) +static int _nfs4_proc_remove(struct inode *dir, const struct qstr *name) { struct nfs_server *server = NFS_SERVER(dir); struct nfs_removeargs args = { @@ -3770,7 +3770,7 @@ static int _nfs4_proc_remove(struct inode *dir, struct qstr *name) return status; } -static int nfs4_proc_remove(struct inode *dir, struct qstr *name) +static int nfs4_proc_remove(struct inode *dir, const struct qstr *name) { struct nfs4_exception exception = { }; int err; @@ -3853,7 +3853,7 @@ static int nfs4_proc_rename_done(struct rpc_task *task, struct inode *old_dir, return 1; } -static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) +static int _nfs4_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name) { struct nfs_server *server = NFS_SERVER(inode); struct nfs4_link_arg arg = { @@ -3900,7 +3900,7 @@ out: return status; } -static int nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) +static int nfs4_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name) { struct nfs4_exception exception = { }; int err; @@ -3922,7 +3922,7 @@ struct nfs4_createdata { }; static struct nfs4_createdata *nfs4_alloc_createdata(struct inode *dir, - struct qstr *name, struct iattr *sattr, u32 ftype) + const struct qstr *name, struct iattr *sattr, u32 ftype) { struct nfs4_createdata *data; diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c index b417bbcd9704..b7bca8303989 100644 --- a/fs/nfs/proc.c +++ b/fs/nfs/proc.c @@ -145,7 +145,7 @@ nfs_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, } static int -nfs_proc_lookup(struct inode *dir, struct qstr *name, +nfs_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr, struct nfs4_label *label) { @@ -299,7 +299,7 @@ out: } static int -nfs_proc_remove(struct inode *dir, struct qstr *name) +nfs_proc_remove(struct inode *dir, const struct qstr *name) { struct nfs_removeargs arg = { .fh = NFS_FH(dir), @@ -357,7 +357,7 @@ nfs_proc_rename_done(struct rpc_task *task, struct inode *old_dir, } static int -nfs_proc_link(struct inode *inode, struct inode *dir, struct qstr *name) +nfs_proc_link(struct inode *inode, struct inode *dir, const struct qstr *name) { struct nfs_linkargs arg = { .fromfh = NFS_FH(inode), @@ -456,7 +456,7 @@ out: } static int -nfs_proc_rmdir(struct inode *dir, struct qstr *name) +nfs_proc_rmdir(struct inode *dir, const struct qstr *name) { struct nfs_diropargs arg = { .fh = NFS_FH(dir), diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c index 1868246f56e6..191aa577dd1f 100644 --- a/fs/nfs/unlink.c +++ b/fs/nfs/unlink.c @@ -162,7 +162,7 @@ static int nfs_call_unlink(struct dentry *dentry, struct nfs_unlinkdata *data) * @dentry: dentry to unlink */ static int -nfs_async_unlink(struct dentry *dentry, struct qstr *name) +nfs_async_unlink(struct dentry *dentry, const struct qstr *name) { struct nfs_unlinkdata *data; int status = -ENOMEM; diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index c304a11b5b1a..890b4aba2711 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1543,7 +1543,7 @@ struct nfs_rpc_ops { struct nfs_fattr *, struct nfs4_label *); int (*setattr) (struct dentry *, struct nfs_fattr *, struct iattr *); - int (*lookup) (struct inode *, struct qstr *, + int (*lookup) (struct inode *, const struct qstr *, struct nfs_fh *, struct nfs_fattr *, struct nfs4_label *); int (*access) (struct inode *, struct nfs_access_entry *); @@ -1551,18 +1551,18 @@ struct nfs_rpc_ops { unsigned int); int (*create) (struct inode *, struct dentry *, struct iattr *, int); - int (*remove) (struct inode *, struct qstr *); + int (*remove) (struct inode *, const struct qstr *); void (*unlink_setup) (struct rpc_message *, struct inode *dir); void (*unlink_rpc_prepare) (struct rpc_task *, struct nfs_unlinkdata *); int (*unlink_done) (struct rpc_task *, struct inode *); void (*rename_setup) (struct rpc_message *msg, struct inode *dir); void (*rename_rpc_prepare)(struct rpc_task *task, struct nfs_renamedata *); int (*rename_done) (struct rpc_task *task, struct inode *old_dir, struct inode *new_dir); - int (*link) (struct inode *, struct inode *, struct qstr *); + int (*link) (struct inode *, struct inode *, const struct qstr *); int (*symlink) (struct inode *, struct dentry *, struct page *, unsigned int, struct iattr *); int (*mkdir) (struct inode *, struct dentry *, struct iattr *); - int (*rmdir) (struct inode *, struct qstr *); + int (*rmdir) (struct inode *, const struct qstr *); int (*readdir) (struct dentry *, struct rpc_cred *, u64, struct page **, unsigned int, int); int (*mknod) (struct inode *, struct dentry *, struct iattr *, -- cgit v1.2.3 From 9aba36dea59264d11cd36219937296b4903c89b6 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 22:28:45 -0400 Subject: qstr constify instances in fs/dcache.c Signed-off-by: Al Viro --- fs/dcache.c | 4 ++-- include/linux/dcache.h | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index d6847d7b123d..f12d9d09604d 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2708,7 +2708,7 @@ EXPORT_SYMBOL(d_exact_alias); * Parent inode i_mutex must be held over d_lookup and into this call (to * keep renames and concurrent inserts, and readdir(2) away). */ -void dentry_update_name_case(struct dentry *dentry, struct qstr *name) +void dentry_update_name_case(struct dentry *dentry, const struct qstr *name) { BUG_ON(!inode_is_locked(dentry->d_parent->d_inode)); BUG_ON(dentry->d_name.len != name->len); /* d_lookup gives this */ @@ -3111,7 +3111,7 @@ static int prepend(char **buffer, int *buflen, const char *str, int namelen) * Data dependency barrier is needed to make sure that we see that terminating * NUL. Alpha strikes again, film at 11... */ -static int prepend_name(char **buffer, int *buflen, struct qstr *name) +static int prepend_name(char **buffer, int *buflen, const struct qstr *name) { const char *dname = ACCESS_ONCE(name->name); u32 dlen = ACCESS_ONCE(name->len); diff --git a/include/linux/dcache.h b/include/linux/dcache.h index f53fa055021a..f912a1e5c7cb 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -264,7 +264,7 @@ extern void d_rehash(struct dentry *); extern void d_add(struct dentry *, struct inode *); -extern void dentry_update_name_case(struct dentry *, struct qstr *); +extern void dentry_update_name_case(struct dentry *, const struct qstr *); /* used for rename() and baskets */ extern void d_move(struct dentry *, struct dentry *); -- cgit v1.2.3 From 8b8b08cbfb9021af4b54b4175fc4c51d655aac8c Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Tue, 19 Jul 2016 05:52:36 -0700 Subject: Btrfs: fix delalloc accounting after copy_from_user faults Commit 56244ef151c3cd11 was almost but not quite enough to fix the reservation math after btrfs_copy_from_user returned partial copies. Some users are still seeing warnings in btrfs_destroy_inode, and with a long enough test run I'm able to trigger them as well. This patch fixes the accounting math again, bringing it much closer to the way it was before the sectorsize conversion Chandan did. The problem is accounting for the offset into the page/sector when we do a partial copy. This one just uses the dirty_sectors variable which should already be updated properly. Signed-off-by: Chris Mason cc: stable@vger.kernel.org # v4.6+ --- fs/btrfs/file.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index f3f61d1ad18a..bcfb4a27ddd4 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1629,13 +1629,11 @@ again: * managed to copy. */ if (num_sectors > dirty_sectors) { - /* - * we round down because we don't want to count - * any partial blocks actually sent through the - * IO machines - */ - release_bytes = round_down(release_bytes - copied, - root->sectorsize); + + /* release everything except the sectors we dirtied */ + release_bytes -= dirty_sectors << + root->fs_info->sb->s_blocksize_bits; + if (copied > 0) { spin_lock(&BTRFS_I(inode)->lock); BTRFS_I(inode)->outstanding_extents++; -- cgit v1.2.3 From e033fb51ebb2983ee17b4a1b96ccbaedb137d9e9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 21 Jul 2016 09:43:43 -0400 Subject: pNFS/files: filelayout_write_done_cb must call nfs_writeback_update_inode() All write callbacks are required to call nfs_writeback_update_inode() upon success to ensure that file size changes are recorded, and the attribute cache is invalidated. Signed-off-by: Trond Myklebust --- fs/nfs/filelayout/filelayout.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/nfs/filelayout/filelayout.c b/fs/nfs/filelayout/filelayout.c index 25bd91a6e088..a3fc48ba4931 100644 --- a/fs/nfs/filelayout/filelayout.c +++ b/fs/nfs/filelayout/filelayout.c @@ -357,6 +357,12 @@ static int filelayout_write_done_cb(struct rpc_task *task, } filelayout_set_layoutcommit(hdr); + + /* zero out the fattr */ + hdr->fattr.valid = 0; + if (task->tk_status >= 0) + nfs_writeback_update_inode(hdr); + return 0; } -- cgit v1.2.3 From 297fae4d0bee5d683533f3324baf1b363e7b48bf Mon Sep 17 00:00:00 2001 From: Artem Savkov Date: Thu, 21 Jul 2016 13:32:04 +0200 Subject: Fix NULL pointer dereference in bl_free_device(). When bl_parse_deviceid() fails in bl_alloc_deviceid_node() on blkdev_get_by_*() step we get an pnfs_block_dev struct that is uninitialized except for bdev field which is set to whatever error blkdev_get_by_*() returns. bl_free_device() then tries to call blkdev_put() if bdev is not 0 resulting in a wrong pointer dereference. Fixing this by setting bdev in struct pnfs_block_dev only if we didn't get an error from blkdev_get_by_*(). Signed-off-by: Artem Savkov Reviewed-by: Benjamin Coddington Signed-off-by: Trond Myklebust --- fs/nfs/blocklayout/dev.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c index 118252fd1d64..a69ef4e9c24c 100644 --- a/fs/nfs/blocklayout/dev.c +++ b/fs/nfs/blocklayout/dev.c @@ -235,18 +235,20 @@ bl_parse_simple(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) { struct pnfs_block_volume *v = &volumes[idx]; + struct block_device *bdev; dev_t dev; dev = bl_resolve_deviceid(server, v, gfp_mask); if (!dev) return -EIO; - d->bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL); - if (IS_ERR(d->bdev)) { + bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL); + if (IS_ERR(bdev)) { printk(KERN_WARNING "pNFS: failed to open device %d:%d (%ld)\n", - MAJOR(dev), MINOR(dev), PTR_ERR(d->bdev)); - return PTR_ERR(d->bdev); + MAJOR(dev), MINOR(dev), PTR_ERR(bdev)); + return PTR_ERR(bdev); } + d->bdev = bdev; d->len = i_size_read(d->bdev->bd_inode); @@ -350,17 +352,19 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d, struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask) { struct pnfs_block_volume *v = &volumes[idx]; + struct block_device *bdev; const struct pr_ops *ops; int error; if (!bl_validate_designator(v)) return -EINVAL; - d->bdev = bl_open_dm_mpath_udev_path(v); - if (IS_ERR(d->bdev)) - d->bdev = bl_open_udev_path(v); - if (IS_ERR(d->bdev)) - return PTR_ERR(d->bdev); + bdev = bl_open_dm_mpath_udev_path(v); + if (IS_ERR(bdev)) + bdev = bl_open_udev_path(v); + if (IS_ERR(bdev)) + return PTR_ERR(bdev); + d->bdev = bdev; d->len = i_size_read(d->bdev->bd_inode); d->map = bl_map_simple; -- cgit v1.2.3 From 149a4fddd0a72d526abbeac0c8deaab03559836a Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Mon, 18 Jul 2016 10:41:57 -0400 Subject: nfs: don't create zero-length requests NFS doesn't expect requests with wb_bytes set to zero and may make unexpected decisions about how to handle that request at the page IO layer. Skip request creation if we won't have any wb_bytes in the request. Signed-off-by: Benjamin Coddington Signed-off-by: Alexey Dobriyan Reviewed-by: Weston Andros Adamson Cc: stable@vger.kernel.org Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 0b949a06b297..b5f3da346f1d 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1291,6 +1291,9 @@ int nfs_updatepage(struct file *file, struct page *page, dprintk("NFS: nfs_updatepage(%pD2 %d@%lld)\n", file, count, (long long)(page_file_offset(page) + offset)); + if (!count) + goto out; + if (nfs_can_extend_write(file, page, inode)) { count = max(count + offset, nfs_page_length(page)); offset = 0; @@ -1301,7 +1304,7 @@ int nfs_updatepage(struct file *file, struct page *page, nfs_set_pageerror(page); else __set_page_dirty_nobuffers(page); - +out: dprintk("NFS: nfs_updatepage returns %d (isize %lld)\n", status, (long long)i_size_read(inode)); return status; -- cgit v1.2.3 From 45fcc7bca7004687e9ba28e08b3dfb6787a0b466 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 12:26:34 -0400 Subject: pNFS: LAYOUTRETURN should only update the stateid if the layout is valid If the layout was completely returned, then ignore the returned layout stateid. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- fs/nfs/pnfs.h | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index ff416d0e24bc..dc50ba4d84ad 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8130,7 +8130,7 @@ static void nfs4_layoutreturn_release(void *calldata) pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range, be32_to_cpu(lrp->args.stateid.seqid)); pnfs_mark_layout_returned_if_empty(lo); - if (lrp->res.lrs_present) + if (lrp->res.lrs_present && pnfs_layout_is_valid(lo)) pnfs_set_layout_stateid(lo, &lrp->res.stateid, true); pnfs_clear_layoutreturn_waitbit(lo); spin_unlock(&lo->plh_inode->i_lock); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index b21bd0bee784..2f4f26905c03 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -375,6 +375,11 @@ static inline bool nfs_have_layout(struct inode *inode) return NFS_I(inode)->layout != NULL; } +static inline bool pnfs_layout_is_valid(const struct pnfs_layout_hdr *lo) +{ + return test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) == 0; +} + static inline struct nfs4_deviceid_node * nfs4_get_deviceid(struct nfs4_deviceid_node *d) { -- cgit v1.2.3 From 8e0acf9046b868a02e9afdddd25c8a132d50e99c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 21 Jul 2016 11:53:29 -0400 Subject: pNFS: Clear the layout return tracking on layout reinitialisation Ensure that we don't carry over layoutreturn info from a previous incarnation of this layout. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 0fbe734cc38c..dcccded1aeed 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -873,15 +873,21 @@ void pnfs_clear_layoutreturn_waitbit(struct pnfs_layout_hdr *lo) rpc_wake_up(&NFS_SERVER(lo->plh_inode)->roc_rpcwaitq); } +static void +pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo) +{ + lo->plh_return_iomode = 0; + lo->plh_return_seq = 0; + clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); +} + static bool pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo) { if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) return false; - lo->plh_return_iomode = 0; - lo->plh_return_seq = 0; pnfs_get_layout_hdr(lo); - clear_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); + pnfs_clear_layoutreturn_info(lo); return true; } @@ -1764,10 +1770,13 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) lo->plh_barrier = be32_to_cpu(res->stateid.seqid); } - clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_get_lseg(lseg); pnfs_layout_insert_lseg(lo, lseg, &free_me); + if (!pnfs_layout_is_valid(lo)) { + pnfs_clear_layoutreturn_info(lo); + clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); + } + if (res->return_on_close) set_bit(NFS_LSEG_ROC, &lseg->pls_flags); -- cgit v1.2.3 From 13bede18de41e2cfe8f67c1cd8b6d10be42ef473 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 11:39:03 -0400 Subject: pNFS: Always update the layout stateid if NFS_LAYOUT_INVALID_STID is set If the layout stateid is invalid, then pnfs_set_layout_stateid() must always initialise it. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index dcccded1aeed..eef844785bc6 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -762,7 +762,7 @@ pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new, bool update_barrier) { u32 oldseq, newseq, new_barrier; - int empty = list_empty(&lo->plh_segs); + bool empty = !pnfs_layout_is_valid(lo); oldseq = be32_to_cpu(lo->plh_stateid.seqid); newseq = be32_to_cpu(new->seqid); -- cgit v1.2.3 From ecebb80bf3ee8c5f3172f00bb17ba55f9e3ae24f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 11:46:06 -0400 Subject: pNFS: Always update the layout barrier seqid on LAYOUTGET Currently, pnfs_set_layout_stateid() will update the layout sequence id barrier only if the stateid itself is newer than the current layout stateid. However in a situation where multiple LAYOUTGET calls and a LAYOUTRETURN raced, it is entirely possible for one of the LAYOUTGET to set the current stateid to something newer than the LAYOUTRETURN that needs to set the barrier. The fix is to allow the "update_barrier" flag to force a check as to whether or not the barrier needs to be updated. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index eef844785bc6..85c3e7b47ddb 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -761,24 +761,25 @@ void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new, bool update_barrier) { - u32 oldseq, newseq, new_barrier; - bool empty = !pnfs_layout_is_valid(lo); + u32 oldseq, newseq, new_barrier = 0; + bool invalid = !pnfs_layout_is_valid(lo); oldseq = be32_to_cpu(lo->plh_stateid.seqid); newseq = be32_to_cpu(new->seqid); - if (empty || pnfs_seqid_is_newer(newseq, oldseq)) { + if (invalid || pnfs_seqid_is_newer(newseq, oldseq)) { nfs4_stateid_copy(&lo->plh_stateid, new); - if (update_barrier) { - new_barrier = be32_to_cpu(new->seqid); - } else { - /* Because of wraparound, we want to keep the barrier - * "close" to the current seqids. - */ - new_barrier = newseq - atomic_read(&lo->plh_outstanding); - } - if (empty || pnfs_seqid_is_newer(new_barrier, lo->plh_barrier)) - lo->plh_barrier = new_barrier; + /* + * Because of wraparound, we want to keep the barrier + * "close" to the current seqids. + */ + new_barrier = newseq - atomic_read(&lo->plh_outstanding); } + if (update_barrier) + new_barrier = be32_to_cpu(new->seqid); + else if (new_barrier == 0) + return; + if (invalid || pnfs_seqid_is_newer(new_barrier, lo->plh_barrier)) + lo->plh_barrier = new_barrier; } static bool -- cgit v1.2.3 From 793b7fe55858dca1f5bd3e42185b541a9eddc144 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 23 Jul 2016 21:11:43 -0400 Subject: pNFS: Fix CB_LAYOUTRECALL stateid verification We want to evaluate in this order: If the client holds no layout for this inode, then return NFS4ERR_NOMATCHING_LAYOUT; it probably forgot the layout. If the client finds the inode among the list of layouts, but the corresponding stateid has not yet been initialised, then return NFS4ERR_DELAY to ask the server to retry once the outstanding LAYOUTGET is complete. If the current layout stateid's "other" field does not match the recalled stateid, return NFS4ERR_BAD_STATEID. If already processing a layout recall with a newer stateid, return NFS4ERR_OLD_STATEID. This can only happens for servers that are non-compliant with the NFSv4.1 protocol. If already processing a layout recall with an older stateid, return NFS4ERR_DELAY to ask the server to retry once the outstanding LAYOUTRETURN is complete. Again, this is technically incompliant with the NFSv4.1 protocol. If the current layout sequence id is newer than the recalled stateid's sequence id, return NFS4ERR_OLD_STATEID. This too implies protocol non-compliance. If the current layout sequence id is older than the recalled stateid's sequence id+1, return NFS4ERR_DELAY. Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 63 +++++++++++++++++++++++++++++++++++--------------- 1 file changed, 44 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index aaa2e8d3df6f..837da8a02d35 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -119,27 +119,30 @@ out: * hashed by filehandle. */ static struct pnfs_layout_hdr * get_layout_by_fh_locked(struct nfs_client *clp, - struct nfs_fh *fh, nfs4_stateid *stateid) + struct nfs_fh *fh) { struct nfs_server *server; + struct nfs_inode *nfsi; struct inode *ino; struct pnfs_layout_hdr *lo; +restart: list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) { list_for_each_entry(lo, &server->layouts, plh_layouts) { - if (!nfs4_stateid_match_other(&lo->plh_stateid, stateid)) + nfsi = NFS_I(lo->plh_inode); + if (nfs_compare_fh(fh, &nfsi->fh)) continue; - if (nfs_compare_fh(fh, &NFS_I(lo->plh_inode)->fh)) + if (nfsi->layout != lo) continue; ino = igrab(lo->plh_inode); if (!ino) break; spin_lock(&ino->i_lock); /* Is this layout in the process of being freed? */ - if (NFS_I(ino)->layout != lo) { + if (nfsi->layout != lo) { spin_unlock(&ino->i_lock); iput(ino); - break; + goto restart; } pnfs_get_layout_hdr(lo); spin_unlock(&ino->i_lock); @@ -151,13 +154,13 @@ static struct pnfs_layout_hdr * get_layout_by_fh_locked(struct nfs_client *clp, } static struct pnfs_layout_hdr * get_layout_by_fh(struct nfs_client *clp, - struct nfs_fh *fh, nfs4_stateid *stateid) + struct nfs_fh *fh) { struct pnfs_layout_hdr *lo; spin_lock(&clp->cl_lock); rcu_read_lock(); - lo = get_layout_by_fh_locked(clp, fh, stateid); + lo = get_layout_by_fh_locked(clp, fh); rcu_read_unlock(); spin_unlock(&clp->cl_lock); @@ -167,17 +170,39 @@ static struct pnfs_layout_hdr * get_layout_by_fh(struct nfs_client *clp, /* * Enforce RFC5661 section 12.5.5.2.1. (Layout Recall and Return Sequencing) */ -static bool pnfs_check_stateid_sequence(struct pnfs_layout_hdr *lo, +static u32 pnfs_check_callback_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new) { u32 oldseq, newseq; - oldseq = be32_to_cpu(lo->plh_stateid.seqid); + /* Is the stateid still not initialised? */ + if (!pnfs_layout_is_valid(lo)) + return NFS4ERR_DELAY; + + /* Mismatched stateid? */ + if (!nfs4_stateid_match_other(&lo->plh_stateid, new)) + return NFS4ERR_BAD_STATEID; + newseq = be32_to_cpu(new->seqid); + /* Are we already in a layout recall situation? */ + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags) && + lo->plh_return_seq != 0) { + if (newseq < lo->plh_return_seq) + return NFS4ERR_OLD_STATEID; + if (newseq > lo->plh_return_seq) + return NFS4ERR_DELAY; + goto out; + } + /* Check that the stateid matches what we think it should be. */ + oldseq = be32_to_cpu(lo->plh_stateid.seqid); if (newseq > oldseq + 1) - return false; - return true; + return NFS4ERR_DELAY; + /* Crazy server! */ + if (newseq <= oldseq) + return NFS4ERR_OLD_STATEID; +out: + return NFS_OK; } static u32 initiate_file_draining(struct nfs_client *clp, @@ -188,7 +213,7 @@ static u32 initiate_file_draining(struct nfs_client *clp, u32 rv = NFS4ERR_NOMATCHING_LAYOUT; LIST_HEAD(free_me_list); - lo = get_layout_by_fh(clp, &args->cbl_fh, &args->cbl_stateid); + lo = get_layout_by_fh(clp, &args->cbl_fh); if (!lo) { trace_nfs4_cb_layoutrecall_file(clp, &args->cbl_fh, NULL, &args->cbl_stateid, -rv); @@ -196,18 +221,15 @@ static u32 initiate_file_draining(struct nfs_client *clp, } ino = lo->plh_inode; + pnfs_layoutcommit_inode(ino, false); + spin_lock(&ino->i_lock); - if (!pnfs_check_stateid_sequence(lo, &args->cbl_stateid)) { - rv = NFS4ERR_DELAY; + rv = pnfs_check_callback_stateid(lo, &args->cbl_stateid); + if (rv != NFS_OK) goto unlock; - } pnfs_set_layout_stateid(lo, &args->cbl_stateid, true); - spin_unlock(&ino->i_lock); - - pnfs_layoutcommit_inode(ino, false); - spin_lock(&ino->i_lock); /* * Enforce RFC5661 Section 12.5.5.2.1.5 (Bulk Recall and Return) */ @@ -223,6 +245,9 @@ static u32 initiate_file_draining(struct nfs_client *clp, goto unlock; } + /* Embrace your forgetfulness! */ + rv = NFS4ERR_NOMATCHING_LAYOUT; + if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) { NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo, &args->cbl_range); -- cgit v1.2.3 From e5fd1904b8422615a2a286777e2b7c881ad53e73 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 21 Jul 2016 12:44:15 -0400 Subject: pNFS: Ensure layoutreturn acts as a completion for layout callbacks When we return NFS_OK to the CB_LAYOUTRECALL, we are required to send a layoutreturn that "completes" that layout recall request, using the correct stateid. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 40 +++++++++++++++++++++++++--------------- 1 file changed, 25 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 85c3e7b47ddb..878dc4b7085a 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -883,12 +883,28 @@ pnfs_clear_layoutreturn_info(struct pnfs_layout_hdr *lo) } static bool -pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo) +pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo, + nfs4_stateid *stateid, + enum pnfs_iomode *iomode) { if (test_and_set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) return false; pnfs_get_layout_hdr(lo); - pnfs_clear_layoutreturn_info(lo); + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags)) { + if (stateid != NULL) { + nfs4_stateid_copy(stateid, &lo->plh_stateid); + if (lo->plh_return_seq != 0) + stateid->seqid = cpu_to_be32(lo->plh_return_seq); + } + if (iomode != NULL) + *iomode = lo->plh_return_iomode; + pnfs_clear_layoutreturn_info(lo); + return true; + } + if (stateid != NULL) + nfs4_stateid_copy(stateid, &lo->plh_stateid); + if (iomode != NULL) + *iomode = IOMODE_ANY; return true; } @@ -956,10 +972,7 @@ static void pnfs_layoutreturn_before_put_layout_hdr(struct pnfs_layout_hdr *lo) enum pnfs_iomode iomode; bool send; - nfs4_stateid_copy(&stateid, &lo->plh_stateid); - stateid.seqid = cpu_to_be32(lo->plh_return_seq); - iomode = lo->plh_return_iomode; - send = pnfs_prepare_layoutreturn(lo); + send = pnfs_prepare_layoutreturn(lo, &stateid, &iomode); spin_unlock(&inode->i_lock); if (send) { /* Send an async layoutreturn so we dont deadlock */ @@ -996,7 +1009,6 @@ _pnfs_return_layout(struct inode *ino) dprintk("NFS: %s no layout to return\n", __func__); goto out; } - nfs4_stateid_copy(&stateid, &nfsi->layout->plh_stateid); /* Reference matched in nfs4_layoutreturn_release */ pnfs_get_layout_hdr(lo); empty = list_empty(&lo->plh_segs); @@ -1020,7 +1032,7 @@ _pnfs_return_layout(struct inode *ino) } set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - send = pnfs_prepare_layoutreturn(lo); + send = pnfs_prepare_layoutreturn(lo, &stateid, NULL); spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&tmp_list); if (send) @@ -1087,11 +1099,10 @@ bool pnfs_roc(struct inode *ino) goto out_noroc; } - nfs4_stateid_copy(&stateid, &lo->plh_stateid); /* always send layoutreturn if being marked so */ - if (test_and_clear_bit(NFS_LAYOUT_RETURN_REQUESTED, - &lo->plh_flags)) - layoutreturn = pnfs_prepare_layoutreturn(lo); + if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags)) + layoutreturn = pnfs_prepare_layoutreturn(lo, + &stateid, NULL); list_for_each_entry_safe(lseg, tmp, &lo->plh_segs, pls_list) /* If we are sending layoutreturn, invalidate all valid lsegs */ @@ -1874,10 +1885,9 @@ void pnfs_error_mark_layout_for_return(struct inode *inode, if (!pnfs_mark_matching_lsegs_return(lo, &free_me, &range, lseg->pls_seq)) { nfs4_stateid stateid; - enum pnfs_iomode iomode = lo->plh_return_iomode; + enum pnfs_iomode iomode; - nfs4_stateid_copy(&stateid, &lo->plh_stateid); - return_now = pnfs_prepare_layoutreturn(lo); + return_now = pnfs_prepare_layoutreturn(lo, &stateid, &iomode); spin_unlock(&inode->i_lock); if (return_now) pnfs_send_layoutreturn(lo, &stateid, iomode, false); -- cgit v1.2.3 From 2d6cf5ab0b5d13d06c4b7920d6a12dbedf003190 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 21 Jul 2016 13:06:18 -0400 Subject: pNFS: Do not set plh_return_seq for non-callback related layoutreturns In cases where we need to send a layoutreturn in order to propagate an error, we should not tie that to a specific layout stateid. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 878dc4b7085a..c57cbddca760 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1808,14 +1808,14 @@ static void pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode, u32 seq) { - if (lo->plh_return_iomode == iomode) - return; - if (lo->plh_return_iomode != 0) + if (lo->plh_return_iomode != 0 && lo->plh_return_iomode != iomode) iomode = IOMODE_ANY; lo->plh_return_iomode = iomode; set_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags); - if (!lo->plh_return_seq || pnfs_seqid_is_newer(seq, lo->plh_return_seq)) + if (seq != 0) { + WARN_ON_ONCE(lo->plh_return_seq != 0 && lo->plh_return_seq != seq); lo->plh_return_seq = seq; + } } /** @@ -1876,14 +1876,13 @@ void pnfs_error_mark_layout_for_return(struct inode *inode, bool return_now = false; spin_lock(&inode->i_lock); - pnfs_set_plh_return_info(lo, range.iomode, lseg->pls_seq); + pnfs_set_plh_return_info(lo, range.iomode, 0); /* * mark all matching lsegs so that we are sure to have no live * segments at hand when sending layoutreturn. See pnfs_put_lseg() * for how it works. */ - if (!pnfs_mark_matching_lsegs_return(lo, &free_me, - &range, lseg->pls_seq)) { + if (!pnfs_mark_matching_lsegs_return(lo, &free_me, &range, 0)) { nfs4_stateid stateid; enum pnfs_iomode iomode; -- cgit v1.2.3 From e036f46453f252539cb62bf91d82c3d08e37e73c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 22 Jul 2016 11:13:22 -0400 Subject: NFS: pnfs_mark_matching_lsegs_return() should match the layout sequence id When determining which layout segments to return, we do want pnfs_mark_matching_lsegs_return to check that they match the layout sequence id. This ensures that we don't waste time if the server is replaying a layout recall that has already been satisfied. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 37 +++++++++++++++++++++++-------------- 1 file changed, 23 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index c57cbddca760..52b2a4dfdcb0 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -486,15 +486,6 @@ pnfs_lseg_range_intersecting(const struct pnfs_layout_range *l1, (end2 == NFS4_MAX_UINT64 || end2 > start1); } -static bool -should_free_lseg(const struct pnfs_layout_range *lseg_range, - const struct pnfs_layout_range *recall_range) -{ - return (recall_range->iomode == IOMODE_ANY || - lseg_range->iomode == recall_range->iomode) && - pnfs_lseg_range_intersecting(lseg_range, recall_range); -} - static bool pnfs_lseg_dec_and_remove_zero(struct pnfs_layout_segment *lseg, struct list_head *tmp_list) { @@ -533,6 +524,27 @@ static bool pnfs_seqid_is_newer(u32 s1, u32 s2) return (s32)(s1 - s2) > 0; } +static bool +pnfs_should_free_range(const struct pnfs_layout_range *lseg_range, + const struct pnfs_layout_range *recall_range) +{ + return (recall_range->iomode == IOMODE_ANY || + lseg_range->iomode == recall_range->iomode) && + pnfs_lseg_range_intersecting(lseg_range, recall_range); +} + +static bool +pnfs_match_lseg_recall(const struct pnfs_layout_segment *lseg, + const struct pnfs_layout_range *recall_range, + u32 seq) +{ + if (seq != 0 && pnfs_seqid_is_newer(lseg->pls_seq, seq)) + return false; + if (recall_range == NULL) + return true; + return pnfs_should_free_range(&lseg->pls_range, recall_range); +} + /** * pnfs_mark_matching_lsegs_invalid - tear down lsegs or mark them for later * @lo: layout header containing the lsegs @@ -562,10 +574,7 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, if (list_empty(&lo->plh_segs)) return 0; list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) - if (!recall_range || - should_free_lseg(&lseg->pls_range, recall_range)) { - if (seq && pnfs_seqid_is_newer(lseg->pls_seq, seq)) - continue; + if (pnfs_match_lseg_recall(lseg, recall_range, seq)) { dprintk("%s: freeing lseg %p iomode %d seq %u" "offset %llu length %llu\n", __func__, lseg, lseg->pls_range.iomode, lseg->pls_seq, @@ -1845,7 +1854,7 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, assert_spin_locked(&lo->plh_inode->i_lock); list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) - if (should_free_lseg(&lseg->pls_range, return_range)) { + if (pnfs_match_lseg_recall(lseg, return_range, seq)) { dprintk("%s: marking lseg %p iomode %d " "offset %llu length %llu\n", __func__, lseg, lseg->pls_range.iomode, -- cgit v1.2.3 From 5f46be049b0dfdd36188058f044c165e1d4b8f56 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 22 Jul 2016 11:25:27 -0400 Subject: pNFS: Cleanup - don't open code pnfs_mark_layout_stateid_invalid() Ensure nfs42_layoutstat_done() layoutget don't open code layout stateid invalidation. Signed-off-by: Trond Myklebust --- fs/nfs/nfs42proc.c | 3 +-- fs/nfs/nfs4proc.c | 3 +-- fs/nfs/pnfs.c | 2 +- fs/nfs/pnfs.h | 2 ++ 4 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c index aa03ed09ba06..6ea5ad6f0d44 100644 --- a/fs/nfs/nfs42proc.c +++ b/fs/nfs/nfs42proc.c @@ -336,8 +336,7 @@ nfs42_layoutstat_done(struct rpc_task *task, void *calldata) * Mark the bad layout state as invalid, then retry * with the current stateid. */ - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); + pnfs_mark_layout_stateid_invalid(lo, &head); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); } else diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index dc50ba4d84ad..8c453515d98c 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7931,8 +7931,7 @@ nfs4_layoutget_handle_exception(struct rpc_task *task, * Mark the bad layout state as invalid, then retry * with the current stateid. */ - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); - pnfs_mark_matching_lsegs_invalid(lo, &head, NULL, 0); + pnfs_mark_layout_stateid_invalid(lo, &head); spin_unlock(&inode->i_lock); pnfs_free_lseg_list(&head); status = -EAGAIN; diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 52b2a4dfdcb0..1374fcd40484 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -259,7 +259,7 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo) * is required. * Note that caller must hold inode->i_lock. */ -static int +int pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo, struct list_head *lseg_list) { diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 2f4f26905c03..d71c9493693a 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -268,6 +268,8 @@ int pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, const struct pnfs_layout_range *recall_range, u32 seq); +int pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo, + struct list_head *lseg_list); bool pnfs_roc(struct inode *ino); void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); -- cgit v1.2.3 From d9b61708fe6dcab7b50abfb87cb47690d00269a1 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 15:04:07 -0400 Subject: pNFS: Clear the layout metadata if the server changed the layout stateid If the server changed the layout stateid's "other" field, then we should treat the old layout as being completely gone. In that case, we want to clear the metadata such as scheduled layoutreturns. Do this by calling pnfs_mark_layout_stateid_invalid(). Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 1374fcd40484..e16997da157f 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1785,7 +1785,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) * inode invalid, and don't bother validating the stateid * sequence number. */ - pnfs_mark_matching_lsegs_invalid(lo, &free_me, NULL, 0); + pnfs_mark_layout_stateid_invalid(lo, &free_me); nfs4_stateid_copy(&lo->plh_stateid, &res->stateid); lo->plh_barrier = be32_to_cpu(res->stateid.seqid); -- cgit v1.2.3 From f71dfe8fc947e04a9e3d10723d4ec4c5c44ce0f5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 12:45:47 -0400 Subject: pNFS: Remove redundant pnfs_mark_layout_returned_if_empty() That's already being taken care of in pnfs_layout_remove_lseg(). Signed-off-by: Trond Myklebust --- fs/nfs/callback_proc.c | 1 - fs/nfs/nfs4proc.c | 1 - fs/nfs/pnfs.c | 1 - fs/nfs/pnfs.h | 13 ------------- 4 files changed, 16 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 837da8a02d35..c92a75e066a6 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -252,7 +252,6 @@ static u32 initiate_file_draining(struct nfs_client *clp, NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo, &args->cbl_range); } - pnfs_mark_layout_returned_if_empty(lo); unlock: spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&free_me_list); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 8c453515d98c..d38cc5f0282f 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8128,7 +8128,6 @@ static void nfs4_layoutreturn_release(void *calldata) spin_lock(&lo->plh_inode->i_lock); pnfs_mark_matching_lsegs_invalid(lo, &freeme, &lrp->args.range, be32_to_cpu(lrp->args.stateid.seqid)); - pnfs_mark_layout_returned_if_empty(lo); if (lrp->res.lrs_present && pnfs_layout_is_valid(lo)) pnfs_set_layout_stateid(lo, &lrp->res.stateid, true); pnfs_clear_layoutreturn_waitbit(lo); diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index e16997da157f..49e952968ede 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1159,7 +1159,6 @@ void pnfs_roc_set_barrier(struct inode *ino, u32 barrier) spin_lock(&ino->i_lock); lo = NFS_I(ino)->layout; - pnfs_mark_layout_returned_if_empty(lo); if (pnfs_seqid_is_newer(barrier, lo->plh_barrier)) lo->plh_barrier = barrier; spin_unlock(&ino->i_lock); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index d71c9493693a..595648a3ad7a 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -552,19 +552,6 @@ pnfs_calc_offset_length(u64 offset, u64 end) return 1 + end - offset; } -/** - * pnfs_mark_layout_returned_if_empty - marks the layout as returned - * @lo: layout header - * - * Note: Caller must hold inode->i_lock - */ -static inline void -pnfs_mark_layout_returned_if_empty(struct pnfs_layout_hdr *lo) -{ - if (list_empty(&lo->plh_segs)) - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); -} - static inline void pnfs_copy_range(struct pnfs_layout_range *dst, const struct pnfs_layout_range *src) -- cgit v1.2.3 From 28c1acffead59a461e552708cc9daa9c1cb5a085 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 21 Jul 2016 14:45:19 -0400 Subject: pNFS: Remove redundant stateid invalidation The layout stateid will be invalidated once it holds no more layout segments anyway. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 49e952968ede..c9a684c128b8 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1040,7 +1040,6 @@ _pnfs_return_layout(struct inode *ino) goto out_put_layout_hdr; } - set_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags); send = pnfs_prepare_layoutreturn(lo, &stateid, NULL); spin_unlock(&ino->i_lock); pnfs_free_lseg_list(&tmp_list); -- cgit v1.2.3 From 119cef97a46e2a4b4fbebcf6655358a0b277732d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 15:10:12 -0400 Subject: pNFS: Cleanup - do layout segment initialisation in one place ...instead of splitting the initialisation over init_lseg() and pnfs_layout_process(). Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index c9a684c128b8..09b77a68422f 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -334,7 +334,9 @@ pnfs_layout_io_test_failed(struct pnfs_layout_hdr *lo, u32 iomode) } static void -init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg) +pnfs_init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg, + const struct pnfs_layout_range *range, + const nfs4_stateid *stateid) { INIT_LIST_HEAD(&lseg->pls_list); INIT_LIST_HEAD(&lseg->pls_lc_list); @@ -342,6 +344,8 @@ init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg) smp_mb(); set_bit(NFS_LSEG_VALID, &lseg->pls_flags); lseg->pls_layout = lo; + lseg->pls_range = *range; + lseg->pls_seq = be32_to_cpu(stateid->seqid); } static void pnfs_free_lseg(struct pnfs_layout_segment *lseg) @@ -1760,9 +1764,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) return lseg; } - init_lseg(lo, lseg); - lseg->pls_range = res->range; - lseg->pls_seq = be32_to_cpu(res->stateid.seqid); + pnfs_init_lseg(lo, lseg, &res->range, &res->stateid); spin_lock(&ino->i_lock); if (pnfs_layoutgets_blocked(lo)) { -- cgit v1.2.3 From 01d7b29f0edc1a19ff2b960dae1c7dade3bb1753 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 15:14:44 -0400 Subject: pNFS: Remove redundant smp_mb() from pnfs_init_lseg() It's not visible yet, and won't be until after we grab the inode->i_lock. Signed-off-by: Trond Myklebust --- fs/nfs/pnfs.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 09b77a68422f..1d47a845d77a 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -341,7 +341,6 @@ pnfs_init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg, INIT_LIST_HEAD(&lseg->pls_list); INIT_LIST_HEAD(&lseg->pls_lc_list); atomic_set(&lseg->pls_refcount, 1); - smp_mb(); set_bit(NFS_LSEG_VALID, &lseg->pls_flags); lseg->pls_layout = lo; lseg->pls_range = *range; -- cgit v1.2.3 From 139978239b53a981b45c53b835c020015c6c819e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 17:10:52 -0400 Subject: NFSv4: Fix warning "no previous prototype for ‘nfs4_listxattr’" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Make it static Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index d30f88c667b7..4b1491da16eb 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -8848,7 +8848,7 @@ const struct nfs4_minor_version_ops *nfs_v4_minor_ops[] = { #endif }; -ssize_t nfs4_listxattr(struct dentry *dentry, char *list, size_t size) +static ssize_t nfs4_listxattr(struct dentry *dentry, char *list, size_t size) { ssize_t error, error2; -- cgit v1.2.3 From 6fdf339b0ca73abd879394ad03a9e4695d644e13 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sun, 24 Jul 2016 17:17:16 -0400 Subject: NFSv4.2: Fix warning "variable ‘stateids’ set but not used" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Replace it with a test for whether or not the sent a stateid in violation of what we asked for. Signed-off-by: Trond Myklebust --- fs/nfs/nfs42xdr.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs42xdr.c b/fs/nfs/nfs42xdr.c index 6dc6f2aea0d6..8b2605882a20 100644 --- a/fs/nfs/nfs42xdr.c +++ b/fs/nfs/nfs42xdr.c @@ -330,13 +330,21 @@ static int decode_write_response(struct xdr_stream *xdr, struct nfs42_write_res *res) { __be32 *p; - int stateids; p = xdr_inline_decode(xdr, 4 + 8 + 4); if (unlikely(!p)) goto out_overflow; - stateids = be32_to_cpup(p++); + /* + * We never use asynchronous mode, so warn if a server returns + * a stateid. + */ + if (unlikely(*p != 0)) { + pr_err_once("%s: server has set unrequested " + "asynchronous mode\n", __func__); + return -EREMOTEIO; + } + p++; p = xdr_decode_hyper(p, &res->count); res->verifier.committed = be32_to_cpup(p); return decode_verifier(xdr, &res->verifier.verifier); -- cgit v1.2.3 From 13c3f50c914e6a51d11e1aba4a85c2223e197e13 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:15 -0400 Subject: binfmt_flat: assorted cleanups Remove excessive casts, do some code grouping, fix most important checkpatch.pl complaints, etc. No functional changes. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 230 ++++++++++++++++++++++++++----------------------------- 1 file changed, 109 insertions(+), 121 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index caf9e39bb82b..892dba62bf2e 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -15,7 +15,6 @@ * JAN/99 -- coded full program relocation (gerg@snapgear.com) */ -#include #include #include #include @@ -25,8 +24,6 @@ #include #include #include -#include -#include #include #include #include @@ -34,10 +31,9 @@ #include #include #include -#include +#include #include -#include #include #include #include @@ -80,7 +76,7 @@ struct lib_info { unsigned long text_len; /* Length of text segment */ unsigned long entry; /* Start address for this module */ unsigned long build_date; /* When this one was compiled */ - short loaded; /* Has this library been loaded? */ + bool loaded; /* Has this library been loaded? */ } lib_list[MAX_SHARED_LIBS]; }; @@ -107,8 +103,8 @@ static struct linux_binfmt flat_format = { static int flat_core_dump(struct coredump_params *cprm) { printk("Process %s:%d received signr %d and should have core dumped\n", - current->comm, current->pid, (int) cprm->siginfo->si_signo); - return(1); + current->comm, current->pid, cprm->siginfo->si_signo); + return 1; } /****************************************************************************/ @@ -120,11 +116,11 @@ static int flat_core_dump(struct coredump_params *cprm) static unsigned long create_flat_tables( unsigned long pp, - struct linux_binprm * bprm) + struct linux_binprm *bprm) { - unsigned long *argv,*envp; - unsigned long * sp; - char * p = (char*)pp; + unsigned long *argv, *envp; + unsigned long *sp; + char *p = (char *)pp; int argc = bprm->argc; int envc = bprm->envc; char uninitialized_var(dummy); @@ -142,7 +138,7 @@ static unsigned long create_flat_tables( put_user(argc, sp); current->mm->arg_start = (unsigned long) p; - while (argc-->0) { + while (argc-- > 0) { put_user((unsigned long) p, argv++); do { get_user(dummy, p); p++; @@ -150,7 +146,7 @@ static unsigned long create_flat_tables( } put_user((unsigned long) NULL, argv); current->mm->arg_end = current->mm->env_start = (unsigned long) p; - while (envc-->0) { + while (envc-- > 0) { put_user((unsigned long)p, envp); envp++; do { get_user(dummy, p); p++; @@ -190,7 +186,7 @@ static int decompress_exec( loff_t fpos; int ret, retval; - DBG_FLT("decompress_exec(offset=%x,buf=%x,len=%x)\n",(int)offset, (int)dst, (int)len); + DBG_FLT("decompress_exec(offset=%lx,buf=%p,len=%lx)\n", offset, dst, len); memset(&strm, 0, sizeof(strm)); strm.workspace = kmalloc(zlib_inflate_workspacesize(), GFP_KERNEL); @@ -243,7 +239,7 @@ static int decompress_exec( ret = 10; if (buf[3] & EXTRA_FIELD) { ret += 2 + buf[10] + (buf[11] << 8); - if (unlikely(LBUFSIZE <= ret)) { + if (unlikely(ret >= LBUFSIZE)) { DBG_FLT("binfmt_flat: buffer overflow (EXTRA)?\n"); goto out_free_buf; } @@ -251,7 +247,7 @@ static int decompress_exec( if (buf[3] & ORIG_NAME) { while (ret < LBUFSIZE && buf[ret++] != 0) ; - if (unlikely(LBUFSIZE == ret)) { + if (unlikely(ret == LBUFSIZE)) { DBG_FLT("binfmt_flat: buffer overflow (ORIG_NAME)?\n"); goto out_free_buf; } @@ -259,7 +255,7 @@ static int decompress_exec( if (buf[3] & COMMENT) { while (ret < LBUFSIZE && buf[ret++] != 0) ; - if (unlikely(LBUFSIZE == ret)) { + if (unlikely(ret == LBUFSIZE)) { DBG_FLT("binfmt_flat: buffer overflow (COMMENT)?\n"); goto out_free_buf; } @@ -327,17 +323,17 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) r &= 0x00ffffff; /* Trim ID off here */ } if (id >= MAX_SHARED_LIBS) { - printk("BINFMT_FLAT: reference 0x%x to shared library %d", - (unsigned) r, id); + printk("BINFMT_FLAT: reference 0x%lx to shared library %d", + r, id); goto failed; } if (curid != id) { if (internalp) { - printk("BINFMT_FLAT: reloc address 0x%x not in same module " - "(%d != %d)", (unsigned) r, curid, id); + printk("BINFMT_FLAT: reloc address 0x%lx not in same module " + "(%d != %d)", r, curid, id); goto failed; - } else if ( ! p->lib_list[id].loaded && - load_flat_shared_library(id, p) < 0) { + } else if (!p->lib_list[id].loaded && + load_flat_shared_library(id, p) < 0) { printk("BINFMT_FLAT: failed to load library %d", id); goto failed; } @@ -358,8 +354,8 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) text_len = p->lib_list[id].text_len; if (!flat_reloc_valid(r, start_brk - start_data + text_len)) { - printk("BINFMT_FLAT: reloc outside program 0x%x (0 - 0x%x/0x%x)", - (int) r,(int)(start_brk-start_data+text_len),(int)text_len); + printk("BINFMT_FLAT: reloc outside program 0x%lx (0 - 0x%lx/0x%lx)", + r, start_brk-start_data+text_len, text_len); goto failed; } @@ -369,7 +365,7 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) addr = r - text_len + start_data; /* Range checked already above so doing the range tests is redundant...*/ - return(addr); + return addr; failed: printk(", killing %s!\n", current->comm); @@ -383,11 +379,11 @@ failed: static void old_reloc(unsigned long rl) { #ifdef DEBUG - char *segment[] = { "TEXT", "DATA", "BSS", "*UNKNOWN*" }; + static const char *segment[] = { "TEXT", "DATA", "BSS", "*UNKNOWN*" }; #endif flat_v2_reloc_t r; unsigned long *ptr; - + r.value = rl; #if defined(CONFIG_COLDFIRE) ptr = (unsigned long *) (current->mm->start_code + r.reloc.offset); @@ -397,10 +393,10 @@ static void old_reloc(unsigned long rl) #ifdef DEBUG printk("Relocation of variable at DATASEG+%x " - "(address %p, currently %x) into segment %s\n", - r.reloc.offset, ptr, (int)*ptr, segment[r.reloc.type]); + "(address %p, currently %lx) into segment %s\n", + r.reloc.offset, ptr, *ptr, segment[r.reloc.type]); #endif - + switch (r.reloc.type) { case OLD_FLAT_RELOC_TYPE_TEXT: *ptr += current->mm->start_code; @@ -417,27 +413,25 @@ static void old_reloc(unsigned long rl) } #ifdef DEBUG - printk("Relocation became %x\n", (int)*ptr); + printk("Relocation became %lx\n", *ptr); #endif -} +} /****************************************************************************/ -static int load_flat_file(struct linux_binprm * bprm, +static int load_flat_file(struct linux_binprm *bprm, struct lib_info *libinfo, int id, unsigned long *extra_stack) { - struct flat_hdr * hdr; - unsigned long textpos = 0, datapos = 0, result; - unsigned long realdatastart = 0; - unsigned long text_len, data_len, bss_len, stack_len, flags; - unsigned long full_data; - unsigned long len, memp = 0; - unsigned long memp_size, extra, rlim; - unsigned long *reloc = 0, *rp; + struct flat_hdr *hdr; + unsigned long textpos, datapos, realdatastart; + unsigned long text_len, data_len, bss_len, stack_len, full_data, flags; + unsigned long len, memp, memp_size, extra, rlim; + unsigned long *reloc, *rp; struct inode *inode; - int i, rev, relocs = 0; + int i, rev, relocs; loff_t fpos; unsigned long start_code, end_code; + ssize_t result; int ret; hdr = ((struct flat_hdr *) bprm->buf); /* exec-header */ @@ -478,11 +472,11 @@ static int load_flat_file(struct linux_binprm * bprm, ret = -ENOEXEC; goto err; } - + /* Don't allow old format executables to use shared libraries */ if (rev == OLD_FLAT_VERSION && id != 0) { - printk("BINFMT_FLAT: shared libraries are not available before rev 0x%x\n", - (int) FLAT_VERSION); + printk("BINFMT_FLAT: shared libraries are not available before rev 0x%lx\n", + FLAT_VERSION); ret = -ENOEXEC; goto err; } @@ -517,11 +511,9 @@ static int load_flat_file(struct linux_binprm * bprm, /* Flush all traces of the currently running executable */ if (id == 0) { - result = flush_old_exec(bprm); - if (result) { - ret = result; + ret = flush_old_exec(bprm); + if (ret) goto err; - } /* OK, This is the point of no return */ set_personality(PER_LINUX_32BIT); @@ -549,38 +541,38 @@ static int load_flat_file(struct linux_binprm * bprm, textpos = vm_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_EXECUTABLE, 0); if (!textpos || IS_ERR_VALUE(textpos)) { - if (!textpos) - textpos = (unsigned long) -ENOMEM; - printk("Unable to mmap process text, errno %d\n", (int)-textpos); ret = textpos; + if (!textpos) + ret = -ENOMEM; + printk("Unable to mmap process text, errno %d\n", ret); goto err; } len = data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long); len = PAGE_ALIGN(len); - realdatastart = vm_mmap(0, 0, len, + realdatastart = vm_mmap(NULL, 0, len, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, 0); if (realdatastart == 0 || IS_ERR_VALUE(realdatastart)) { + ret = realdatastart; if (!realdatastart) - realdatastart = (unsigned long) -ENOMEM; - printk("Unable to allocate RAM for process data, errno %d\n", - (int)-realdatastart); + ret = -ENOMEM; + printk("Unable to allocate RAM for process data, " + "errno %d\n", ret); vm_munmap(textpos, text_len); - ret = realdatastart; goto err; } datapos = ALIGN(realdatastart + MAX_SHARED_LIBS * sizeof(unsigned long), FLAT_DATA_ALIGN); - DBG_FLT("BINFMT_FLAT: Allocated data+bss+stack (%d bytes): %x\n", - (int)(data_len + bss_len + stack_len), (int)datapos); + DBG_FLT("BINFMT_FLAT: Allocated data+bss+stack (%ld bytes): %lx\n", + data_len + bss_len + stack_len, datapos); fpos = ntohl(hdr->data_start); #ifdef CONFIG_BINFMT_ZFLAT if (flags & FLAT_FLAG_GZDATA) { - result = decompress_exec(bprm, fpos, (char *) datapos, + result = decompress_exec(bprm, fpos, (char *)datapos, full_data, 0); } else #endif @@ -589,29 +581,30 @@ static int load_flat_file(struct linux_binprm * bprm, full_data); } if (IS_ERR_VALUE(result)) { - printk("Unable to read data+bss, errno %d\n", (int)-result); + ret = result; + printk("Unable to read data+bss, errno %d\n", ret); vm_munmap(textpos, text_len); vm_munmap(realdatastart, len); - ret = result; goto err; } - reloc = (unsigned long *) (datapos+(ntohl(hdr->reloc_start)-text_len)); + reloc = (unsigned long *) + (datapos + (ntohl(hdr->reloc_start) - text_len)); memp = realdatastart; memp_size = len; } else { len = text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long); len = PAGE_ALIGN(len); - textpos = vm_mmap(0, 0, len, + textpos = vm_mmap(NULL, 0, len, PROT_READ | PROT_EXEC | PROT_WRITE, MAP_PRIVATE, 0); if (!textpos || IS_ERR_VALUE(textpos)) { - if (!textpos) - textpos = (unsigned long) -ENOMEM; - printk("Unable to allocate RAM for process text/data, errno %d\n", - (int)-textpos); ret = textpos; + if (!textpos) + ret = -ENOMEM; + printk("Unable to allocate RAM for process text/data, " + "errno %d\n", ret); goto err; } @@ -629,10 +622,10 @@ static int load_flat_file(struct linux_binprm * bprm, * load it all in and treat it like a RAM load from now on */ if (flags & FLAT_FLAG_GZIP) { - result = decompress_exec(bprm, sizeof (struct flat_hdr), - (((char *) textpos) + sizeof (struct flat_hdr)), + result = decompress_exec(bprm, sizeof(struct flat_hdr), + (((char *)textpos) + sizeof(struct flat_hdr)), (text_len + full_data - - sizeof (struct flat_hdr)), + - sizeof(struct flat_hdr)), 0); memmove((void *) datapos, (void *) realdatastart, full_data); @@ -641,8 +634,7 @@ static int load_flat_file(struct linux_binprm * bprm, if (!IS_ERR_VALUE(result)) result = decompress_exec(bprm, text_len, (char *) datapos, full_data, 0); - } - else + } else #endif { result = read_code(bprm->file, textpos, 0, text_len); @@ -652,21 +644,19 @@ static int load_flat_file(struct linux_binprm * bprm, full_data); } if (IS_ERR_VALUE(result)) { - printk("Unable to read code+data+bss, errno %d\n",(int)-result); + ret = result; + printk("Unable to read code+data+bss, errno %d\n", ret); vm_munmap(textpos, text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long)); - ret = result; goto err; } } - if (flags & FLAT_FLAG_KTRACE) - printk("Mapping is %x, Entry point is %x, data_start is %x\n", - (int)textpos, 0x00ffffff&ntohl(hdr->entry), ntohl(hdr->data_start)); + start_code = textpos + sizeof(struct flat_hdr); + end_code = textpos + text_len; + text_len -= sizeof(struct flat_hdr); /* the real code len */ /* The main program needs a little extra setup in the task structure */ - start_code = textpos + sizeof (struct flat_hdr); - end_code = textpos + text_len; if (id == 0) { current->mm->start_code = start_code; current->mm->end_code = end_code; @@ -684,16 +674,14 @@ static int load_flat_file(struct linux_binprm * bprm, current->mm->context.end_brk = memp + memp_size - stack_len; } - if (flags & FLAT_FLAG_KTRACE) - printk("%s %s: TEXT=%x-%x DATA=%x-%x BSS=%x-%x\n", - id ? "Lib" : "Load", bprm->filename, - (int) start_code, (int) end_code, - (int) datapos, - (int) (datapos + data_len), - (int) (datapos + data_len), - (int) (((datapos + data_len + bss_len) + 3) & ~3)); - - text_len -= sizeof(struct flat_hdr); /* the real code len */ + if (flags & FLAT_FLAG_KTRACE) { + printk("Mapping is %lx, Entry point is %x, data_start is %x\n", + textpos, 0x00ffffff&ntohl(hdr->entry), ntohl(hdr->data_start)); + printk("%s %s: TEXT=%lx-%lx DATA=%lx-%lx BSS=%lx-%lx\n", + id ? "Lib" : "Load", bprm->filename, + start_code, end_code, datapos, datapos + data_len, + datapos + data_len, (datapos + data_len + bss_len + 3) & ~3); + } /* Store the current module values into the global library structure */ libinfo->lib_list[id].start_code = start_code; @@ -703,7 +691,7 @@ static int load_flat_file(struct linux_binprm * bprm, libinfo->lib_list[id].loaded = 1; libinfo->lib_list[id].entry = (0x00ffffff & ntohl(hdr->entry)) + textpos; libinfo->lib_list[id].build_date = ntohl(hdr->build_date); - + /* * We just load the allocations into some temporary memory to * help simplify all this mumbo jumbo @@ -743,14 +731,16 @@ static int load_flat_file(struct linux_binprm * bprm, */ if (rev > OLD_FLAT_VERSION) { unsigned long persistent = 0; - for (i=0; i < relocs; i++) { + for (i = 0; i < relocs; i++) { unsigned long addr, relval; - /* Get the address of the pointer to be - relocated (of course, the address has to be - relocated first). */ + /* + * Get the address of the pointer to be + * relocated (of course, the address has to be + * relocated first). + */ relval = ntohl(reloc[i]); - if (flat_set_persistent (relval, &persistent)) + if (flat_set_persistent(relval, &persistent)) continue; addr = flat_get_relocate_addr(relval); rp = (unsigned long *) calc_reloc(addr, libinfo, id, 1); @@ -780,14 +770,14 @@ static int load_flat_file(struct linux_binprm * bprm, } } } else { - for (i=0; i < relocs; i++) + for (i = 0; i < relocs; i++) old_reloc(ntohl(reloc[i])); } - + flush_icache_range(start_code, end_code); /* zero the BSS, BRK and stack areas */ - memset((void*)(datapos + data_len), 0, bss_len + + memset((void *)(datapos + data_len), 0, bss_len + (memp + memp_size - stack_len - /* end brk */ libinfo->lib_list[id].start_brk) + /* start brk */ stack_len); @@ -846,7 +836,7 @@ out: allow_write_access(bprm.file); fput(bprm.file); - return(res); + return res; } #endif /* CONFIG_BINFMT_SHARED_FLAT */ @@ -857,7 +847,7 @@ out: * libraries. There is no binary dependent code anywhere else. */ -static int load_flat_binary(struct linux_binprm * bprm) +static int load_flat_binary(struct linux_binprm *bprm) { struct lib_info libinfo; struct pt_regs *regs = current_pt_regs(); @@ -869,6 +859,7 @@ static int load_flat_binary(struct linux_binprm * bprm) int i, j; memset(&libinfo, 0, sizeof(libinfo)); + /* * We have to add the size of our arguments to our stack size * otherwise it's too easy for users to create stack overflows @@ -881,33 +872,33 @@ static int load_flat_binary(struct linux_binprm * bprm) stack_len += (bprm->argc + 1) * sizeof(char *); /* the argv array */ stack_len += (bprm->envc + 1) * sizeof(char *); /* the envp array */ stack_len += FLAT_STACK_ALIGN - 1; /* reserve for upcoming alignment */ - + res = load_flat_file(bprm, &libinfo, 0, &stack_len); if (res < 0) return res; - + /* Update data segment pointers for all libraries */ - for (i=0; imm->context.end_brk + stack_len + 3) & ~3) - 4; - DBG_FLT("p=%x\n", (int)p); + DBG_FLT("p=%lx\n", p); /* copy the arg pages onto the stack, this could be more efficient :-) */ for (i = TOP_OF_ARGS - 1; i >= bprm->p; i--) - * (char *) --p = + *(char *) --p = ((char *) page_address(bprm->page[i/PAGE_SIZE]))[i % PAGE_SIZE]; sp = (unsigned long *) create_flat_tables(p, bprm); - + /* Fake some return addresses to ensure the call chain will * initialise library in order for us. We are required to call * lib 1 first, then 2, ... and finally the main program (id 0). @@ -915,7 +906,7 @@ static int load_flat_binary(struct linux_binprm * bprm) start_addr = libinfo.lib_list[0].entry; #ifdef CONFIG_BINFMT_SHARED_FLAT - for (i = MAX_SHARED_LIBS-1; i>0; i--) { + for (i = MAX_SHARED_LIBS-1; i > 0; i--) { if (libinfo.lib_list[i].loaded) { /* Push previos first to call address */ --sp; put_user(start_addr, sp); @@ -923,16 +914,16 @@ static int load_flat_binary(struct linux_binprm * bprm) } } #endif - + /* Stash our initial stack pointer into the mm structure */ - current->mm->start_stack = (unsigned long )sp; + current->mm->start_stack = (unsigned long)sp; #ifdef FLAT_PLAT_INIT FLAT_PLAT_INIT(regs); #endif - DBG_FLT("start_thread(regs=0x%x, entry=0x%x, start_stack=0x%x)\n", - (int)regs, (int)start_addr, (int)current->mm->start_stack); - + + DBG_FLT("start_thread(regs=0x%p, entry=0x%lx, start_stack=0x%lx)\n", + regs, start_addr, current->mm->start_stack); start_thread(regs, start_addr, current->mm->start_stack); return 0; @@ -945,9 +936,6 @@ static int __init init_flat_binfmt(void) register_binfmt(&flat_format); return 0; } - -/****************************************************************************/ - core_initcall(init_flat_binfmt); /****************************************************************************/ -- cgit v1.2.3 From 4adbb6ac4b807e2f3f07887cf2a925fa012a7825 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:16 -0400 Subject: binfmt_flat: convert printk invocations to their modern form Signed-off-by: Nicolas Pitre Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 118 ++++++++++++++++++++++++------------------------------- 1 file changed, 51 insertions(+), 67 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 892dba62bf2e..c3ccdefdeac2 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -15,6 +15,8 @@ * JAN/99 -- coded full program relocation (gerg@snapgear.com) */ +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + #include #include #include @@ -40,16 +42,6 @@ /****************************************************************************/ -#if 0 -#define DEBUG 1 -#endif - -#ifdef DEBUG -#define DBG_FLT(a...) printk(a) -#else -#define DBG_FLT(a...) -#endif - /* * User data (data section and bss) needs to be aligned. * We pick 0x20 here because it is the max value elf2flt has always @@ -102,8 +94,8 @@ static struct linux_binfmt flat_format = { static int flat_core_dump(struct coredump_params *cprm) { - printk("Process %s:%d received signr %d and should have core dumped\n", - current->comm, current->pid, cprm->siginfo->si_signo); + pr_warn("Process %s:%d received signr %d and should have core dumped\n", + current->comm, current->pid, cprm->siginfo->si_signo); return 1; } @@ -186,17 +178,17 @@ static int decompress_exec( loff_t fpos; int ret, retval; - DBG_FLT("decompress_exec(offset=%lx,buf=%p,len=%lx)\n", offset, dst, len); + pr_debug("decompress_exec(offset=%lx,buf=%p,len=%lx)\n", offset, dst, len); memset(&strm, 0, sizeof(strm)); strm.workspace = kmalloc(zlib_inflate_workspacesize(), GFP_KERNEL); if (strm.workspace == NULL) { - DBG_FLT("binfmt_flat: no memory for decompress workspace\n"); + pr_debug("no memory for decompress workspace\n"); return -ENOMEM; } buf = kmalloc(LBUFSIZE, GFP_KERNEL); if (buf == NULL) { - DBG_FLT("binfmt_flat: no memory for read buffer\n"); + pr_debug("no memory for read buffer\n"); retval = -ENOMEM; goto out_free; } @@ -214,25 +206,25 @@ static int decompress_exec( /* Check minimum size -- gzip header */ if (ret < 10) { - DBG_FLT("binfmt_flat: file too small?\n"); + pr_debug("file too small?\n"); goto out_free_buf; } /* Check gzip magic number */ if ((buf[0] != 037) || ((buf[1] != 0213) && (buf[1] != 0236))) { - DBG_FLT("binfmt_flat: unknown compression magic?\n"); + pr_debug("unknown compression magic?\n"); goto out_free_buf; } /* Check gzip method */ if (buf[2] != 8) { - DBG_FLT("binfmt_flat: unknown compression method?\n"); + pr_debug("unknown compression method?\n"); goto out_free_buf; } /* Check gzip flags */ if ((buf[3] & ENCRYPTED) || (buf[3] & CONTINUATION) || (buf[3] & RESERVED)) { - DBG_FLT("binfmt_flat: unknown flags?\n"); + pr_debug("unknown flags?\n"); goto out_free_buf; } @@ -240,7 +232,7 @@ static int decompress_exec( if (buf[3] & EXTRA_FIELD) { ret += 2 + buf[10] + (buf[11] << 8); if (unlikely(ret >= LBUFSIZE)) { - DBG_FLT("binfmt_flat: buffer overflow (EXTRA)?\n"); + pr_debug("buffer overflow (EXTRA)?\n"); goto out_free_buf; } } @@ -248,7 +240,7 @@ static int decompress_exec( while (ret < LBUFSIZE && buf[ret++] != 0) ; if (unlikely(ret == LBUFSIZE)) { - DBG_FLT("binfmt_flat: buffer overflow (ORIG_NAME)?\n"); + pr_debug("buffer overflow (ORIG_NAME)?\n"); goto out_free_buf; } } @@ -256,7 +248,7 @@ static int decompress_exec( while (ret < LBUFSIZE && buf[ret++] != 0) ; if (unlikely(ret == LBUFSIZE)) { - DBG_FLT("binfmt_flat: buffer overflow (COMMENT)?\n"); + pr_debug("buffer overflow (COMMENT)?\n"); goto out_free_buf; } } @@ -269,7 +261,7 @@ static int decompress_exec( strm.total_out = 0; if (zlib_inflateInit2(&strm, -MAX_WBITS) != Z_OK) { - DBG_FLT("binfmt_flat: zlib init failed?\n"); + pr_debug("zlib init failed?\n"); goto out_free_buf; } @@ -286,7 +278,7 @@ static int decompress_exec( } if (ret < 0) { - DBG_FLT("binfmt_flat: decompression failed (%d), %s\n", + pr_debug("decompression failed (%d), %s\n", ret, strm.msg); goto out_zlib; } @@ -323,24 +315,23 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) r &= 0x00ffffff; /* Trim ID off here */ } if (id >= MAX_SHARED_LIBS) { - printk("BINFMT_FLAT: reference 0x%lx to shared library %d", - r, id); + pr_err("reference 0x%lx to shared library %d", r, id); goto failed; } if (curid != id) { if (internalp) { - printk("BINFMT_FLAT: reloc address 0x%lx not in same module " - "(%d != %d)", r, curid, id); + pr_err("reloc address 0x%lx not in same module " + "(%d != %d)", r, curid, id); goto failed; } else if (!p->lib_list[id].loaded && load_flat_shared_library(id, p) < 0) { - printk("BINFMT_FLAT: failed to load library %d", id); + pr_err("failed to load library %d", id); goto failed; } /* Check versioning information (i.e. time stamps) */ if (p->lib_list[id].build_date && p->lib_list[curid].build_date && p->lib_list[curid].build_date < p->lib_list[id].build_date) { - printk("BINFMT_FLAT: library %d is younger than %d", id, curid); + pr_err("library %d is younger than %d", id, curid); goto failed; } } @@ -354,7 +345,7 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) text_len = p->lib_list[id].text_len; if (!flat_reloc_valid(r, start_brk - start_data + text_len)) { - printk("BINFMT_FLAT: reloc outside program 0x%lx (0 - 0x%lx/0x%lx)", + pr_err("reloc outside program 0x%lx (0 - 0x%lx/0x%lx)", r, start_brk-start_data+text_len, text_len); goto failed; } @@ -368,7 +359,7 @@ calc_reloc(unsigned long r, struct lib_info *p, int curid, int internalp) return addr; failed: - printk(", killing %s!\n", current->comm); + pr_cont(", killing %s!\n", current->comm); send_sig(SIGSEGV, current, 0); return RELOC_FAILED; @@ -378,9 +369,7 @@ failed: static void old_reloc(unsigned long rl) { -#ifdef DEBUG static const char *segment[] = { "TEXT", "DATA", "BSS", "*UNKNOWN*" }; -#endif flat_v2_reloc_t r; unsigned long *ptr; @@ -391,11 +380,9 @@ static void old_reloc(unsigned long rl) ptr = (unsigned long *) (current->mm->start_data + r.reloc.offset); #endif -#ifdef DEBUG - printk("Relocation of variable at DATASEG+%x " - "(address %p, currently %lx) into segment %s\n", - r.reloc.offset, ptr, *ptr, segment[r.reloc.type]); -#endif + pr_debug("Relocation of variable at DATASEG+%x " + "(address %p, currently %lx) into segment %s\n", + r.reloc.offset, ptr, *ptr, segment[r.reloc.type]); switch (r.reloc.type) { case OLD_FLAT_RELOC_TYPE_TEXT: @@ -408,13 +395,11 @@ static void old_reloc(unsigned long rl) *ptr += current->mm->end_data; break; default: - printk("BINFMT_FLAT: Unknown relocation type=%x\n", r.reloc.type); + pr_err("Unknown relocation type=%x\n", r.reloc.type); break; } -#ifdef DEBUG - printk("Relocation became %lx\n", *ptr); -#endif + pr_debug("Relocation became %lx\n", *ptr); } /****************************************************************************/ @@ -463,20 +448,19 @@ static int load_flat_file(struct linux_binprm *bprm, } if (flags & FLAT_FLAG_KTRACE) - printk("BINFMT_FLAT: Loading file: %s\n", bprm->filename); + pr_info("Loading file: %s\n", bprm->filename); if (rev != FLAT_VERSION && rev != OLD_FLAT_VERSION) { - printk("BINFMT_FLAT: bad flat file version 0x%x (supported " - "0x%lx and 0x%lx)\n", - rev, FLAT_VERSION, OLD_FLAT_VERSION); + pr_err("bad flat file version 0x%x (supported 0x%lx and 0x%lx)\n", + rev, FLAT_VERSION, OLD_FLAT_VERSION); ret = -ENOEXEC; goto err; } /* Don't allow old format executables to use shared libraries */ if (rev == OLD_FLAT_VERSION && id != 0) { - printk("BINFMT_FLAT: shared libraries are not available before rev 0x%lx\n", - FLAT_VERSION); + pr_err("shared libraries are not available before rev 0x%lx\n", + FLAT_VERSION); ret = -ENOEXEC; goto err; } @@ -490,7 +474,7 @@ static int load_flat_file(struct linux_binprm *bprm, #ifndef CONFIG_BINFMT_ZFLAT if (flags & (FLAT_FLAG_GZIP|FLAT_FLAG_GZDATA)) { - printk("Support for ZFLAT executables is not enabled.\n"); + pr_err("Support for ZFLAT executables is not enabled.\n"); ret = -ENOEXEC; goto err; } @@ -536,7 +520,7 @@ static int load_flat_file(struct linux_binprm *bprm, * this should give us a ROM ptr, but if it doesn't we don't * really care */ - DBG_FLT("BINFMT_FLAT: ROM mapping of file (we hope)\n"); + pr_debug("ROM mapping of file (we hope)\n"); textpos = vm_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_EXECUTABLE, 0); @@ -544,7 +528,7 @@ static int load_flat_file(struct linux_binprm *bprm, ret = textpos; if (!textpos) ret = -ENOMEM; - printk("Unable to mmap process text, errno %d\n", ret); + pr_err("Unable to mmap process text, errno %d\n", ret); goto err; } @@ -557,7 +541,7 @@ static int load_flat_file(struct linux_binprm *bprm, ret = realdatastart; if (!realdatastart) ret = -ENOMEM; - printk("Unable to allocate RAM for process data, " + pr_err("Unable to allocate RAM for process data, " "errno %d\n", ret); vm_munmap(textpos, text_len); goto err; @@ -566,8 +550,8 @@ static int load_flat_file(struct linux_binprm *bprm, MAX_SHARED_LIBS * sizeof(unsigned long), FLAT_DATA_ALIGN); - DBG_FLT("BINFMT_FLAT: Allocated data+bss+stack (%ld bytes): %lx\n", - data_len + bss_len + stack_len, datapos); + pr_debug("Allocated data+bss+stack (%ld bytes): %lx\n", + data_len + bss_len + stack_len, datapos); fpos = ntohl(hdr->data_start); #ifdef CONFIG_BINFMT_ZFLAT @@ -582,7 +566,7 @@ static int load_flat_file(struct linux_binprm *bprm, } if (IS_ERR_VALUE(result)) { ret = result; - printk("Unable to read data+bss, errno %d\n", ret); + pr_err("Unable to read data+bss, errno %d\n", ret); vm_munmap(textpos, text_len); vm_munmap(realdatastart, len); goto err; @@ -603,7 +587,7 @@ static int load_flat_file(struct linux_binprm *bprm, ret = textpos; if (!textpos) ret = -ENOMEM; - printk("Unable to allocate RAM for process text/data, " + pr_err("Unable to allocate RAM for process text/data, " "errno %d\n", ret); goto err; } @@ -645,7 +629,7 @@ static int load_flat_file(struct linux_binprm *bprm, } if (IS_ERR_VALUE(result)) { ret = result; - printk("Unable to read code+data+bss, errno %d\n", ret); + pr_err("Unable to read code+data+bss, errno %d\n", ret); vm_munmap(textpos, text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long)); goto err; @@ -675,12 +659,12 @@ static int load_flat_file(struct linux_binprm *bprm, } if (flags & FLAT_FLAG_KTRACE) { - printk("Mapping is %lx, Entry point is %x, data_start is %x\n", - textpos, 0x00ffffff&ntohl(hdr->entry), ntohl(hdr->data_start)); - printk("%s %s: TEXT=%lx-%lx DATA=%lx-%lx BSS=%lx-%lx\n", - id ? "Lib" : "Load", bprm->filename, - start_code, end_code, datapos, datapos + data_len, - datapos + data_len, (datapos + data_len + bss_len + 3) & ~3); + pr_info("Mapping is %lx, Entry point is %x, data_start is %x\n", + textpos, 0x00ffffff&ntohl(hdr->entry), ntohl(hdr->data_start)); + pr_info("%s %s: TEXT=%lx-%lx DATA=%lx-%lx BSS=%lx-%lx\n", + id ? "Lib" : "Load", bprm->filename, + start_code, end_code, datapos, datapos + data_len, + datapos + data_len, (datapos + data_len + bss_len + 3) & ~3); } /* Store the current module values into the global library structure */ @@ -890,7 +874,7 @@ static int load_flat_binary(struct linux_binprm *bprm) set_binfmt(&flat_format); p = ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; - DBG_FLT("p=%lx\n", p); + pr_debug("p=%lx\n", p); /* copy the arg pages onto the stack, this could be more efficient :-) */ for (i = TOP_OF_ARGS - 1; i >= bprm->p; i--) @@ -922,8 +906,8 @@ static int load_flat_binary(struct linux_binprm *bprm) FLAT_PLAT_INIT(regs); #endif - DBG_FLT("start_thread(regs=0x%p, entry=0x%lx, start_stack=0x%lx)\n", - regs, start_addr, current->mm->start_stack); + pr_debug("start_thread(regs=0x%p, entry=0x%lx, start_stack=0x%lx)\n", + regs, start_addr, current->mm->start_stack); start_thread(regs, start_addr, current->mm->start_stack); return 0; -- cgit v1.2.3 From c995ee28d29d6f256c3a8a6c4e66469554374f25 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:17 -0400 Subject: binfmt_flat: prevent kernel dammage from corrupted executable headers Signed-off-by: Nicolas Pitre Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index c3ccdefdeac2..03301bad1f9b 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -465,6 +465,17 @@ static int load_flat_file(struct linux_binprm *bprm, goto err; } + /* + * Make sure the header params are sane. + * 28 bits (256 MB) is way more than reasonable in this case. + * If some top bits are set we have probable binary corruption. + */ + if ((text_len | data_len | bss_len | stack_len | full_data) >> 28) { + pr_err("bad header\n"); + ret = -ENOEXEC; + goto err; + } + /* * fix up the flags for the older format, there were all kinds * of endian hacks, this only works for the simple cases -- cgit v1.2.3 From 7e7ec6a934349ef6983f06f7ac0da09cc8a42983 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:18 -0400 Subject: elf_fdpic_transfer_args_to_stack(): make it generic This copying of arguments and environment is common to both NOMMU binary formats we support. Let's make the elf_fdpic version available to the flat format as well. While at it, improve the code a bit not to copy below the actual data area. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_elf_fdpic.c | 38 ++------------------------------------ fs/exec.c | 33 +++++++++++++++++++++++++++++++++ include/linux/binfmts.h | 2 ++ 3 files changed, 37 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 203589311bf8..464a972e88c1 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -67,8 +67,6 @@ static int create_elf_fdpic_tables(struct linux_binprm *, struct mm_struct *, struct elf_fdpic_params *); #ifndef CONFIG_MMU -static int elf_fdpic_transfer_args_to_stack(struct linux_binprm *, - unsigned long *); static int elf_fdpic_map_file_constdisp_on_uclinux(struct elf_fdpic_params *, struct file *, struct mm_struct *); @@ -515,8 +513,9 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm, sp = mm->start_stack; /* stack the program arguments and environment */ - if (elf_fdpic_transfer_args_to_stack(bprm, &sp) < 0) + if (transfer_args_to_stack(bprm, &sp) < 0) return -EFAULT; + sp &= ~15; #endif /* @@ -709,39 +708,6 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm, return 0; } -/*****************************************************************************/ -/* - * transfer the program arguments and environment from the holding pages onto - * the stack - */ -#ifndef CONFIG_MMU -static int elf_fdpic_transfer_args_to_stack(struct linux_binprm *bprm, - unsigned long *_sp) -{ - unsigned long index, stop, sp; - char *src; - int ret = 0; - - stop = bprm->p >> PAGE_SHIFT; - sp = *_sp; - - for (index = MAX_ARG_PAGES - 1; index >= stop; index--) { - src = kmap(bprm->page[index]); - sp -= PAGE_SIZE; - if (copy_to_user((void *) sp, src, PAGE_SIZE) != 0) - ret = -EFAULT; - kunmap(bprm->page[index]); - if (ret < 0) - goto out; - } - - *_sp = (*_sp - (MAX_ARG_PAGES * PAGE_SIZE - bprm->p)) & ~15; - -out: - return ret; -} -#endif - /*****************************************************************************/ /* * load the appropriate binary image (executable or interpreter) into memory diff --git a/fs/exec.c b/fs/exec.c index 887c1c955df8..ef0df2f09257 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -762,6 +762,39 @@ out_unlock: } EXPORT_SYMBOL(setup_arg_pages); +#else + +/* + * Transfer the program arguments and environment from the holding pages + * onto the stack. The provided stack pointer is adjusted accordingly. + */ +int transfer_args_to_stack(struct linux_binprm *bprm, + unsigned long *sp_location) +{ + unsigned long index, stop, sp; + int ret = 0; + + stop = bprm->p >> PAGE_SHIFT; + sp = *sp_location; + + for (index = MAX_ARG_PAGES - 1; index >= stop; index--) { + unsigned int offset = index == stop ? bprm->p & ~PAGE_MASK : 0; + char *src = kmap(bprm->page[index]) + offset; + sp -= PAGE_SIZE - offset; + if (copy_to_user((void *) sp, src, PAGE_SIZE - offset) != 0) + ret = -EFAULT; + kunmap(bprm->page[index]); + if (ret) + goto out; + } + + *sp_location = sp; + +out: + return ret; +} +EXPORT_SYMBOL(transfer_args_to_stack); + #endif /* CONFIG_MMU */ static struct file *do_open_execat(int fd, struct filename *name, int flags) diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index 314b3caa701c..1303b570b18c 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -113,6 +113,8 @@ extern int suid_dumpable; extern int setup_arg_pages(struct linux_binprm * bprm, unsigned long stack_top, int executable_stack); +extern int transfer_args_to_stack(struct linux_binprm *bprm, + unsigned long *sp_location); extern int bprm_change_interp(char *interp, struct linux_binprm *bprm); extern int copy_strings_kernel(int argc, const char *const *argv, struct linux_binprm *bprm); -- cgit v1.2.3 From 687fd7738ec322ea5994a692c20301eed315899d Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:19 -0400 Subject: binfmt_flat: use generic transfer_args_to_stack() This gets rid of the rather ugly, open coded and suboptimal copy code. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 03301bad1f9b..a002e1a3b9e8 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -846,10 +846,8 @@ static int load_flat_binary(struct linux_binprm *bprm) { struct lib_info libinfo; struct pt_regs *regs = current_pt_regs(); - unsigned long p = bprm->p; - unsigned long stack_len; + unsigned long sp, stack_len; unsigned long start_addr; - unsigned long *sp; int res; int i, j; @@ -884,15 +882,15 @@ static int load_flat_binary(struct linux_binprm *bprm) set_binfmt(&flat_format); - p = ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; - pr_debug("p=%lx\n", p); + sp = ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; + pr_debug("sp=%lx\n", sp); - /* copy the arg pages onto the stack, this could be more efficient :-) */ - for (i = TOP_OF_ARGS - 1; i >= bprm->p; i--) - *(char *) --p = - ((char *) page_address(bprm->page[i/PAGE_SIZE]))[i % PAGE_SIZE]; + /* copy the arg pages onto the stack */ + res = transfer_args_to_stack(bprm, &sp); + if (res) + return res; - sp = (unsigned long *) create_flat_tables(p, bprm); + sp = create_flat_tables(sp, bprm); /* Fake some return addresses to ensure the call chain will * initialise library in order for us. We are required to call @@ -904,14 +902,14 @@ static int load_flat_binary(struct linux_binprm *bprm) for (i = MAX_SHARED_LIBS-1; i > 0; i--) { if (libinfo.lib_list[i].loaded) { /* Push previos first to call address */ - --sp; put_user(start_addr, sp); + --sp; put_user(start_addr, (unsigned long *)sp); start_addr = libinfo.lib_list[i].entry; } } #endif /* Stash our initial stack pointer into the mm structure */ - current->mm->start_stack = (unsigned long)sp; + current->mm->start_stack = sp; #ifdef FLAT_PLAT_INIT FLAT_PLAT_INIT(regs); -- cgit v1.2.3 From a97d157d00cd0ccbcfde3da7bbf34ab7c68cf4eb Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:20 -0400 Subject: binfmt_flat: clean up create_flat_tables() and stack accesses In addition to better code clarity, this brings proper usage of user memory accessors everywhere the stack is touched. This is essential for making this work on MMU systems. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 117 ++++++++++++++++++++++++++++++------------------------- 1 file changed, 63 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index a002e1a3b9e8..5dc7968a424f 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -103,50 +103,58 @@ static int flat_core_dump(struct coredump_params *cprm) /* * create_flat_tables() parses the env- and arg-strings in new user * memory and creates the pointer tables from them, and puts their - * addresses on the "stack", returning the new stack pointer value. + * addresses on the "stack", recording the new stack pointer value. */ -static unsigned long create_flat_tables( - unsigned long pp, - struct linux_binprm *bprm) +static int create_flat_tables(struct linux_binprm *bprm, unsigned long arg_start) { - unsigned long *argv, *envp; - unsigned long *sp; - char *p = (char *)pp; - int argc = bprm->argc; - int envc = bprm->envc; - char uninitialized_var(dummy); - - sp = (unsigned long *)p; - sp -= (envc + argc + 2) + 1 + (flat_argvp_envp_on_stack() ? 2 : 0); - sp = (unsigned long *) ((unsigned long)sp & -FLAT_STACK_ALIGN); - argv = sp + 1 + (flat_argvp_envp_on_stack() ? 2 : 0); - envp = argv + (argc + 1); + char __user *p; + unsigned long __user *sp; + long i, len; + p = (char __user *)arg_start; + sp = (unsigned long __user *)current->mm->start_stack; + + sp -= bprm->envc + 1; + sp -= bprm->argc + 1; + sp -= flat_argvp_envp_on_stack() ? 2 : 0; + sp -= 1; /* &argc */ + + current->mm->start_stack = (unsigned long)sp & -FLAT_STACK_ALIGN; + sp = (unsigned long __user *)current->mm->start_stack; + + __put_user(bprm->argc, sp++); if (flat_argvp_envp_on_stack()) { - put_user((unsigned long) envp, sp + 2); - put_user((unsigned long) argv, sp + 1); - } - - put_user(argc, sp); - current->mm->arg_start = (unsigned long) p; - while (argc-- > 0) { - put_user((unsigned long) p, argv++); - do { - get_user(dummy, p); p++; - } while (dummy); - } - put_user((unsigned long) NULL, argv); - current->mm->arg_end = current->mm->env_start = (unsigned long) p; - while (envc-- > 0) { - put_user((unsigned long)p, envp); envp++; - do { - get_user(dummy, p); p++; - } while (dummy); - } - put_user((unsigned long) NULL, envp); - current->mm->env_end = (unsigned long) p; - return (unsigned long)sp; + unsigned long argv, envp; + argv = (unsigned long)(sp + 2); + envp = (unsigned long)(sp + 2 + bprm->argc + 1); + __put_user(argv, sp++); + __put_user(envp, sp++); + } + + current->mm->arg_start = (unsigned long)p; + for (i = bprm->argc; i > 0; i--) { + __put_user((unsigned long)p, sp++); + len = strnlen_user(p, MAX_ARG_STRLEN); + if (!len || len > MAX_ARG_STRLEN) + return -EINVAL; + p += len; + } + __put_user(0, sp++); + current->mm->arg_end = (unsigned long)p; + + current->mm->env_start = (unsigned long) p; + for (i = bprm->envc; i > 0; i--) { + __put_user((unsigned long)p, sp++); + len = strnlen_user(p, MAX_ARG_STRLEN); + if (!len || len > MAX_ARG_STRLEN) + return -EINVAL; + p += len; + } + __put_user(0, sp++); + current->mm->env_end = (unsigned long)p; + + return 0; } /****************************************************************************/ @@ -846,7 +854,7 @@ static int load_flat_binary(struct linux_binprm *bprm) { struct lib_info libinfo; struct pt_regs *regs = current_pt_regs(); - unsigned long sp, stack_len; + unsigned long stack_len; unsigned long start_addr; int res; int i, j; @@ -860,11 +868,10 @@ static int load_flat_binary(struct linux_binprm *bprm) * pedantic and include space for the argv/envp array as it may have * a lot of entries. */ -#define TOP_OF_ARGS (PAGE_SIZE * MAX_ARG_PAGES - sizeof(void *)) - stack_len = TOP_OF_ARGS - bprm->p; /* the strings */ - stack_len += (bprm->argc + 1) * sizeof(char *); /* the argv array */ - stack_len += (bprm->envc + 1) * sizeof(char *); /* the envp array */ - stack_len += FLAT_STACK_ALIGN - 1; /* reserve for upcoming alignment */ + stack_len = PAGE_SIZE * MAX_ARG_PAGES - bprm->p; /* the strings */ + stack_len += (bprm->argc + 1) * sizeof(char *); /* the argv array */ + stack_len += (bprm->envc + 1) * sizeof(char *); /* the envp array */ + stack_len = ALIGN(stack_len, FLAT_STACK_ALIGN); res = load_flat_file(bprm, &libinfo, 0, &stack_len); if (res < 0) @@ -882,16 +889,18 @@ static int load_flat_binary(struct linux_binprm *bprm) set_binfmt(&flat_format); - sp = ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; - pr_debug("sp=%lx\n", sp); + /* Stash our initial stack pointer into the mm structure */ + current->mm->start_stack = + ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; + pr_debug("sp=%lx\n", current->mm->start_stack); /* copy the arg pages onto the stack */ - res = transfer_args_to_stack(bprm, &sp); + res = transfer_args_to_stack(bprm, ¤t->mm->start_stack); + if (!res) + res = create_flat_tables(bprm, current->mm->start_stack); if (res) return res; - sp = create_flat_tables(sp, bprm); - /* Fake some return addresses to ensure the call chain will * initialise library in order for us. We are required to call * lib 1 first, then 2, ... and finally the main program (id 0). @@ -902,15 +911,15 @@ static int load_flat_binary(struct linux_binprm *bprm) for (i = MAX_SHARED_LIBS-1; i > 0; i--) { if (libinfo.lib_list[i].loaded) { /* Push previos first to call address */ - --sp; put_user(start_addr, (unsigned long *)sp); + unsigned long __user *sp; + current->mm->start_stack -= sizeof(unsigned long); + sp = (unsigned long __user *)current->mm->start_stack; + __put_user(start_addr, sp); start_addr = libinfo.lib_list[i].entry; } } #endif - /* Stash our initial stack pointer into the mm structure */ - current->mm->start_stack = sp; - #ifdef FLAT_PLAT_INIT FLAT_PLAT_INIT(regs); #endif -- cgit v1.2.3 From 6e572ffe3feb8a02ed0044987654e07670aeffc2 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:21 -0400 Subject: binfmt_flat: use proper user space accessors with relocs processing code Relocs are fixed up in place in user space memory. The appropriate accessors are required for this code to work with an active MMU. The architecture specific handlers flat_get_addr_from_rp() and flat_put_addr_at_rp() for ARM and M68K are adjusted with separate patches. SuperH and Xtensa are left out as they doesn't implement __get_user_unaligned() and __put_user_unaligned() yet. The other architectures that use BFLT don't have any MMU. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 31 +++++++++++++++++++------------ 1 file changed, 19 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 5dc7968a424f..9deafb282db5 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -419,7 +419,7 @@ static int load_flat_file(struct linux_binprm *bprm, unsigned long textpos, datapos, realdatastart; unsigned long text_len, data_len, bss_len, stack_len, full_data, flags; unsigned long len, memp, memp_size, extra, rlim; - unsigned long *reloc, *rp; + unsigned long __user *reloc, *rp; struct inode *inode; int i, rev, relocs; loff_t fpos; @@ -591,7 +591,7 @@ static int load_flat_file(struct linux_binprm *bprm, goto err; } - reloc = (unsigned long *) + reloc = (unsigned long __user *) (datapos + (ntohl(hdr->reloc_start) - text_len)); memp = realdatastart; memp_size = len; @@ -616,7 +616,7 @@ static int load_flat_file(struct linux_binprm *bprm, MAX_SHARED_LIBS * sizeof(unsigned long), FLAT_DATA_ALIGN); - reloc = (unsigned long *) + reloc = (unsigned long __user *) (datapos + (ntohl(hdr->reloc_start) - text_len)); memp = textpos; memp_size = len; @@ -708,15 +708,20 @@ static int load_flat_file(struct linux_binprm *bprm, * image. */ if (flags & FLAT_FLAG_GOTPIC) { - for (rp = (unsigned long *)datapos; *rp != 0xffffffff; rp++) { - unsigned long addr; - if (*rp) { - addr = calc_reloc(*rp, libinfo, id, 0); + for (rp = (unsigned long __user *)datapos; ; rp++) { + unsigned long addr, rp_val; + if (get_user(rp_val, rp)) + return -EFAULT; + if (rp_val == 0xffffffff) + break; + if (rp_val) { + addr = calc_reloc(rp_val, libinfo, id, 0); if (addr == RELOC_FAILED) { ret = -ENOEXEC; goto err; } - *rp = addr; + if (put_user(addr, rp)) + return -EFAULT; } } } @@ -733,7 +738,7 @@ static int load_flat_file(struct linux_binprm *bprm, * __start to address 4 so that is okay). */ if (rev > OLD_FLAT_VERSION) { - unsigned long persistent = 0; + unsigned long __maybe_unused persistent = 0; for (i = 0; i < relocs; i++) { unsigned long addr, relval; @@ -742,12 +747,14 @@ static int load_flat_file(struct linux_binprm *bprm, * relocated (of course, the address has to be * relocated first). */ - relval = ntohl(reloc[i]); + if (get_user(relval, reloc + i)) + return -EFAULT; + relval = ntohl(relval); if (flat_set_persistent(relval, &persistent)) continue; addr = flat_get_relocate_addr(relval); - rp = (unsigned long *) calc_reloc(addr, libinfo, id, 1); - if (rp == (unsigned long *)RELOC_FAILED) { + rp = (unsigned long __user *)calc_reloc(addr, libinfo, id, 1); + if (rp == (unsigned long __user *)RELOC_FAILED) { ret = -ENOEXEC; goto err; } -- cgit v1.2.3 From c1fd5c30d1328e0a0c3aafc3768b9cea94d7b28f Mon Sep 17 00:00:00 2001 From: Wang Xiaoguang Date: Tue, 21 Jun 2016 11:12:56 +0800 Subject: btrfs: add missing bytes_readonly attribute file in sysfs Signed-off-by: Wang Xiaoguang Signed-off-by: David Sterba --- fs/btrfs/sysfs.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 4879656bda3c..c6569905d3d1 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -326,6 +326,7 @@ SPACE_INFO_ATTR(bytes_used); SPACE_INFO_ATTR(bytes_pinned); SPACE_INFO_ATTR(bytes_reserved); SPACE_INFO_ATTR(bytes_may_use); +SPACE_INFO_ATTR(bytes_readonly); SPACE_INFO_ATTR(disk_used); SPACE_INFO_ATTR(disk_total); BTRFS_ATTR(total_bytes_pinned, btrfs_space_info_show_total_bytes_pinned); @@ -337,6 +338,7 @@ static struct attribute *space_info_attrs[] = { BTRFS_ATTR_PTR(bytes_pinned), BTRFS_ATTR_PTR(bytes_reserved), BTRFS_ATTR_PTR(bytes_may_use), + BTRFS_ATTR_PTR(bytes_readonly), BTRFS_ATTR_PTR(disk_used), BTRFS_ATTR_PTR(disk_total), BTRFS_ATTR_PTR(total_bytes_pinned), -- cgit v1.2.3 From 6fb37b756acce6d6e045f79c3764206033f617b4 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 22 Jun 2016 18:31:27 -0700 Subject: Btrfs: check inconsistence between chunk and block group With btrfs-corrupt-block, one can drop one chunk item and mounting will end up with a panic in btrfs_full_stripe_len(). This doesn't not remove the BUG_ON, but instead checks it a bit earlier when we find the block group item. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 9fcb8c97083b..cab2f9c8d321 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -9887,7 +9887,22 @@ static int find_first_block_group(struct btrfs_root *root, if (found_key.objectid >= key->objectid && found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) { - ret = 0; + struct extent_map_tree *em_tree; + struct extent_map *em; + + em_tree = &root->fs_info->mapping_tree.map_tree; + read_lock(&em_tree->lock); + em = lookup_extent_mapping(em_tree, found_key.objectid, + found_key.offset); + read_unlock(&em_tree->lock); + if (!em) { + btrfs_err(root->fs_info, + "logical %llu len %llu found bg but no related chunk", + found_key.objectid, found_key.offset); + ret = -ENOENT; + } else { + ret = 0; + } goto out; } path->slots[0]++; -- cgit v1.2.3 From 5e24e9af01abcb151173bb133f1a72b94239c670 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Thu, 23 Jun 2016 16:32:45 -0700 Subject: Btrfs: error out if generic_bin_search get invalid arguments With btrfs-corrupt-block, one can set btree node/leaf's field, if we assign a negative value to node/leaf, we can get various hangs, eg. if extent_root's nritems is -2ULL, then we get stuck in btrfs_read_block_groups() because it has a while loop and btrfs_search_slot() on extent_root will always return the first child. This lets us know what's happening and returns a EINVAL to callers instead of returning the first item. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index a85cf7d23309..362879da4f0d 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1771,6 +1771,14 @@ static noinline int generic_bin_search(struct extent_buffer *eb, unsigned long map_len = 0; int err; + if (low > high) { + btrfs_err(eb->fs_info, + "%s: low (%d) > high (%d) eb %llu owner %llu level %d", + __func__, low, high, eb->start, + btrfs_header_owner(eb), btrfs_header_level(eb)); + return -EINVAL; + } + while (low < high) { mid = (low + high) / 2; offset = p + mid * item_size; -- cgit v1.2.3 From 876d2cf141b42f9be70a383502a324b64b23f33a Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 28 Jun 2016 13:44:38 -0700 Subject: Btrfs: fix double free of fs root I got this warning while mounting a btrfs image, [ 3020.509606] ------------[ cut here ]------------ [ 3020.510107] WARNING: CPU: 3 PID: 5581 at lib/idr.c:1051 ida_remove+0xca/0x190 [ 3020.510853] ida_remove called for id=42 which is not allocated. [ 3020.511466] Modules linked in: [ 3020.511802] CPU: 3 PID: 5581 Comm: mount Not tainted 4.7.0-rc5+ #274 [ 3020.512438] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.8.2-20150714_191134- 04/01/2014 [ 3020.513385] 0000000000000286 0000000021295d86 ffff88006c66b8f0 ffffffff8182ba5a [ 3020.514153] 0000000000000000 0000000000000009 ffff88006c66b930 ffffffff810e0ed7 [ 3020.514928] 0000041b00000000 ffffffff8289a8c0 ffff88007f437880 0000000000000000 [ 3020.515717] Call Trace: [ 3020.515965] [] dump_stack+0xc9/0x13f [ 3020.516487] [] __warn+0x147/0x160 [ 3020.517005] [] warn_slowpath_fmt+0x5f/0x80 [ 3020.517572] [] ida_remove+0xca/0x190 [ 3020.518075] [] free_anon_bdev+0x2c/0x60 [ 3020.518609] [] free_fs_root+0x13f/0x160 [ 3020.519138] [] btrfs_get_fs_root+0x379/0x3d0 [ 3020.519710] [] ? __mutex_unlock_slowpath+0x155/0x2c0 [ 3020.520366] [] open_ctree+0x2e91/0x3200 [ 3020.520965] [] btrfs_mount+0x1322/0x15b0 [ 3020.521536] [] ? kmemleak_alloc_percpu+0x44/0x170 [ 3020.522167] [] ? lockdep_init_map+0x61/0x210 [ 3020.522780] [] mount_fs+0x49/0x2c0 [ 3020.523305] [] vfs_kern_mount+0xac/0x1b0 [ 3020.523872] [] btrfs_mount+0x421/0x15b0 [ 3020.524402] [] ? kmemleak_alloc_percpu+0x44/0x170 [ 3020.525045] [] ? lockdep_init_map+0x61/0x210 [ 3020.525657] [] ? lockdep_init_map+0x61/0x210 [ 3020.526289] [] mount_fs+0x49/0x2c0 [ 3020.526803] [] vfs_kern_mount+0xac/0x1b0 [ 3020.527365] [] do_mount+0x41a/0x1770 [ 3020.527899] [] ? strndup_user+0x6d/0xc0 [ 3020.528447] [] ? memdup_user+0x78/0xb0 [ 3020.528987] [] SyS_mount+0x150/0x160 [ 3020.529493] [] entry_SYSCALL_64_fastpath+0x1f/0xbd It turns out that we free fs root twice, btrfs_init_fs_root() calls free_anon_bdev(root->anon_dev) and later then btrfs_get_fs_root() cals free_fs_root which does another free_anon_bdev() and it ends up with the above warning. Instead of reset root->anon_dev to 0 after free_anon_bdev(), we can let btrfs_init_fs_root() return directly since its callers have already done the free job by calling free_fs_root(). Signed-off-by: Liu Bo Reviewed-by: David Sterba Reviewed-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 12 +++--------- 1 file changed, 3 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 60ce1190307b..6c88c63f3be7 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1600,14 +1600,14 @@ int btrfs_init_fs_root(struct btrfs_root *root) ret = get_anon_bdev(&root->anon_dev); if (ret) - goto free_writers; + goto fail; mutex_lock(&root->objectid_mutex); ret = btrfs_find_highest_objectid(root, &root->highest_objectid); if (ret) { mutex_unlock(&root->objectid_mutex); - goto free_root_dev; + goto fail; } ASSERT(root->highest_objectid <= BTRFS_LAST_FREE_OBJECTID); @@ -1615,14 +1615,8 @@ int btrfs_init_fs_root(struct btrfs_root *root) mutex_unlock(&root->objectid_mutex); return 0; - -free_root_dev: - free_anon_bdev(root->anon_dev); -free_writers: - btrfs_free_subvolume_writers(root->subv_writers); fail: - kfree(root->free_ino_ctl); - kfree(root->free_ino_pinned); + /* the caller is responsible to call free_fs_root */ return ret; } -- cgit v1.2.3 From fb770ae414d018255afa7a70b14ba1f8620762dd Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 5 Jul 2016 12:10:14 -0700 Subject: Btrfs: fix read_node_slot to return errors We use read_node_slot() to read btree node and it has two cases, a) slot is out of range, which means 'no such entry' b) we fail to read the block, due to checksum fails or corrupted content or not with uptodate flag. But we're returning NULL in both cases, this makes it return -ENOENT in case a) and return -EIO in case b), and this fixes its callers as well as btrfs_search_forward() 's caller to catch the new errors. The problem is reported by Peter Becker, and I can manage to hit the same BUG_ON by mounting my fuzz image. Reported-by: Peter Becker Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 69 +++++++++++++++++++++++++++++++++++++---------------- fs/btrfs/tree-log.c | 4 ++++ 2 files changed, 52 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 362879da4f0d..0abed1a0caa7 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1866,7 +1866,6 @@ static void root_sub_used(struct btrfs_root *root, u32 size) /* given a node and slot number, this reads the blocks it points to. The * extent buffer is returned with a reference taken (but unlocked). - * NULL is returned on error. */ static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root, struct extent_buffer *parent, int slot) @@ -1874,19 +1873,16 @@ static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root, int level = btrfs_header_level(parent); struct extent_buffer *eb; - if (slot < 0) - return NULL; - if (slot >= btrfs_header_nritems(parent)) - return NULL; + if (slot < 0 || slot >= btrfs_header_nritems(parent)) + return ERR_PTR(-ENOENT); BUG_ON(level == 0); eb = read_tree_block(root, btrfs_node_blockptr(parent, slot), btrfs_node_ptr_generation(parent, slot)); - if (IS_ERR(eb) || !extent_buffer_uptodate(eb)) { - if (!IS_ERR(eb)) - free_extent_buffer(eb); - eb = NULL; + if (!IS_ERR(eb) && !extent_buffer_uptodate(eb)) { + free_extent_buffer(eb); + eb = ERR_PTR(-EIO); } return eb; @@ -1939,8 +1935,8 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, /* promote the child to a root */ child = read_node_slot(root, mid, 0); - if (!child) { - ret = -EROFS; + if (IS_ERR(child)) { + ret = PTR_ERR(child); btrfs_handle_fs_error(root->fs_info, ret, NULL); goto enospc; } @@ -1978,6 +1974,9 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, return 0; left = read_node_slot(root, parent, pslot - 1); + if (IS_ERR(left)) + left = NULL; + if (left) { btrfs_tree_lock(left); btrfs_set_lock_blocking(left); @@ -1988,7 +1987,11 @@ static noinline int balance_level(struct btrfs_trans_handle *trans, goto enospc; } } + right = read_node_slot(root, parent, pslot + 1); + if (IS_ERR(right)) + right = NULL; + if (right) { btrfs_tree_lock(right); btrfs_set_lock_blocking(right); @@ -2143,6 +2146,8 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, return 1; left = read_node_slot(root, parent, pslot - 1); + if (IS_ERR(left)) + left = NULL; /* first, try to make some room in the middle buffer */ if (left) { @@ -2193,6 +2198,8 @@ static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, free_extent_buffer(left); } right = read_node_slot(root, parent, pslot + 1); + if (IS_ERR(right)) + right = NULL; /* * then try to empty the right most buffer into the middle @@ -3781,7 +3788,11 @@ static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root btrfs_assert_tree_locked(path->nodes[1]); right = read_node_slot(root, upper, slot + 1); - if (right == NULL) + /* + * slot + 1 is not valid or we fail to read the right node, + * no big deal, just return. + */ + if (IS_ERR(right)) return 1; btrfs_tree_lock(right); @@ -4011,7 +4022,11 @@ static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root btrfs_assert_tree_locked(path->nodes[1]); left = read_node_slot(root, path->nodes[1], slot - 1); - if (left == NULL) + /* + * slot - 1 is not valid or we fail to read the left node, + * no big deal, just return. + */ + if (IS_ERR(left)) return 1; btrfs_tree_lock(left); @@ -5218,7 +5233,10 @@ find_next_key: } btrfs_set_path_blocking(path); cur = read_node_slot(root, cur, slot); - BUG_ON(!cur); /* -ENOMEM */ + if (IS_ERR(cur)) { + ret = PTR_ERR(cur); + goto out; + } btrfs_tree_read_lock(cur); @@ -5237,15 +5255,21 @@ out: return ret; } -static void tree_move_down(struct btrfs_root *root, +static int tree_move_down(struct btrfs_root *root, struct btrfs_path *path, int *level, int root_level) { + struct extent_buffer *eb; + BUG_ON(*level == 0); - path->nodes[*level - 1] = read_node_slot(root, path->nodes[*level], - path->slots[*level]); + eb = read_node_slot(root, path->nodes[*level], path->slots[*level]); + if (IS_ERR(eb)) + return PTR_ERR(eb); + + path->nodes[*level - 1] = eb; path->slots[*level - 1] = 0; (*level)--; + return 0; } static int tree_move_next_or_upnext(struct btrfs_root *root, @@ -5290,8 +5314,7 @@ static int tree_advance(struct btrfs_root *root, if (*level == 0 || !allow_down) { ret = tree_move_next_or_upnext(root, path, level, root_level); } else { - tree_move_down(root, path, level, root_level); - ret = 0; + ret = tree_move_down(root, path, level, root_level); } if (ret >= 0) { if (*level == 0) @@ -5465,8 +5488,10 @@ int btrfs_compare_trees(struct btrfs_root *left_root, left_root_level, advance_left != ADVANCE_ONLY_NEXT, &left_key); - if (ret < 0) + if (ret == -1) left_end_reached = ADVANCE; + else if (ret < 0) + goto out; advance_left = 0; } if (advance_right && !right_end_reached) { @@ -5474,8 +5499,10 @@ int btrfs_compare_trees(struct btrfs_root *left_root, right_root_level, advance_right != ADVANCE_ONLY_NEXT, &right_key); - if (ret < 0) + if (ret == -1) right_end_reached = ADVANCE; + else if (ret < 0) + goto out; advance_right = 0; } diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index c05f69a8ec42..aa4475e3046b 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -4703,6 +4703,10 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, ins_nr = 0; ret = btrfs_search_forward(root, &min_key, path, trans->transid); + if (ret < 0) { + err = ret; + goto out_unlock; + } if (ret != 0) break; again: -- cgit v1.2.3 From 3d48d9810de4ee1bca4a1c905efd0618944904cd Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Wed, 29 Jun 2016 09:46:41 +0300 Subject: btrfs: Handle uninitialised inode eviction The code flow in btrfs_new_inode allows for btrfs_evict_inode to be called with not fully initialised inode (e.g. ->root member not being set). This can happen when btrfs_set_inode_index in btrfs_new_inode fails, which in turn would call iput for the newly allocated inode. This in turn leads to vfs calling into btrfs_evict_inode. This leads to null pointer dereference. To handle this situation check whether the passed inode has root set and just free it in case it doesn't. Signed-off-by: Nikolay Borisov Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/inode.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3e61bd1f3f65..5c2c076bc552 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5159,11 +5159,18 @@ void btrfs_evict_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *rsv, *global_rsv; int steal_from_global = 0; - u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); + u64 min_size; int ret; trace_btrfs_inode_evict(inode); + if (!root) { + kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); + return; + } + + min_size = btrfs_calc_trunc_metadata_size(root, 1); + evict_inode_truncate_pages(inode); if (inode->i_nlink && -- cgit v1.2.3 From a60617d0ae51c22f652b6fcf4ca56558c76e5aad Mon Sep 17 00:00:00 2001 From: Salah Triki Date: Sun, 3 Jul 2016 05:40:10 +0100 Subject: btrfs: Replace -ENOENT by -ERANGE in btrfs_get_acl() size contains the value returned by posix_acl_from_xattr(), which returns -ERANGE, -ENODATA, zero, or an integer greater than zero. So replace -ENOENT by -ERANGE. Signed-off-by: Salah Triki Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/acl.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index 67a607709d4f..53bb7af4e5f0 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -55,8 +55,7 @@ struct posix_acl *btrfs_get_acl(struct inode *inode, int type) } if (size > 0) { acl = posix_acl_from_xattr(&init_user_ns, value, size); - } else if (size == -ENOENT || size == -ENODATA || size == 0) { - /* FIXME, who returns -ENOENT? I think nobody */ + } else if (size == -ERANGE || size == -ENODATA || size == 0) { acl = NULL; } else { acl = ERR_PTR(-EIO); -- cgit v1.2.3 From fba4b697710eb2a4bee456b9d39e9239c66f8bee Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Thu, 23 Jun 2016 21:17:08 +0300 Subject: btrfs: Fix slab accounting flags BTRFS is using a variety of slab caches to satisfy internal needs. Those slab caches are always allocated with the SLAB_RECLAIM_ACCOUNT, meaning allocations from the caches are going to be accounted as SReclaimable. At the same time btrfs is not registering any shrinkers whatsoever, thus preventing memory from the slabs to be shrunk. This means those caches are not in fact reclaimable. To fix this remove the SLAB_RECLAIM_ACCOUNT on all caches apart from the inode cache, since this one is being freed by the generic VFS super_block shrinker. Also set the transaction related caches as SLAB_TEMPORARY, to better document the lifetime of the objects (it just translates to SLAB_RECLAIM_ACCOUNT). Signed-off-by: Nikolay Borisov Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/backref.c | 2 +- fs/btrfs/delayed-inode.c | 2 +- fs/btrfs/delayed-ref.c | 8 ++++---- fs/btrfs/disk-io.c | 2 +- fs/btrfs/extent_io.c | 4 ++-- fs/btrfs/extent_map.c | 2 +- fs/btrfs/file.c | 2 +- fs/btrfs/inode.c | 8 ++++---- fs/btrfs/ordered-data.c | 2 +- 9 files changed, 16 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 8bb3509099e8..1ac8a3c7f68e 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -139,7 +139,7 @@ int __init btrfs_prelim_ref_init(void) btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref", sizeof(struct __prelim_ref), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + SLAB_MEM_SPREAD, NULL); if (!btrfs_prelim_ref_cache) return -ENOMEM; diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index dd3c040139a2..9e2a907d9ddb 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -34,7 +34,7 @@ int __init btrfs_delayed_inode_init(void) delayed_node_cache = kmem_cache_create("btrfs_delayed_node", sizeof(struct btrfs_delayed_node), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + SLAB_MEM_SPREAD, NULL); if (!delayed_node_cache) return -ENOMEM; diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index 430b3689b112..08e452599249 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -940,28 +940,28 @@ int btrfs_delayed_ref_init(void) btrfs_delayed_ref_head_cachep = kmem_cache_create( "btrfs_delayed_ref_head", sizeof(struct btrfs_delayed_ref_head), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_delayed_ref_head_cachep) goto fail; btrfs_delayed_tree_ref_cachep = kmem_cache_create( "btrfs_delayed_tree_ref", sizeof(struct btrfs_delayed_tree_ref), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_delayed_tree_ref_cachep) goto fail; btrfs_delayed_data_ref_cachep = kmem_cache_create( "btrfs_delayed_data_ref", sizeof(struct btrfs_delayed_data_ref), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_delayed_data_ref_cachep) goto fail; btrfs_delayed_extent_op_cachep = kmem_cache_create( "btrfs_delayed_extent_op", sizeof(struct btrfs_delayed_extent_op), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_delayed_extent_op_cachep) goto fail; diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 6c88c63f3be7..49565626321e 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -101,7 +101,7 @@ int __init btrfs_end_io_wq_init(void) btrfs_end_io_wq_cache = kmem_cache_create("btrfs_end_io_wq", sizeof(struct btrfs_end_io_wq), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + SLAB_MEM_SPREAD, NULL); if (!btrfs_end_io_wq_cache) return -ENOMEM; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 75533adef998..9aaa544e65a9 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -163,13 +163,13 @@ int __init extent_io_init(void) { extent_state_cache = kmem_cache_create("btrfs_extent_state", sizeof(struct extent_state), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!extent_state_cache) return -ENOMEM; extent_buffer_cache = kmem_cache_create("btrfs_extent_buffer", sizeof(struct extent_buffer), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!extent_buffer_cache) goto free_state_cache; diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c index e0715fcfb11e..26f9ac719d20 100644 --- a/fs/btrfs/extent_map.c +++ b/fs/btrfs/extent_map.c @@ -13,7 +13,7 @@ int __init extent_map_init(void) { extent_map_cache = kmem_cache_create("btrfs_extent_map", sizeof(struct extent_map), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!extent_map_cache) return -ENOMEM; return 0; diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index bcfb4a27ddd4..38bea7bb0987 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -2975,7 +2975,7 @@ int btrfs_auto_defrag_init(void) { btrfs_inode_defrag_cachep = kmem_cache_create("btrfs_inode_defrag", sizeof(struct inode_defrag), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + SLAB_MEM_SPREAD, NULL); if (!btrfs_inode_defrag_cachep) return -ENOMEM; diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 5c2c076bc552..01c6fca925f7 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -9392,25 +9392,25 @@ int btrfs_init_cachep(void) btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle", sizeof(struct btrfs_trans_handle), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_TEMPORARY | SLAB_MEM_SPREAD, NULL); if (!btrfs_trans_handle_cachep) goto fail; btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction", sizeof(struct btrfs_transaction), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_TEMPORARY | SLAB_MEM_SPREAD, NULL); if (!btrfs_transaction_cachep) goto fail; btrfs_path_cachep = kmem_cache_create("btrfs_path", sizeof(struct btrfs_path), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_path_cachep) goto fail; btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space", sizeof(struct btrfs_free_space), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + SLAB_MEM_SPREAD, NULL); if (!btrfs_free_space_cachep) goto fail; diff --git a/fs/btrfs/ordered-data.c b/fs/btrfs/ordered-data.c index aca8264f4a49..3b78d38173b3 100644 --- a/fs/btrfs/ordered-data.c +++ b/fs/btrfs/ordered-data.c @@ -1122,7 +1122,7 @@ int __init ordered_data_init(void) { btrfs_ordered_extent_cache = kmem_cache_create("btrfs_ordered_extent", sizeof(struct btrfs_ordered_extent), 0, - SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + SLAB_MEM_SPREAD, NULL); if (!btrfs_ordered_extent_cache) return -ENOMEM; -- cgit v1.2.3 From 6f034ece3422d8d4fbe0afe7cd8c2ac6438e651e Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 22 Jun 2016 18:31:49 -0700 Subject: Btrfs: cleanup BUG_ON in merge_bio One can use btrfs-corrupt-block to hit BUG_ON() in merge_bio(), thus this aims to stop anyone to panic the whole system by using their btrfs. Since the error in merge_bio can only come from __btrfs_map_block() when chunk tree mapping has something insane and __btrfs_map_block() has already had printed the reason, we can just return errors in merge_bio. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 1 - fs/btrfs/inode.c | 8 ++++++-- 2 files changed, 6 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 9aaa544e65a9..ea12d0ee684f 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2756,7 +2756,6 @@ static int merge_bio(int rw, struct extent_io_tree *tree, struct page *page, if (tree->ops && tree->ops->merge_bio_hook) ret = tree->ops->merge_bio_hook(rw, page, offset, size, bio, bio_flags); - BUG_ON(ret < 0); return ret; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 01c6fca925f7..f935fad17838 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1822,6 +1822,10 @@ static void btrfs_clear_bit_hook(struct inode *inode, /* * extent_io.c merge_bio_hook, this must check the chunk tree to make sure * we don't create bios that span stripes or chunks + * + * return 1 if page cannot be merged to bio + * return 0 if page can be merged to bio + * return error otherwise */ int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset, size_t size, struct bio *bio, @@ -1840,8 +1844,8 @@ int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset, map_length = length; ret = btrfs_map_block(root->fs_info, rw, logical, &map_length, NULL, 0); - /* Will always return 0 with map_multi == NULL */ - BUG_ON(ret < 0); + if (ret < 0) + return ret; if (map_length < length + size) return 1; return 0; -- cgit v1.2.3 From c8bb0c8bd22a4b9d1c457a9c2b1c192d8271831a Mon Sep 17 00:00:00 2001 From: Ashish Samant Date: Fri, 25 Mar 2016 19:01:33 -0700 Subject: btrfs: Cleanup compress_file_range() Remove unnecessary checks in compress_file_range(). Signed-off-by: Ashish Samant [ minor coding style fixups ] Signed-off-by: David Sterba --- fs/btrfs/inode.c | 72 +++++++++++++++++++++++++++----------------------------- 1 file changed, 35 insertions(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f935fad17838..be854649c51b 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -585,9 +585,27 @@ cont: will_compress = 0; } else { num_bytes = total_in; + *num_added += 1; + + /* + * The async work queues will take care of doing actual + * allocation on disk for these compressed pages, and + * will submit them to the elevator. + */ + add_async_extent(async_cow, start, num_bytes, + total_compressed, pages, nr_pages_ret, + compress_type); + + if (start + num_bytes < end) { + start += num_bytes; + pages = NULL; + cond_resched(); + goto again; + } + return; } } - if (!will_compress && pages) { + if (pages) { /* * the compression code ran but failed to make things smaller, * free any pages it allocated and our page pointer array @@ -607,43 +625,23 @@ cont: BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; } } - if (will_compress) { - *num_added += 1; - - /* the async work queues will take care of doing actual - * allocation on disk for these compressed pages, - * and will submit them to the elevator. - */ - add_async_extent(async_cow, start, num_bytes, - total_compressed, pages, nr_pages_ret, - compress_type); - - if (start + num_bytes < end) { - start += num_bytes; - pages = NULL; - cond_resched(); - goto again; - } - } else { cleanup_and_bail_uncompressed: - /* - * No compression, but we still need to write the pages in - * the file we've been given so far. redirty the locked - * page if it corresponds to our extent and set things up - * for the async work queue to run cow_file_range to do - * the normal delalloc dance - */ - if (page_offset(locked_page) >= start && - page_offset(locked_page) <= end) { - __set_page_dirty_nobuffers(locked_page); - /* unlocked later on in the async handlers */ - } - if (redirty) - extent_range_redirty_for_io(inode, start, end); - add_async_extent(async_cow, start, end - start + 1, - 0, NULL, 0, BTRFS_COMPRESS_NONE); - *num_added += 1; - } + /* + * No compression, but we still need to write the pages in the file + * we've been given so far. redirty the locked page if it corresponds + * to our extent and set things up for the async work queue to run + * cow_file_range to do the normal delalloc dance. + */ + if (page_offset(locked_page) >= start && + page_offset(locked_page) <= end) + __set_page_dirty_nobuffers(locked_page); + /* unlocked later on in the async handlers */ + + if (redirty) + extent_range_redirty_for_io(inode, start, end); + add_async_extent(async_cow, start, end - start + 1, 0, NULL, 0, + BTRFS_COMPRESS_NONE); + *num_added += 1; return; -- cgit v1.2.3 From f448341af9cf50b9f405f5192f184d398e1812e0 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Tue, 14 Jun 2016 18:55:25 +0800 Subject: btrfs: reorg btrfs_close_one_device() Moves closer to the caller and removes declaration Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 71 +++++++++++++++++++++++++++--------------------------- 1 file changed, 35 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 589f128173b1..e54b3fe92e90 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -140,7 +140,6 @@ static int btrfs_relocate_sys_chunks(struct btrfs_root *root); static void __btrfs_reset_dev_stats(struct btrfs_device *dev); static void btrfs_dev_stat_print_on_error(struct btrfs_device *dev); static void btrfs_dev_stat_print_on_load(struct btrfs_device *device); -static void btrfs_close_one_device(struct btrfs_device *device); DEFINE_MUTEX(uuid_mutex); static LIST_HEAD(fs_uuids); @@ -853,6 +852,41 @@ static void free_device(struct rcu_head *head) schedule_work(&device->rcu_work); } +static void btrfs_close_one_device(struct btrfs_device *device) +{ + struct btrfs_fs_devices *fs_devices = device->fs_devices; + struct btrfs_device *new_device; + struct rcu_string *name; + + if (device->bdev) + fs_devices->open_devices--; + + if (device->writeable && + device->devid != BTRFS_DEV_REPLACE_DEVID) { + list_del_init(&device->dev_alloc_list); + fs_devices->rw_devices--; + } + + if (device->missing) + fs_devices->missing_devices--; + + new_device = btrfs_alloc_device(NULL, &device->devid, + device->uuid); + BUG_ON(IS_ERR(new_device)); /* -ENOMEM */ + + /* Safe because we are under uuid_mutex */ + if (device->name) { + name = rcu_string_strdup(device->name->str, GFP_NOFS); + BUG_ON(!name); /* -ENOMEM */ + rcu_assign_pointer(new_device->name, name); + } + + list_replace_rcu(&device->dev_list, &new_device->dev_list); + new_device->fs_devices = device->fs_devices; + + call_rcu(&device->rcu, free_device); +} + static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) { struct btrfs_device *device, *tmp; @@ -7138,38 +7172,3 @@ void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info) fs_devices = fs_devices->seed; } } - -static void btrfs_close_one_device(struct btrfs_device *device) -{ - struct btrfs_fs_devices *fs_devices = device->fs_devices; - struct btrfs_device *new_device; - struct rcu_string *name; - - if (device->bdev) - fs_devices->open_devices--; - - if (device->writeable && - device->devid != BTRFS_DEV_REPLACE_DEVID) { - list_del_init(&device->dev_alloc_list); - fs_devices->rw_devices--; - } - - if (device->missing) - fs_devices->missing_devices--; - - new_device = btrfs_alloc_device(NULL, &device->devid, - device->uuid); - BUG_ON(IS_ERR(new_device)); /* -ENOMEM */ - - /* Safe because we are under uuid_mutex */ - if (device->name) { - name = rcu_string_strdup(device->name->str, GFP_NOFS); - BUG_ON(!name); /* -ENOMEM */ - rcu_assign_pointer(new_device->name, name); - } - - list_replace_rcu(&device->dev_list, &new_device->dev_list); - new_device->fs_devices = device->fs_devices; - - call_rcu(&device->rcu, free_device); -} -- cgit v1.2.3 From e2bf6e89b4c71ea08223f13a700cfd0b576f7e7a Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Thu, 23 Jun 2016 20:54:07 +0800 Subject: btrfs: make sure device is synced before return An inconsistent behavior due to stale reads from the disk was reported mail-archive.com/linux-btrfs@vger.kernel.org/msg54188.html This patch will make sure devices are synced before return in the unmount thread. Signed-off-by: Anand Jain Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index e54b3fe92e90..51bd3ee8e64f 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -870,6 +870,11 @@ static void btrfs_close_one_device(struct btrfs_device *device) if (device->missing) fs_devices->missing_devices--; + if (device->bdev && device->writeable) { + sync_blockdev(device->bdev); + invalidate_bdev(device->bdev); + } + new_device = btrfs_alloc_device(NULL, &device->devid, device->uuid); BUG_ON(IS_ERR(new_device)); /* -ENOMEM */ -- cgit v1.2.3 From f5daf2c780f2ddfddbf98f2eb668f026c4dd2e5d Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 22 Jun 2016 18:32:06 -0700 Subject: Btrfs: fix BUG_ON in btrfs_submit_compressed_write This is similar to btrfs_submit_compressed_read(), if we fail after bio is allocated, then we can use bio_endio() and errors are saved in bio->bi_error. But please note that we don't return errors to its caller because the caller assumes it won't call endio to cleanup on error. Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/compression.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 658c39b70fba..7a4d9c81aa2a 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -402,7 +402,10 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); - BUG_ON(ret); /* -ENOMEM */ + if (ret) { + bio->bi_error = ret; + bio_endio(bio); + } bio_put(bio); @@ -432,7 +435,10 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, } ret = btrfs_map_bio(root, WRITE, bio, 0, 1); - BUG_ON(ret); /* -ENOMEM */ + if (ret) { + bio->bi_error = ret; + bio_endio(bio); + } bio_put(bio); return 0; -- cgit v1.2.3 From dda3245eca18c73413a854a834c41fc770feb0dd Mon Sep 17 00:00:00 2001 From: Wang Xiaoguang Date: Mon, 11 Jul 2016 11:05:29 +0800 Subject: btrfs: expand cow_file_range() to support in-band dedup and subpage-blocksize Extract cow_file_range() new parameters for both in-band dedupe and subpage sector size patchset. This should make conflict of both patchset to minimal, and reduce the effort needed to rebase them. Cc: Chandan Rajendra Cc: David Sterba Signed-off-by: Wang Xiaoguang Signed-off-by: Qu Wenruo Signed-off-by: David Sterba --- fs/btrfs/dedupe.h | 24 ++++++++++++++++++++++++ fs/btrfs/inode.c | 28 +++++++++++++++++----------- 2 files changed, 41 insertions(+), 11 deletions(-) create mode 100644 fs/btrfs/dedupe.h (limited to 'fs') diff --git a/fs/btrfs/dedupe.h b/fs/btrfs/dedupe.h new file mode 100644 index 000000000000..83ebfe28da9e --- /dev/null +++ b/fs/btrfs/dedupe.h @@ -0,0 +1,24 @@ +/* + * Copyright (C) 2016 Fujitsu. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#ifndef __BTRFS_DEDUPE__ +#define __BTRFS_DEDUPE__ + +/* later in-band dedupe will expand this struct */ +struct btrfs_dedupe_hash; +#endif diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index be854649c51b..0612231aaeba 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -60,6 +60,7 @@ #include "hash.h" #include "props.h" #include "qgroup.h" +#include "dedupe.h" struct btrfs_iget_args { struct btrfs_key *location; @@ -105,8 +106,9 @@ static int btrfs_truncate(struct inode *inode); static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent); static noinline int cow_file_range(struct inode *inode, struct page *locked_page, - u64 start, u64 end, int *page_started, - unsigned long *nr_written, int unlock); + u64 start, u64 end, u64 delalloc_end, + int *page_started, unsigned long *nr_written, + int unlock, struct btrfs_dedupe_hash *hash); static struct extent_map *create_pinned_em(struct inode *inode, u64 start, u64 len, u64 orig_start, u64 block_start, u64 block_len, @@ -710,7 +712,10 @@ retry: async_extent->start, async_extent->start + async_extent->ram_size - 1, - &page_started, &nr_written, 0); + async_extent->start + + async_extent->ram_size - 1, + &page_started, &nr_written, 0, + NULL); /* JDM XXX */ @@ -923,9 +928,9 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start, */ static noinline int cow_file_range(struct inode *inode, struct page *locked_page, - u64 start, u64 end, int *page_started, - unsigned long *nr_written, - int unlock) + u64 start, u64 end, u64 delalloc_end, + int *page_started, unsigned long *nr_written, + int unlock, struct btrfs_dedupe_hash *hash) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 alloc_hint = 0; @@ -1416,7 +1421,8 @@ out_check: if (cow_start != (u64)-1) { ret = cow_file_range(inode, locked_page, cow_start, found_key.offset - 1, - page_started, nr_written, 1); + end, page_started, nr_written, 1, + NULL); if (ret) { if (!nolock && nocow) btrfs_end_write_no_snapshoting(root); @@ -1499,8 +1505,8 @@ out_check: } if (cow_start != (u64)-1) { - ret = cow_file_range(inode, locked_page, cow_start, end, - page_started, nr_written, 1); + ret = cow_file_range(inode, locked_page, cow_start, end, end, + page_started, nr_written, 1, NULL); if (ret) goto error; } @@ -1559,8 +1565,8 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 0, nr_written); } else if (!inode_need_compress(inode)) { - ret = cow_file_range(inode, locked_page, start, end, - page_started, nr_written, 1); + ret = cow_file_range(inode, locked_page, start, end, end, + page_started, nr_written, 1, NULL); } else { set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, &BTRFS_I(inode)->runtime_flags); -- cgit v1.2.3 From 751bebbe0ad261b3ac5d9e75eaec4d5fe9e23160 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Mon, 4 Jul 2016 10:04:39 +0530 Subject: Btrfs: subpage-blocksize: Rate limit scrub error message btrfs/073 invokes scrub ioctl in a tight loop. In subpage-blocksize scenario this results in a lot of "scrub: size assumption sectorsize != PAGE_SIZE " messages being printed on the console. To reduce the number of such messages this commit uses btrfs_err_rl() instead of btrfs_err(). Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 70427ef66b04..acfe72004646 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3856,7 +3856,7 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, if (fs_info->chunk_root->sectorsize != PAGE_SIZE) { /* not supported for data w/o checksums */ - btrfs_err(fs_info, + btrfs_err_rl(fs_info, "scrub: size assumption sectorsize != PAGE_SIZE " "(%d != %lu) fails", fs_info->chunk_root->sectorsize, PAGE_SIZE); -- cgit v1.2.3 From 39581a3a1af7e03af41dde62a6485a934deb109b Mon Sep 17 00:00:00 2001 From: Wang Xiaoguang Date: Mon, 11 Jul 2016 19:30:04 +0800 Subject: btrfs: fix free space calculation in dump_space_info() In btrfs, btrfs_space_info's bytes_may_use is treated as fs used space, as what we do in reserve_metadata_bytes() or btrfs_alloc_data_chunk_ondemand(), so in dump_space_info(), when calculating free space, we should also subtract btrfs_space_info's bytes_may_use. Signed-off-by: Wang Xiaoguang Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index cab2f9c8d321..3bb9607edc5d 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -7906,8 +7906,8 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes, printk(KERN_INFO "BTRFS: space_info %llu has %llu free, is %sfull\n", info->flags, info->total_bytes - info->bytes_used - info->bytes_pinned - - info->bytes_reserved - info->bytes_readonly, - (info->full) ? "" : "not "); + info->bytes_reserved - info->bytes_readonly - + info->bytes_may_use, (info->full) ? "" : "not "); printk(KERN_INFO "BTRFS: space_info total=%llu, used=%llu, pinned=%llu, " "reserved=%llu, may_use=%llu, readonly=%llu\n", info->total_bytes, info->bytes_used, info->bytes_pinned, -- cgit v1.2.3 From f49070957ffed84feb7944550f7edd53672b5201 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 11 Jul 2016 18:52:57 -0700 Subject: Btrfs: change BUG_ON()'s to ASSERT()'s in backref_cache_cleanup() Since it is just an in-memory building of the backrefs of several btree blocks, nothing is fatal other than memory leaks, so this changes BUG_ON()'s to ASSERT()'s. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/relocation.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index fc067b07e31f..13c9a48c83b0 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -235,12 +235,12 @@ static void backref_cache_cleanup(struct backref_cache *cache) cache->last_trans = 0; for (i = 0; i < BTRFS_MAX_LEVEL; i++) - BUG_ON(!list_empty(&cache->pending[i])); - BUG_ON(!list_empty(&cache->changed)); - BUG_ON(!list_empty(&cache->detached)); - BUG_ON(!RB_EMPTY_ROOT(&cache->rb_root)); - BUG_ON(cache->nr_nodes); - BUG_ON(cache->nr_edges); + ASSERT(list_empty(&cache->pending[i])); + ASSERT(list_empty(&cache->changed)); + ASSERT(list_empty(&cache->detached)); + ASSERT(RB_EMPTY_ROOT(&cache->rb_root)); + ASSERT(!cache->nr_nodes); + ASSERT(!cache->nr_edges); } static struct backref_node *alloc_backref_node(struct backref_cache *cache) -- cgit v1.2.3 From baf863b9c29617cc9eaf24e039f58846e700db48 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 11 Jul 2016 10:39:07 -0700 Subject: Btrfs: fix eb memory leak due to readpage failure eb->io_pages is set in read_extent_buffer_pages(). In case of readpage failure, for pages that have been added to bio, it calls bio_endio and later readpage_io_failed_hook() does the work. When this eb's page (couldn't be the 1st page) fails to add itself to bio due to failure in merge_bio(), it cannot decrease eb->io_pages via bio_endio, and ends up with a memory leak eventually. This lets __do_readpage propagate errors to callers and adds the 'atomic_dec(&eb->io_pages)'. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 25 ++++++++++++++++++++++--- 1 file changed, 22 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index ea12d0ee684f..dae2f8470e04 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2878,6 +2878,7 @@ __get_extent_map(struct inode *inode, struct page *page, size_t pg_offset, * into the tree that are removed when the IO is done (by the end_io * handlers) * XXX JDM: This needs looking at to ensure proper page locking + * return 0 on success, otherwise return error */ static int __do_readpage(struct extent_io_tree *tree, struct page *page, @@ -2899,7 +2900,7 @@ static int __do_readpage(struct extent_io_tree *tree, sector_t sector; struct extent_map *em; struct block_device *bdev; - int ret; + int ret = 0; int nr = 0; size_t pg_offset = 0; size_t iosize; @@ -3080,6 +3081,7 @@ static int __do_readpage(struct extent_io_tree *tree, } else { SetPageError(page); unlock_extent(tree, cur, cur + iosize - 1); + goto out; } cur = cur + iosize; pg_offset += iosize; @@ -3090,7 +3092,7 @@ out: SetPageUptodate(page); unlock_page(page); } - return 0; + return ret; } static inline void __do_contiguous_readpages(struct extent_io_tree *tree, @@ -5230,14 +5232,31 @@ int read_extent_buffer_pages(struct extent_io_tree *tree, atomic_set(&eb->io_pages, num_reads); for (i = start_i; i < num_pages; i++) { page = eb->pages[i]; + if (!PageUptodate(page)) { + if (ret) { + atomic_dec(&eb->io_pages); + unlock_page(page); + continue; + } + ClearPageError(page); err = __extent_read_full_page(tree, page, get_extent, &bio, mirror_num, &bio_flags, READ | REQ_META); - if (err) + if (err) { ret = err; + /* + * We use &bio in above __extent_read_full_page, + * so we ensure that if it returns error, the + * current page fails to add itself to bio and + * it's been unlocked. + * + * We must dec io_pages by ourselves. + */ + atomic_dec(&eb->io_pages); + } } else { unlock_page(page); } -- cgit v1.2.3 From 0fd8c3dae14fb64947842472940b807ca0781da9 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 12 Jul 2016 10:29:37 -0700 Subject: Btrfs: fix panic in balance due to EIO During build_backref_tree(), if we fail to read a btree node, we can eventually run into BUG_ON(cache->nr_nodes) that we put in backref_cache_cleanup(), meaning we have at least one memory leak. This frees the backref_node that we's allocated at the very beginning of build_backref_tree(). Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/relocation.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 13c9a48c83b0..97cafe7291b8 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -1171,8 +1171,12 @@ out: lower = list_entry(useless.next, struct backref_node, list); list_del_init(&lower->list); + if (lower == node) + node = NULL; free_backref_node(cache, lower); } + + free_backref_node(cache, node); return ERR_PTR(err); } ASSERT(!node || !node->detached); -- cgit v1.2.3 From 5a488b9d2c25d98cdd11d09de311bfc83ba09fbd Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 12 Jul 2016 11:24:21 -0700 Subject: Btrfs: fix unexpected balance crash due to BUG_ON Mounting a btrfs can resume previous balance operations asynchronously. An user got a crash when one drive has some corrupt sectors. Since balance can cancel itself in case of any error, we can gracefully return errors to upper layers and let balance do the cancel job. Reported-by: sash Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 28 ++++++++++++++++++++++++---- 1 file changed, 24 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 51bd3ee8e64f..05aec96e4c96 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3460,7 +3460,7 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) u64 size_to_free; u64 chunk_type; struct btrfs_chunk *chunk; - struct btrfs_path *path; + struct btrfs_path *path = NULL; struct btrfs_key key; struct btrfs_key found_key; struct btrfs_trans_handle *trans; @@ -3494,13 +3494,33 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) ret = btrfs_shrink_device(device, old_size - size_to_free); if (ret == -ENOSPC) break; - BUG_ON(ret); + if (ret) { + /* btrfs_shrink_device never returns ret > 0 */ + WARN_ON(ret > 0); + goto error; + } trans = btrfs_start_transaction(dev_root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + btrfs_info_in_rcu(fs_info, + "resize: unable to start transaction after shrinking device %s (error %d), old size %llu, new size %llu", + rcu_str_deref(device->name), ret, + old_size, old_size - size_to_free); + goto error; + } ret = btrfs_grow_device(trans, device, old_size); - BUG_ON(ret); + if (ret) { + btrfs_end_transaction(trans, dev_root); + /* btrfs_grow_device never returns ret > 0 */ + WARN_ON(ret > 0); + btrfs_info_in_rcu(fs_info, + "resize: unable to grow device after shrinking device %s (error %d), old size %llu, new size %llu", + rcu_str_deref(device->name), ret, + old_size, old_size - size_to_free); + goto error; + } btrfs_end_transaction(trans, dev_root); } -- cgit v1.2.3 From 35f4e5e6f198a11c11606a742b309aa80271c66d Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Wed, 13 Jul 2016 16:19:15 +0300 Subject: btrfs: Add ratelimit to btrfs printing This patch adds ratelimiting to all messages which are not using the _rl version of the various printing APIs in btrfs. This is designed to be used as a safety net, since a flood messages might cause the softlockup detector to trigger. To reduce interference between different classes of messages use a separate ratelimit state for every class of message. Signed-off-by: Nikolay Borisov Signed-off-by: David Sterba --- fs/btrfs/super.c | 26 ++++++++++++++++++++++++-- 1 file changed, 24 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 60e7179ed4b7..8766d560a48f 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -184,6 +184,22 @@ static const char * const logtypes[] = { "debug", }; + +/* + * Use one ratelimit state per log level so that a flood of less important + * messages doesn't cause more important ones to be dropped. + */ +static struct ratelimit_state printk_limits[] = { + RATELIMIT_STATE_INIT(printk_limits[0], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[1], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[2], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[3], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[4], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[5], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[6], DEFAULT_RATELIMIT_INTERVAL, 100), + RATELIMIT_STATE_INIT(printk_limits[7], DEFAULT_RATELIMIT_INTERVAL, 100), +}; + void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...) { struct super_block *sb = fs_info->sb; @@ -192,6 +208,7 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...) va_list args; const char *type = logtypes[4]; int kern_level; + struct ratelimit_state *ratelimit; va_start(args, fmt); @@ -202,13 +219,18 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...) lvl[size] = '\0'; fmt += size; type = logtypes[kern_level - '0']; - } else + ratelimit = &printk_limits[kern_level - '0']; + } else { *lvl = '\0'; + /* Default to debug output */ + ratelimit = &printk_limits[7]; + } vaf.fmt = fmt; vaf.va = &args; - printk("%sBTRFS %s (device %s): %pV\n", lvl, type, sb->s_id, &vaf); + if (__ratelimit(ratelimit)) + printk("%sBTRFS %s (device %s): %pV\n", lvl, type, sb->s_id, &vaf); va_end(args); } -- cgit v1.2.3 From aee133afcdf77641be66c5c9f32975feca8e6bd0 Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Wed, 13 Jul 2016 16:19:14 +0300 Subject: btrfs: Ratelimit "no csum found" info message Recently during a crash it became apparent that this particular message can be printed so many times that it causes the softlockup detector to trigger. Fix it by ratelimiting it. Signed-off-by: Nikolay Borisov Signed-off-by: David Sterba --- fs/btrfs/file-item.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 62a81ee13a5f..6b58d0620e2f 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -250,7 +250,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, offset + root->sectorsize - 1, EXTENT_NODATASUM); } else { - btrfs_info(BTRFS_I(inode)->root->fs_info, + btrfs_info_rl(BTRFS_I(inode)->root->fs_info, "no csum found for inode %llu start %llu", btrfs_ino(inode), offset); } -- cgit v1.2.3 From 05653ef38659edc46a886c9bda105e64edba6896 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 15 Jul 2016 15:23:37 +0200 Subject: btrfs: hide test-only member under ifdef Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 2 ++ fs/btrfs/extent-tree.c | 2 ++ 2 files changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 83a6a931af09..8627b7f26a60 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1181,8 +1181,10 @@ struct btrfs_root { u64 highest_objectid; +#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS /* only used with CONFIG_BTRFS_FS_RUN_SANITY_TESTS is enabled */ u64 alloc_bytenr; +#endif u64 defrag_trans_start; struct btrfs_key defrag_progress; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 3bb9607edc5d..81d83300c99b 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -8354,6 +8354,7 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, bool skinny_metadata = btrfs_fs_incompat(root->fs_info, SKINNY_METADATA); +#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS if (btrfs_test_is_dummy_root(root)) { buf = btrfs_init_new_buffer(trans, root, root->alloc_bytenr, level); @@ -8361,6 +8362,7 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, root->alloc_bytenr += blocksize; return buf; } +#endif block_rsv = use_block_rsv(trans, root, blocksize); if (IS_ERR(block_rsv)) -- cgit v1.2.3 From 9f8d49095b86d3494ee184f6ede4a5b179d5dc5c Mon Sep 17 00:00:00 2001 From: David Sterba Date: Fri, 15 Jul 2016 15:31:56 +0200 Subject: btrfs: remove obsolete part of comment in statfs The mixed blockgroup reporting has been fixed by commit ae02d1bd070767e109f4a6f1bb1f466e9698a355 "btrfs: fix mixed block count of available space" Signed-off-by: David Sterba --- fs/btrfs/super.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 8766d560a48f..208b15e7a171 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2052,9 +2052,6 @@ static int btrfs_calc_avail_data_space(struct btrfs_root *root, u64 *free_bytes) * chunk). * * If metadata is exhausted, f_bavail will be 0. - * - * FIXME: not accurate for mixed block groups, total and free/used are ok, - * available appears slightly larger. */ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf) { -- cgit v1.2.3 From cb001095ca705dcd95f57fe98867e38a4889916d Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Thu, 9 Jun 2016 16:22:11 -0400 Subject: btrfs: plumb fs_info into btrfs_work In order to provide an fsid for trace events, we'll need a btrfs_fs_info pointer. The most lightweight way to do that for btrfs_work structures is to associate it with the __btrfs_workqueue structure. Each queued btrfs_work structure has a workqueue associated with it, so that's a natural fit. It's a privately defined structures, so we add accessors to retrieve the fs_info pointer. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/async-thread.c | 31 +++++++++++++++++++++++++------ fs/btrfs/async-thread.h | 6 +++++- fs/btrfs/disk-io.c | 47 ++++++++++++++++++++++++++++------------------- fs/btrfs/scrub.c | 10 +++++----- 4 files changed, 63 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/async-thread.c b/fs/btrfs/async-thread.c index 5fb60ea7eee2..e0f071f6b5a7 100644 --- a/fs/btrfs/async-thread.c +++ b/fs/btrfs/async-thread.c @@ -34,6 +34,10 @@ struct __btrfs_workqueue { struct workqueue_struct *normal_wq; + + /* File system this workqueue services */ + struct btrfs_fs_info *fs_info; + /* List head pointing to ordered work list */ struct list_head ordered_list; @@ -70,6 +74,18 @@ void btrfs_##name(struct work_struct *arg) \ normal_work_helper(work); \ } +struct btrfs_fs_info * +btrfs_workqueue_owner(struct __btrfs_workqueue *wq) +{ + return wq->fs_info; +} + +struct btrfs_fs_info * +btrfs_work_owner(struct btrfs_work *work) +{ + return work->wq->fs_info; +} + BTRFS_WORK_HELPER(worker_helper); BTRFS_WORK_HELPER(delalloc_helper); BTRFS_WORK_HELPER(flush_delalloc_helper); @@ -94,14 +110,15 @@ BTRFS_WORK_HELPER(scrubnc_helper); BTRFS_WORK_HELPER(scrubparity_helper); static struct __btrfs_workqueue * -__btrfs_alloc_workqueue(const char *name, unsigned int flags, int limit_active, - int thresh) +__btrfs_alloc_workqueue(struct btrfs_fs_info *fs_info, const char *name, + unsigned int flags, int limit_active, int thresh) { struct __btrfs_workqueue *ret = kzalloc(sizeof(*ret), GFP_KERNEL); if (!ret) return NULL; + ret->fs_info = fs_info; ret->limit_active = limit_active; atomic_set(&ret->pending, 0); if (thresh == 0) @@ -143,7 +160,8 @@ __btrfs_alloc_workqueue(const char *name, unsigned int flags, int limit_active, static inline void __btrfs_destroy_workqueue(struct __btrfs_workqueue *wq); -struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name, +struct btrfs_workqueue *btrfs_alloc_workqueue(struct btrfs_fs_info *fs_info, + const char *name, unsigned int flags, int limit_active, int thresh) @@ -153,7 +171,8 @@ struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name, if (!ret) return NULL; - ret->normal = __btrfs_alloc_workqueue(name, flags & ~WQ_HIGHPRI, + ret->normal = __btrfs_alloc_workqueue(fs_info, name, + flags & ~WQ_HIGHPRI, limit_active, thresh); if (!ret->normal) { kfree(ret); @@ -161,8 +180,8 @@ struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name, } if (flags & WQ_HIGHPRI) { - ret->high = __btrfs_alloc_workqueue(name, flags, limit_active, - thresh); + ret->high = __btrfs_alloc_workqueue(fs_info, name, flags, + limit_active, thresh); if (!ret->high) { __btrfs_destroy_workqueue(ret->normal); kfree(ret); diff --git a/fs/btrfs/async-thread.h b/fs/btrfs/async-thread.h index ad4d0647d1a6..8e52484cd461 100644 --- a/fs/btrfs/async-thread.h +++ b/fs/btrfs/async-thread.h @@ -21,6 +21,7 @@ #define __BTRFS_ASYNC_THREAD_ #include +struct btrfs_fs_info; struct btrfs_workqueue; /* Internal use only */ struct __btrfs_workqueue; @@ -67,7 +68,8 @@ BTRFS_WORK_HELPER_PROTO(scrubnc_helper); BTRFS_WORK_HELPER_PROTO(scrubparity_helper); -struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name, +struct btrfs_workqueue *btrfs_alloc_workqueue(struct btrfs_fs_info *fs_info, + const char *name, unsigned int flags, int limit_active, int thresh); @@ -80,4 +82,6 @@ void btrfs_queue_work(struct btrfs_workqueue *wq, void btrfs_destroy_workqueue(struct btrfs_workqueue *wq); void btrfs_workqueue_set_max(struct btrfs_workqueue *wq, int max); void btrfs_set_work_high_priority(struct btrfs_work *work); +struct btrfs_fs_info *btrfs_work_owner(struct btrfs_work *work); +struct btrfs_fs_info *btrfs_workqueue_owner(struct __btrfs_workqueue *wq); #endif diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 49565626321e..cb6d13c39bee 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2310,17 +2310,19 @@ static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info, unsigned int flags = WQ_MEM_RECLAIM | WQ_FREEZABLE | WQ_UNBOUND; fs_info->workers = - btrfs_alloc_workqueue("worker", flags | WQ_HIGHPRI, - max_active, 16); + btrfs_alloc_workqueue(fs_info, "worker", + flags | WQ_HIGHPRI, max_active, 16); fs_info->delalloc_workers = - btrfs_alloc_workqueue("delalloc", flags, max_active, 2); + btrfs_alloc_workqueue(fs_info, "delalloc", + flags, max_active, 2); fs_info->flush_workers = - btrfs_alloc_workqueue("flush_delalloc", flags, max_active, 0); + btrfs_alloc_workqueue(fs_info, "flush_delalloc", + flags, max_active, 0); fs_info->caching_workers = - btrfs_alloc_workqueue("cache", flags, max_active, 0); + btrfs_alloc_workqueue(fs_info, "cache", flags, max_active, 0); /* * a higher idle thresh on the submit workers makes it much more @@ -2328,41 +2330,48 @@ static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info, * devices */ fs_info->submit_workers = - btrfs_alloc_workqueue("submit", flags, + btrfs_alloc_workqueue(fs_info, "submit", flags, min_t(u64, fs_devices->num_devices, max_active), 64); fs_info->fixup_workers = - btrfs_alloc_workqueue("fixup", flags, 1, 0); + btrfs_alloc_workqueue(fs_info, "fixup", flags, 1, 0); /* * endios are largely parallel and should have a very * low idle thresh */ fs_info->endio_workers = - btrfs_alloc_workqueue("endio", flags, max_active, 4); + btrfs_alloc_workqueue(fs_info, "endio", flags, max_active, 4); fs_info->endio_meta_workers = - btrfs_alloc_workqueue("endio-meta", flags, max_active, 4); + btrfs_alloc_workqueue(fs_info, "endio-meta", flags, + max_active, 4); fs_info->endio_meta_write_workers = - btrfs_alloc_workqueue("endio-meta-write", flags, max_active, 2); + btrfs_alloc_workqueue(fs_info, "endio-meta-write", flags, + max_active, 2); fs_info->endio_raid56_workers = - btrfs_alloc_workqueue("endio-raid56", flags, max_active, 4); + btrfs_alloc_workqueue(fs_info, "endio-raid56", flags, + max_active, 4); fs_info->endio_repair_workers = - btrfs_alloc_workqueue("endio-repair", flags, 1, 0); + btrfs_alloc_workqueue(fs_info, "endio-repair", flags, 1, 0); fs_info->rmw_workers = - btrfs_alloc_workqueue("rmw", flags, max_active, 2); + btrfs_alloc_workqueue(fs_info, "rmw", flags, max_active, 2); fs_info->endio_write_workers = - btrfs_alloc_workqueue("endio-write", flags, max_active, 2); + btrfs_alloc_workqueue(fs_info, "endio-write", flags, + max_active, 2); fs_info->endio_freespace_worker = - btrfs_alloc_workqueue("freespace-write", flags, max_active, 0); + btrfs_alloc_workqueue(fs_info, "freespace-write", flags, + max_active, 0); fs_info->delayed_workers = - btrfs_alloc_workqueue("delayed-meta", flags, max_active, 0); + btrfs_alloc_workqueue(fs_info, "delayed-meta", flags, + max_active, 0); fs_info->readahead_workers = - btrfs_alloc_workqueue("readahead", flags, max_active, 2); + btrfs_alloc_workqueue(fs_info, "readahead", flags, + max_active, 2); fs_info->qgroup_rescan_workers = - btrfs_alloc_workqueue("qgroup-rescan", flags, 1, 0); + btrfs_alloc_workqueue(fs_info, "qgroup-rescan", flags, 1, 0); fs_info->extent_workers = - btrfs_alloc_workqueue("extent-refs", flags, + btrfs_alloc_workqueue(fs_info, "extent-refs", flags, min_t(u64, fs_devices->num_devices, max_active), 8); diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index acfe72004646..68c8a09ae7e5 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3781,27 +3781,27 @@ static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info, if (fs_info->scrub_workers_refcnt == 0) { if (is_dev_replace) fs_info->scrub_workers = - btrfs_alloc_workqueue("scrub", flags, + btrfs_alloc_workqueue(fs_info, "scrub", flags, 1, 4); else fs_info->scrub_workers = - btrfs_alloc_workqueue("scrub", flags, + btrfs_alloc_workqueue(fs_info, "scrub", flags, max_active, 4); if (!fs_info->scrub_workers) goto fail_scrub_workers; fs_info->scrub_wr_completion_workers = - btrfs_alloc_workqueue("scrubwrc", flags, + btrfs_alloc_workqueue(fs_info, "scrubwrc", flags, max_active, 2); if (!fs_info->scrub_wr_completion_workers) goto fail_scrub_wr_completion_workers; fs_info->scrub_nocow_workers = - btrfs_alloc_workqueue("scrubnc", flags, 1, 0); + btrfs_alloc_workqueue(fs_info, "scrubnc", flags, 1, 0); if (!fs_info->scrub_nocow_workers) goto fail_scrub_nocow_workers; fs_info->scrub_parity_workers = - btrfs_alloc_workqueue("scrubparity", flags, + btrfs_alloc_workqueue(fs_info, "scrubparity", flags, max_active, 2); if (!fs_info->scrub_parity_workers) goto fail_scrub_parity_workers; -- cgit v1.2.3 From bc074524e123ded281cde25ebc5661910f9679e3 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Thu, 9 Jun 2016 17:27:55 -0400 Subject: btrfs: prefix fsid to all trace events When using trace events to debug a problem, it's impossible to determine which file system generated a particular event. This patch adds a macro to prefix standard information to the head of a trace event. The extent_state alloc/free events are all that's left without an fs_info available. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/delayed-ref.c | 9 +- fs/btrfs/extent-tree.c | 10 +- fs/btrfs/qgroup.c | 19 +-- fs/btrfs/qgroup.h | 9 +- fs/btrfs/super.c | 2 +- include/trace/events/btrfs.h | 295 +++++++++++++++++++++++-------------------- 6 files changed, 186 insertions(+), 158 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index 08e452599249..b6d210e7a993 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -606,7 +606,8 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info, qrecord->num_bytes = num_bytes; qrecord->old_roots = NULL; - qexisting = btrfs_qgroup_insert_dirty_extent(delayed_refs, + qexisting = btrfs_qgroup_insert_dirty_extent(fs_info, + delayed_refs, qrecord); if (qexisting) kfree(qrecord); @@ -615,7 +616,7 @@ add_delayed_ref_head(struct btrfs_fs_info *fs_info, spin_lock_init(&head_ref->lock); mutex_init(&head_ref->mutex); - trace_add_delayed_ref_head(ref, head_ref, action); + trace_add_delayed_ref_head(fs_info, ref, head_ref, action); existing = htree_insert(&delayed_refs->href_root, &head_ref->href_node); @@ -682,7 +683,7 @@ add_delayed_tree_ref(struct btrfs_fs_info *fs_info, ref->type = BTRFS_TREE_BLOCK_REF_KEY; full_ref->level = level; - trace_add_delayed_tree_ref(ref, full_ref, action); + trace_add_delayed_tree_ref(fs_info, ref, full_ref, action); ret = add_delayed_ref_tail_merge(trans, delayed_refs, head_ref, ref); @@ -739,7 +740,7 @@ add_delayed_data_ref(struct btrfs_fs_info *fs_info, full_ref->objectid = owner; full_ref->offset = offset; - trace_add_delayed_data_ref(ref, full_ref, action); + trace_add_delayed_data_ref(fs_info, ref, full_ref, action); ret = add_delayed_ref_tail_merge(trans, delayed_refs, head_ref, ref); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 81d83300c99b..73f6dd2cf3cb 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2204,7 +2204,7 @@ static int run_delayed_data_ref(struct btrfs_trans_handle *trans, ins.type = BTRFS_EXTENT_ITEM_KEY; ref = btrfs_delayed_node_to_data_ref(node); - trace_run_delayed_data_ref(node, ref, node->action); + trace_run_delayed_data_ref(root->fs_info, node, ref, node->action); if (node->type == BTRFS_SHARED_DATA_REF_KEY) parent = ref->parent; @@ -2359,7 +2359,7 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans, SKINNY_METADATA); ref = btrfs_delayed_node_to_tree_ref(node); - trace_run_delayed_tree_ref(node, ref, node->action); + trace_run_delayed_tree_ref(root->fs_info, node, ref, node->action); if (node->type == BTRFS_SHARED_BLOCK_REF_KEY) parent = ref->parent; @@ -2423,7 +2423,8 @@ static int run_one_delayed_ref(struct btrfs_trans_handle *trans, */ BUG_ON(extent_op); head = btrfs_delayed_node_to_head(node); - trace_run_delayed_ref_head(node, head, node->action); + trace_run_delayed_ref_head(root->fs_info, node, head, + node->action); if (insert_reserved) { btrfs_pin_extent(root, node->bytenr, @@ -8542,7 +8543,8 @@ static int record_one_subtree_extent(struct btrfs_trans_handle *trans, delayed_refs = &trans->transaction->delayed_refs; spin_lock(&delayed_refs->lock); - if (btrfs_qgroup_insert_dirty_extent(delayed_refs, qrecord)) + if (btrfs_qgroup_insert_dirty_extent(trans->root->fs_info, + delayed_refs, qrecord)) kfree(qrecord); spin_unlock(&delayed_refs->lock); diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 9d4c05b14f6e..13e28d8a7d42 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -1453,9 +1453,10 @@ int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans, return ret; } -struct btrfs_qgroup_extent_record -*btrfs_qgroup_insert_dirty_extent(struct btrfs_delayed_ref_root *delayed_refs, - struct btrfs_qgroup_extent_record *record) +struct btrfs_qgroup_extent_record * +btrfs_qgroup_insert_dirty_extent(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_qgroup_extent_record *record) { struct rb_node **p = &delayed_refs->dirty_extent_root.rb_node; struct rb_node *parent_node = NULL; @@ -1463,7 +1464,7 @@ struct btrfs_qgroup_extent_record u64 bytenr = record->bytenr; assert_spin_locked(&delayed_refs->lock); - trace_btrfs_qgroup_insert_dirty_extent(record); + trace_btrfs_qgroup_insert_dirty_extent(fs_info, record); while (*p) { parent_node = *p; @@ -1595,8 +1596,8 @@ static int qgroup_update_counters(struct btrfs_fs_info *fs_info, cur_old_count = btrfs_qgroup_get_old_refcnt(qg, seq); cur_new_count = btrfs_qgroup_get_new_refcnt(qg, seq); - trace_qgroup_update_counters(qg->qgroupid, cur_old_count, - cur_new_count); + trace_qgroup_update_counters(fs_info, qg->qgroupid, + cur_old_count, cur_new_count); /* Rfer update part */ if (cur_old_count == 0 && cur_new_count > 0) { @@ -1687,8 +1688,8 @@ btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, goto out_free; BUG_ON(!fs_info->quota_root); - trace_btrfs_qgroup_account_extent(bytenr, num_bytes, nr_old_roots, - nr_new_roots); + trace_btrfs_qgroup_account_extent(fs_info, bytenr, num_bytes, + nr_old_roots, nr_new_roots); qgroups = ulist_alloc(GFP_NOFS); if (!qgroups) { @@ -1759,7 +1760,7 @@ int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans, record = rb_entry(node, struct btrfs_qgroup_extent_record, node); - trace_btrfs_qgroup_account_extents(record); + trace_btrfs_qgroup_account_extents(fs_info, record); if (!ret) { /* diff --git a/fs/btrfs/qgroup.h b/fs/btrfs/qgroup.h index ecb2c143ef75..710887c06aaf 100644 --- a/fs/btrfs/qgroup.h +++ b/fs/btrfs/qgroup.h @@ -63,9 +63,10 @@ void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info); struct btrfs_delayed_extent_op; int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info); -struct btrfs_qgroup_extent_record -*btrfs_qgroup_insert_dirty_extent(struct btrfs_delayed_ref_root *delayed_refs, - struct btrfs_qgroup_extent_record *record); +struct btrfs_qgroup_extent_record * +btrfs_qgroup_insert_dirty_extent(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_root *delayed_refs, + struct btrfs_qgroup_extent_record *record); int btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, struct btrfs_fs_info *fs_info, @@ -88,7 +89,7 @@ static inline void btrfs_qgroup_free_delayed_ref(struct btrfs_fs_info *fs_info, u64 ref_root, u64 num_bytes) { btrfs_qgroup_free_refroot(fs_info, ref_root, num_bytes); - trace_btrfs_qgroup_free_delayed_ref(ref_root, num_bytes); + trace_btrfs_qgroup_free_delayed_ref(fs_info, ref_root, num_bytes); } void assert_qgroups_uptodate(struct btrfs_trans_handle *trans); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 208b15e7a171..d965abbcac8e 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1171,7 +1171,7 @@ int btrfs_sync_fs(struct super_block *sb, int wait) struct btrfs_fs_info *fs_info = btrfs_sb(sb); struct btrfs_root *root = fs_info->tree_root; - trace_btrfs_sync_fs(wait); + trace_btrfs_sync_fs(fs_info, wait); if (!wait) { filemap_flush(fs_info->btree_inode->i_mapping); diff --git a/include/trace/events/btrfs.h b/include/trace/events/btrfs.h index 5b81ef304388..e030d6f6c19a 100644 --- a/include/trace/events/btrfs.h +++ b/include/trace/events/btrfs.h @@ -66,6 +66,21 @@ struct btrfs_qgroup_extent_record; { BTRFS_BLOCK_GROUP_RAID6, "RAID6"} #define BTRFS_UUID_SIZE 16 +#define TP_STRUCT__entry_fsid __array(u8, fsid, BTRFS_UUID_SIZE) + +#define TP_fast_assign_fsid(fs_info) \ + memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE) + +#define TP_STRUCT__entry_btrfs(args...) \ + TP_STRUCT__entry( \ + TP_STRUCT__entry_fsid \ + args) +#define TP_fast_assign_btrfs(fs_info, args...) \ + TP_fast_assign( \ + TP_fast_assign_fsid(fs_info); \ + args) +#define TP_printk_btrfs(fmt, args...) \ + TP_printk("%pU: " fmt, __entry->fsid, args) TRACE_EVENT(btrfs_transaction_commit, @@ -73,17 +88,17 @@ TRACE_EVENT(btrfs_transaction_commit, TP_ARGS(root), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, generation ) __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(root->fs_info, __entry->generation = root->fs_info->generation; __entry->root_objectid = root->root_key.objectid; ), - TP_printk("root = %llu(%s), gen = %llu", + TP_printk_btrfs("root = %llu(%s), gen = %llu", show_root_type(__entry->root_objectid), (unsigned long long)__entry->generation) ); @@ -94,7 +109,7 @@ DECLARE_EVENT_CLASS(btrfs__inode, TP_ARGS(inode), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( ino_t, ino ) __field( blkcnt_t, blocks ) __field( u64, disk_i_size ) @@ -104,7 +119,7 @@ DECLARE_EVENT_CLASS(btrfs__inode, __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(inode->i_sb), __entry->ino = inode->i_ino; __entry->blocks = inode->i_blocks; __entry->disk_i_size = BTRFS_I(inode)->disk_i_size; @@ -115,7 +130,7 @@ DECLARE_EVENT_CLASS(btrfs__inode, BTRFS_I(inode)->root->root_key.objectid; ), - TP_printk("root = %llu(%s), gen = %llu, ino = %lu, blocks = %llu, " + TP_printk_btrfs("root = %llu(%s), gen = %llu, ino = %lu, blocks = %llu, " "disk_i_size = %llu, last_trans = %llu, logged_trans = %llu", show_root_type(__entry->root_objectid), (unsigned long long)__entry->generation, @@ -175,7 +190,7 @@ TRACE_EVENT_CONDITION(btrfs_get_extent, TP_CONDITION(map), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, root_objectid ) __field( u64, start ) __field( u64, len ) @@ -187,7 +202,7 @@ TRACE_EVENT_CONDITION(btrfs_get_extent, __field( unsigned int, compress_type ) ), - TP_fast_assign( + TP_fast_assign_btrfs(root->fs_info, __entry->root_objectid = root->root_key.objectid; __entry->start = map->start; __entry->len = map->len; @@ -199,7 +214,7 @@ TRACE_EVENT_CONDITION(btrfs_get_extent, __entry->compress_type = map->compress_type; ), - TP_printk("root = %llu(%s), start = %llu, len = %llu, " + TP_printk_btrfs("root = %llu(%s), start = %llu, len = %llu, " "orig_start = %llu, block_start = %llu(%s), " "block_len = %llu, flags = %s, refs = %u, " "compress_type = %u", @@ -233,7 +248,7 @@ DECLARE_EVENT_CLASS(btrfs__ordered_extent, TP_ARGS(inode, ordered), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( ino_t, ino ) __field( u64, file_offset ) __field( u64, start ) @@ -246,7 +261,7 @@ DECLARE_EVENT_CLASS(btrfs__ordered_extent, __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(inode->i_sb), __entry->ino = inode->i_ino; __entry->file_offset = ordered->file_offset; __entry->start = ordered->start; @@ -260,7 +275,7 @@ DECLARE_EVENT_CLASS(btrfs__ordered_extent, BTRFS_I(inode)->root->root_key.objectid; ), - TP_printk("root = %llu(%s), ino = %llu, file_offset = %llu, " + TP_printk_btrfs("root = %llu(%s), ino = %llu, file_offset = %llu, " "start = %llu, len = %llu, disk_len = %llu, " "bytes_left = %llu, flags = %s, compress_type = %d, " "refs = %d", @@ -310,7 +325,7 @@ DECLARE_EVENT_CLASS(btrfs__writepage, TP_ARGS(page, inode, wbc), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( ino_t, ino ) __field( pgoff_t, index ) __field( long, nr_to_write ) @@ -324,7 +339,7 @@ DECLARE_EVENT_CLASS(btrfs__writepage, __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(inode->i_sb), __entry->ino = inode->i_ino; __entry->index = page->index; __entry->nr_to_write = wbc->nr_to_write; @@ -339,7 +354,7 @@ DECLARE_EVENT_CLASS(btrfs__writepage, BTRFS_I(inode)->root->root_key.objectid; ), - TP_printk("root = %llu(%s), ino = %lu, page_index = %lu, " + TP_printk_btrfs("root = %llu(%s), ino = %lu, page_index = %lu, " "nr_to_write = %ld, pages_skipped = %ld, range_start = %llu, " "range_end = %llu, for_kupdate = %d, " "for_reclaim = %d, range_cyclic = %d, writeback_index = %lu", @@ -366,7 +381,7 @@ TRACE_EVENT(btrfs_writepage_end_io_hook, TP_ARGS(page, start, end, uptodate), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( ino_t, ino ) __field( pgoff_t, index ) __field( u64, start ) @@ -375,7 +390,7 @@ TRACE_EVENT(btrfs_writepage_end_io_hook, __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(page->mapping->host->i_sb), __entry->ino = page->mapping->host->i_ino; __entry->index = page->index; __entry->start = start; @@ -385,7 +400,7 @@ TRACE_EVENT(btrfs_writepage_end_io_hook, BTRFS_I(page->mapping->host)->root->root_key.objectid; ), - TP_printk("root = %llu(%s), ino = %lu, page_index = %lu, start = %llu, " + TP_printk_btrfs("root = %llu(%s), ino = %lu, page_index = %lu, start = %llu, " "end = %llu, uptodate = %d", show_root_type(__entry->root_objectid), (unsigned long)__entry->ino, (unsigned long)__entry->index, @@ -399,7 +414,7 @@ TRACE_EVENT(btrfs_sync_file, TP_ARGS(file, datasync), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( ino_t, ino ) __field( ino_t, parent ) __field( int, datasync ) @@ -410,6 +425,7 @@ TRACE_EVENT(btrfs_sync_file, struct dentry *dentry = file->f_path.dentry; struct inode *inode = d_inode(dentry); + TP_fast_assign_fsid(btrfs_sb(file->f_path.dentry->d_sb)); __entry->ino = inode->i_ino; __entry->parent = d_inode(dentry->d_parent)->i_ino; __entry->datasync = datasync; @@ -417,7 +433,7 @@ TRACE_EVENT(btrfs_sync_file, BTRFS_I(inode)->root->root_key.objectid; ), - TP_printk("root = %llu(%s), ino = %ld, parent = %ld, datasync = %d", + TP_printk_btrfs("root = %llu(%s), ino = %ld, parent = %ld, datasync = %d", show_root_type(__entry->root_objectid), (unsigned long)__entry->ino, (unsigned long)__entry->parent, __entry->datasync) @@ -425,19 +441,19 @@ TRACE_EVENT(btrfs_sync_file, TRACE_EVENT(btrfs_sync_fs, - TP_PROTO(int wait), + TP_PROTO(struct btrfs_fs_info *fs_info, int wait), - TP_ARGS(wait), + TP_ARGS(fs_info, wait), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( int, wait ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->wait = wait; ), - TP_printk("wait = %d", __entry->wait) + TP_printk_btrfs("wait = %d", __entry->wait) ); TRACE_EVENT(btrfs_add_block_group, @@ -490,13 +506,14 @@ TRACE_EVENT(btrfs_add_block_group, DECLARE_EVENT_CLASS(btrfs_delayed_tree_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_tree_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action), + TP_ARGS(fs_info, ref, full_ref, action), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bytenr ) __field( u64, num_bytes ) __field( int, action ) @@ -507,7 +524,7 @@ DECLARE_EVENT_CLASS(btrfs_delayed_tree_ref, __field( u64, seq ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->bytenr = ref->bytenr; __entry->num_bytes = ref->num_bytes; __entry->action = action; @@ -518,7 +535,7 @@ DECLARE_EVENT_CLASS(btrfs_delayed_tree_ref, __entry->seq = ref->seq; ), - TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, " + TP_printk_btrfs("bytenr = %llu, num_bytes = %llu, action = %s, " "parent = %llu(%s), ref_root = %llu(%s), level = %d, " "type = %s, seq = %llu", (unsigned long long)__entry->bytenr, @@ -532,31 +549,34 @@ DECLARE_EVENT_CLASS(btrfs_delayed_tree_ref, DEFINE_EVENT(btrfs_delayed_tree_ref, add_delayed_tree_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_tree_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action) + TP_ARGS(fs_info, ref, full_ref, action) ); DEFINE_EVENT(btrfs_delayed_tree_ref, run_delayed_tree_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_tree_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action) + TP_ARGS(fs_info, ref, full_ref, action) ); DECLARE_EVENT_CLASS(btrfs_delayed_data_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_data_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action), + TP_ARGS(fs_info, ref, full_ref, action), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bytenr ) __field( u64, num_bytes ) __field( int, action ) @@ -568,7 +588,7 @@ DECLARE_EVENT_CLASS(btrfs_delayed_data_ref, __field( u64, seq ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->bytenr = ref->bytenr; __entry->num_bytes = ref->num_bytes; __entry->action = action; @@ -580,7 +600,7 @@ DECLARE_EVENT_CLASS(btrfs_delayed_data_ref, __entry->seq = ref->seq; ), - TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, " + TP_printk_btrfs("bytenr = %llu, num_bytes = %llu, action = %s, " "parent = %llu(%s), ref_root = %llu(%s), owner = %llu, " "offset = %llu, type = %s, seq = %llu", (unsigned long long)__entry->bytenr, @@ -596,45 +616,48 @@ DECLARE_EVENT_CLASS(btrfs_delayed_data_ref, DEFINE_EVENT(btrfs_delayed_data_ref, add_delayed_data_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_data_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action) + TP_ARGS(fs_info, ref, full_ref, action) ); DEFINE_EVENT(btrfs_delayed_data_ref, run_delayed_data_ref, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_data_ref *full_ref, int action), - TP_ARGS(ref, full_ref, action) + TP_ARGS(fs_info, ref, full_ref, action) ); DECLARE_EVENT_CLASS(btrfs_delayed_ref_head, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_ref_head *head_ref, int action), - TP_ARGS(ref, head_ref, action), + TP_ARGS(fs_info, ref, head_ref, action), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bytenr ) __field( u64, num_bytes ) __field( int, action ) __field( int, is_data ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->bytenr = ref->bytenr; __entry->num_bytes = ref->num_bytes; __entry->action = action; __entry->is_data = head_ref->is_data; ), - TP_printk("bytenr = %llu, num_bytes = %llu, action = %s, is_data = %d", + TP_printk_btrfs("bytenr = %llu, num_bytes = %llu, action = %s, is_data = %d", (unsigned long long)__entry->bytenr, (unsigned long long)__entry->num_bytes, show_ref_action(__entry->action), @@ -643,20 +666,22 @@ DECLARE_EVENT_CLASS(btrfs_delayed_ref_head, DEFINE_EVENT(btrfs_delayed_ref_head, add_delayed_ref_head, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_ref_head *head_ref, int action), - TP_ARGS(ref, head_ref, action) + TP_ARGS(fs_info, ref, head_ref, action) ); DEFINE_EVENT(btrfs_delayed_ref_head, run_delayed_ref_head, - TP_PROTO(struct btrfs_delayed_ref_node *ref, + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_delayed_ref_node *ref, struct btrfs_delayed_ref_head *head_ref, int action), - TP_ARGS(ref, head_ref, action) + TP_ARGS(fs_info, ref, head_ref, action) ); #define show_chunk_type(type) \ @@ -678,7 +703,7 @@ DECLARE_EVENT_CLASS(btrfs__chunk, TP_ARGS(root, map, offset, size), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( int, num_stripes ) __field( u64, type ) __field( int, sub_stripes ) @@ -687,7 +712,7 @@ DECLARE_EVENT_CLASS(btrfs__chunk, __field( u64, root_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(root->fs_info, __entry->num_stripes = map->num_stripes; __entry->type = map->type; __entry->sub_stripes = map->sub_stripes; @@ -696,7 +721,7 @@ DECLARE_EVENT_CLASS(btrfs__chunk, __entry->root_objectid = root->root_key.objectid; ), - TP_printk("root = %llu(%s), offset = %llu, size = %llu, " + TP_printk_btrfs("root = %llu(%s), offset = %llu, size = %llu, " "num_stripes = %d, sub_stripes = %d, type = %s", show_root_type(__entry->root_objectid), (unsigned long long)__entry->offset, @@ -728,7 +753,7 @@ TRACE_EVENT(btrfs_cow_block, TP_ARGS(root, buf, cow), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, root_objectid ) __field( u64, buf_start ) __field( int, refs ) @@ -737,7 +762,7 @@ TRACE_EVENT(btrfs_cow_block, __field( int, cow_level ) ), - TP_fast_assign( + TP_fast_assign_btrfs(root->fs_info, __entry->root_objectid = root->root_key.objectid; __entry->buf_start = buf->start; __entry->refs = atomic_read(&buf->refs); @@ -746,7 +771,7 @@ TRACE_EVENT(btrfs_cow_block, __entry->cow_level = btrfs_header_level(cow); ), - TP_printk("root = %llu(%s), refs = %d, orig_buf = %llu " + TP_printk_btrfs("root = %llu(%s), refs = %d, orig_buf = %llu " "(orig_level = %d), cow_buf = %llu (cow_level = %d)", show_root_type(__entry->root_objectid), __entry->refs, @@ -763,25 +788,23 @@ TRACE_EVENT(btrfs_space_reservation, TP_ARGS(fs_info, type, val, bytes, reserve), - TP_STRUCT__entry( - __array( u8, fsid, BTRFS_UUID_SIZE ) + TP_STRUCT__entry_btrfs( __string( type, type ) __field( u64, val ) __field( u64, bytes ) __field( int, reserve ) ), - TP_fast_assign( - memcpy(__entry->fsid, fs_info->fsid, BTRFS_UUID_SIZE); + TP_fast_assign_btrfs(fs_info, __assign_str(type, type); __entry->val = val; __entry->bytes = bytes; __entry->reserve = reserve; ), - TP_printk("%pU: %s: %Lu %s %Lu", __entry->fsid, __get_str(type), - __entry->val, __entry->reserve ? "reserve" : "release", - __entry->bytes) + TP_printk_btrfs("%s: %Lu %s %Lu", __get_str(type), __entry->val, + __entry->reserve ? "reserve" : "release", + __entry->bytes) ); #define show_flush_action(action) \ @@ -872,22 +895,19 @@ DECLARE_EVENT_CLASS(btrfs__reserved_extent, TP_ARGS(root, start, len), - TP_STRUCT__entry( - __array( u8, fsid, BTRFS_UUID_SIZE ) - __field( u64, root_objectid ) - __field( u64, start ) - __field( u64, len ) + TP_STRUCT__entry_btrfs( + __field( u64, root_objectid ) + __field( u64, start ) + __field( u64, len ) ), - TP_fast_assign( - memcpy(__entry->fsid, root->fs_info->fsid, BTRFS_UUID_SIZE); + TP_fast_assign_btrfs(root->fs_info, __entry->root_objectid = root->root_key.objectid; __entry->start = start; __entry->len = len; ), - TP_printk("%pU: root = %llu(%s), start = %llu, len = %llu", - __entry->fsid, + TP_printk_btrfs("root = %llu(%s), start = %llu, len = %llu", show_root_type(__entry->root_objectid), (unsigned long long)__entry->start, (unsigned long long)__entry->len) @@ -914,21 +934,21 @@ TRACE_EVENT(find_free_extent, TP_ARGS(root, num_bytes, empty_size, data), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, root_objectid ) __field( u64, num_bytes ) __field( u64, empty_size ) __field( u64, data ) ), - TP_fast_assign( + TP_fast_assign_btrfs(root->fs_info, __entry->root_objectid = root->root_key.objectid; __entry->num_bytes = num_bytes; __entry->empty_size = empty_size; __entry->data = data; ), - TP_printk("root = %Lu(%s), len = %Lu, empty_size = %Lu, " + TP_printk_btrfs("root = %Lu(%s), len = %Lu, empty_size = %Lu, " "flags = %Lu(%s)", show_root_type(__entry->root_objectid), __entry->num_bytes, __entry->empty_size, __entry->data, __print_flags((unsigned long)__entry->data, "|", @@ -943,8 +963,7 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent, TP_ARGS(root, block_group, start, len), - TP_STRUCT__entry( - __array( u8, fsid, BTRFS_UUID_SIZE ) + TP_STRUCT__entry_btrfs( __field( u64, root_objectid ) __field( u64, bg_objectid ) __field( u64, flags ) @@ -952,8 +971,7 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent, __field( u64, len ) ), - TP_fast_assign( - memcpy(__entry->fsid, root->fs_info->fsid, BTRFS_UUID_SIZE); + TP_fast_assign_btrfs(root->fs_info, __entry->root_objectid = root->root_key.objectid; __entry->bg_objectid = block_group->key.objectid; __entry->flags = block_group->flags; @@ -961,8 +979,8 @@ DECLARE_EVENT_CLASS(btrfs__reserve_extent, __entry->len = len; ), - TP_printk("%pU: root = %Lu(%s), block_group = %Lu, flags = %Lu(%s), " - "start = %Lu, len = %Lu", __entry->fsid, + TP_printk_btrfs("root = %Lu(%s), block_group = %Lu, flags = %Lu(%s), " + "start = %Lu, len = %Lu", show_root_type(__entry->root_objectid), __entry->bg_objectid, __entry->flags, __print_flags((unsigned long)__entry->flags, "|", BTRFS_GROUP_FLAGS), @@ -994,7 +1012,7 @@ TRACE_EVENT(btrfs_find_cluster, TP_ARGS(block_group, start, bytes, empty_size, min_bytes), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bg_objectid ) __field( u64, flags ) __field( u64, start ) @@ -1003,7 +1021,7 @@ TRACE_EVENT(btrfs_find_cluster, __field( u64, min_bytes ) ), - TP_fast_assign( + TP_fast_assign_btrfs(block_group->fs_info, __entry->bg_objectid = block_group->key.objectid; __entry->flags = block_group->flags; __entry->start = start; @@ -1012,7 +1030,7 @@ TRACE_EVENT(btrfs_find_cluster, __entry->min_bytes = min_bytes; ), - TP_printk("block_group = %Lu, flags = %Lu(%s), start = %Lu, len = %Lu," + TP_printk_btrfs("block_group = %Lu, flags = %Lu(%s), start = %Lu, len = %Lu," " empty_size = %Lu, min_bytes = %Lu", __entry->bg_objectid, __entry->flags, __print_flags((unsigned long)__entry->flags, "|", @@ -1026,15 +1044,15 @@ TRACE_EVENT(btrfs_failed_cluster_setup, TP_ARGS(block_group), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bg_objectid ) ), - TP_fast_assign( + TP_fast_assign_btrfs(block_group->fs_info, __entry->bg_objectid = block_group->key.objectid; ), - TP_printk("block_group = %Lu", __entry->bg_objectid) + TP_printk_btrfs("block_group = %Lu", __entry->bg_objectid) ); TRACE_EVENT(btrfs_setup_cluster, @@ -1044,7 +1062,7 @@ TRACE_EVENT(btrfs_setup_cluster, TP_ARGS(block_group, cluster, size, bitmap), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bg_objectid ) __field( u64, flags ) __field( u64, start ) @@ -1053,7 +1071,7 @@ TRACE_EVENT(btrfs_setup_cluster, __field( int, bitmap ) ), - TP_fast_assign( + TP_fast_assign_btrfs(block_group->fs_info, __entry->bg_objectid = block_group->key.objectid; __entry->flags = block_group->flags; __entry->start = cluster->window_start; @@ -1062,7 +1080,7 @@ TRACE_EVENT(btrfs_setup_cluster, __entry->bitmap = bitmap; ), - TP_printk("block_group = %Lu, flags = %Lu(%s), window_start = %Lu, " + TP_printk_btrfs("block_group = %Lu, flags = %Lu(%s), window_start = %Lu, " "size = %Lu, max_size = %Lu, bitmap = %d", __entry->bg_objectid, __entry->flags, @@ -1120,7 +1138,7 @@ DECLARE_EVENT_CLASS(btrfs__work, TP_ARGS(work), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( void *, work ) __field( void *, wq ) __field( void *, func ) @@ -1129,7 +1147,7 @@ DECLARE_EVENT_CLASS(btrfs__work, __field( void *, normal_work ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_work_owner(work), __entry->work = work; __entry->wq = work->wq; __entry->func = work->func; @@ -1138,7 +1156,7 @@ DECLARE_EVENT_CLASS(btrfs__work, __entry->normal_work = &work->normal_work; ), - TP_printk("work=%p (normal_work=%p), wq=%p, func=%pf, ordered_func=%p," + TP_printk_btrfs("work=%p (normal_work=%p), wq=%p, func=%pf, ordered_func=%p," " ordered_free=%p", __entry->work, __entry->normal_work, __entry->wq, __entry->func, __entry->ordered_func, __entry->ordered_free) @@ -1151,15 +1169,15 @@ DECLARE_EVENT_CLASS(btrfs__work__done, TP_ARGS(work), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( void *, work ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_work_owner(work), __entry->work = work; ), - TP_printk("work->%p", __entry->work) + TP_printk_btrfs("work->%p", __entry->work) ); DEFINE_EVENT(btrfs__work, btrfs_work_queued, @@ -1196,19 +1214,19 @@ DECLARE_EVENT_CLASS(btrfs__workqueue, TP_ARGS(wq, name, high), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( void *, wq ) __string( name, name ) __field( int , high ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_workqueue_owner(wq), __entry->wq = wq; __assign_str(name, name); __entry->high = high; ), - TP_printk("name=%s%s, wq=%p", __get_str(name), + TP_printk_btrfs("name=%s%s, wq=%p", __get_str(name), __print_flags(__entry->high, "", {(WQ_HIGHPRI), "-high"}), __entry->wq) @@ -1227,15 +1245,15 @@ DECLARE_EVENT_CLASS(btrfs__workqueue_done, TP_ARGS(wq), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( void *, wq ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_workqueue_owner(wq), __entry->wq = wq; ), - TP_printk("wq=%p", __entry->wq) + TP_printk_btrfs("wq=%p", __entry->wq) ); DEFINE_EVENT(btrfs__workqueue_done, btrfs_workqueue_destroy, @@ -1251,19 +1269,19 @@ DECLARE_EVENT_CLASS(btrfs__qgroup_data_map, TP_ARGS(inode, free_reserved), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, rootid ) __field( unsigned long, ino ) __field( u64, free_reserved ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(inode->i_sb), __entry->rootid = BTRFS_I(inode)->root->objectid; __entry->ino = inode->i_ino; __entry->free_reserved = free_reserved; ), - TP_printk("rootid=%llu, ino=%lu, free_reserved=%llu", + TP_printk_btrfs("rootid=%llu, ino=%lu, free_reserved=%llu", __entry->rootid, __entry->ino, __entry->free_reserved) ); @@ -1292,7 +1310,7 @@ DECLARE_EVENT_CLASS(btrfs__qgroup_rsv_data, TP_ARGS(inode, start, len, reserved, op), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, rootid ) __field( unsigned long, ino ) __field( u64, start ) @@ -1301,7 +1319,7 @@ DECLARE_EVENT_CLASS(btrfs__qgroup_rsv_data, __field( int, op ) ), - TP_fast_assign( + TP_fast_assign_btrfs(btrfs_sb(inode->i_sb), __entry->rootid = BTRFS_I(inode)->root->objectid; __entry->ino = inode->i_ino; __entry->start = start; @@ -1310,7 +1328,7 @@ DECLARE_EVENT_CLASS(btrfs__qgroup_rsv_data, __entry->op = op; ), - TP_printk("root=%llu, ino=%lu, start=%llu, len=%llu, reserved=%llu, op=%s", + TP_printk_btrfs("root=%llu, ino=%lu, start=%llu, len=%llu, reserved=%llu, op=%s", __entry->rootid, __entry->ino, __entry->start, __entry->len, __entry->reserved, __print_flags((unsigned long)__entry->op, "", @@ -1334,86 +1352,90 @@ DEFINE_EVENT(btrfs__qgroup_rsv_data, btrfs_qgroup_release_data, DECLARE_EVENT_CLASS(btrfs__qgroup_delayed_ref, - TP_PROTO(u64 ref_root, u64 reserved), + TP_PROTO(struct btrfs_fs_info *fs_info, u64 ref_root, u64 reserved), - TP_ARGS(ref_root, reserved), + TP_ARGS(fs_info, ref_root, reserved), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, ref_root ) __field( u64, reserved ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->ref_root = ref_root; __entry->reserved = reserved; ), - TP_printk("root=%llu, reserved=%llu, op=free", + TP_printk_btrfs("root=%llu, reserved=%llu, op=free", __entry->ref_root, __entry->reserved) ); DEFINE_EVENT(btrfs__qgroup_delayed_ref, btrfs_qgroup_free_delayed_ref, - TP_PROTO(u64 ref_root, u64 reserved), + TP_PROTO(struct btrfs_fs_info *fs_info, u64 ref_root, u64 reserved), - TP_ARGS(ref_root, reserved) + TP_ARGS(fs_info, ref_root, reserved) ); DECLARE_EVENT_CLASS(btrfs_qgroup_extent, - TP_PROTO(struct btrfs_qgroup_extent_record *rec), + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_qgroup_extent_record *rec), - TP_ARGS(rec), + TP_ARGS(fs_info, rec), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bytenr ) __field( u64, num_bytes ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->bytenr = rec->bytenr, __entry->num_bytes = rec->num_bytes; ), - TP_printk("bytenr = %llu, num_bytes = %llu", + TP_printk_btrfs("bytenr = %llu, num_bytes = %llu", (unsigned long long)__entry->bytenr, (unsigned long long)__entry->num_bytes) ); DEFINE_EVENT(btrfs_qgroup_extent, btrfs_qgroup_account_extents, - TP_PROTO(struct btrfs_qgroup_extent_record *rec), + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_qgroup_extent_record *rec), - TP_ARGS(rec) + TP_ARGS(fs_info, rec) ); DEFINE_EVENT(btrfs_qgroup_extent, btrfs_qgroup_insert_dirty_extent, - TP_PROTO(struct btrfs_qgroup_extent_record *rec), + TP_PROTO(struct btrfs_fs_info *fs_info, + struct btrfs_qgroup_extent_record *rec), - TP_ARGS(rec) + TP_ARGS(fs_info, rec) ); TRACE_EVENT(btrfs_qgroup_account_extent, - TP_PROTO(u64 bytenr, u64 num_bytes, u64 nr_old_roots, u64 nr_new_roots), + TP_PROTO(struct btrfs_fs_info *fs_info, u64 bytenr, + u64 num_bytes, u64 nr_old_roots, u64 nr_new_roots), - TP_ARGS(bytenr, num_bytes, nr_old_roots, nr_new_roots), + TP_ARGS(fs_info, bytenr, num_bytes, nr_old_roots, nr_new_roots), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, bytenr ) __field( u64, num_bytes ) __field( u64, nr_old_roots ) __field( u64, nr_new_roots ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->bytenr = bytenr; __entry->num_bytes = num_bytes; __entry->nr_old_roots = nr_old_roots; __entry->nr_new_roots = nr_new_roots; ), - TP_printk("bytenr = %llu, num_bytes = %llu, nr_old_roots = %llu, " + TP_printk_btrfs("bytenr = %llu, num_bytes = %llu, nr_old_roots = %llu, " "nr_new_roots = %llu", __entry->bytenr, __entry->num_bytes, @@ -1423,23 +1445,24 @@ TRACE_EVENT(btrfs_qgroup_account_extent, TRACE_EVENT(qgroup_update_counters, - TP_PROTO(u64 qgid, u64 cur_old_count, u64 cur_new_count), + TP_PROTO(struct btrfs_fs_info *fs_info, u64 qgid, + u64 cur_old_count, u64 cur_new_count), - TP_ARGS(qgid, cur_old_count, cur_new_count), + TP_ARGS(fs_info, qgid, cur_old_count, cur_new_count), - TP_STRUCT__entry( + TP_STRUCT__entry_btrfs( __field( u64, qgid ) __field( u64, cur_old_count ) __field( u64, cur_new_count ) ), - TP_fast_assign( + TP_fast_assign_btrfs(fs_info, __entry->qgid = qgid; __entry->cur_old_count = cur_old_count; __entry->cur_new_count = cur_new_count; ), - TP_printk("qgid = %llu, cur_old_count = %llu, cur_new_count = %llu", + TP_printk_btrfs("qgid = %llu, cur_old_count = %llu, cur_new_count = %llu", __entry->qgid, __entry->cur_old_count, __entry->cur_new_count) -- cgit v1.2.3 From 3cdde2240d4533ff71fbb8dc9c32d5d57d3cdeed Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Thu, 9 Jun 2016 21:38:35 -0400 Subject: btrfs: btrfs_test_opt and friends should take a btrfs_fs_info btrfs_test_opt and friends only use the root pointer to access the fs_info. Let's pass the fs_info directly in preparation to eliminate similar patterns all over btrfs. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 22 ++++---- fs/btrfs/dev-replace.c | 4 +- fs/btrfs/disk-io.c | 22 ++++---- fs/btrfs/extent-tree.c | 30 +++++----- fs/btrfs/file.c | 2 +- fs/btrfs/free-space-cache.c | 6 +- fs/btrfs/inode-map.c | 12 ++-- fs/btrfs/inode.c | 12 ++-- fs/btrfs/ioctl.c | 2 +- fs/btrfs/super.c | 132 +++++++++++++++++++++++--------------------- fs/btrfs/transaction.c | 6 +- fs/btrfs/tree-log.c | 4 +- fs/btrfs/volumes.c | 11 ++-- 13 files changed, 135 insertions(+), 130 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 8627b7f26a60..61a53cca5963 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1301,21 +1301,21 @@ struct btrfs_root { #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) #define btrfs_raw_test_opt(o, opt) ((o) & BTRFS_MOUNT_##opt) -#define btrfs_test_opt(root, opt) ((root)->fs_info->mount_opt & \ +#define btrfs_test_opt(fs_info, opt) ((fs_info)->mount_opt & \ BTRFS_MOUNT_##opt) -#define btrfs_set_and_info(root, opt, fmt, args...) \ +#define btrfs_set_and_info(fs_info, opt, fmt, args...) \ { \ - if (!btrfs_test_opt(root, opt)) \ - btrfs_info(root->fs_info, fmt, ##args); \ - btrfs_set_opt(root->fs_info->mount_opt, opt); \ + if (!btrfs_test_opt(fs_info, opt)) \ + btrfs_info(fs_info, fmt, ##args); \ + btrfs_set_opt(fs_info->mount_opt, opt); \ } -#define btrfs_clear_and_info(root, opt, fmt, args...) \ +#define btrfs_clear_and_info(fs_info, opt, fmt, args...) \ { \ - if (btrfs_test_opt(root, opt)) \ - btrfs_info(root->fs_info, fmt, ##args); \ - btrfs_clear_opt(root->fs_info->mount_opt, opt); \ + if (btrfs_test_opt(fs_info, opt)) \ + btrfs_info(fs_info, fmt, ##args); \ + btrfs_clear_opt(fs_info->mount_opt, opt); \ } #ifdef CONFIG_BTRFS_DEBUG @@ -1323,9 +1323,9 @@ static inline int btrfs_should_fragment_free_space(struct btrfs_root *root, struct btrfs_block_group_cache *block_group) { - return (btrfs_test_opt(root, FRAGMENT_METADATA) && + return (btrfs_test_opt(root->fs_info, FRAGMENT_METADATA) && block_group->flags & BTRFS_BLOCK_GROUP_METADATA) || - (btrfs_test_opt(root, FRAGMENT_DATA) && + (btrfs_test_opt(root->fs_info, FRAGMENT_DATA) && block_group->flags & BTRFS_BLOCK_GROUP_DATA); } #endif diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c index 63ef9cdf0144..e9bbff3c0029 100644 --- a/fs/btrfs/dev-replace.c +++ b/fs/btrfs/dev-replace.c @@ -142,7 +142,7 @@ no_valid_dev_replace_entry_found: * missing */ if (!dev_replace->srcdev && - !btrfs_test_opt(dev_root, DEGRADED)) { + !btrfs_test_opt(dev_root->fs_info, DEGRADED)) { ret = -EIO; btrfs_warn(fs_info, "cannot mount because device replace operation is ongoing and"); @@ -151,7 +151,7 @@ no_valid_dev_replace_entry_found: src_devid); } if (!dev_replace->tgtdev && - !btrfs_test_opt(dev_root, DEGRADED)) { + !btrfs_test_opt(dev_root->fs_info, DEGRADED)) { ret = -EIO; btrfs_warn(fs_info, "cannot mount because device replace operation is ongoing and"); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index cb6d13c39bee..802a3113854f 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3019,8 +3019,8 @@ retry_root_backup: if (IS_ERR(fs_info->transaction_kthread)) goto fail_cleaner; - if (!btrfs_test_opt(tree_root, SSD) && - !btrfs_test_opt(tree_root, NOSSD) && + if (!btrfs_test_opt(tree_root->fs_info, SSD) && + !btrfs_test_opt(tree_root->fs_info, NOSSD) && !fs_info->fs_devices->rotating) { btrfs_info(fs_info, "detected SSD devices, enabling SSD mode"); btrfs_set_opt(fs_info->mount_opt, SSD); @@ -3033,9 +3033,9 @@ retry_root_backup: btrfs_apply_pending_changes(fs_info); #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY - if (btrfs_test_opt(tree_root, CHECK_INTEGRITY)) { + if (btrfs_test_opt(tree_root->fs_info, CHECK_INTEGRITY)) { ret = btrfsic_mount(tree_root, fs_devices, - btrfs_test_opt(tree_root, + btrfs_test_opt(tree_root->fs_info, CHECK_INTEGRITY_INCLUDING_EXTENT_DATA) ? 1 : 0, fs_info->check_integrity_print_mask); @@ -3051,7 +3051,7 @@ retry_root_backup: /* do not make disk changes in broken FS or nologreplay is given */ if (btrfs_super_log_root(disk_super) != 0 && - !btrfs_test_opt(tree_root, NOLOGREPLAY)) { + !btrfs_test_opt(tree_root->fs_info, NOLOGREPLAY)) { ret = btrfs_replay_log(fs_info, fs_devices); if (ret) { err = ret; @@ -3092,7 +3092,7 @@ retry_root_backup: if (sb->s_flags & MS_RDONLY) return 0; - if (btrfs_test_opt(tree_root, FREE_SPACE_TREE) && + if (btrfs_test_opt(tree_root->fs_info, FREE_SPACE_TREE) && !btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) { btrfs_info(fs_info, "creating free space tree"); ret = btrfs_create_free_space_tree(fs_info); @@ -3129,7 +3129,7 @@ retry_root_backup: btrfs_qgroup_rescan_resume(fs_info); - if (btrfs_test_opt(tree_root, CLEAR_CACHE) && + if (btrfs_test_opt(tree_root->fs_info, CLEAR_CACHE) && btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) { btrfs_info(fs_info, "clearing free space tree"); ret = btrfs_clear_free_space_tree(fs_info); @@ -3150,7 +3150,7 @@ retry_root_backup: close_ctree(tree_root); return ret; } - } else if (btrfs_test_opt(tree_root, RESCAN_UUID_TREE) || + } else if (btrfs_test_opt(tree_root->fs_info, RESCAN_UUID_TREE) || fs_info->generation != btrfs_super_uuid_tree_generation(disk_super)) { btrfs_info(fs_info, "checking UUID tree"); @@ -3227,7 +3227,7 @@ fail: return err; recovery_tree_root: - if (!btrfs_test_opt(tree_root, USEBACKUPROOT)) + if (!btrfs_test_opt(tree_root->fs_info, USEBACKUPROOT)) goto fail_tree_roots; free_root_pointers(fs_info, 0); @@ -3642,7 +3642,7 @@ static int write_all_supers(struct btrfs_root *root, int max_mirrors) int total_errors = 0; u64 flags; - do_barriers = !btrfs_test_opt(root, NOBARRIER); + do_barriers = !btrfs_test_opt(root->fs_info, NOBARRIER); backup_super_roots(root->fs_info); sb = root->fs_info->super_for_commit; @@ -3926,7 +3926,7 @@ void close_ctree(struct btrfs_root *root) iput(fs_info->btree_inode); #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY - if (btrfs_test_opt(root, CHECK_INTEGRITY)) + if (btrfs_test_opt(root->fs_info, CHECK_INTEGRITY)) btrfsic_unmount(root, fs_info->fs_devices); #endif diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 73f6dd2cf3cb..44b529f183e3 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3448,7 +3448,7 @@ again: spin_lock(&block_group->lock); if (block_group->cached != BTRFS_CACHE_FINISHED || - !btrfs_test_opt(root, SPACE_CACHE)) { + !btrfs_test_opt(root->fs_info, SPACE_CACHE)) { /* * don't bother trying to write stuff out _if_ * a) we're not cached, @@ -3525,7 +3525,7 @@ int btrfs_setup_space_cache(struct btrfs_trans_handle *trans, struct btrfs_path *path; if (list_empty(&cur_trans->dirty_bgs) || - !btrfs_test_opt(root, SPACE_CACHE)) + !btrfs_test_opt(root->fs_info, SPACE_CACHE)) return 0; path = btrfs_alloc_path(); @@ -4444,7 +4444,7 @@ void check_system_chunk(struct btrfs_trans_handle *trans, thresh = btrfs_calc_trunc_metadata_size(root, num_devs) + btrfs_calc_trans_metadata_size(root, 1); - if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) { + if (left < thresh && btrfs_test_opt(root->fs_info, ENOSPC_DEBUG)) { btrfs_info(root->fs_info, "left=%llu, need=%llu, flags=%llu", left, thresh, type); dump_space_info(info, 0, 0); @@ -6216,7 +6216,7 @@ static int update_block_group(struct btrfs_trans_handle *trans, spin_lock(&cache->space_info->lock); spin_lock(&cache->lock); - if (btrfs_test_opt(root, SPACE_CACHE) && + if (btrfs_test_opt(root->fs_info, SPACE_CACHE) && cache->disk_cache_state < BTRFS_DC_CLEAR) cache->disk_cache_state = BTRFS_DC_CLEAR; @@ -6598,7 +6598,7 @@ fetch_cluster_info(struct btrfs_root *root, struct btrfs_space_info *space_info, u64 *empty_cluster) { struct btrfs_free_cluster *ret = NULL; - bool ssd = btrfs_test_opt(root, SSD); + bool ssd = btrfs_test_opt(root->fs_info, SSD); *empty_cluster = 0; if (btrfs_mixed_space_info(space_info)) @@ -6743,7 +6743,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, break; } - if (btrfs_test_opt(root, DISCARD)) + if (btrfs_test_opt(root->fs_info, DISCARD)) ret = btrfs_discard_extent(root, start, end + 1 - start, NULL); @@ -7962,7 +7962,7 @@ again: if (num_bytes == min_alloc_size) final_tried = true; goto again; - } else if (btrfs_test_opt(root, ENOSPC_DEBUG)) { + } else if (btrfs_test_opt(root->fs_info, ENOSPC_DEBUG)) { struct btrfs_space_info *sinfo; sinfo = __find_space_info(root->fs_info, flags); @@ -7993,7 +7993,7 @@ static int __btrfs_free_reserved_extent(struct btrfs_root *root, if (pin) pin_down_extent(root, cache, start, len, 1); else { - if (btrfs_test_opt(root, DISCARD)) + if (btrfs_test_opt(root->fs_info, DISCARD)) ret = btrfs_discard_extent(root, start, len, NULL); btrfs_add_free_space(cache, start, len); btrfs_update_reserved_bytes(cache, len, RESERVE_FREE, delalloc); @@ -8301,7 +8301,7 @@ again: goto again; } - if (btrfs_test_opt(root, ENOSPC_DEBUG)) { + if (btrfs_test_opt(root->fs_info, ENOSPC_DEBUG)) { static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL * 10, /*DEFAULT_RATELIMIT_BURST*/ 1); @@ -9735,7 +9735,7 @@ int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) int full = 0; int ret = 0; - debug = btrfs_test_opt(root, ENOSPC_DEBUG); + debug = btrfs_test_opt(root->fs_info, ENOSPC_DEBUG); block_group = btrfs_lookup_block_group(root->fs_info, bytenr); @@ -10148,10 +10148,10 @@ int btrfs_read_block_groups(struct btrfs_root *root) path->reada = READA_FORWARD; cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); - if (btrfs_test_opt(root, SPACE_CACHE) && + if (btrfs_test_opt(root->fs_info, SPACE_CACHE) && btrfs_super_generation(root->fs_info->super_copy) != cache_gen) need_clear = 1; - if (btrfs_test_opt(root, CLEAR_CACHE)) + if (btrfs_test_opt(root->fs_info, CLEAR_CACHE)) need_clear = 1; while (1) { @@ -10182,7 +10182,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) * b) Setting 'dirty flag' makes sure that we flush * the new space cache info onto disk. */ - if (btrfs_test_opt(root, SPACE_CACHE)) + if (btrfs_test_opt(root->fs_info, SPACE_CACHE)) cache->disk_cache_state = BTRFS_DC_CLEAR; } @@ -10641,7 +10641,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, spin_lock(&block_group->space_info->lock); list_del_init(&block_group->ro_list); - if (btrfs_test_opt(root, ENOSPC_DEBUG)) { + if (btrfs_test_opt(root->fs_info, ENOSPC_DEBUG)) { WARN_ON(block_group->space_info->total_bytes < block_group->key.offset); WARN_ON(block_group->space_info->bytes_readonly @@ -10909,7 +10909,7 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info) spin_unlock(&space_info->lock); /* DISCARD can flip during remount */ - trimming = btrfs_test_opt(root, DISCARD); + trimming = btrfs_test_opt(root->fs_info, DISCARD); /* Implicit trim during transaction commit. */ if (trimming) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 38bea7bb0987..48b94904a519 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -132,7 +132,7 @@ static int __btrfs_add_inode_defrag(struct inode *inode, static inline int __need_auto_defrag(struct btrfs_root *root) { - if (!btrfs_test_opt(root, AUTO_DEFRAG)) + if (!btrfs_test_opt(root->fs_info, AUTO_DEFRAG)) return 0; if (btrfs_fs_closing(root->fs_info)) diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 69d270f6602c..cee2049ef64e 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -3026,7 +3026,7 @@ int btrfs_find_space_cluster(struct btrfs_root *root, * For metadata, allow allocates with smaller extents. For * data, keep it dense. */ - if (btrfs_test_opt(root, SSD_SPREAD)) { + if (btrfs_test_opt(root->fs_info, SSD_SPREAD)) { cont1_bytes = min_bytes = bytes + empty_size; } else if (block_group->flags & BTRFS_BLOCK_GROUP_METADATA) { cont1_bytes = bytes; @@ -3470,7 +3470,7 @@ int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root) int ret = 0; u64 root_gen = btrfs_root_generation(&root->root_item); - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return 0; /* @@ -3514,7 +3514,7 @@ int btrfs_write_out_ino_cache(struct btrfs_root *root, struct btrfs_io_ctl io_ctl; bool release_metadata = true; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return 0; memset(&io_ctl, 0, sizeof(io_ctl)); diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index 70107f7c9307..e3ad8c197997 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -38,7 +38,7 @@ static int caching_kthread(void *data) int slot; int ret; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return 0; path = btrfs_alloc_path(); @@ -141,7 +141,7 @@ static void start_caching(struct btrfs_root *root) int ret; u64 objectid; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return; spin_lock(&root->ino_cache_lock); @@ -185,7 +185,7 @@ static void start_caching(struct btrfs_root *root) int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid) { - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return btrfs_find_free_objectid(root, objectid); again: @@ -211,7 +211,7 @@ void btrfs_return_ino(struct btrfs_root *root, u64 objectid) { struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return; again: if (root->ino_cache_state == BTRFS_CACHE_FINISHED) { @@ -251,7 +251,7 @@ void btrfs_unpin_free_ino(struct btrfs_root *root) struct rb_node *n; u64 count; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return; while (1) { @@ -412,7 +412,7 @@ int btrfs_save_ino_cache(struct btrfs_root *root, if (btrfs_root_refs(&root->root_item) == 0) return 0; - if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + if (!btrfs_test_opt(root->fs_info, INODE_MAP_CACHE)) return 0; path = btrfs_alloc_path(); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 0612231aaeba..b166ee0763b6 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -376,12 +376,12 @@ static inline int inode_need_compress(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; /* force compress */ - if (btrfs_test_opt(root, FORCE_COMPRESS)) + if (btrfs_test_opt(root->fs_info, FORCE_COMPRESS)) return 1; /* bad compression ratios */ if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) return 0; - if (btrfs_test_opt(root, COMPRESS) || + if (btrfs_test_opt(root->fs_info, COMPRESS) || BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS || BTRFS_I(inode)->force_compress) return 1; @@ -622,7 +622,7 @@ cont: nr_pages_ret = 0; /* flag the file so we don't compress in the future */ - if (!btrfs_test_opt(root, FORCE_COMPRESS) && + if (!btrfs_test_opt(root->fs_info, FORCE_COMPRESS) && !(BTRFS_I(inode)->force_compress)) { BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; } @@ -1159,7 +1159,7 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page, async_cow->start = start; if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS && - !btrfs_test_opt(root, FORCE_COMPRESS)) + !btrfs_test_opt(root->fs_info, FORCE_COMPRESS)) cur_end = end; else cur_end = min(end, start + SZ_512K - 1); @@ -6255,9 +6255,9 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, btrfs_inherit_iflags(inode, dir); if (S_ISREG(mode)) { - if (btrfs_test_opt(root, NODATASUM)) + if (btrfs_test_opt(root->fs_info, NODATASUM)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(root->fs_info, NODATACOW)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW | BTRFS_INODE_NODATASUM; } diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 05173563e4a6..ffb1628cd1a1 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2406,7 +2406,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, * rmdir(2). */ err = -EPERM; - if (!btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED)) + if (!btrfs_test_opt(root->fs_info, USER_SUBVOL_RM_ALLOWED)) goto out_dput; /* diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index d965abbcac8e..4a0c0aeb0041 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -454,12 +454,12 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, */ break; case Opt_nodatasum: - btrfs_set_and_info(root, NODATASUM, + btrfs_set_and_info(info, NODATASUM, "setting nodatasum"); break; case Opt_datasum: - if (btrfs_test_opt(root, NODATASUM)) { - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(info, NODATASUM)) { + if (btrfs_test_opt(info, NODATACOW)) btrfs_info(root->fs_info, "setting datasum, datacow enabled"); else btrfs_info(root->fs_info, "setting datasum"); @@ -468,9 +468,9 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, btrfs_clear_opt(info->mount_opt, NODATASUM); break; case Opt_nodatacow: - if (!btrfs_test_opt(root, NODATACOW)) { - if (!btrfs_test_opt(root, COMPRESS) || - !btrfs_test_opt(root, FORCE_COMPRESS)) { + if (!btrfs_test_opt(info, NODATACOW)) { + if (!btrfs_test_opt(info, COMPRESS) || + !btrfs_test_opt(info, FORCE_COMPRESS)) { btrfs_info(root->fs_info, "setting nodatacow, compression disabled"); } else { @@ -483,7 +483,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, btrfs_set_opt(info->mount_opt, NODATASUM); break; case Opt_datacow: - btrfs_clear_and_info(root, NODATACOW, + btrfs_clear_and_info(info, NODATACOW, "setting datacow"); break; case Opt_compress_force: @@ -492,10 +492,11 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, /* Fallthrough */ case Opt_compress: case Opt_compress_type: - saved_compress_type = btrfs_test_opt(root, COMPRESS) ? + saved_compress_type = btrfs_test_opt(info, + COMPRESS) ? info->compress_type : BTRFS_COMPRESS_NONE; saved_compress_force = - btrfs_test_opt(root, FORCE_COMPRESS); + btrfs_test_opt(info, FORCE_COMPRESS); if (token == Opt_compress || token == Opt_compress_force || strcmp(args[0].from, "zlib") == 0) { @@ -535,10 +536,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, */ btrfs_clear_opt(info->mount_opt, FORCE_COMPRESS); } - if ((btrfs_test_opt(root, COMPRESS) && + if ((btrfs_test_opt(info, COMPRESS) && (info->compress_type != saved_compress_type || compress_force != saved_compress_force)) || - (!btrfs_test_opt(root, COMPRESS) && + (!btrfs_test_opt(info, COMPRESS) && no_compress == 1)) { btrfs_info(root->fs_info, "%s %s compression", @@ -548,25 +549,25 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, compress_force = false; break; case Opt_ssd: - btrfs_set_and_info(root, SSD, + btrfs_set_and_info(info, SSD, "use ssd allocation scheme"); break; case Opt_ssd_spread: - btrfs_set_and_info(root, SSD_SPREAD, + btrfs_set_and_info(info, SSD_SPREAD, "use spread ssd allocation scheme"); btrfs_set_opt(info->mount_opt, SSD); break; case Opt_nossd: - btrfs_set_and_info(root, NOSSD, + btrfs_set_and_info(info, NOSSD, "not using ssd allocation scheme"); btrfs_clear_opt(info->mount_opt, SSD); break; case Opt_barrier: - btrfs_clear_and_info(root, NOBARRIER, + btrfs_clear_and_info(info, NOBARRIER, "turning on barriers"); break; case Opt_nobarrier: - btrfs_set_and_info(root, NOBARRIER, + btrfs_set_and_info(info, NOBARRIER, "turning off barriers"); break; case Opt_thread_pool: @@ -626,24 +627,24 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, root->fs_info->sb->s_flags &= ~MS_POSIXACL; break; case Opt_notreelog: - btrfs_set_and_info(root, NOTREELOG, + btrfs_set_and_info(info, NOTREELOG, "disabling tree log"); break; case Opt_treelog: - btrfs_clear_and_info(root, NOTREELOG, + btrfs_clear_and_info(info, NOTREELOG, "enabling tree log"); break; case Opt_norecovery: case Opt_nologreplay: - btrfs_set_and_info(root, NOLOGREPLAY, + btrfs_set_and_info(info, NOLOGREPLAY, "disabling log replay at mount time"); break; case Opt_flushoncommit: - btrfs_set_and_info(root, FLUSHONCOMMIT, + btrfs_set_and_info(info, FLUSHONCOMMIT, "turning on flush-on-commit"); break; case Opt_noflushoncommit: - btrfs_clear_and_info(root, FLUSHONCOMMIT, + btrfs_clear_and_info(info, FLUSHONCOMMIT, "turning off flush-on-commit"); break; case Opt_ratio: @@ -660,11 +661,11 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, } break; case Opt_discard: - btrfs_set_and_info(root, DISCARD, + btrfs_set_and_info(info, DISCARD, "turning on discard"); break; case Opt_nodiscard: - btrfs_clear_and_info(root, DISCARD, + btrfs_clear_and_info(info, DISCARD, "turning off discard"); break; case Opt_space_cache: @@ -673,12 +674,13 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, strcmp(args[0].from, "v1") == 0) { btrfs_clear_opt(root->fs_info->mount_opt, FREE_SPACE_TREE); - btrfs_set_and_info(root, SPACE_CACHE, + btrfs_set_and_info(info, SPACE_CACHE, "enabling disk space caching"); } else if (strcmp(args[0].from, "v2") == 0) { btrfs_clear_opt(root->fs_info->mount_opt, SPACE_CACHE); - btrfs_set_and_info(root, FREE_SPACE_TREE, + btrfs_set_and_info(info, + FREE_SPACE_TREE, "enabling free space tree"); } else { ret = -EINVAL; @@ -689,12 +691,14 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, btrfs_set_opt(info->mount_opt, RESCAN_UUID_TREE); break; case Opt_no_space_cache: - if (btrfs_test_opt(root, SPACE_CACHE)) { - btrfs_clear_and_info(root, SPACE_CACHE, + if (btrfs_test_opt(info, SPACE_CACHE)) { + btrfs_clear_and_info(info, + SPACE_CACHE, "disabling disk space caching"); } - if (btrfs_test_opt(root, FREE_SPACE_TREE)) { - btrfs_clear_and_info(root, FREE_SPACE_TREE, + if (btrfs_test_opt(info, FREE_SPACE_TREE)) { + btrfs_clear_and_info(info, + FREE_SPACE_TREE, "disabling free space tree"); } break; @@ -707,7 +711,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, "disabling inode map caching"); break; case Opt_clear_cache: - btrfs_set_and_info(root, CLEAR_CACHE, + btrfs_set_and_info(info, CLEAR_CACHE, "force clearing of disk cache"); break; case Opt_user_subvol_rm_allowed: @@ -720,11 +724,11 @@ int btrfs_parse_options(struct btrfs_root *root, char *options, btrfs_clear_opt(info->mount_opt, ENOSPC_DEBUG); break; case Opt_defrag: - btrfs_set_and_info(root, AUTO_DEFRAG, + btrfs_set_and_info(info, AUTO_DEFRAG, "enabling auto defrag"); break; case Opt_nodefrag: - btrfs_clear_and_info(root, AUTO_DEFRAG, + btrfs_clear_and_info(info, AUTO_DEFRAG, "disabling auto defrag"); break; case Opt_recovery: @@ -832,22 +836,22 @@ check: /* * Extra check for current option against current flag */ - if (btrfs_test_opt(root, NOLOGREPLAY) && !(new_flags & MS_RDONLY)) { + if (btrfs_test_opt(info, NOLOGREPLAY) && !(new_flags & MS_RDONLY)) { btrfs_err(root->fs_info, "nologreplay must be used with ro mount option"); ret = -EINVAL; } out: if (btrfs_fs_compat_ro(root->fs_info, FREE_SPACE_TREE) && - !btrfs_test_opt(root, FREE_SPACE_TREE) && - !btrfs_test_opt(root, CLEAR_CACHE)) { + !btrfs_test_opt(info, FREE_SPACE_TREE) && + !btrfs_test_opt(info, CLEAR_CACHE)) { btrfs_err(root->fs_info, "cannot disable free space tree"); ret = -EINVAL; } - if (!ret && btrfs_test_opt(root, SPACE_CACHE)) + if (!ret && btrfs_test_opt(info, SPACE_CACHE)) btrfs_info(root->fs_info, "disk space caching is enabled"); - if (!ret && btrfs_test_opt(root, FREE_SPACE_TREE)) + if (!ret && btrfs_test_opt(info, FREE_SPACE_TREE)) btrfs_info(root->fs_info, "using free space tree"); kfree(orig); return ret; @@ -1214,13 +1218,13 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry) struct btrfs_root *root = info->tree_root; char *compress_type; - if (btrfs_test_opt(root, DEGRADED)) + if (btrfs_test_opt(info, DEGRADED)) seq_puts(seq, ",degraded"); - if (btrfs_test_opt(root, NODATASUM)) + if (btrfs_test_opt(info, NODATASUM)) seq_puts(seq, ",nodatasum"); - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(info, NODATACOW)) seq_puts(seq, ",nodatacow"); - if (btrfs_test_opt(root, NOBARRIER)) + if (btrfs_test_opt(info, NOBARRIER)) seq_puts(seq, ",nobarrier"); if (info->max_inline != BTRFS_DEFAULT_MAX_INLINE) seq_printf(seq, ",max_inline=%llu", info->max_inline); @@ -1229,56 +1233,56 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry) if (info->thread_pool_size != min_t(unsigned long, num_online_cpus() + 2, 8)) seq_printf(seq, ",thread_pool=%d", info->thread_pool_size); - if (btrfs_test_opt(root, COMPRESS)) { + if (btrfs_test_opt(info, COMPRESS)) { if (info->compress_type == BTRFS_COMPRESS_ZLIB) compress_type = "zlib"; else compress_type = "lzo"; - if (btrfs_test_opt(root, FORCE_COMPRESS)) + if (btrfs_test_opt(info, FORCE_COMPRESS)) seq_printf(seq, ",compress-force=%s", compress_type); else seq_printf(seq, ",compress=%s", compress_type); } - if (btrfs_test_opt(root, NOSSD)) + if (btrfs_test_opt(info, NOSSD)) seq_puts(seq, ",nossd"); - if (btrfs_test_opt(root, SSD_SPREAD)) + if (btrfs_test_opt(info, SSD_SPREAD)) seq_puts(seq, ",ssd_spread"); - else if (btrfs_test_opt(root, SSD)) + else if (btrfs_test_opt(info, SSD)) seq_puts(seq, ",ssd"); - if (btrfs_test_opt(root, NOTREELOG)) + if (btrfs_test_opt(info, NOTREELOG)) seq_puts(seq, ",notreelog"); - if (btrfs_test_opt(root, NOLOGREPLAY)) + if (btrfs_test_opt(info, NOLOGREPLAY)) seq_puts(seq, ",nologreplay"); - if (btrfs_test_opt(root, FLUSHONCOMMIT)) + if (btrfs_test_opt(info, FLUSHONCOMMIT)) seq_puts(seq, ",flushoncommit"); - if (btrfs_test_opt(root, DISCARD)) + if (btrfs_test_opt(info, DISCARD)) seq_puts(seq, ",discard"); if (!(root->fs_info->sb->s_flags & MS_POSIXACL)) seq_puts(seq, ",noacl"); - if (btrfs_test_opt(root, SPACE_CACHE)) + if (btrfs_test_opt(info, SPACE_CACHE)) seq_puts(seq, ",space_cache"); - else if (btrfs_test_opt(root, FREE_SPACE_TREE)) + else if (btrfs_test_opt(info, FREE_SPACE_TREE)) seq_puts(seq, ",space_cache=v2"); else seq_puts(seq, ",nospace_cache"); - if (btrfs_test_opt(root, RESCAN_UUID_TREE)) + if (btrfs_test_opt(info, RESCAN_UUID_TREE)) seq_puts(seq, ",rescan_uuid_tree"); - if (btrfs_test_opt(root, CLEAR_CACHE)) + if (btrfs_test_opt(info, CLEAR_CACHE)) seq_puts(seq, ",clear_cache"); - if (btrfs_test_opt(root, USER_SUBVOL_RM_ALLOWED)) + if (btrfs_test_opt(info, USER_SUBVOL_RM_ALLOWED)) seq_puts(seq, ",user_subvol_rm_allowed"); - if (btrfs_test_opt(root, ENOSPC_DEBUG)) + if (btrfs_test_opt(info, ENOSPC_DEBUG)) seq_puts(seq, ",enospc_debug"); - if (btrfs_test_opt(root, AUTO_DEFRAG)) + if (btrfs_test_opt(info, AUTO_DEFRAG)) seq_puts(seq, ",autodefrag"); - if (btrfs_test_opt(root, INODE_MAP_CACHE)) + if (btrfs_test_opt(info, INODE_MAP_CACHE)) seq_puts(seq, ",inode_cache"); - if (btrfs_test_opt(root, SKIP_BALANCE)) + if (btrfs_test_opt(info, SKIP_BALANCE)) seq_puts(seq, ",skip_balance"); #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY - if (btrfs_test_opt(root, CHECK_INTEGRITY_INCLUDING_EXTENT_DATA)) + if (btrfs_test_opt(info, CHECK_INTEGRITY_INCLUDING_EXTENT_DATA)) seq_puts(seq, ",check_int_data"); - else if (btrfs_test_opt(root, CHECK_INTEGRITY)) + else if (btrfs_test_opt(info, CHECK_INTEGRITY)) seq_puts(seq, ",check_int"); if (info->check_integrity_print_mask) seq_printf(seq, ",check_int_print_mask=%d", @@ -1287,14 +1291,14 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry) if (info->metadata_ratio) seq_printf(seq, ",metadata_ratio=%d", info->metadata_ratio); - if (btrfs_test_opt(root, PANIC_ON_FATAL_ERROR)) + if (btrfs_test_opt(info, PANIC_ON_FATAL_ERROR)) seq_puts(seq, ",fatal_errors=panic"); if (info->commit_interval != BTRFS_DEFAULT_COMMIT_INTERVAL) seq_printf(seq, ",commit=%d", info->commit_interval); #ifdef CONFIG_BTRFS_DEBUG - if (btrfs_test_opt(root, FRAGMENT_DATA)) + if (btrfs_test_opt(info, FRAGMENT_DATA)) seq_puts(seq, ",fragment=data"); - if (btrfs_test_opt(root, FRAGMENT_METADATA)) + if (btrfs_test_opt(info, FRAGMENT_METADATA)) seq_puts(seq, ",fragment=metadata"); #endif seq_printf(seq, ",subvolid=%llu", diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 948aa186b353..d986447d19a4 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1709,7 +1709,7 @@ static void update_super_roots(struct btrfs_root *root) super->root = root_item->bytenr; super->generation = root_item->generation; super->root_level = root_item->level; - if (btrfs_test_opt(root, SPACE_CACHE)) + if (btrfs_test_opt(root->fs_info, SPACE_CACHE)) super->cache_generation = root_item->generation; if (root->fs_info->update_uuid_tree_gen) super->uuid_tree_generation = root_item->generation; @@ -1895,14 +1895,14 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans, static inline int btrfs_start_delalloc_flush(struct btrfs_fs_info *fs_info) { - if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT)) + if (btrfs_test_opt(fs_info, FLUSHONCOMMIT)) return btrfs_start_delalloc_roots(fs_info, 1, -1); return 0; } static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info) { - if (btrfs_test_opt(fs_info->tree_root, FLUSHONCOMMIT)) + if (btrfs_test_opt(fs_info, FLUSHONCOMMIT)) btrfs_wait_ordered_roots(fs_info, -1, 0, (u64)-1); } diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index aa4475e3046b..391bd3ddaed3 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -2757,7 +2757,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, while (1) { int batch = atomic_read(&root->log_batch); /* when we're on an ssd, just kick the log commit out */ - if (!btrfs_test_opt(root, SSD) && + if (!btrfs_test_opt(root->fs_info, SSD) && test_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state)) { mutex_unlock(&root->log_mutex); schedule_timeout_uninterruptible(1); @@ -5305,7 +5305,7 @@ static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, sb = inode->i_sb; - if (btrfs_test_opt(root, NOTREELOG)) { + if (btrfs_test_opt(root->fs_info, NOTREELOG)) { ret = 1; goto end_no_trans; } diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 05aec96e4c96..5e409e28a5c0 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3944,7 +3944,7 @@ int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info) } spin_unlock(&fs_info->balance_lock); - if (btrfs_test_opt(fs_info->tree_root, SKIP_BALANCE)) { + if (btrfs_test_opt(fs_info, SKIP_BALANCE)) { btrfs_info(fs_info, "force skipping balance"); return 0; } @@ -6455,7 +6455,8 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, BTRFS_UUID_SIZE); map->stripes[i].dev = btrfs_find_device(root->fs_info, devid, uuid, NULL); - if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) { + if (!map->stripes[i].dev && + !btrfs_test_opt(root->fs_info, DEGRADED)) { free_extent_map(em); return -EIO; } @@ -6523,7 +6524,7 @@ static struct btrfs_fs_devices *open_seed_devices(struct btrfs_root *root, fs_devices = find_fsid(fsid); if (!fs_devices) { - if (!btrfs_test_opt(root, DEGRADED)) + if (!btrfs_test_opt(root->fs_info, DEGRADED)) return ERR_PTR(-ENOENT); fs_devices = alloc_fs_devices(fsid); @@ -6585,7 +6586,7 @@ static int read_one_dev(struct btrfs_root *root, device = btrfs_find_device(root->fs_info, devid, dev_uuid, fs_uuid); if (!device) { - if (!btrfs_test_opt(root, DEGRADED)) + if (!btrfs_test_opt(root->fs_info, DEGRADED)) return -EIO; device = add_missing_dev(root, fs_devices, devid, dev_uuid); @@ -6594,7 +6595,7 @@ static int read_one_dev(struct btrfs_root *root, btrfs_warn(root->fs_info, "devid %llu uuid %pU missing", devid, dev_uuid); } else { - if (!device->bdev && !btrfs_test_opt(root, DEGRADED)) + if (!device->bdev && !btrfs_test_opt(root->fs_info, DEGRADED)) return -EIO; if(!device->bdev && !device->missing) { -- cgit v1.2.3 From 8632daae40c1a5b8fcc1d66d788c143353fd70b7 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Mon, 20 Jun 2016 13:16:40 -0400 Subject: btrfs: tests, move initialization into tests/ We have all these stubs that only exist because they're called from btrfs_run_sanity_tests, which is a static inside super.c. Let's just move it all into tests/btrfs-tests.c and only have one stub. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/super.c | 43 ---------------------------------------- fs/btrfs/tests/btrfs-tests.c | 47 ++++++++++++++++++++++++++++++++++++++++++-- fs/btrfs/tests/btrfs-tests.h | 35 +++------------------------------ 3 files changed, 48 insertions(+), 77 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 4a0c0aeb0041..959823a12c42 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2342,49 +2342,6 @@ static void btrfs_print_mod_info(void) btrfs_crc32c_impl()); } -static int btrfs_run_sanity_tests(void) -{ - int ret, i; - u32 sectorsize, nodesize; - u32 test_sectorsize[] = { - PAGE_SIZE, - }; - ret = btrfs_init_test_fs(); - if (ret) - return ret; - for (i = 0; i < ARRAY_SIZE(test_sectorsize); i++) { - sectorsize = test_sectorsize[i]; - for (nodesize = sectorsize; - nodesize <= BTRFS_MAX_METADATA_BLOCKSIZE; - nodesize <<= 1) { - pr_info("BTRFS: selftest: sectorsize: %u nodesize: %u\n", - sectorsize, nodesize); - ret = btrfs_test_free_space_cache(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_extent_buffer_operations(sectorsize, - nodesize); - if (ret) - goto out; - ret = btrfs_test_extent_io(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_inodes(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_qgroups(sectorsize, nodesize); - if (ret) - goto out; - ret = btrfs_test_free_space_tree(sectorsize, nodesize); - if (ret) - goto out; - } - } -out: - btrfs_destroy_test_fs(); - return ret; -} - static int __init init_btrfs_fs(void) { int err; diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c index 02223f3f78f4..71d9a0129347 100644 --- a/fs/btrfs/tests/btrfs-tests.c +++ b/fs/btrfs/tests/btrfs-tests.c @@ -54,7 +54,7 @@ struct inode *btrfs_new_test_inode(void) return new_inode(test_mnt->mnt_sb); } -int btrfs_init_test_fs(void) +static int btrfs_init_test_fs(void) { int ret; @@ -73,7 +73,7 @@ int btrfs_init_test_fs(void) return 0; } -void btrfs_destroy_test_fs(void) +static void btrfs_destroy_test_fs(void) { kern_unmount(test_mnt); unregister_filesystem(&test_type); @@ -220,3 +220,46 @@ void btrfs_init_dummy_trans(struct btrfs_trans_handle *trans) INIT_LIST_HEAD(&trans->qgroup_ref_list); trans->type = __TRANS_DUMMY; } + +int btrfs_run_sanity_tests(void) +{ + int ret, i; + u32 sectorsize, nodesize; + u32 test_sectorsize[] = { + PAGE_SIZE, + }; + ret = btrfs_init_test_fs(); + if (ret) + return ret; + for (i = 0; i < ARRAY_SIZE(test_sectorsize); i++) { + sectorsize = test_sectorsize[i]; + for (nodesize = sectorsize; + nodesize <= BTRFS_MAX_METADATA_BLOCKSIZE; + nodesize <<= 1) { + pr_info("BTRFS: selftest: sectorsize: %u nodesize: %u\n", + sectorsize, nodesize); + ret = btrfs_test_free_space_cache(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_extent_buffer_operations(sectorsize, + nodesize); + if (ret) + goto out; + ret = btrfs_test_extent_io(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_inodes(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_qgroups(sectorsize, nodesize); + if (ret) + goto out; + ret = btrfs_test_free_space_tree(sectorsize, nodesize); + if (ret) + goto out; + } + } +out: + btrfs_destroy_test_fs(); + return ret; +} diff --git a/fs/btrfs/tests/btrfs-tests.h b/fs/btrfs/tests/btrfs-tests.h index 66fb6b701eb7..e7d364f315db 100644 --- a/fs/btrfs/tests/btrfs-tests.h +++ b/fs/btrfs/tests/btrfs-tests.h @@ -20,20 +20,19 @@ #define __BTRFS_TESTS #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS +int btrfs_run_sanity_tests(void); #define test_msg(fmt, ...) pr_info("BTRFS: selftest: " fmt, ##__VA_ARGS__) struct btrfs_root; struct btrfs_trans_handle; -int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize); int btrfs_test_extent_buffer_operations(u32 sectorsize, u32 nodesize); +int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize); int btrfs_test_extent_io(u32 sectorsize, u32 nodesize); int btrfs_test_inodes(u32 sectorsize, u32 nodesize); int btrfs_test_qgroups(u32 sectorsize, u32 nodesize); int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize); -int btrfs_init_test_fs(void); -void btrfs_destroy_test_fs(void); struct inode *btrfs_new_test_inode(void); struct btrfs_fs_info *btrfs_alloc_dummy_fs_info(void); void btrfs_free_dummy_root(struct btrfs_root *root); @@ -42,35 +41,7 @@ btrfs_alloc_dummy_block_group(unsigned long length, u32 sectorsize); void btrfs_free_dummy_block_group(struct btrfs_block_group_cache *cache); void btrfs_init_dummy_trans(struct btrfs_trans_handle *trans); #else -static inline int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) -{ - return 0; -} -static inline int btrfs_test_extent_buffer_operations(u32 sectorsize, - u32 nodesize) -{ - return 0; -} -static inline int btrfs_init_test_fs(void) -{ - return 0; -} -static inline void btrfs_destroy_test_fs(void) -{ -} -static inline int btrfs_test_extent_io(u32 sectorsize, u32 nodesize) -{ - return 0; -} -static inline int btrfs_test_inodes(u32 sectorsize, u32 nodesize) -{ - return 0; -} -static inline int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) -{ - return 0; -} -static inline int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize) +static inline int btrfs_run_sanity_tests(void) { return 0; } -- cgit v1.2.3 From 7c0260ee098db7a05fd68812b2e21ce2e19dfcf0 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Mon, 20 Jun 2016 14:14:09 -0400 Subject: btrfs: tests, require fs_info for root This allows the upcoming patchset to push nodesize and sectorsize into fs_info. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 1 + fs/btrfs/disk-io.c | 15 +++++++---- fs/btrfs/disk-io.h | 3 ++- fs/btrfs/tests/btrfs-tests.c | 20 ++++++++++++--- fs/btrfs/tests/btrfs-tests.h | 1 + fs/btrfs/tests/extent-buffer-tests.c | 23 +++++++++++------ fs/btrfs/tests/free-space-tests.c | 14 +++++++---- fs/btrfs/tests/free-space-tree-tests.c | 18 +++++++------ fs/btrfs/tests/inode-tests.c | 46 ++++++++++++++++++---------------- fs/btrfs/tests/qgroup-tests.c | 23 +++++++++-------- 10 files changed, 103 insertions(+), 61 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 61a53cca5963..1cb84e01f2fd 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -117,6 +117,7 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes) #define BTRFS_FS_STATE_REMOUNTING 1 #define BTRFS_FS_STATE_TRANS_ABORTED 2 #define BTRFS_FS_STATE_DEV_REPLACING 3 +#define BTRFS_FS_STATE_DUMMY_FS_INFO 4 #define BTRFS_BACKREF_REV_MAX 256 #define BTRFS_BACKREF_REV_SHIFT 56 diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 802a3113854f..8e838b1cc806 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1233,6 +1233,7 @@ static void __setup_root(u32 nodesize, u32 sectorsize, u32 stripesize, struct btrfs_root *root, struct btrfs_fs_info *fs_info, u64 objectid) { + bool dummy = test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state); root->node = NULL; root->commit_root = NULL; root->sectorsize = sectorsize; @@ -1287,14 +1288,14 @@ static void __setup_root(u32 nodesize, u32 sectorsize, u32 stripesize, root->log_transid = 0; root->log_transid_committed = -1; root->last_log_commit = 0; - if (fs_info) + if (!dummy) extent_io_tree_init(&root->dirty_log_pages, fs_info->btree_inode->i_mapping); memset(&root->root_key, 0, sizeof(root->root_key)); memset(&root->root_item, 0, sizeof(root->root_item)); memset(&root->defrag_progress, 0, sizeof(root->defrag_progress)); - if (fs_info) + if (!dummy) root->defrag_trans_start = fs_info->generation; else root->defrag_trans_start = 0; @@ -1315,15 +1316,19 @@ static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info, #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS /* Should only be used by the testing infrastructure */ -struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize) +struct btrfs_root *btrfs_alloc_dummy_root(struct btrfs_fs_info *fs_info, + u32 sectorsize, u32 nodesize) { struct btrfs_root *root; - root = btrfs_alloc_root(NULL, GFP_KERNEL); + if (!fs_info) + return ERR_PTR(-EINVAL); + + root = btrfs_alloc_root(fs_info, GFP_KERNEL); if (!root) return ERR_PTR(-ENOMEM); /* We don't use the stripesize in selftest, set it as sectorsize */ - __setup_root(nodesize, sectorsize, sectorsize, root, NULL, + __setup_root(nodesize, sectorsize, sectorsize, root, fs_info, BTRFS_ROOT_TREE_OBJECTID); set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index acba821499a9..c9d42c92da2b 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -90,7 +90,8 @@ void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info, void btrfs_free_fs_root(struct btrfs_root *root); #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS -struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize); +struct btrfs_root *btrfs_alloc_dummy_root(struct btrfs_fs_info *fs_info, + u32 sectorsize, u32 nodesize); #endif /* diff --git a/fs/btrfs/tests/btrfs-tests.c b/fs/btrfs/tests/btrfs-tests.c index 71d9a0129347..bf62ad919a95 100644 --- a/fs/btrfs/tests/btrfs-tests.c +++ b/fs/btrfs/tests/btrfs-tests.c @@ -128,14 +128,27 @@ struct btrfs_fs_info *btrfs_alloc_dummy_fs_info(void) extent_io_tree_init(&fs_info->freed_extents[0], NULL); extent_io_tree_init(&fs_info->freed_extents[1], NULL); fs_info->pinned_extents = &fs_info->freed_extents[0]; + set_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, &fs_info->fs_state); + + test_mnt->mnt_sb->s_fs_info = fs_info; + return fs_info; } -static void btrfs_free_dummy_fs_info(struct btrfs_fs_info *fs_info) +void btrfs_free_dummy_fs_info(struct btrfs_fs_info *fs_info) { struct radix_tree_iter iter; void **slot; + if (!fs_info) + return; + + if (WARN_ON(!test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, + &fs_info->fs_state))) + return; + + test_mnt->mnt_sb->s_fs_info = NULL; + spin_lock(&fs_info->buffer_lock); radix_tree_for_each_slot(slot, &fs_info->buffer_radix, &iter, 0) { struct extent_buffer *eb; @@ -167,10 +180,11 @@ void btrfs_free_dummy_root(struct btrfs_root *root) { if (!root) return; + /* Will be freed by btrfs_free_fs_roots */ + if (WARN_ON(test_bit(BTRFS_ROOT_IN_RADIX, &root->state))) + return; if (root->node) free_extent_buffer(root->node); - if (root->fs_info) - btrfs_free_dummy_fs_info(root->fs_info); kfree(root); } diff --git a/fs/btrfs/tests/btrfs-tests.h b/fs/btrfs/tests/btrfs-tests.h index e7d364f315db..b17ffbe8f9f3 100644 --- a/fs/btrfs/tests/btrfs-tests.h +++ b/fs/btrfs/tests/btrfs-tests.h @@ -35,6 +35,7 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize); int btrfs_test_free_space_tree(u32 sectorsize, u32 nodesize); struct inode *btrfs_new_test_inode(void); struct btrfs_fs_info *btrfs_alloc_dummy_fs_info(void); +void btrfs_free_dummy_fs_info(struct btrfs_fs_info *fs_info); void btrfs_free_dummy_root(struct btrfs_root *root); struct btrfs_block_group_cache * btrfs_alloc_dummy_block_group(unsigned long length, u32 sectorsize); diff --git a/fs/btrfs/tests/extent-buffer-tests.c b/fs/btrfs/tests/extent-buffer-tests.c index 4f8cbd1ec5ee..199569174637 100644 --- a/fs/btrfs/tests/extent-buffer-tests.c +++ b/fs/btrfs/tests/extent-buffer-tests.c @@ -24,8 +24,9 @@ static int test_btrfs_split_item(u32 sectorsize, u32 nodesize) { - struct btrfs_path *path; - struct btrfs_root *root; + struct btrfs_fs_info *fs_info; + struct btrfs_path *path = NULL; + struct btrfs_root *root = NULL; struct extent_buffer *eb; struct btrfs_item *item; char *value = "mary had a little lamb"; @@ -40,17 +41,24 @@ static int test_btrfs_split_item(u32 sectorsize, u32 nodesize) test_msg("Running btrfs_split_item tests\n"); - root = btrfs_alloc_dummy_root(sectorsize, nodesize); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Could not allocate fs_info\n"); + return -ENOMEM; + } + + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); if (IS_ERR(root)) { test_msg("Could not allocate root\n"); - return PTR_ERR(root); + ret = PTR_ERR(root); + goto out; } path = btrfs_alloc_path(); if (!path) { test_msg("Could not allocate path\n"); - kfree(root); - return -ENOMEM; + ret = -ENOMEM; + goto out; } path->nodes[0] = eb = alloc_dummy_extent_buffer(NULL, nodesize, @@ -219,7 +227,8 @@ static int test_btrfs_split_item(u32 sectorsize, u32 nodesize) } out: btrfs_free_path(path); - kfree(root); + btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } diff --git a/fs/btrfs/tests/free-space-tests.c b/fs/btrfs/tests/free-space-tests.c index 3956bb2ff84c..3221c8dee272 100644 --- a/fs/btrfs/tests/free-space-tests.c +++ b/fs/btrfs/tests/free-space-tests.c @@ -837,6 +837,7 @@ test_steal_space_from_bitmap_to_extent(struct btrfs_block_group_cache *cache, int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info; struct btrfs_block_group_cache *cache; struct btrfs_root *root = NULL; int ret = -ENOMEM; @@ -855,15 +856,17 @@ int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) return 0; } - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - ret = PTR_ERR(root); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + ret = -ENOMEM; goto out; } - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + ret = PTR_ERR(root); goto out; + } root->fs_info->extent_root = root; cache->fs_info = root->fs_info; @@ -882,6 +885,7 @@ int btrfs_test_free_space_cache(u32 sectorsize, u32 nodesize) out: btrfs_free_dummy_block_group(cache); btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); test_msg("Free space cache tests finished\n"); return ret; } diff --git a/fs/btrfs/tests/free-space-tree-tests.c b/fs/btrfs/tests/free-space-tree-tests.c index aac507085ab0..7508d3b42780 100644 --- a/fs/btrfs/tests/free-space-tree-tests.c +++ b/fs/btrfs/tests/free-space-tree-tests.c @@ -443,23 +443,24 @@ typedef int (*test_func_t)(struct btrfs_trans_handle *, static int run_test(test_func_t test_func, int bitmaps, u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info; struct btrfs_root *root = NULL; struct btrfs_block_group_cache *cache = NULL; struct btrfs_trans_handle trans; struct btrfs_path *path = NULL; int ret; - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - test_msg("Couldn't allocate dummy root\n"); - ret = PTR_ERR(root); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Couldn't allocate dummy fs info\n"); + ret = -ENOMEM; goto out; } - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) { - test_msg("Couldn't allocate dummy fs info\n"); - ret = -ENOMEM; + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + test_msg("Couldn't allocate dummy root\n"); + ret = PTR_ERR(root); goto out; } @@ -534,6 +535,7 @@ out: btrfs_free_path(path); btrfs_free_dummy_block_group(cache); btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } diff --git a/fs/btrfs/tests/inode-tests.c b/fs/btrfs/tests/inode-tests.c index 29648c0a39f1..9f72aeda9220 100644 --- a/fs/btrfs/tests/inode-tests.c +++ b/fs/btrfs/tests/inode-tests.c @@ -230,6 +230,7 @@ static unsigned long vacancy_only = 0; static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info = NULL; struct inode *inode = NULL; struct btrfs_root *root = NULL; struct extent_map *em = NULL; @@ -248,19 +249,15 @@ static noinline int test_btrfs_get_extent(u32 sectorsize, u32 nodesize) BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID; BTRFS_I(inode)->location.offset = 0; - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - test_msg("Couldn't allocate root\n"); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Couldn't allocate dummy fs info\n"); goto out; } - /* - * We do this since btrfs_get_extent wants to assign em->bdev to - * root->fs_info->fs_devices->latest_bdev. - */ - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) { - test_msg("Couldn't allocate dummy fs info\n"); + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + test_msg("Couldn't allocate root\n"); goto out; } @@ -835,11 +832,13 @@ out: free_extent_map(em); iput(inode); btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } static int test_hole_first(u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info = NULL; struct inode *inode = NULL; struct btrfs_root *root = NULL; struct extent_map *em = NULL; @@ -855,15 +854,15 @@ static int test_hole_first(u32 sectorsize, u32 nodesize) BTRFS_I(inode)->location.objectid = BTRFS_FIRST_FREE_OBJECTID; BTRFS_I(inode)->location.offset = 0; - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - test_msg("Couldn't allocate root\n"); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Couldn't allocate dummy fs info\n"); goto out; } - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) { - test_msg("Couldn't allocate dummy fs info\n"); + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + test_msg("Couldn't allocate root\n"); goto out; } @@ -934,11 +933,13 @@ out: free_extent_map(em); iput(inode); btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } static int test_extent_accounting(u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info = NULL; struct inode *inode = NULL; struct btrfs_root *root = NULL; int ret = -ENOMEM; @@ -949,15 +950,15 @@ static int test_extent_accounting(u32 sectorsize, u32 nodesize) return ret; } - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - test_msg("Couldn't allocate root\n"); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Couldn't allocate dummy fs info\n"); goto out; } - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) { - test_msg("Couldn't allocate dummy fs info\n"); + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + test_msg("Couldn't allocate root\n"); goto out; } @@ -1132,6 +1133,7 @@ out: NULL, GFP_KERNEL); iput(inode); btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } diff --git a/fs/btrfs/tests/qgroup-tests.c b/fs/btrfs/tests/qgroup-tests.c index 57a12c0d680b..4407fef7c16c 100644 --- a/fs/btrfs/tests/qgroup-tests.c +++ b/fs/btrfs/tests/qgroup-tests.c @@ -453,22 +453,24 @@ static int test_multiple_refs(struct btrfs_root *root, int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) { + struct btrfs_fs_info *fs_info = NULL; struct btrfs_root *root; struct btrfs_root *tmp_root; int ret = 0; - root = btrfs_alloc_dummy_root(sectorsize, nodesize); - if (IS_ERR(root)) { - test_msg("Couldn't allocate root\n"); - return PTR_ERR(root); + fs_info = btrfs_alloc_dummy_fs_info(); + if (!fs_info) { + test_msg("Couldn't allocate dummy fs info\n"); + return -ENOMEM; } - root->fs_info = btrfs_alloc_dummy_fs_info(); - if (!root->fs_info) { - test_msg("Couldn't allocate dummy fs info\n"); - ret = -ENOMEM; + root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); + if (IS_ERR(root)) { + test_msg("Couldn't allocate root\n"); + ret = PTR_ERR(root); goto out; } + /* We are using this root as our extent root */ root->fs_info->extent_root = root; @@ -495,7 +497,7 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) btrfs_set_header_nritems(root->node, 0); root->alloc_bytenr += 2 * nodesize; - tmp_root = btrfs_alloc_dummy_root(sectorsize, nodesize); + tmp_root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); if (IS_ERR(tmp_root)) { test_msg("Couldn't allocate a fs root\n"); ret = PTR_ERR(tmp_root); @@ -510,7 +512,7 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) goto out; } - tmp_root = btrfs_alloc_dummy_root(sectorsize, nodesize); + tmp_root = btrfs_alloc_dummy_root(fs_info, sectorsize, nodesize); if (IS_ERR(tmp_root)) { test_msg("Couldn't allocate a fs root\n"); ret = PTR_ERR(tmp_root); @@ -531,5 +533,6 @@ int btrfs_test_qgroups(u32 sectorsize, u32 nodesize) ret = test_multiple_refs(root, sectorsize, nodesize); out: btrfs_free_dummy_root(root); + btrfs_free_dummy_fs_info(fs_info); return ret; } -- cgit v1.2.3 From f5ee5c9ac56cd328fcc915582f81226affebd81c Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 21 Jun 2016 09:52:41 -0400 Subject: btrfs: tests, use BTRFS_FS_STATE_DUMMY_FS_INFO instead of dummy root Now that we have a dummy fs_info associated with each test that uses a root, we don't need the DUMMY_ROOT bit anymore. This lets us make choices without needing an actual root like in e.g. btrfs_find_create_tree_block. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/backref.c | 2 +- fs/btrfs/ctree.c | 2 +- fs/btrfs/ctree.h | 17 ++++++++--------- fs/btrfs/disk-io.c | 3 +-- fs/btrfs/extent-tree.c | 8 ++++---- fs/btrfs/inode.c | 4 ++-- fs/btrfs/qgroup.c | 4 ++-- 7 files changed, 19 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 1ac8a3c7f68e..2b88439c2ee8 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -361,7 +361,7 @@ static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, goto out; } - if (btrfs_test_is_dummy_root(root)) { + if (btrfs_is_testing(fs_info)) { srcu_read_unlock(&fs_info->subvol_srcu, index); ret = -ENOENT; goto out; diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 0abed1a0caa7..e926b9fdb64a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1505,7 +1505,7 @@ static inline int should_cow_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf) { - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return 0; /* ensure we can see the force_cow */ diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 1cb84e01f2fd..41d5ca986c0c 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1115,12 +1115,11 @@ struct btrfs_subvolume_writers { #define BTRFS_ROOT_REF_COWS 1 #define BTRFS_ROOT_TRACK_DIRTY 2 #define BTRFS_ROOT_IN_RADIX 3 -#define BTRFS_ROOT_DUMMY_ROOT 4 -#define BTRFS_ROOT_ORPHAN_ITEM_INSERTED 5 -#define BTRFS_ROOT_DEFRAG_RUNNING 6 -#define BTRFS_ROOT_FORCE_COW 7 -#define BTRFS_ROOT_MULTI_LOG_TASKS 8 -#define BTRFS_ROOT_DIRTY 9 +#define BTRFS_ROOT_ORPHAN_ITEM_INSERTED 4 +#define BTRFS_ROOT_DEFRAG_RUNNING 5 +#define BTRFS_ROOT_FORCE_COW 6 +#define BTRFS_ROOT_MULTI_LOG_TASKS 7 +#define BTRFS_ROOT_DIRTY 8 /* * in ram representation of the tree. extent_root is used for all allocations @@ -3613,13 +3612,13 @@ static inline int btrfs_defrag_cancelled(struct btrfs_fs_info *fs_info) void btrfs_test_destroy_inode(struct inode *inode); #endif -static inline int btrfs_test_is_dummy_root(struct btrfs_root *root) +static inline int btrfs_is_testing(struct btrfs_fs_info *fs_info) { #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS - if (unlikely(test_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state))) + if (unlikely(test_bit(BTRFS_FS_STATE_DUMMY_FS_INFO, + &fs_info->fs_state))) return 1; #endif return 0; } - #endif diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 8e838b1cc806..ff2362dca91a 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1146,7 +1146,7 @@ struct extent_buffer *btrfs_find_tree_block(struct btrfs_fs_info *fs_info, struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr) { - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return alloc_test_extent_buffer(root->fs_info, bytenr, root->nodesize); return alloc_extent_buffer(root->fs_info, bytenr); @@ -1330,7 +1330,6 @@ struct btrfs_root *btrfs_alloc_dummy_root(struct btrfs_fs_info *fs_info, /* We don't use the stripesize in selftest, set it as sectorsize */ __setup_root(nodesize, sectorsize, sectorsize, root, fs_info, BTRFS_ROOT_TREE_OBJECTID); - set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; return root; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 44b529f183e3..c4487e177234 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -3235,7 +3235,7 @@ static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, u64, u64, u64, u64, u64, u64); - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return 0; ref_root = btrfs_header_owner(buf); @@ -6101,7 +6101,7 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) if (dropped > 0) to_free += btrfs_calc_trans_metadata_size(root, dropped); - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return; trace_btrfs_space_reservation(root->fs_info, "delalloc", @@ -7217,7 +7217,7 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, int ret; struct btrfs_fs_info *fs_info = root->fs_info; - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(fs_info)) return 0; add_pinned_bytes(root->fs_info, num_bytes, owner, root_objectid); @@ -8356,7 +8356,7 @@ struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, SKINNY_METADATA); #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS - if (btrfs_test_is_dummy_root(root)) { + if (btrfs_is_testing(root->fs_info)) { buf = btrfs_init_new_buffer(trans, root, root->alloc_bytenr, level); if (!IS_ERR(buf)) diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b166ee0763b6..3f0311bb4d34 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1744,7 +1744,7 @@ static void btrfs_set_bit_hook(struct inode *inode, } /* For sanity tests */ - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return; __percpu_counter_add(&root->fs_info->delalloc_bytes, len, @@ -1803,7 +1803,7 @@ static void btrfs_clear_bit_hook(struct inode *inode, btrfs_delalloc_release_metadata(inode, len); /* For sanity tests. */ - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return; if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 13e28d8a7d42..3a83631c1176 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -571,7 +571,7 @@ static int add_qgroup_item(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; struct btrfs_key key; - if (btrfs_test_is_dummy_root(quota_root)) + if (btrfs_is_testing(quota_root->fs_info)) return 0; path = btrfs_alloc_path(); @@ -728,7 +728,7 @@ static int update_qgroup_info_item(struct btrfs_trans_handle *trans, int ret; int slot; - if (btrfs_test_is_dummy_root(root)) + if (btrfs_is_testing(root->fs_info)) return 0; key.objectid = 0; -- cgit v1.2.3 From bd6c57dda666868abb222c1fa48e72f81db20e70 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 10 Jun 2016 17:03:59 -0400 Subject: btrfs: simpilify btrfs_subvol_inherit_props We just need a superblock, but we look it up using two different roots depending on the call site. Let's just use a superblock pointer initialized at the outset. This is mostly for Coccinelle not to choke on my root push up set. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/props.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/props.c b/fs/btrfs/props.c index 36992128c746..cf0b444ac4f3 100644 --- a/fs/btrfs/props.c +++ b/fs/btrfs/props.c @@ -350,6 +350,7 @@ int btrfs_subvol_inherit_props(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_root *parent_root) { + struct super_block *sb = root->fs_info->sb; struct btrfs_key key; struct inode *parent_inode, *child_inode; int ret; @@ -358,12 +359,11 @@ int btrfs_subvol_inherit_props(struct btrfs_trans_handle *trans, key.type = BTRFS_INODE_ITEM_KEY; key.offset = 0; - parent_inode = btrfs_iget(parent_root->fs_info->sb, &key, - parent_root, NULL); + parent_inode = btrfs_iget(sb, &key, parent_root, NULL); if (IS_ERR(parent_inode)) return PTR_ERR(parent_inode); - child_inode = btrfs_iget(root->fs_info->sb, &key, root, NULL); + child_inode = btrfs_iget(sb, &key, root, NULL); if (IS_ERR(child_inode)) { iput(parent_inode); return PTR_ERR(child_inode); -- cgit v1.2.3 From df3975652f23dcfb72aa043a32ecaf587e75b158 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 21 Jun 2016 20:18:21 -0400 Subject: btrfs: copy_to_sk drop unused root parameter The root parameter for copy_to_sk is not used at all. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index ffb1628cd1a1..81413e600d9b 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -1948,8 +1948,7 @@ static noinline int key_in_sk(struct btrfs_key *key, return 1; } -static noinline int copy_to_sk(struct btrfs_root *root, - struct btrfs_path *path, +static noinline int copy_to_sk(struct btrfs_path *path, struct btrfs_key *key, struct btrfs_ioctl_search_key *sk, size_t *buf_size, @@ -2120,7 +2119,7 @@ static noinline int search_ioctl(struct inode *inode, ret = 0; goto err; } - ret = copy_to_sk(root, path, &key, sk, buf_size, ubuf, + ret = copy_to_sk(path, &key, sk, buf_size, ubuf, &sk_offset, &num_found); btrfs_release_path(path); if (ret) -- cgit v1.2.3 From 0c83b62e2222a8872f0b65d50540fdfa53b675ff Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 21 Jun 2016 20:49:29 -0400 Subject: btrfs: cleanup, remove prototype for btrfs_find_root_ref The function isn't implemented anywhere. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 41d5ca986c0c..950e4ee6b66e 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -2888,9 +2888,6 @@ void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info, int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq); /* root-item.c */ -int btrfs_find_root_ref(struct btrfs_root *tree_root, - struct btrfs_path *path, - u64 root_id, u64 ref_id); int btrfs_add_root_ref(struct btrfs_trans_handle *trans, struct btrfs_root *tree_root, u64 root_id, u64 ref_id, u64 dirid, u64 sequence, -- cgit v1.2.3 From 14a1e067b45614d6236e3c82b36f62caef44ac62 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Wed, 15 Jun 2016 10:25:38 -0400 Subject: btrfs: introduce BTRFS_MAX_ITEM_SIZE We use BTRFS_LEAF_DATA_SIZE - sizeof(struct btrfs_item) in several places. This introduces a BTRFS_MAX_ITEM_SIZE macro to do the same. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 8 ++++---- fs/btrfs/extent-tree.c | 2 +- fs/btrfs/file-item.c | 5 ++--- fs/btrfs/volumes.c | 3 +-- 4 files changed, 8 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 950e4ee6b66e..1d46ceec1fc0 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -150,13 +150,13 @@ struct btrfs_header { sizeof(struct btrfs_key_ptr)) #define __BTRFS_LEAF_DATA_SIZE(bs) ((bs) - sizeof(struct btrfs_header)) #define BTRFS_LEAF_DATA_SIZE(r) (__BTRFS_LEAF_DATA_SIZE(r->nodesize)) +#define BTRFS_MAX_ITEM_SIZE(r) \ + (BTRFS_LEAF_DATA_SIZE(r) - sizeof(struct btrfs_item)) #define BTRFS_FILE_EXTENT_INLINE_DATA_START \ (offsetof(struct btrfs_file_extent_item, disk_bytenr)) -#define BTRFS_MAX_INLINE_DATA_SIZE(r) (BTRFS_LEAF_DATA_SIZE(r) - \ - sizeof(struct btrfs_item) - \ +#define BTRFS_MAX_INLINE_DATA_SIZE(r) (BTRFS_MAX_ITEM_SIZE(r) - \ BTRFS_FILE_EXTENT_INLINE_DATA_START) -#define BTRFS_MAX_XATTR_SIZE(r) (BTRFS_LEAF_DATA_SIZE(r) - \ - sizeof(struct btrfs_item) -\ +#define BTRFS_MAX_XATTR_SIZE(r) (BTRFS_MAX_ITEM_SIZE(r) - \ sizeof(struct btrfs_dir_item)) diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index c4487e177234..44549d706dcd 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2779,7 +2779,7 @@ u64 btrfs_csum_bytes_to_leaves(struct btrfs_root *root, u64 csum_bytes) u64 num_csums_per_leaf; u64 num_csums; - csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); + csum_size = BTRFS_MAX_ITEM_SIZE(root); num_csums_per_leaf = div64_u64(csum_size, (u64)btrfs_super_csum_size(root->fs_info->super_copy)); num_csums = div64_u64(csum_bytes, root->sectorsize); diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 6b58d0620e2f..cdbbf85a269b 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -27,9 +27,8 @@ #include "print-tree.h" #include "compression.h" -#define __MAX_CSUM_ITEMS(r, size) ((unsigned long)(((BTRFS_LEAF_DATA_SIZE(r) - \ - sizeof(struct btrfs_item) * 2) / \ - size) - 1)) +#define __MAX_CSUM_ITEMS(r, size) \ + ((unsigned long)(((BTRFS_MAX_ITEM_SIZE(r) * 2) / size) - 1)) #define MAX_CSUM_ITEMS(r, size) (min_t(u32, __MAX_CSUM_ITEMS(r, size), \ PAGE_SIZE)) diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 5e409e28a5c0..104a2e42ed71 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -4573,8 +4573,7 @@ static void check_raid56_incompat_flag(struct btrfs_fs_info *info, u64 type) btrfs_set_fs_incompat(info, RAID56); } -#define BTRFS_MAX_DEVS(r) ((BTRFS_LEAF_DATA_SIZE(r) \ - - sizeof(struct btrfs_item) \ +#define BTRFS_MAX_DEVS(r) ((BTRFS_MAX_ITEM_SIZE(r) \ - sizeof(struct btrfs_chunk)) \ / sizeof(struct btrfs_stripe) + 1) -- cgit v1.2.3 From 1db1ff92b6ce2247999787480c2eeb63a1811e79 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Wed, 15 Jun 2016 10:33:06 -0400 Subject: btrfs: convert nodesize macros to static inlines This patch converts the macros used to calculate various node size limits to static inlines. That way we get type checking for free. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 48 +++++++++++++++++++++++++++++++++--------------- 1 file changed, 33 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 1d46ceec1fc0..85e0b608b7c0 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -145,21 +145,6 @@ struct btrfs_header { u8 level; } __attribute__ ((__packed__)); -#define BTRFS_NODEPTRS_PER_BLOCK(r) (((r)->nodesize - \ - sizeof(struct btrfs_header)) / \ - sizeof(struct btrfs_key_ptr)) -#define __BTRFS_LEAF_DATA_SIZE(bs) ((bs) - sizeof(struct btrfs_header)) -#define BTRFS_LEAF_DATA_SIZE(r) (__BTRFS_LEAF_DATA_SIZE(r->nodesize)) -#define BTRFS_MAX_ITEM_SIZE(r) \ - (BTRFS_LEAF_DATA_SIZE(r) - sizeof(struct btrfs_item)) -#define BTRFS_FILE_EXTENT_INLINE_DATA_START \ - (offsetof(struct btrfs_file_extent_item, disk_bytenr)) -#define BTRFS_MAX_INLINE_DATA_SIZE(r) (BTRFS_MAX_ITEM_SIZE(r) - \ - BTRFS_FILE_EXTENT_INLINE_DATA_START) -#define BTRFS_MAX_XATTR_SIZE(r) (BTRFS_MAX_ITEM_SIZE(r) - \ - sizeof(struct btrfs_dir_item)) - - /* * this is a very generous portion of the super block, giving us * room to translate 14 chunks with 3 stripes each. @@ -1261,6 +1246,39 @@ struct btrfs_root { atomic_t qgroup_meta_rsv; }; +static inline u32 __BTRFS_LEAF_DATA_SIZE(u32 blocksize) +{ + return blocksize - sizeof(struct btrfs_header); +} + +static inline u32 BTRFS_LEAF_DATA_SIZE(const struct btrfs_root *root) +{ + return __BTRFS_LEAF_DATA_SIZE(root->nodesize); +} + +static inline u32 BTRFS_MAX_ITEM_SIZE(const struct btrfs_root *root) +{ + return BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); +} + +static inline u32 BTRFS_NODEPTRS_PER_BLOCK(const struct btrfs_root *root) +{ + return BTRFS_LEAF_DATA_SIZE(root) / sizeof(struct btrfs_key_ptr); +} + +#define BTRFS_FILE_EXTENT_INLINE_DATA_START \ + (offsetof(struct btrfs_file_extent_item, disk_bytenr)) +static inline u32 BTRFS_MAX_INLINE_DATA_SIZE(const struct btrfs_root *root) +{ + return BTRFS_MAX_ITEM_SIZE(root) - + BTRFS_FILE_EXTENT_INLINE_DATA_START; +} + +static inline u32 BTRFS_MAX_XATTR_SIZE(const struct btrfs_root *root) +{ + return BTRFS_MAX_ITEM_SIZE(root) - sizeof(struct btrfs_dir_item); +} + /* * Flags for mount options. * -- cgit v1.2.3 From 05f9a780129c81dfd70ae827a9abb15c7e6321a8 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 24 Jun 2016 16:37:38 -0400 Subject: btrfs: btrfs_relocate_chunk pass extent_root to btrfs_end_transaction In btrfs_relocate_chunk, we get a transaction handle via btrfs_start_trans_remove_block_group, which starts the transaction using the extent root. When we call btrfs_end_transaction, we're calling it using the chunk root. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/volumes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 104a2e42ed71..359ac819a8ed 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2941,7 +2941,7 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, u64 chunk_offset) * chunk tree entries */ ret = btrfs_remove_chunk(trans, root, chunk_offset); - btrfs_end_transaction(trans, root); + btrfs_end_transaction(trans, extent_root); return ret; } -- cgit v1.2.3 From 64b63580728ef19137d35363a1c28794b70ad416 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Mon, 20 Jun 2016 17:23:41 -0400 Subject: btrfs: add btrfs_trans_handle->fs_info pointer btrfs_trans_handle->root is documented as for use for confirming that the root passed in to start the transaction is the same as the one ending it. It's used in several places when an fs_info pointer is needed, so let's just add an fs_info pointer directly. Eventually, the root pointer can be removed. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/extent-tree.c | 6 +++--- fs/btrfs/qgroup.c | 2 +- fs/btrfs/transaction.c | 1 + fs/btrfs/transaction.h | 1 + 4 files changed, 6 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 44549d706dcd..c6c505e2478f 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4589,7 +4589,7 @@ out: */ if (trans->can_flush_pending_bgs && trans->chunk_bytes_reserved >= (u64)SZ_2M) { - btrfs_create_pending_block_groups(trans, trans->root); + btrfs_create_pending_block_groups(trans, extent_root); btrfs_trans_release_chunk_metadata(trans); } return ret; @@ -5730,7 +5730,7 @@ void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, */ void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans) { - struct btrfs_fs_info *fs_info = trans->root->fs_info; + struct btrfs_fs_info *fs_info = trans->fs_info; if (!trans->chunk_bytes_reserved) return; @@ -8543,7 +8543,7 @@ static int record_one_subtree_extent(struct btrfs_trans_handle *trans, delayed_refs = &trans->transaction->delayed_refs; spin_lock(&delayed_refs->lock); - if (btrfs_qgroup_insert_dirty_extent(trans->root->fs_info, + if (btrfs_qgroup_insert_dirty_extent(trans->fs_info, delayed_refs, qrecord)) kfree(qrecord); spin_unlock(&delayed_refs->lock); diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 3a83631c1176..93ee1c18ef9d 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -2196,7 +2196,7 @@ void assert_qgroups_uptodate(struct btrfs_trans_handle *trans) { if (list_empty(&trans->qgroup_ref_list) && !trans->delayed_ref_elem.seq) return; - btrfs_err(trans->root->fs_info, + btrfs_err(trans->fs_info, "qgroups not uptodate in trans handle %p: list is%s empty, " "seq is %#x.%x", trans, list_empty(&trans->qgroup_ref_list) ? "" : " not", diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index d986447d19a4..41e14c632c3f 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -561,6 +561,7 @@ again: h->transaction = cur_trans; h->root = root; h->use_count = 1; + h->fs_info = root->fs_info; h->type = type; h->can_flush_pending_bgs = true; diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index c5abee4f01ad..efb122643380 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -128,6 +128,7 @@ struct btrfs_trans_handle { * Subvolume quota depends on this */ struct btrfs_root *root; + struct btrfs_fs_info *fs_info; struct seq_list delayed_ref_elem; struct list_head qgroup_ref_list; struct list_head new_bgs; -- cgit v1.2.3 From 66642832f06a4351e23cea6cf254967c227f8224 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 10 Jun 2016 18:19:25 -0400 Subject: btrfs: btrfs_abort_transaction, drop root parameter __btrfs_abort_transaction doesn't use its root parameter except to obtain an fs_info pointer. We can obtain that from trans->root->fs_info for now and from trans->fs_info in a later patch. Signed-off-by: Jeff Mahoney Signed-off-by: David Sterba --- fs/btrfs/ctree.c | 12 +++---- fs/btrfs/ctree.h | 8 ++--- fs/btrfs/delayed-inode.c | 2 +- fs/btrfs/extent-tree.c | 47 ++++++++++++++------------- fs/btrfs/file-item.c | 2 +- fs/btrfs/file.c | 8 ++--- fs/btrfs/free-space-cache.c | 2 +- fs/btrfs/free-space-tree.c | 16 +++++----- fs/btrfs/inode-map.c | 4 +-- fs/btrfs/inode.c | 77 ++++++++++++++++++++++----------------------- fs/btrfs/ioctl.c | 33 +++++++++---------- fs/btrfs/relocation.c | 4 +-- fs/btrfs/root-tree.c | 10 +++--- fs/btrfs/super.c | 12 ++++--- fs/btrfs/transaction.c | 30 +++++++++--------- fs/btrfs/tree-log.c | 14 ++++----- fs/btrfs/volumes.c | 18 +++++------ 17 files changed, 147 insertions(+), 152 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index e926b9fdb64a..d1c56c94dd5a 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1153,14 +1153,14 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, ret = update_ref_for_cow(trans, root, buf, cow, &last_ref); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) { ret = btrfs_reloc_cow_block(trans, root, buf, cow); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } } @@ -1198,7 +1198,7 @@ static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, if (last_ref) { ret = tree_mod_log_free_eb(root->fs_info, buf); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } } @@ -3255,7 +3255,7 @@ static int push_node_left(struct btrfs_trans_handle *trans, ret = tree_mod_log_eb_copy(root->fs_info, dst, src, dst_nritems, 0, push_items); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } copy_extent_buffer(dst, src, @@ -3330,7 +3330,7 @@ static int balance_node_right(struct btrfs_trans_handle *trans, ret = tree_mod_log_eb_copy(root->fs_info, dst, src, 0, src_nritems - push_items, push_items); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } copy_extent_buffer(dst, src, @@ -3534,7 +3534,7 @@ static noinline int split_node(struct btrfs_trans_handle *trans, ret = tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } copy_extent_buffer(split, c, diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 85e0b608b7c0..f66a0ba9a2a9 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3379,23 +3379,23 @@ const char *btrfs_decode_error(int errno); __cold void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, - struct btrfs_root *root, const char *function, + const char *function, unsigned int line, int errno); /* * Call btrfs_abort_transaction as early as possible when an error condition is * detected, that way the exact line number is reported. */ -#define btrfs_abort_transaction(trans, root, errno) \ +#define btrfs_abort_transaction(trans, errno) \ do { \ /* Report first abort since mount */ \ if (!test_and_set_bit(BTRFS_FS_STATE_TRANS_ABORTED, \ - &((root)->fs_info->fs_state))) { \ + &((trans)->fs_info->fs_state))) { \ WARN(1, KERN_DEBUG \ "BTRFS: Transaction aborted (error %d)\n", \ (errno)); \ } \ - __btrfs_abort_transaction((trans), (root), __func__, \ + __btrfs_abort_transaction((trans), __func__, \ __LINE__, (errno)); \ } while (0) diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 9e2a907d9ddb..3eeb9cd8cfa5 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -1170,7 +1170,7 @@ static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans, if (ret) { btrfs_release_delayed_node(curr_node); curr_node = NULL; - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); break; } diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index c6c505e2478f..c2b81b0d3fe0 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -2180,7 +2180,7 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, path, bytenr, parent, root_objectid, owner, offset, refs_to_add); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); out: btrfs_free_path(path); return ret; @@ -2971,7 +2971,7 @@ again: trans->can_flush_pending_bgs = false; ret = __btrfs_run_delayed_refs(trans, root, count); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -3430,7 +3430,7 @@ again: * transaction, this only happens in really bad situations * anyway. */ - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_put; } WARN_ON(ret); @@ -3670,7 +3670,7 @@ again: } spin_unlock(&cur_trans->dirty_bgs_lock); } else if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); } } @@ -3816,7 +3816,7 @@ int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans, cache); } if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); } /* if its not on the io list, we need to put the block group */ @@ -6881,7 +6881,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, NULL, refs_to_drop, is_data, &last_ref); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } btrfs_release_path(path); @@ -6930,7 +6930,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, path->nodes[0]); } if (ret < 0) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } extent_slot = path->slots[0]; @@ -6941,10 +6941,10 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, "unable to find ref byte nr %llu parent %llu root %llu owner %llu offset %llu", bytenr, parent, root_objectid, owner_objectid, owner_offset); - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } else { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -6956,7 +6956,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = convert_extent_item_v0(trans, extent_root, path, owner_objectid, 0); if (ret < 0) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -6975,7 +6975,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, btrfs_print_leaf(extent_root, path->nodes[0]); } if (ret < 0) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -7000,7 +7000,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, btrfs_err(info, "trying to drop %d refs but we only have %Lu " "for bytenr %Lu", refs_to_drop, refs, bytenr); ret = -EINVAL; - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } refs -= refs_to_drop; @@ -7023,7 +7023,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, iref, refs_to_drop, is_data, &last_ref); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } @@ -7046,7 +7046,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = btrfs_del_items(trans, extent_root, path, path->slots[0], num_to_del); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } btrfs_release_path(path); @@ -7054,7 +7054,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, if (is_data) { ret = btrfs_del_csums(trans, root, bytenr, num_bytes); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } @@ -7062,13 +7062,13 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = add_to_free_space_tree(trans, root->fs_info, bytenr, num_bytes); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } ret = update_block_group(trans, root, bytenr, num_bytes, 0); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } @@ -7852,8 +7852,7 @@ loop: * can do more things. */ if (ret < 0 && ret != -ENOSPC) - btrfs_abort_transaction(trans, - root, ret); + btrfs_abort_transaction(trans, ret); else ret = 0; if (!exist) @@ -9329,7 +9328,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, &root->root_key, root_item); if (ret) { - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); err = ret; goto out_end_trans; } @@ -9356,7 +9355,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, ret = btrfs_del_root(trans, tree_root, &root->root_key); if (ret) { - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); goto out_end_trans; } @@ -9364,7 +9363,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, ret = btrfs_find_root(tree_root, &root->root_key, path, NULL, NULL); if (ret < 0) { - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); err = ret; goto out_end_trans; } else if (ret > 0) { @@ -10324,11 +10323,11 @@ void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans, ret = btrfs_insert_item(trans, extent_root, &key, &item, sizeof(item)); if (ret) - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); ret = btrfs_finish_chunk_alloc(trans, extent_root, key.objectid, key.offset); if (ret) - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); add_block_group_free_space(trans, root->fs_info, block_group); /* already aborted the transaction if it failed. */ next: diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index cdbbf85a269b..2fc803da4680 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -698,7 +698,7 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, */ ret = btrfs_split_item(trans, root, path, &key, offset); if (ret && ret != -EAGAIN) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 48b94904a519..9404121fd5f7 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -950,7 +950,7 @@ delete_extent_item: ret = btrfs_del_items(trans, root, path, del_slot, del_nr); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); break; } @@ -974,7 +974,7 @@ delete_extent_item: path->slots[0] = del_slot; ret = btrfs_del_items(trans, root, path, del_slot, del_nr); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); } leaf = path->nodes[0]; @@ -1190,7 +1190,7 @@ again: goto again; } if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -1278,7 +1278,7 @@ again: ret = btrfs_del_items(trans, root, path, del_slot, del_nr); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index cee2049ef64e..d571bd2b697b 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -280,7 +280,7 @@ fail: if (locked) mutex_unlock(&trans->transaction->cache_write_mutex); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } diff --git a/fs/btrfs/free-space-tree.c b/fs/btrfs/free-space-tree.c index 53dbeaf6ce94..87e7e3d3e676 100644 --- a/fs/btrfs/free-space-tree.c +++ b/fs/btrfs/free-space-tree.c @@ -305,7 +305,7 @@ int convert_free_space_to_bitmaps(struct btrfs_trans_handle *trans, out: kvfree(bitmap); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -454,7 +454,7 @@ int convert_free_space_to_extents(struct btrfs_trans_handle *trans, out: kvfree(bitmap); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -851,7 +851,7 @@ int remove_from_free_space_tree(struct btrfs_trans_handle *trans, out: btrfs_free_path(path); if (ret) - btrfs_abort_transaction(trans, fs_info->free_space_root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -1047,7 +1047,7 @@ int add_to_free_space_tree(struct btrfs_trans_handle *trans, out: btrfs_free_path(path); if (ret) - btrfs_abort_transaction(trans, fs_info->free_space_root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -1193,7 +1193,7 @@ int btrfs_create_free_space_tree(struct btrfs_fs_info *fs_info) abort: fs_info->creating_free_space_tree = 0; - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, tree_root); return ret; } @@ -1280,7 +1280,7 @@ int btrfs_clear_free_space_tree(struct btrfs_fs_info *fs_info) return 0; abort: - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, tree_root); return ret; } @@ -1333,7 +1333,7 @@ out: btrfs_free_path(path); mutex_unlock(&block_group->free_space_lock); if (ret) - btrfs_abort_transaction(trans, fs_info->free_space_root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -1410,7 +1410,7 @@ int remove_block_group_free_space(struct btrfs_trans_handle *trans, out: btrfs_free_path(path); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index e3ad8c197997..aa6fabaee72e 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -458,7 +458,7 @@ again: BTRFS_I(inode)->generation = 0; ret = btrfs_update_inode(trans, root, inode); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_put; } @@ -466,7 +466,7 @@ again: ret = btrfs_truncate_free_space_cache(root, trans, NULL, inode); if (ret) { if (ret != -ENOSPC) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_put; } } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3f0311bb4d34..72cfd15b1fa8 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -296,7 +296,7 @@ static noinline int cow_file_range_inline(struct btrfs_root *root, start, aligned_end, NULL, 1, 1, extent_item_size, &extent_inserted); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -307,7 +307,7 @@ static noinline int cow_file_range_inline(struct btrfs_root *root, inline_len, compressed_size, compress_type, compressed_pages); if (ret && ret != -ENOSPC) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } else if (ret == -ENOSPC) { ret = 1; @@ -2603,7 +2603,7 @@ again: ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(*extent)); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_free_path; } @@ -2630,7 +2630,7 @@ again: backref->root_id, backref->inum, new->file_pos); /* start - extent_offset */ if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_free_path; } @@ -2899,7 +2899,7 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) trans->block_rsv = &root->fs_info->delalloc_block_rsv; ret = btrfs_update_inode_fallback(trans, root, inode); if (ret) /* -ENOMEM or corruption */ - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -2959,7 +2959,7 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) ordered_extent->file_offset, ordered_extent->len, trans->transid); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_unlock; } @@ -2969,7 +2969,7 @@ static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) btrfs_ordered_update_i_size(inode, 0, ordered_extent); ret = btrfs_update_inode_fallback(trans, root, inode); if (ret) { /* -ENOMEM or corruption */ - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_unlock; } ret = 0; @@ -3213,7 +3213,7 @@ void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root, root->root_key.objectid); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); else clear_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state); @@ -3304,7 +3304,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) if (ret != -EEXIST) { clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, &BTRFS_I(inode)->runtime_flags); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } } @@ -3316,7 +3316,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, root->root_key.objectid); if (ret && ret != -EEXIST) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } } @@ -4015,20 +4015,20 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, btrfs_info(root->fs_info, "failed to delete reference to %.*s, inode %llu parent %llu", name_len, name, ino, dir_ino); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto err; } skip_backref: ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto err; } ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, inode, dir_ino); if (ret != 0 && ret != -ENOENT) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto err; } @@ -4037,7 +4037,7 @@ skip_backref: if (ret == -ENOENT) ret = 0; else if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); err: btrfs_free_path(path); if (ret) @@ -4151,7 +4151,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); ret = btrfs_delete_one_dir_name(trans, root, path, di); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } btrfs_release_path(path); @@ -4161,7 +4161,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, dir_ino, &index, name, name_len); if (ret < 0) { if (ret != -ENOENT) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } di = btrfs_search_dir_index_item(root, path, dir_ino, @@ -4171,7 +4171,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, ret = -ENOENT; else ret = PTR_ERR(di); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -4184,7 +4184,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -4193,7 +4193,7 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, dir->i_mtime = dir->i_ctime = current_fs_time(dir->i_sb); ret = btrfs_update_inode_fallback(trans, root, dir); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); out: btrfs_free_path(path); return ret; @@ -4514,7 +4514,6 @@ search_again: pending_del_nr); if (err) { btrfs_abort_transaction(trans, - root, err); goto error; } @@ -4526,8 +4525,7 @@ search_again: item_end, new_size); if (err) { - btrfs_abort_transaction(trans, - root, err); + btrfs_abort_transaction(trans, err); goto error; } } else if (test_bit(BTRFS_ROOT_REF_COWS, @@ -4591,8 +4589,7 @@ delete: pending_del_slot, pending_del_nr); if (ret) { - btrfs_abort_transaction(trans, - root, ret); + btrfs_abort_transaction(trans, ret); goto error; } pending_del_nr = 0; @@ -4625,7 +4622,7 @@ out: ret = btrfs_del_items(trans, root, path, pending_del_slot, pending_del_nr); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); } error: if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) @@ -4794,7 +4791,7 @@ static int maybe_insert_hole(struct btrfs_root *root, struct inode *inode, ret = btrfs_drop_extents(trans, root, inode, offset, offset + len, 1); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, root); return ret; } @@ -4802,7 +4799,7 @@ static int maybe_insert_hole(struct btrfs_root *root, struct inode *inode, ret = btrfs_insert_file_extent(trans, root, btrfs_ino(inode), offset, 0, 0, len, 0, len, 0, 0, 0); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); else btrfs_update_inode(trans, root, inode); btrfs_end_transaction(trans, root); @@ -5029,7 +5026,7 @@ static int btrfs_setsize(struct inode *inode, struct iattr *attr) i_size_write(inode, BTRFS_I(inode)->disk_i_size); err = btrfs_orphan_del(trans, inode); if (err) - btrfs_abort_transaction(trans, root, err); + btrfs_abort_transaction(trans, err); btrfs_end_transaction(trans, root); } } @@ -6335,7 +6332,7 @@ int btrfs_add_link(struct btrfs_trans_handle *trans, if (ret == -EEXIST || ret == -EOVERFLOW) goto fail_dir_item; else if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; } @@ -6346,7 +6343,7 @@ int btrfs_add_link(struct btrfs_trans_handle *trans, current_fs_time(parent_inode->i_sb); ret = btrfs_update_inode(trans, root, parent_inode); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); return ret; fail_dir_item: @@ -9568,7 +9565,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, ret = btrfs_update_inode(trans, root, old_inode); } if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9588,7 +9585,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, ret = btrfs_update_inode(trans, dest, new_inode); } if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9596,7 +9593,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, new_dentry->d_name.name, new_dentry->d_name.len, 0, old_idx); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9604,7 +9601,7 @@ static int btrfs_rename_exchange(struct inode *old_dir, old_dentry->d_name.name, old_dentry->d_name.len, 0, new_idx); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9843,7 +9840,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, ret = btrfs_update_inode(trans, root, old_inode); } if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9867,7 +9864,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (!ret && new_inode->i_nlink == 0) ret = btrfs_orphan_add(trans, d_inode(new_dentry)); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } } @@ -9876,7 +9873,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, new_dentry->d_name.name, new_dentry->d_name.len, 0, index); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } @@ -9896,7 +9893,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, old_dentry); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_fail; } } @@ -10322,7 +10319,7 @@ static int __btrfs_prealloc_file_range(struct inode *inode, int mode, if (ret) { btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 0); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); if (own_trans) btrfs_end_transaction(trans, root); break; @@ -10382,7 +10379,7 @@ next: ret = btrfs_update_inode(trans, root, inode); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); if (own_trans) btrfs_end_transaction(trans, root); break; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 81413e600d9b..14ed1e9e6bc8 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -561,7 +561,7 @@ static noinline int create_subvol(struct inode *dir, new_root = btrfs_read_fs_root_no_name(root->fs_info, &key); if (IS_ERR(new_root)) { ret = PTR_ERR(new_root); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -570,7 +570,7 @@ static noinline int create_subvol(struct inode *dir, ret = btrfs_create_subvol_root(trans, new_root, root, new_dirid); if (ret) { /* We potentially lose an unused inode item here */ - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -583,7 +583,7 @@ static noinline int create_subvol(struct inode *dir, */ ret = btrfs_set_inode_index(dir, &index); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -591,7 +591,7 @@ static noinline int create_subvol(struct inode *dir, name, namelen, dir, &key, BTRFS_FT_DIR, index); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -608,7 +608,7 @@ static noinline int create_subvol(struct inode *dir, root_item->uuid, BTRFS_UUID_KEY_SUBVOL, objectid); if (ret) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); fail: kfree(root_item); @@ -2488,7 +2488,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, dentry->d_name.len); if (ret) { err = ret; - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_end_trans; } @@ -2504,7 +2504,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, root->fs_info->tree_root, dest->root_key.objectid); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); err = ret; goto out_end_trans; } @@ -2514,7 +2514,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, dest->root_item.uuid, BTRFS_UUID_KEY_SUBVOL, dest->root_key.objectid); if (ret && ret != -ENOENT) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); err = ret; goto out_end_trans; } @@ -2524,7 +2524,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, BTRFS_UUID_KEY_RECEIVED_SUBVOL, dest->root_key.objectid); if (ret && ret != -ENOENT) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); err = ret; goto out_end_trans; } @@ -3291,7 +3291,7 @@ static int clone_finish_inode_update(struct btrfs_trans_handle *trans, ret = btrfs_update_inode(trans, root, inode); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, root); goto out; } @@ -3693,7 +3693,7 @@ process_slot: if (ret) { if (ret != -EOPNOTSUPP) btrfs_abort_transaction(trans, - root, ret); + ret); btrfs_end_transaction(trans, root); goto out; } @@ -3701,8 +3701,7 @@ process_slot: ret = btrfs_insert_empty_item(trans, root, path, &new_key, size); if (ret) { - btrfs_abort_transaction(trans, root, - ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, root); goto out; } @@ -3734,7 +3733,6 @@ process_slot: new_key.offset - datao); if (ret) { btrfs_abort_transaction(trans, - root, ret); btrfs_end_transaction(trans, root); @@ -3771,7 +3769,6 @@ process_slot: if (ret) { if (ret != -EOPNOTSUPP) btrfs_abort_transaction(trans, - root, ret); btrfs_end_transaction(trans, root); goto out; @@ -3827,7 +3824,7 @@ process_slot: last_dest_end, destoff + len, 1); if (ret) { if (ret != -EOPNOTSUPP) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, root); goto out; } @@ -5163,13 +5160,13 @@ static long _btrfs_ioctl_set_received_subvol(struct file *file, BTRFS_UUID_KEY_RECEIVED_SUBVOL, root->root_key.objectid); if (ret < 0 && ret != -EEXIST) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } ret = btrfs_commit_transaction(trans, root); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 97cafe7291b8..b26a5aea41b4 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -1723,7 +1723,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans, btrfs_header_owner(leaf), key.objectid, key.offset); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); break; } @@ -1731,7 +1731,7 @@ int replace_file_extents(struct btrfs_trans_handle *trans, parent, btrfs_header_owner(leaf), key.objectid, key.offset); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); break; } } diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index f1c30861d062..7fd7e1830cfe 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -150,7 +150,7 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root ret = btrfs_search_slot(trans, root, key, path, 0, 1); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -176,20 +176,20 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root ret = btrfs_search_slot(trans, root, key, path, -1, 1); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } ret = btrfs_del_item(trans, root, path); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } btrfs_release_path(path); ret = btrfs_insert_empty_item(trans, root, path, key, sizeof(*item)); if (ret < 0) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } l = path->nodes[0]; @@ -448,7 +448,7 @@ again: ret = btrfs_insert_empty_item(trans, tree_root, path, &key, sizeof(*ref) + name_len); if (ret) { - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); btrfs_free_path(path); return ret; } diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 959823a12c42..864ce334f696 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -251,9 +251,11 @@ void btrfs_printk(const struct btrfs_fs_info *fs_info, const char *fmt, ...) */ __cold void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, - struct btrfs_root *root, const char *function, + const char *function, unsigned int line, int errno) { + struct btrfs_fs_info *fs_info = trans->fs_info; + trans->aborted = errno; /* Nothing used. The other threads that have joined this * transaction may be able to continue. */ @@ -261,16 +263,16 @@ void __btrfs_abort_transaction(struct btrfs_trans_handle *trans, const char *errstr; errstr = btrfs_decode_error(errno); - btrfs_warn(root->fs_info, + btrfs_warn(fs_info, "%s:%d: Aborting unused transaction(%s).", function, line, errstr); return; } ACCESS_ONCE(trans->transaction->aborted) = errno; /* Wake up anybody who may be waiting on this transaction */ - wake_up(&root->fs_info->transaction_wait); - wake_up(&root->fs_info->transaction_blocked_wait); - __btrfs_handle_fs_error(root->fs_info, function, line, errno, NULL); + wake_up(&fs_info->transaction_wait); + wake_up(&fs_info->transaction_blocked_wait); + __btrfs_handle_fs_error(fs_info, function, line, errno, NULL); } /* * __btrfs_panic decodes unexpected, fatal errors from the caller, diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 41e14c632c3f..9cca0a721961 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1492,7 +1492,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto dir_item_existed; } else if (IS_ERR(dir_item)) { ret = PTR_ERR(dir_item); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } btrfs_release_path(path); @@ -1505,7 +1505,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, */ ret = btrfs_run_delayed_items(trans, root); if (ret) { /* Transaction aborted */ - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1544,7 +1544,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, if (ret) { btrfs_tree_unlock(old); free_extent_buffer(old); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1555,7 +1555,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_tree_unlock(old); free_extent_buffer(old); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } /* see comments in should_cow_block() */ @@ -1569,7 +1569,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_tree_unlock(tmp); free_extent_buffer(tmp); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1581,7 +1581,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_ino(parent_inode), index, dentry->d_name.name, dentry->d_name.len); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1589,19 +1589,19 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, pending->snap = btrfs_read_fs_root_no_name(root->fs_info, &key); if (IS_ERR(pending->snap)) { ret = PTR_ERR(pending->snap); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } ret = btrfs_reloc_post_snapshot(trans, pending); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1623,7 +1623,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, /* We have check then name at the beginning, so it is impossible. */ BUG_ON(ret == -EEXIST || ret == -EOVERFLOW); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1633,13 +1633,13 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, current_fs_time(parent_inode->i_sb); ret = btrfs_update_inode_fallback(trans, parent_root, parent_inode); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } ret = btrfs_uuid_tree_add(trans, fs_info->uuid_root, new_uuid.b, BTRFS_UUID_KEY_SUBVOL, objectid); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } if (!btrfs_is_empty_uuid(new_root_item->received_uuid)) { @@ -1648,14 +1648,14 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, BTRFS_UUID_KEY_RECEIVED_SUBVOL, objectid); if (ret && ret != -EEXIST) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } } ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto fail; } @@ -1851,7 +1851,7 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans, WARN_ON(trans->use_count > 1); - btrfs_abort_transaction(trans, root, err); + btrfs_abort_transaction(trans, err); spin_lock(&root->fs_info->trans_lock); diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 391bd3ddaed3..d31a0c4f56be 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -2788,7 +2788,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark); if (ret) { blk_finish_plug(&plug); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_free_logged_extents(log, log_transid); btrfs_set_log_full_commit(root->fs_info, trans); mutex_unlock(&root->log_mutex); @@ -2838,7 +2838,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, btrfs_set_log_full_commit(root->fs_info, trans); if (ret != -ENOSPC) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); mutex_unlock(&log_root_tree->log_mutex); goto out; } @@ -2898,7 +2898,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, blk_finish_plug(&plug); if (ret) { btrfs_set_log_full_commit(root->fs_info, trans); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_free_logged_extents(log, log_transid); mutex_unlock(&log_root_tree->log_mutex); goto out_wake_log_root; @@ -2934,7 +2934,7 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, ret = write_ctree_super(trans, root->fs_info->tree_root, 1); if (ret) { btrfs_set_log_full_commit(root->fs_info, trans); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out_wake_log_root; } @@ -2991,7 +2991,7 @@ static void free_log_tree(struct btrfs_trans_handle *trans, ret = walk_log_tree(trans, log, &wc); /* I don't think this can happen but just in case */ if (ret) - btrfs_abort_transaction(trans, log, ret); + btrfs_abort_transaction(trans, ret); while (1) { ret = find_first_extent_bit(&log->dirty_log_pages, @@ -3160,7 +3160,7 @@ out_unlock: btrfs_set_log_full_commit(root->fs_info, trans); ret = 0; } else if (ret < 0) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_log_trans(root); @@ -3193,7 +3193,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, btrfs_set_log_full_commit(root->fs_info, trans); ret = 0; } else if (ret < 0 && ret != -ENOENT) - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_log_trans(root); return ret; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 359ac819a8ed..e217035d63df 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2438,14 +2438,14 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) ret = init_first_rw_device(trans, root, device); unlock_chunks(root); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto error_trans; } } ret = btrfs_add_device(trans, root, device); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto error_trans; } @@ -2454,7 +2454,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) ret = btrfs_finish_sprout(trans, root); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto error_trans; } @@ -2840,7 +2840,7 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, &dev_extent_len); if (ret) { mutex_unlock(&fs_devices->device_list_mutex); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -2859,7 +2859,7 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, ret = btrfs_update_device(trans, map->stripes[i].dev); if (ret) { mutex_unlock(&fs_devices->device_list_mutex); - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } @@ -2868,7 +2868,7 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, ret = btrfs_free_chunk(trans, root, chunk_objectid, chunk_offset); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -2877,14 +2877,14 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) { ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset); if (ret) { - btrfs_abort_transaction(trans, root, ret); + btrfs_abort_transaction(trans, ret); goto out; } } ret = btrfs_remove_block_group(trans, extent_root, chunk_offset, em); if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); + btrfs_abort_transaction(trans, ret); goto out; } @@ -4299,7 +4299,7 @@ int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info) BTRFS_UUID_TREE_OBJECTID); if (IS_ERR(uuid_root)) { ret = PTR_ERR(uuid_root); - btrfs_abort_transaction(trans, tree_root, ret); + btrfs_abort_transaction(trans, ret); btrfs_end_transaction(trans, tree_root); return ret; } -- cgit v1.2.3 From 698c937b0d17dd55227622b919482fc720cc1095 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 25 Jul 2016 13:31:14 -0400 Subject: NFSv4: Clean up lookup of SECINFO_NO_NAME Use the minor version ops cached in struct nfs_client instead of looking them up again. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 4b1491da16eb..da5c9e58e907 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3281,13 +3281,6 @@ static int nfs4_find_root_sec(struct nfs_server *server, struct nfs_fh *fhandle, return status; } -static int nfs4_do_find_root_sec(struct nfs_server *server, - struct nfs_fh *fhandle, struct nfs_fsinfo *info) -{ - int mv = server->nfs_client->cl_minorversion; - return nfs_v4_minor_ops[mv]->find_root_sec(server, fhandle, info); -} - /** * nfs4_proc_get_rootfh - get file handle for server's pseudoroot * @server: initialized nfs_server handle @@ -3307,7 +3300,8 @@ int nfs4_proc_get_rootfh(struct nfs_server *server, struct nfs_fh *fhandle, status = nfs4_lookup_root(server, fhandle, info); if (auth_probe || status == NFS4ERR_WRONGSEC) - status = nfs4_do_find_root_sec(server, fhandle, info); + status = server->nfs_client->cl_mvops->find_root_sec(server, + fhandle, info); if (status == 0) status = nfs4_server_capabilities(server, fhandle); -- cgit v1.2.3 From 281dbe5db81c6137def9757e07a7aea14b1ed86e Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Tue, 26 Jul 2016 15:22:35 +0200 Subject: libceph: add an ONSTACK initializer for oids An on-stack oid in ceph_ioctl_get_dataloc() is not initialized, resulting in a WARN and a NULL pointer dereference later on. We will have more of these on-stack in the future, so fix it with a convenience macro. Fixes: d30291b985d1 ("libceph: variable-sized ceph_object_id") Signed-off-by: Ilya Dryomov --- fs/ceph/ioctl.c | 2 +- include/linux/ceph/osdmap.h | 5 +++++ 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index be6b1657b1af..0946f2d4a81f 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -183,7 +183,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) struct ceph_osd_client *osdc = &ceph_sb_to_client(inode->i_sb)->client->osdc; struct ceph_object_locator oloc; - struct ceph_object_id oid; + CEPH_DEFINE_OID_ONSTACK(oid); u64 len = 1, olen; u64 tmp; struct ceph_pg pgid; diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 9ccf4dbe55f8..21d7f048959f 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -115,6 +115,11 @@ static inline void ceph_oid_init(struct ceph_object_id *oid) oid->name_len = 0; } +#define CEPH_OID_INIT_ONSTACK(oid) \ + ({ ceph_oid_init(&oid); oid; }) +#define CEPH_DEFINE_OID_ONSTACK(oid) \ + struct ceph_object_id oid = CEPH_OID_INIT_ONSTACK(oid) + static inline bool ceph_oid_empty(const struct ceph_object_id *oid) { return oid->name == oid->inline_name && !oid->name_len; -- cgit v1.2.3 From 7627151ea30bce2051e3cb27d7bb2c30083f86a5 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 3 Feb 2016 21:24:49 +0800 Subject: libceph: define new ceph_file_layout structure Define new ceph_file_layout structure and rename old ceph_file_layout to ceph_file_layout_legacy. This is preparation for adding namespace to ceph_file_layout structure. Signed-off-by: Yan, Zheng --- drivers/block/rbd.c | 14 ++++++------- fs/ceph/addr.c | 18 ++++++++-------- fs/ceph/caps.c | 5 ++++- fs/ceph/file.c | 6 +++--- fs/ceph/inode.c | 7 ++++--- fs/ceph/ioctl.c | 22 +++++++++---------- fs/ceph/mds_client.h | 2 +- fs/ceph/xattr.c | 28 ++++++++++--------------- include/linux/ceph/ceph_fs.h | 50 +++++++++++++++++++------------------------- net/ceph/ceph_fs.c | 30 +++++++++++++++++++++++--- net/ceph/osd_client.c | 4 ++-- net/ceph/osdmap.c | 6 +++--- 12 files changed, 103 insertions(+), 89 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index 81666a56415e..cc272ed46cfd 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -1937,7 +1937,7 @@ static struct ceph_osd_request *rbd_osd_req_create( osd_req->r_callback = rbd_osd_req_callback; osd_req->r_priv = obj_request; - osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + osd_req->r_base_oloc.pool = rbd_dev->layout.pool_id; if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", obj_request->object_name)) goto fail; @@ -1991,7 +1991,7 @@ rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request) osd_req->r_callback = rbd_osd_req_callback; osd_req->r_priv = obj_request; - osd_req->r_base_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + osd_req->r_base_oloc.pool = rbd_dev->layout.pool_id; if (ceph_oid_aprintf(&osd_req->r_base_oid, GFP_NOIO, "%s", obj_request->object_name)) goto fail; @@ -3995,10 +3995,10 @@ static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc, /* Initialize the layout used for all rbd requests */ - rbd_dev->layout.fl_stripe_unit = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); - rbd_dev->layout.fl_stripe_count = cpu_to_le32(1); - rbd_dev->layout.fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER); - rbd_dev->layout.fl_pg_pool = cpu_to_le32((u32) spec->pool_id); + rbd_dev->layout.stripe_unit = 1 << RBD_MAX_OBJ_ORDER; + rbd_dev->layout.stripe_count = 1; + rbd_dev->layout.object_size = 1 << RBD_MAX_OBJ_ORDER; + rbd_dev->layout.pool_id = spec->pool_id; /* * If this is a mapping rbd_dev (as opposed to a parent one), @@ -5187,7 +5187,7 @@ static int rbd_dev_header_name(struct rbd_device *rbd_dev) rbd_assert(rbd_image_format_valid(rbd_dev->image_format)); - rbd_dev->header_oloc.pool = ceph_file_layout_pg_pool(rbd_dev->layout); + rbd_dev->header_oloc.pool = rbd_dev->layout.pool_id; if (rbd_dev->image_format == 1) ret = ceph_oid_aprintf(&rbd_dev->header_oid, GFP_KERNEL, "%s%s", spec->image_name, RBD_SUFFIX); diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 26a9d10d75e9..3f8efd866fec 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1730,7 +1730,7 @@ enum { POOL_WRITE = 2, }; -static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) +static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) { struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); struct ceph_mds_client *mdsc = fsc->mdsc; @@ -1757,7 +1757,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, u32 pool) if (*p) goto out; - dout("__ceph_pool_perm_get pool %u no perm cached\n", pool); + dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool); down_write(&mdsc->pool_perm_rwsem); parent = NULL; @@ -1860,13 +1860,13 @@ out_unlock: out: if (!err) err = have; - dout("__ceph_pool_perm_get pool %u result = %d\n", pool, err); + dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err); return err; } int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) { - u32 pool; + s64 pool; int ret, flags; /* does not support pool namespace yet */ @@ -1879,17 +1879,17 @@ int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) spin_lock(&ci->i_ceph_lock); flags = ci->i_ceph_flags; - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; spin_unlock(&ci->i_ceph_lock); check: if (flags & CEPH_I_POOL_PERM) { if ((need & CEPH_CAP_FILE_RD) && !(flags & CEPH_I_POOL_RD)) { - dout("ceph_pool_perm_check pool %u no read perm\n", + dout("ceph_pool_perm_check pool %lld no read perm\n", pool); return -EPERM; } if ((need & CEPH_CAP_FILE_WR) && !(flags & CEPH_I_POOL_WR)) { - dout("ceph_pool_perm_check pool %u no write perm\n", + dout("ceph_pool_perm_check pool %lld no write perm\n", pool); return -EPERM; } @@ -1907,10 +1907,10 @@ check: flags |= CEPH_I_POOL_WR; spin_lock(&ci->i_ceph_lock); - if (pool == ceph_file_layout_pg_pool(ci->i_layout)) { + if (pool == ci->i_layout.pool_id) { ci->i_ceph_flags = flags; } else { - pool = ceph_file_layout_pg_pool(ci->i_layout); + pool = ci->i_layout.pool_id; flags = ci->i_ceph_flags; } spin_unlock(&ci->i_ceph_lock); diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 6f60d0a3d0f9..f24722dce167 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2895,8 +2895,11 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (newcaps & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR)) { /* file layout may have changed */ - ci->i_layout = grant->layout; + s64 old_pool = ci->i_layout.pool_id; + ceph_file_layout_from_legacy(&ci->i_layout, &grant->layout); ci->i_pool_ns_len = pool_ns_len; + if (ci->i_layout.pool_id != old_pool) + ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; /* size/truncate_seq? */ queue_trunc = ceph_fill_file_size(inode, issued, diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 0daaf7ceedc5..cba5dcf49a65 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -1583,9 +1583,9 @@ static int ceph_zero_objects(struct inode *inode, loff_t offset, loff_t length) { int ret = 0; struct ceph_inode_info *ci = ceph_inode(inode); - s32 stripe_unit = ceph_file_layout_su(ci->i_layout); - s32 stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - s32 object_size = ceph_file_layout_object_size(ci->i_layout); + s32 stripe_unit = ci->i_layout.stripe_unit; + s32 stripe_count = ci->i_layout.stripe_count; + s32 object_size = ci->i_layout.object_size; u64 object_set_size = object_size * stripe_count; u64 nearly, t; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index f059b5997072..6c5903e0a976 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -814,10 +814,11 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (new_version || (new_issued & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { - if (ci->i_layout.fl_pg_pool != info->layout.fl_pg_pool) - ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; - ci->i_layout = info->layout; + s64 old_pool = ci->i_layout.pool_id; + ceph_file_layout_from_legacy(&ci->i_layout, &info->layout); ci->i_pool_ns_len = iinfo->pool_ns_len; + if (ci->i_layout.pool_id != old_pool) + ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; queue_trunc = ceph_fill_file_size(inode, issued, le32_to_cpu(info->truncate_seq), diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 0946f2d4a81f..843dd31a02cd 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -21,10 +21,10 @@ static long ceph_ioctl_get_layout(struct file *file, void __user *arg) err = ceph_do_getattr(file_inode(file), CEPH_STAT_CAP_LAYOUT, false); if (!err) { - l.stripe_unit = ceph_file_layout_su(ci->i_layout); - l.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); - l.object_size = ceph_file_layout_object_size(ci->i_layout); - l.data_pool = le32_to_cpu(ci->i_layout.fl_pg_pool); + l.stripe_unit = ci->i_layout.stripe_unit; + l.stripe_count = ci->i_layout.stripe_count; + l.object_size = ci->i_layout.object_size; + l.data_pool = ci->i_layout.pool_id; l.preferred_osd = (s32)-1; if (copy_to_user(arg, &l, sizeof(l))) return -EFAULT; @@ -82,19 +82,19 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) if (l.stripe_count) nl.stripe_count = l.stripe_count; else - nl.stripe_count = ceph_file_layout_stripe_count(ci->i_layout); + nl.stripe_count = ci->i_layout.stripe_count; if (l.stripe_unit) nl.stripe_unit = l.stripe_unit; else - nl.stripe_unit = ceph_file_layout_su(ci->i_layout); + nl.stripe_unit = ci->i_layout.stripe_unit; if (l.object_size) nl.object_size = l.object_size; else - nl.object_size = ceph_file_layout_object_size(ci->i_layout); + nl.object_size = ci->i_layout.object_size; if (l.data_pool) nl.data_pool = l.data_pool; else - nl.data_pool = ceph_file_layout_pg_pool(ci->i_layout); + nl.data_pool = ci->i_layout.pool_id; /* this is obsolete, and always -1 */ nl.preferred_osd = le64_to_cpu(-1); @@ -202,8 +202,8 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) return -EIO; } dl.file_offset -= dl.object_offset; - dl.object_size = ceph_file_layout_object_size(ci->i_layout); - dl.block_size = ceph_file_layout_su(ci->i_layout); + dl.object_size = ci->i_layout.object_size; + dl.block_size = ci->i_layout.stripe_unit; /* block_offset = object_offset % block_size */ tmp = dl.object_offset; @@ -212,7 +212,7 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) snprintf(dl.object_name, sizeof(dl.object_name), "%llx.%08llx", ceph_ino(inode), dl.object_no); - oloc.pool = ceph_file_layout_pg_pool(ci->i_layout); + oloc.pool = ci->i_layout.pool_id; ceph_oid_printf(&oid, "%s", dl.object_name); r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index e7d38aac7109..75ecf967d0b7 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -275,8 +275,8 @@ struct ceph_mds_request { struct ceph_pool_perm { struct rb_node node; - u32 pool; int perm; + s64 pool; }; /* diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 4870b29df224..5377c9c7a0c5 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -72,7 +72,7 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + s64 pool = ci->i_layout.pool_id; const char *pool_name; char buf[128]; @@ -82,10 +82,9 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, if (pool_name) { size_t len = strlen(pool_name); ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); + "stripe_unit=%u stripe_count=%u object_size=%u pool=", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size); if (!size) { ret += len; } else if (ret + len > size) { @@ -97,11 +96,9 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, } } else { ret = snprintf(buf, sizeof(buf), - "stripe_unit=%lld stripe_count=%lld object_size=%lld pool=%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout), - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout), - (unsigned long long)ceph_file_layout_object_size(ci->i_layout), - (unsigned long long)pool); + "stripe_unit=%u stripe_count=%u object_size=%u pool=%lld", + ci->i_layout.stripe_unit, ci->i_layout.stripe_count, + ci->i_layout.object_size, (unsigned long long)pool); if (size) { if (ret <= size) memcpy(val, buf, ret); @@ -116,22 +113,19 @@ static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, static size_t ceph_vxattrcb_layout_stripe_unit(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_su(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_unit); } static size_t ceph_vxattrcb_layout_stripe_count(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_stripe_count(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.stripe_count); } static size_t ceph_vxattrcb_layout_object_size(struct ceph_inode_info *ci, char *val, size_t size) { - return snprintf(val, size, "%lld", - (unsigned long long)ceph_file_layout_object_size(ci->i_layout)); + return snprintf(val, size, "%u", ci->i_layout.object_size); } static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, @@ -140,7 +134,7 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; - s64 pool = ceph_file_layout_pg_pool(ci->i_layout); + s64 pool = ci->i_layout.pool_id; const char *pool_name; down_read(&osdc->lock); diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index dfce616002ad..e5a5fb9ca3f5 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -34,9 +34,9 @@ #define CEPH_MAX_MON 31 /* - * ceph_file_layout - describe data layout for a file/inode + * legacy ceph_file_layoute */ -struct ceph_file_layout { +struct ceph_file_layout_legacy { /* file -> object mapping */ __le32 fl_stripe_unit; /* stripe unit, in bytes. must be multiple of page size. */ @@ -53,33 +53,25 @@ struct ceph_file_layout { __le32 fl_pg_pool; /* namespace, crush ruleset, rep level */ } __attribute__ ((packed)); -#define ceph_file_layout_su(l) ((__s32)le32_to_cpu((l).fl_stripe_unit)) -#define ceph_file_layout_stripe_count(l) \ - ((__s32)le32_to_cpu((l).fl_stripe_count)) -#define ceph_file_layout_object_size(l) ((__s32)le32_to_cpu((l).fl_object_size)) -#define ceph_file_layout_cas_hash(l) ((__s32)le32_to_cpu((l).fl_cas_hash)) -#define ceph_file_layout_object_su(l) \ - ((__s32)le32_to_cpu((l).fl_object_stripe_unit)) -#define ceph_file_layout_pg_pool(l) \ - ((__s32)le32_to_cpu((l).fl_pg_pool)) - -static inline unsigned ceph_file_layout_stripe_width(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_stripe_unit) * - le32_to_cpu(l->fl_stripe_count); -} - -/* "period" == bytes before i start on a new set of objects */ -static inline unsigned ceph_file_layout_period(struct ceph_file_layout *l) -{ - return le32_to_cpu(l->fl_object_size) * - le32_to_cpu(l->fl_stripe_count); -} +/* + * ceph_file_layout - describe data layout for a file/inode + */ +struct ceph_file_layout { + /* file -> object mapping */ + u32 stripe_unit; /* stripe unit, in bytes */ + u32 stripe_count; /* over this many objects */ + u32 object_size; /* until objects are this big */ + s64 pool_id; /* rados pool id */ +}; + +extern int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); +extern void ceph_file_layout_from_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy); +extern void ceph_file_layout_to_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy); #define CEPH_MIN_STRIPE_UNIT 65536 -int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); - struct ceph_dir_layout { __u8 dl_dir_hash; /* see ceph_hash.h for ids */ __u8 dl_unused1; @@ -399,7 +391,7 @@ union ceph_mds_request_args { __le32 flags; } __attribute__ ((packed)) setxattr; struct { - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; } __attribute__ ((packed)) setlayout; struct { __u8 rule; /* currently fcntl or flock */ @@ -478,7 +470,7 @@ struct ceph_mds_reply_inode { __le64 version; /* inode version */ __le64 xattr_version; /* version for xattr blob */ struct ceph_mds_reply_cap cap; /* caps issued for this inode */ - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; struct ceph_timespec ctime, mtime, atime; __le32 time_warp_seq; __le64 size, max_size, truncate_size; @@ -673,7 +665,7 @@ struct ceph_mds_caps { __le64 size, max_size, truncate_size; __le32 truncate_seq; struct ceph_timespec mtime, atime, ctime; - struct ceph_file_layout layout; + struct ceph_file_layout_legacy layout; __le32 time_warp_seq; } __attribute__ ((packed)); diff --git a/net/ceph/ceph_fs.c b/net/ceph/ceph_fs.c index 41466ccb972a..7d54e944de5e 100644 --- a/net/ceph/ceph_fs.c +++ b/net/ceph/ceph_fs.c @@ -9,9 +9,9 @@ */ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) { - __u32 su = le32_to_cpu(layout->fl_stripe_unit); - __u32 sc = le32_to_cpu(layout->fl_stripe_count); - __u32 os = le32_to_cpu(layout->fl_object_size); + __u32 su = layout->stripe_unit; + __u32 sc = layout->stripe_count; + __u32 os = layout->object_size; /* stripe unit, object size must be non-zero, 64k increment */ if (!su || (su & (CEPH_MIN_STRIPE_UNIT-1))) @@ -27,6 +27,30 @@ int ceph_file_layout_is_valid(const struct ceph_file_layout *layout) return 1; } +void ceph_file_layout_from_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy) +{ + fl->stripe_unit = le32_to_cpu(legacy->fl_stripe_unit); + fl->stripe_count = le32_to_cpu(legacy->fl_stripe_count); + fl->object_size = le32_to_cpu(legacy->fl_object_size); + fl->pool_id = le32_to_cpu(legacy->fl_pg_pool); + if (fl->pool_id == 0) + fl->pool_id = -1; +} +EXPORT_SYMBOL(ceph_file_layout_from_legacy); + +void ceph_file_layout_to_legacy(struct ceph_file_layout *fl, + struct ceph_file_layout_legacy *legacy) +{ + legacy->fl_stripe_unit = cpu_to_le32(fl->stripe_unit); + legacy->fl_stripe_count = cpu_to_le32(fl->stripe_count); + legacy->fl_object_size = cpu_to_le32(fl->object_size); + if (fl->pool_id >= 0) + legacy->fl_pg_pool = cpu_to_le32(fl->pool_id); + else + legacy->fl_pg_pool = 0; +} +EXPORT_SYMBOL(ceph_file_layout_to_legacy); int ceph_flags_to_mode(int flags) { diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 89469592076c..23efcac80072 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -932,7 +932,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, if (opcode == CEPH_OSD_OP_CREATE || opcode == CEPH_OSD_OP_DELETE) { osd_req_op_init(req, which, opcode, 0); } else { - u32 object_size = le32_to_cpu(layout->fl_object_size); + u32 object_size = layout->object_size; u32 object_base = off - objoff; if (!(truncate_seq == 1 && truncate_size == -1ULL)) { if (truncate_size <= object_base) { @@ -948,7 +948,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, } req->r_flags = flags; - req->r_base_oloc.pool = ceph_file_layout_pg_pool(*layout); + req->r_base_oloc.pool = layout->pool_id; ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum); req->r_snapid = vino.snap; diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 7e480bf75bcf..5947b2e9fb8e 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1770,9 +1770,9 @@ int ceph_calc_file_object_mapping(struct ceph_file_layout *layout, u64 *ono, u64 *oxoff, u64 *oxlen) { - u32 osize = le32_to_cpu(layout->fl_object_size); - u32 su = le32_to_cpu(layout->fl_stripe_unit); - u32 sc = le32_to_cpu(layout->fl_stripe_count); + u32 osize = layout->object_size; + u32 su = layout->stripe_unit; + u32 sc = layout->stripe_count; u32 bl, stripeno, stripepos, objsetno; u32 su_per_object; u64 t, su_offset; -- cgit v1.2.3 From 30c156d9951e0aa88202707d80c583b0a09d3167 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Sun, 14 Feb 2016 11:24:31 +0800 Subject: libceph: rados pool namespace support Add pool namesapce pointer to struct ceph_file_layout and struct ceph_object_locator. Pool namespace is used by when mapping object to PG, it's also used when composing OSD request. The namespace pointer in struct ceph_file_layout is RCU protected. So libceph can read namespace without taking lock. Signed-off-by: Yan, Zheng [idryomov@gmail.com: ceph_oloc_destroy(), misc minor changes] Signed-off-by: Ilya Dryomov --- drivers/block/rbd.c | 1 + fs/ceph/inode.c | 3 +++ include/linux/ceph/ceph_fs.h | 2 ++ include/linux/ceph/osdmap.h | 10 ++++----- net/ceph/debugfs.c | 12 ++++++++-- net/ceph/osd_client.c | 32 +++++++++++++++++++++------ net/ceph/osdmap.c | 52 +++++++++++++++++++++++++++++++++++++++----- 7 files changed, 92 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c index cc272ed46cfd..58fd02d4e534 100644 --- a/drivers/block/rbd.c +++ b/drivers/block/rbd.c @@ -3999,6 +3999,7 @@ static struct rbd_device *rbd_dev_create(struct rbd_client *rbdc, rbd_dev->layout.stripe_count = 1; rbd_dev->layout.object_size = 1 << RBD_MAX_OBJ_ORDER; rbd_dev->layout.pool_id = spec->pool_id; + RCU_INIT_POINTER(rbd_dev->layout.pool_ns, NULL); /* * If this is a mapping rbd_dev (as opposed to a parent one), diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 6c5903e0a976..d035e0ab6029 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -446,6 +446,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_symlink = NULL; memset(&ci->i_dir_layout, 0, sizeof(ci->i_dir_layout)); + RCU_INIT_POINTER(ci->i_layout.pool_ns, NULL); ci->i_pool_ns_len = 0; ci->i_fragtree = RB_ROOT; @@ -570,6 +571,8 @@ void ceph_destroy_inode(struct inode *inode) if (ci->i_xattrs.prealloc_blob) ceph_buffer_put(ci->i_xattrs.prealloc_blob); + ceph_put_string(ci->i_layout.pool_ns); + call_rcu(&inode->i_rcu, ceph_i_callback); } diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index e5a5fb9ca3f5..08b8fd72261e 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -53,6 +53,7 @@ struct ceph_file_layout_legacy { __le32 fl_pg_pool; /* namespace, crush ruleset, rep level */ } __attribute__ ((packed)); +struct ceph_string; /* * ceph_file_layout - describe data layout for a file/inode */ @@ -62,6 +63,7 @@ struct ceph_file_layout { u32 stripe_count; /* over this many objects */ u32 object_size; /* until objects are this big */ s64 pool_id; /* rados pool id */ + struct ceph_string __rcu *pool_ns; /* rados pool namespace */ }; extern int ceph_file_layout_is_valid(const struct ceph_file_layout *layout); diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index 21d7f048959f..9a9041784dcf 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -63,11 +63,13 @@ static inline bool ceph_can_shift_osds(struct ceph_pg_pool_info *pool) struct ceph_object_locator { s64 pool; + struct ceph_string *pool_ns; }; static inline void ceph_oloc_init(struct ceph_object_locator *oloc) { oloc->pool = -1; + oloc->pool_ns = NULL; } static inline bool ceph_oloc_empty(const struct ceph_object_locator *oloc) @@ -75,11 +77,9 @@ static inline bool ceph_oloc_empty(const struct ceph_object_locator *oloc) return oloc->pool == -1; } -static inline void ceph_oloc_copy(struct ceph_object_locator *dest, - const struct ceph_object_locator *src) -{ - dest->pool = src->pool; -} +void ceph_oloc_copy(struct ceph_object_locator *dest, + const struct ceph_object_locator *src); +void ceph_oloc_destroy(struct ceph_object_locator *oloc); /* * Maximum supported by kernel client object name length diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index e77b04ca7802..c62b2b029a6e 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -156,8 +156,16 @@ static void dump_target(struct seq_file *s, struct ceph_osd_request_target *t) seq_printf(s, "]/%d\t[", t->up.primary); for (i = 0; i < t->acting.size; i++) seq_printf(s, "%s%d", (!i ? "" : ","), t->acting.osds[i]); - seq_printf(s, "]/%d\t%*pE\t0x%x", t->acting.primary, - t->target_oid.name_len, t->target_oid.name, t->flags); + seq_printf(s, "]/%d\t", t->acting.primary); + if (t->target_oloc.pool_ns) { + seq_printf(s, "%*pE/%*pE\t0x%x", + (int)t->target_oloc.pool_ns->len, + t->target_oloc.pool_ns->str, + t->target_oid.name_len, t->target_oid.name, t->flags); + } else { + seq_printf(s, "%*pE\t0x%x", t->target_oid.name_len, + t->target_oid.name, t->flags); + } if (t->paused) seq_puts(s, "\tP"); } diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 23efcac80072..b68cc15e9cdc 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -387,7 +387,9 @@ static void target_copy(struct ceph_osd_request_target *dest, static void target_destroy(struct ceph_osd_request_target *t) { ceph_oid_destroy(&t->base_oid); + ceph_oloc_destroy(&t->base_oloc); ceph_oid_destroy(&t->target_oid); + ceph_oloc_destroy(&t->target_oloc); } /* @@ -533,6 +535,11 @@ struct ceph_osd_request *ceph_osdc_alloc_request(struct ceph_osd_client *osdc, } EXPORT_SYMBOL(ceph_osdc_alloc_request); +static int ceph_oloc_encoding_size(struct ceph_object_locator *oloc) +{ + return 8 + 4 + 4 + 4 + (oloc->pool_ns ? oloc->pool_ns->len : 0); +} + int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp) { struct ceph_osd_client *osdc = req->r_osdc; @@ -540,11 +547,13 @@ int ceph_osdc_alloc_messages(struct ceph_osd_request *req, gfp_t gfp) int msg_size; WARN_ON(ceph_oid_empty(&req->r_base_oid)); + WARN_ON(ceph_oloc_empty(&req->r_base_oloc)); /* create request message */ msg_size = 4 + 4 + 4; /* client_inc, osdmap_epoch, flags */ msg_size += 4 + 4 + 4 + 8; /* mtime, reassert_version */ - msg_size += 2 + 4 + 8 + 4 + 4; /* oloc */ + msg_size += CEPH_ENCODING_START_BLK_LEN + + ceph_oloc_encoding_size(&req->r_base_oloc); /* oloc */ msg_size += 1 + 8 + 4 + 4; /* pgid */ msg_size += 4 + req->r_base_oid.name_len; /* oid */ msg_size += 2 + req->r_num_ops * sizeof(struct ceph_osd_op); @@ -949,6 +958,7 @@ struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *osdc, req->r_flags = flags; req->r_base_oloc.pool = layout->pool_id; + req->r_base_oloc.pool_ns = ceph_try_get_string(layout->pool_ns); ceph_oid_printf(&req->r_base_oid, "%llx.%08llx", vino.ino, objnum); req->r_snapid = vino.snap; @@ -1489,12 +1499,16 @@ static void encode_request(struct ceph_osd_request *req, struct ceph_msg *msg) p += sizeof(req->r_replay_version); /* oloc */ - ceph_encode_8(&p, 4); - ceph_encode_8(&p, 4); - ceph_encode_32(&p, 8 + 4 + 4); + ceph_start_encoding(&p, 5, 4, + ceph_oloc_encoding_size(&req->r_t.target_oloc)); ceph_encode_64(&p, req->r_t.target_oloc.pool); ceph_encode_32(&p, -1); /* preferred */ ceph_encode_32(&p, 0); /* key len */ + if (req->r_t.target_oloc.pool_ns) + ceph_encode_string(&p, end, req->r_t.target_oloc.pool_ns->str, + req->r_t.target_oloc.pool_ns->len); + else + ceph_encode_32(&p, 0); /* pgid */ ceph_encode_8(&p, 1); @@ -2596,8 +2610,8 @@ static int ceph_oloc_decode(void **p, void *end, if (struct_v >= 5) { len = ceph_decode_32(p); if (len > 0) { - pr_warn("ceph_object_locator::nspace is set\n"); - goto e_inval; + ceph_decode_need(p, end, len, e_inval); + *p += len; } } @@ -2835,7 +2849,11 @@ static void handle_reply(struct ceph_osd *osd, struct ceph_msg *msg) unlink_request(osd, req); mutex_unlock(&osd->lock); - ceph_oloc_copy(&req->r_t.target_oloc, &m.redirect.oloc); + /* + * Not ceph_oloc_copy() - changing pool_ns is not + * supported. + */ + req->r_t.target_oloc.pool = m.redirect.oloc.pool; req->r_flags |= CEPH_OSD_FLAG_REDIRECTED; req->r_tid = 0; __submit_request(req, false); diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 5947b2e9fb8e..d2436880b305 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -1510,6 +1510,24 @@ bad: return ERR_PTR(err); } +void ceph_oloc_copy(struct ceph_object_locator *dest, + const struct ceph_object_locator *src) +{ + WARN_ON(!ceph_oloc_empty(dest)); + WARN_ON(dest->pool_ns); /* empty() only covers ->pool */ + + dest->pool = src->pool; + if (src->pool_ns) + dest->pool_ns = ceph_get_string(src->pool_ns); +} +EXPORT_SYMBOL(ceph_oloc_copy); + +void ceph_oloc_destroy(struct ceph_object_locator *oloc) +{ + ceph_put_string(oloc->pool_ns); +} +EXPORT_SYMBOL(ceph_oloc_destroy); + void ceph_oid_copy(struct ceph_object_id *dest, const struct ceph_object_id *src) { @@ -1844,12 +1862,34 @@ int ceph_object_locator_to_pg(struct ceph_osdmap *osdmap, if (!pi) return -ENOENT; - raw_pgid->pool = oloc->pool; - raw_pgid->seed = ceph_str_hash(pi->object_hash, oid->name, - oid->name_len); - - dout("%s %s -> raw_pgid %llu.%x\n", __func__, oid->name, - raw_pgid->pool, raw_pgid->seed); + if (!oloc->pool_ns) { + raw_pgid->pool = oloc->pool; + raw_pgid->seed = ceph_str_hash(pi->object_hash, oid->name, + oid->name_len); + dout("%s %s -> raw_pgid %llu.%x\n", __func__, oid->name, + raw_pgid->pool, raw_pgid->seed); + } else { + char stack_buf[256]; + char *buf = stack_buf; + int nsl = oloc->pool_ns->len; + size_t total = nsl + 1 + oid->name_len; + + if (total > sizeof(stack_buf)) { + buf = kmalloc(total, GFP_NOIO); + if (!buf) + return -ENOMEM; + } + memcpy(buf, oloc->pool_ns->str, nsl); + buf[nsl] = '\037'; + memcpy(buf + nsl + 1, oid->name, oid->name_len); + raw_pgid->pool = oloc->pool; + raw_pgid->seed = ceph_str_hash(pi->object_hash, buf, total); + if (buf != stack_buf) + kfree(buf); + dout("%s %s ns %.*s -> raw_pgid %llu.%x\n", __func__, + oid->name, nsl, oloc->pool_ns->str, + raw_pgid->pool, raw_pgid->seed); + } return 0; } EXPORT_SYMBOL(ceph_object_locator_to_pg); -- cgit v1.2.3 From 779fe0fb8e1883d5c479ac6bd85fbd237deed1f7 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 7 Mar 2016 09:35:06 +0800 Subject: ceph: rados pool namespace support This patch adds codes that decode pool namespace information in cap message and request reply. Pool namespace is saved in i_layout, it will be passed to libceph when doing read/write. Signed-off-by: Yan, Zheng --- fs/ceph/addr.c | 67 +++++++++++++++++++++++++++++++++++---------- fs/ceph/caps.c | 46 +++++++++++++++++-------------- fs/ceph/inode.c | 20 +++++++++++--- fs/ceph/ioctl.c | 3 ++ fs/ceph/mds_client.c | 19 +++++-------- fs/ceph/mds_client.h | 3 ++ fs/ceph/super.h | 1 - fs/ceph/xattr.c | 77 +++++++++++++++++++++++++++++++++++----------------- 8 files changed, 159 insertions(+), 77 deletions(-) (limited to 'fs') diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 3f8efd866fec..d5b6f959a3c3 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -1730,7 +1730,8 @@ enum { POOL_WRITE = 2, }; -static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) +static int __ceph_pool_perm_get(struct ceph_inode_info *ci, + s64 pool, struct ceph_string *pool_ns) { struct ceph_fs_client *fsc = ceph_inode_to_client(&ci->vfs_inode); struct ceph_mds_client *mdsc = fsc->mdsc; @@ -1738,6 +1739,7 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) struct rb_node **p, *parent; struct ceph_pool_perm *perm; struct page **pages; + size_t pool_ns_len; int err = 0, err2 = 0, have = 0; down_read(&mdsc->pool_perm_rwsem); @@ -1749,17 +1751,31 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) else if (pool > perm->pool) p = &(*p)->rb_right; else { - have = perm->perm; - break; + int ret = ceph_compare_string(pool_ns, + perm->pool_ns, + perm->pool_ns_len); + if (ret < 0) + p = &(*p)->rb_left; + else if (ret > 0) + p = &(*p)->rb_right; + else { + have = perm->perm; + break; + } } } up_read(&mdsc->pool_perm_rwsem); if (*p) goto out; - dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool); + if (pool_ns) + dout("__ceph_pool_perm_get pool %lld ns %.*s no perm cached\n", + pool, (int)pool_ns->len, pool_ns->str); + else + dout("__ceph_pool_perm_get pool %lld no perm cached\n", pool); down_write(&mdsc->pool_perm_rwsem); + p = &mdsc->pool_perm_tree.rb_node; parent = NULL; while (*p) { parent = *p; @@ -1769,8 +1785,17 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) else if (pool > perm->pool) p = &(*p)->rb_right; else { - have = perm->perm; - break; + int ret = ceph_compare_string(pool_ns, + perm->pool_ns, + perm->pool_ns_len); + if (ret < 0) + p = &(*p)->rb_left; + else if (ret > 0) + p = &(*p)->rb_right; + else { + have = perm->perm; + break; + } } } if (*p) { @@ -1788,6 +1813,8 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) rd_req->r_flags = CEPH_OSD_FLAG_READ; osd_req_op_init(rd_req, 0, CEPH_OSD_OP_STAT, 0); rd_req->r_base_oloc.pool = pool; + if (pool_ns) + rd_req->r_base_oloc.pool_ns = ceph_get_string(pool_ns); ceph_oid_printf(&rd_req->r_base_oid, "%llx.00000000", ci->i_vino.ino); err = ceph_osdc_alloc_messages(rd_req, GFP_NOFS); @@ -1841,7 +1868,8 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) goto out_unlock; } - perm = kmalloc(sizeof(*perm), GFP_NOFS); + pool_ns_len = pool_ns ? pool_ns->len : 0; + perm = kmalloc(sizeof(*perm) + pool_ns_len + 1, GFP_NOFS); if (!perm) { err = -ENOMEM; goto out_unlock; @@ -1849,6 +1877,11 @@ static int __ceph_pool_perm_get(struct ceph_inode_info *ci, s64 pool) perm->pool = pool; perm->perm = have; + perm->pool_ns_len = pool_ns_len; + if (pool_ns_len > 0) + memcpy(perm->pool_ns, pool_ns->str, pool_ns_len); + perm->pool_ns[pool_ns_len] = 0; + rb_link_node(&perm->node, parent, p); rb_insert_color(&perm->node, &mdsc->pool_perm_tree); err = 0; @@ -1860,19 +1893,20 @@ out_unlock: out: if (!err) err = have; - dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err); + if (pool_ns) + dout("__ceph_pool_perm_get pool %lld ns %.*s result = %d\n", + pool, (int)pool_ns->len, pool_ns->str, err); + else + dout("__ceph_pool_perm_get pool %lld result = %d\n", pool, err); return err; } int ceph_pool_perm_check(struct ceph_inode_info *ci, int need) { s64 pool; + struct ceph_string *pool_ns; int ret, flags; - /* does not support pool namespace yet */ - if (ci->i_pool_ns_len) - return -EIO; - if (ceph_test_mount_opt(ceph_inode_to_client(&ci->vfs_inode), NOPOOLPERM)) return 0; @@ -1896,7 +1930,9 @@ check: return 0; } - ret = __ceph_pool_perm_get(ci, pool); + pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); + ret = __ceph_pool_perm_get(ci, pool, pool_ns); + ceph_put_string(pool_ns); if (ret < 0) return ret; @@ -1907,8 +1943,9 @@ check: flags |= CEPH_I_POOL_WR; spin_lock(&ci->i_ceph_lock); - if (pool == ci->i_layout.pool_id) { - ci->i_ceph_flags = flags; + if (pool == ci->i_layout.pool_id && + pool_ns == rcu_dereference_raw(ci->i_layout.pool_ns)) { + ci->i_ceph_flags |= flags; } else { pool = ci->i_layout.pool_id; flags = ci->i_ceph_flags; diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index f24722dce167..0a9406a8a794 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2779,12 +2779,11 @@ static void invalidate_aliases(struct inode *inode) */ static void handle_cap_grant(struct ceph_mds_client *mdsc, struct inode *inode, struct ceph_mds_caps *grant, - u64 inline_version, - void *inline_data, int inline_len, + struct ceph_string **pns, u64 inline_version, + void *inline_data, u32 inline_len, struct ceph_buffer *xattr_buf, struct ceph_mds_session *session, - struct ceph_cap *cap, int issued, - u32 pool_ns_len) + struct ceph_cap *cap, int issued) __releases(ci->i_ceph_lock) __releases(mdsc->snap_rwsem) { @@ -2896,11 +2895,18 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, if (newcaps & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR)) { /* file layout may have changed */ s64 old_pool = ci->i_layout.pool_id; + struct ceph_string *old_ns; + ceph_file_layout_from_legacy(&ci->i_layout, &grant->layout); - ci->i_pool_ns_len = pool_ns_len; - if (ci->i_layout.pool_id != old_pool) + old_ns = rcu_dereference_protected(ci->i_layout.pool_ns, + lockdep_is_held(&ci->i_ceph_lock)); + rcu_assign_pointer(ci->i_layout.pool_ns, *pns); + + if (ci->i_layout.pool_id != old_pool || *pns != old_ns) ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; + *pns = old_ns; + /* size/truncate_seq? */ queue_trunc = ceph_fill_file_size(inode, issued, le32_to_cpu(grant->truncate_seq), @@ -3423,20 +3429,18 @@ void ceph_handle_caps(struct ceph_mds_session *session, struct ceph_cap *cap; struct ceph_mds_caps *h; struct ceph_mds_cap_peer *peer = NULL; - struct ceph_snap_realm *realm; + struct ceph_snap_realm *realm = NULL; + struct ceph_string *pool_ns = NULL; int mds = session->s_mds; int op, issued; u32 seq, mseq; struct ceph_vino vino; - u64 cap_id; - u64 size, max_size; u64 tid; u64 inline_version = 0; void *inline_data = NULL; u32 inline_len = 0; void *snaptrace; size_t snaptrace_len; - u32 pool_ns_len = 0; void *p, *end; dout("handle_caps from mds%d\n", mds); @@ -3450,11 +3454,8 @@ void ceph_handle_caps(struct ceph_mds_session *session, op = le32_to_cpu(h->op); vino.ino = le64_to_cpu(h->ino); vino.snap = CEPH_NOSNAP; - cap_id = le64_to_cpu(h->cap_id); seq = le32_to_cpu(h->seq); mseq = le32_to_cpu(h->migrate_seq); - size = le64_to_cpu(h->size); - max_size = le64_to_cpu(h->max_size); snaptrace = h + 1; snaptrace_len = le32_to_cpu(h->snap_trace_len); @@ -3493,6 +3494,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, u64 flush_tid; u32 caller_uid, caller_gid; u32 osd_epoch_barrier; + u32 pool_ns_len; /* version >= 5 */ ceph_decode_32_safe(&p, end, osd_epoch_barrier, bad); /* version >= 6 */ @@ -3502,6 +3504,11 @@ void ceph_handle_caps(struct ceph_mds_session *session, ceph_decode_32_safe(&p, end, caller_gid, bad); /* version >= 8 */ ceph_decode_32_safe(&p, end, pool_ns_len, bad); + if (pool_ns_len > 0) { + ceph_decode_need(&p, end, pool_ns_len, bad); + pool_ns = ceph_find_or_create_string(p, pool_ns_len); + p += pool_ns_len; + } } /* lookup ino */ @@ -3522,7 +3529,7 @@ void ceph_handle_caps(struct ceph_mds_session *session, cap = ceph_get_cap(mdsc, NULL); cap->cap_ino = vino.ino; cap->queue_release = 1; - cap->cap_id = cap_id; + cap->cap_id = le64_to_cpu(h->cap_id); cap->mseq = mseq; cap->seq = seq; spin_lock(&session->s_cap_lock); @@ -3557,10 +3564,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, } handle_cap_import(mdsc, inode, h, peer, session, &cap, &issued); - handle_cap_grant(mdsc, inode, h, + handle_cap_grant(mdsc, inode, h, &pool_ns, inline_version, inline_data, inline_len, - msg->middle, session, cap, issued, - pool_ns_len); + msg->middle, session, cap, issued); if (realm) ceph_put_snap_realm(mdsc, realm); goto done_unlocked; @@ -3582,10 +3588,9 @@ void ceph_handle_caps(struct ceph_mds_session *session, case CEPH_CAP_OP_GRANT: __ceph_caps_issued(ci, &issued); issued |= __ceph_caps_dirty(ci); - handle_cap_grant(mdsc, inode, h, + handle_cap_grant(mdsc, inode, h, &pool_ns, inline_version, inline_data, inline_len, - msg->middle, session, cap, issued, - pool_ns_len); + msg->middle, session, cap, issued); goto done_unlocked; case CEPH_CAP_OP_FLUSH_ACK: @@ -3616,6 +3621,7 @@ done: mutex_unlock(&session->s_mutex); done_unlocked: iput(inode); + ceph_put_string(pool_ns); return; bad: diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index d035e0ab6029..dc032566ed71 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -447,7 +447,6 @@ struct inode *ceph_alloc_inode(struct super_block *sb) memset(&ci->i_dir_layout, 0, sizeof(ci->i_dir_layout)); RCU_INIT_POINTER(ci->i_layout.pool_ns, NULL); - ci->i_pool_ns_len = 0; ci->i_fragtree = RB_ROOT; mutex_init(&ci->i_fragtree_mutex); @@ -571,7 +570,7 @@ void ceph_destroy_inode(struct inode *inode) if (ci->i_xattrs.prealloc_blob) ceph_buffer_put(ci->i_xattrs.prealloc_blob); - ceph_put_string(ci->i_layout.pool_ns); + ceph_put_string(rcu_dereference_raw(ci->i_layout.pool_ns)); call_rcu(&inode->i_rcu, ceph_i_callback); } @@ -736,6 +735,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page, int issued = 0, implemented, new_issued; struct timespec mtime, atime, ctime; struct ceph_buffer *xattr_blob = NULL; + struct ceph_string *pool_ns = NULL; struct ceph_cap *new_cap = NULL; int err = 0; bool wake = false; @@ -763,6 +763,10 @@ static int fill_inode(struct inode *inode, struct page *locked_page, iinfo->xattr_len); } + if (iinfo->pool_ns_len > 0) + pool_ns = ceph_find_or_create_string(iinfo->pool_ns_data, + iinfo->pool_ns_len); + spin_lock(&ci->i_ceph_lock); /* @@ -818,11 +822,18 @@ static int fill_inode(struct inode *inode, struct page *locked_page, if (new_version || (new_issued & (CEPH_CAP_ANY_FILE_RD | CEPH_CAP_ANY_FILE_WR))) { s64 old_pool = ci->i_layout.pool_id; + struct ceph_string *old_ns; + ceph_file_layout_from_legacy(&ci->i_layout, &info->layout); - ci->i_pool_ns_len = iinfo->pool_ns_len; - if (ci->i_layout.pool_id != old_pool) + old_ns = rcu_dereference_protected(ci->i_layout.pool_ns, + lockdep_is_held(&ci->i_ceph_lock)); + rcu_assign_pointer(ci->i_layout.pool_ns, pool_ns); + + if (ci->i_layout.pool_id != old_pool || pool_ns != old_ns) ci->i_ceph_flags &= ~CEPH_I_POOL_PERM; + pool_ns = old_ns; + queue_trunc = ceph_fill_file_size(inode, issued, le32_to_cpu(info->truncate_seq), le64_to_cpu(info->truncate_size), @@ -989,6 +1000,7 @@ out: ceph_put_cap(mdsc, new_cap); if (xattr_blob) ceph_buffer_put(xattr_blob); + ceph_put_string(pool_ns); return err; } diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 843dd31a02cd..6a30101b55ef 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -213,9 +213,12 @@ static long ceph_ioctl_get_dataloc(struct file *file, void __user *arg) ceph_ino(inode), dl.object_no); oloc.pool = ci->i_layout.pool_id; + oloc.pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); ceph_oid_printf(&oid, "%s", dl.object_name); r = ceph_object_locator_to_pg(osdc->osdmap, &oid, &oloc, &pgid); + + ceph_oloc_destroy(&oloc); if (r < 0) { up_read(&osdc->lock); return r; diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 2103b823bec0..46641bbc8056 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -100,12 +100,15 @@ static int parse_reply_info_in(void **p, void *end, } else info->inline_version = CEPH_INLINE_NONE; + info->pool_ns_len = 0; + info->pool_ns_data = NULL; if (features & CEPH_FEATURE_FS_FILE_LAYOUT_V2) { ceph_decode_32_safe(p, end, info->pool_ns_len, bad); - ceph_decode_need(p, end, info->pool_ns_len, bad); - *p += info->pool_ns_len; - } else { - info->pool_ns_len = 0; + if (info->pool_ns_len > 0) { + ceph_decode_need(p, end, info->pool_ns_len, bad); + info->pool_ns_data = *p; + *p += info->pool_ns_len; + } } return 0; @@ -2292,14 +2295,6 @@ int ceph_mdsc_do_request(struct ceph_mds_client *mdsc, ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir), CEPH_CAP_PIN); - /* deny access to directories with pool_ns layouts */ - if (req->r_inode && S_ISDIR(req->r_inode->i_mode) && - ceph_inode(req->r_inode)->i_pool_ns_len) - return -EIO; - if (req->r_locked_dir && - ceph_inode(req->r_locked_dir)->i_pool_ns_len) - return -EIO; - /* issue */ mutex_lock(&mdsc->mutex); __register_request(mdsc, req, dir); diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 75ecf967d0b7..2ce8e9f9bfc9 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -45,6 +45,7 @@ struct ceph_mds_reply_info_in { u32 inline_len; char *inline_data; u32 pool_ns_len; + char *pool_ns_data; }; struct ceph_mds_reply_dir_entry { @@ -277,6 +278,8 @@ struct ceph_pool_perm { struct rb_node node; int perm; s64 pool; + size_t pool_ns_len; + char pool_ns[]; }; /* diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 0168b49fb6ad..7ceab18c8ee2 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -287,7 +287,6 @@ struct ceph_inode_info { struct ceph_dir_layout i_dir_layout; struct ceph_file_layout i_layout; - size_t i_pool_ns_len; char *i_symlink; /* for dirs */ diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index 5377c9c7a0c5..adc231892b0d 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -57,56 +57,69 @@ struct ceph_vxattr { static bool ceph_vxattrcb_layout_exists(struct ceph_inode_info *ci) { - size_t s; - char *p = (char *)&ci->i_layout; - - for (s = 0; s < sizeof(ci->i_layout); s++, p++) - if (*p) - return true; - return false; + struct ceph_file_layout *fl = &ci->i_layout; + return (fl->stripe_unit > 0 || fl->stripe_count > 0 || + fl->object_size > 0 || fl->pool_id >= 0 || + rcu_dereference_raw(fl->pool_ns) != NULL); } static size_t ceph_vxattrcb_layout(struct ceph_inode_info *ci, char *val, size_t size) { - int ret; struct ceph_fs_client *fsc = ceph_sb_to_client(ci->vfs_inode.i_sb); struct ceph_osd_client *osdc = &fsc->client->osdc; + struct ceph_string *pool_ns; s64 pool = ci->i_layout.pool_id; const char *pool_name; + const char *ns_field = " pool_namespace="; char buf[128]; + size_t len, total_len = 0; + int ret; + + pool_ns = ceph_try_get_string(ci->i_layout.pool_ns); dout("ceph_vxattrcb_layout %p\n", &ci->vfs_inode); down_read(&osdc->lock); pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, pool); if (pool_name) { - size_t len = strlen(pool_name); - ret = snprintf(buf, sizeof(buf), + len = snprintf(buf, sizeof(buf), "stripe_unit=%u stripe_count=%u object_size=%u pool=", ci->i_layout.stripe_unit, ci->i_layout.stripe_count, ci->i_layout.object_size); - if (!size) { - ret += len; - } else if (ret + len > size) { - ret = -ERANGE; - } else { - memcpy(val, buf, ret); - memcpy(val + ret, pool_name, len); - ret += len; - } + total_len = len + strlen(pool_name); } else { - ret = snprintf(buf, sizeof(buf), + len = snprintf(buf, sizeof(buf), "stripe_unit=%u stripe_count=%u object_size=%u pool=%lld", ci->i_layout.stripe_unit, ci->i_layout.stripe_count, ci->i_layout.object_size, (unsigned long long)pool); - if (size) { - if (ret <= size) - memcpy(val, buf, ret); - else - ret = -ERANGE; + total_len = len; + } + + if (pool_ns) + total_len += strlen(ns_field) + pool_ns->len; + + if (!size) { + ret = total_len; + } else if (total_len > size) { + ret = -ERANGE; + } else { + memcpy(val, buf, len); + ret = len; + if (pool_name) { + len = strlen(pool_name); + memcpy(val + ret, pool_name, len); + ret += len; + } + if (pool_ns) { + len = strlen(ns_field); + memcpy(val + ret, ns_field, len); + ret += len; + memcpy(val + ret, pool_ns->str, pool_ns->len); + ret += pool_ns->len; } } up_read(&osdc->lock); + ceph_put_string(pool_ns); return ret; } @@ -147,6 +160,18 @@ static size_t ceph_vxattrcb_layout_pool(struct ceph_inode_info *ci, return ret; } +static size_t ceph_vxattrcb_layout_pool_namespace(struct ceph_inode_info *ci, + char *val, size_t size) +{ + int ret = 0; + struct ceph_string *ns = ceph_try_get_string(ci->i_layout.pool_ns); + if (ns) { + ret = snprintf(val, size, "%.*s", (int)ns->len, ns->str); + ceph_put_string(ns); + } + return ret; +} + /* directories */ static size_t ceph_vxattrcb_dir_entries(struct ceph_inode_info *ci, char *val, @@ -235,6 +260,7 @@ static struct ceph_vxattr ceph_dir_vxattrs[] = { XATTR_LAYOUT_FIELD(dir, layout, stripe_count), XATTR_LAYOUT_FIELD(dir, layout, object_size), XATTR_LAYOUT_FIELD(dir, layout, pool), + XATTR_LAYOUT_FIELD(dir, layout, pool_namespace), XATTR_NAME_CEPH(dir, entries), XATTR_NAME_CEPH(dir, files), XATTR_NAME_CEPH(dir, subdirs), @@ -262,6 +288,7 @@ static struct ceph_vxattr ceph_file_vxattrs[] = { XATTR_LAYOUT_FIELD(file, layout, stripe_count), XATTR_LAYOUT_FIELD(file, layout, object_size), XATTR_LAYOUT_FIELD(file, layout, pool), + XATTR_LAYOUT_FIELD(file, layout, pool_namespace), { .name = NULL, 0 } /* Required table terminator */ }; static size_t ceph_file_vxattrs_name_size; /* total size of all names */ -- cgit v1.2.3 From a22bd5ffae2d22c054c832fe0d60976ed9e4a49d Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 26 May 2016 10:30:13 +0800 Subject: ceph: set user pages dirty after direct IO read Signed-off-by: Yan, Zheng --- fs/ceph/file.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index cba5dcf49a65..ac75fa9fd858 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -708,7 +708,7 @@ static void ceph_aio_complete_req(struct ceph_osd_request *req) } } - ceph_put_page_vector(osd_data->pages, num_pages, false); + ceph_put_page_vector(osd_data->pages, num_pages, !aio_req->write); ceph_osdc_put_request(req); if (rc < 0) @@ -964,7 +964,7 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, len = ret; } - ceph_put_page_vector(pages, num_pages, false); + ceph_put_page_vector(pages, num_pages, !write); ceph_osdc_put_request(req); if (ret < 0) -- cgit v1.2.3 From 774a6a118c70f8c11fcfe636032b5016ad71a746 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 6 Jun 2016 16:01:39 +0800 Subject: ceph: reduce i_nr_by_mode array size Track usage count for individual fmode bit. This can reduce the array size by half. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 45 ++++++++++++++++++++++++++++++-------------- fs/ceph/inode.c | 2 +- fs/ceph/ioctl.c | 3 +-- fs/ceph/super.h | 7 ++----- include/linux/ceph/ceph_fs.h | 2 +- 5 files changed, 36 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 0a9406a8a794..a08d245f16f5 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -849,12 +849,14 @@ int __ceph_caps_used(struct ceph_inode_info *ci) */ int __ceph_caps_file_wanted(struct ceph_inode_info *ci) { - int want = 0; - int mode; - for (mode = 0; mode < CEPH_FILE_MODE_NUM; mode++) - if (ci->i_nr_by_mode[mode]) - want |= ceph_caps_for_mode(mode); - return want; + int i, bits = 0; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (ci->i_nr_by_mode[i]) + bits |= 1 << i; + } + if (bits == 0) + return 0; + return ceph_caps_for_mode(bits >> 1); } /* @@ -3682,6 +3684,16 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) dout("flush_dirty_caps done\n"); } +void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode) +{ + int i; + int bits = (fmode << 1) | 1; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (bits & (1 << i)) + ci->i_nr_by_mode[i]++; + } +} + /* * Drop open file reference. If we were the last open file, * we may need to release capabilities to the MDS (or schedule @@ -3689,15 +3701,20 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) */ void ceph_put_fmode(struct ceph_inode_info *ci, int fmode) { - struct inode *inode = &ci->vfs_inode; - int last = 0; - + int i, last = 0; + int bits = (fmode << 1) | 1; spin_lock(&ci->i_ceph_lock); - dout("put_fmode %p fmode %d %d -> %d\n", inode, fmode, - ci->i_nr_by_mode[fmode], ci->i_nr_by_mode[fmode]-1); - BUG_ON(ci->i_nr_by_mode[fmode] == 0); - if (--ci->i_nr_by_mode[fmode] == 0) - last++; + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) { + if (bits & (1 << i)) { + BUG_ON(ci->i_nr_by_mode[i] == 0); + if (--ci->i_nr_by_mode[i] == 0) + last++; + } + } + dout("put_fmode %p fmode %d {%d,%d,%d,%d}\n", + &ci->vfs_inode, fmode, + ci->i_nr_by_mode[0], ci->i_nr_by_mode[1], + ci->i_nr_by_mode[2], ci->i_nr_by_mode[3]); spin_unlock(&ci->i_ceph_lock); if (last && ci->i_vino.snap == CEPH_NOSNAP) diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index dc032566ed71..8ca843371d4b 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -477,7 +477,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) ci->i_head_snapc = NULL; ci->i_snap_caps = 0; - for (i = 0; i < CEPH_FILE_MODE_NUM; i++) + for (i = 0; i < CEPH_FILE_MODE_BITS; i++) ci->i_nr_by_mode[i] = 0; mutex_init(&ci->i_truncate_mutex); diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 6a30101b55ef..7d752d53353a 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -250,9 +250,8 @@ static long ceph_ioctl_lazyio(struct file *file) if ((fi->fmode & CEPH_FILE_MODE_LAZY) == 0) { spin_lock(&ci->i_ceph_lock); - ci->i_nr_by_mode[fi->fmode]--; fi->fmode |= CEPH_FILE_MODE_LAZY; - ci->i_nr_by_mode[fi->fmode]++; + ci->i_nr_by_mode[ffs(CEPH_FILE_MODE_LAZY)]++; spin_unlock(&ci->i_ceph_lock); dout("ioctl_layzio: file %p marked lazy\n", file); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 7ceab18c8ee2..50846e6f6a8c 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -321,7 +321,7 @@ struct ceph_inode_info { dirty|flushing caps */ unsigned i_snap_caps; /* cap bits for snapped files */ - int i_nr_by_mode[CEPH_FILE_MODE_NUM]; /* open file counts */ + int i_nr_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */ struct mutex i_truncate_mutex; u32 i_truncate_seq; /* last truncate to smaller size */ @@ -906,10 +906,7 @@ extern int ceph_get_caps(struct ceph_inode_info *ci, int need, int want, loff_t endoff, int *got, struct page **pinned_page); /* for counting open files by mode */ -static inline void __ceph_get_fmode(struct ceph_inode_info *ci, int mode) -{ - ci->i_nr_by_mode[mode]++; -} +extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode); extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode); /* addr.c */ diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index 08b8fd72261e..6bc7730d22ac 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -525,7 +525,7 @@ struct ceph_filelock { #define CEPH_FILE_MODE_WR 2 #define CEPH_FILE_MODE_RDWR 3 /* RD | WR */ #define CEPH_FILE_MODE_LAZY 4 /* lazy io */ -#define CEPH_FILE_MODE_NUM 8 /* bc these are bit fields.. mostly */ +#define CEPH_FILE_MODE_BITS 4 int ceph_flags_to_mode(int flags); -- cgit v1.2.3 From fc8c3892f30c39f28fdb835f7c8598ac4cf5ed1e Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 14 Jun 2016 11:13:59 +0800 Subject: ceph: fix use-after-free bug in ceph_direct_read_write() ceph_aio_complete() can free the ceph_aio_request struct before the code exits the while loop. Signed-off-by: Yan, Zheng --- fs/ceph/file.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index ac75fa9fd858..033e88753875 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -985,6 +985,8 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, } if (aio_req) { + LIST_HEAD(osd_reqs); + if (aio_req->num_reqs == 0) { kfree(aio_req); return ret; @@ -993,8 +995,9 @@ ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, ceph_get_cap_refs(ci, write ? CEPH_CAP_FILE_WR : CEPH_CAP_FILE_RD); - while (!list_empty(&aio_req->osd_reqs)) { - req = list_first_entry(&aio_req->osd_reqs, + list_splice(&aio_req->osd_reqs, &osd_reqs); + while (!list_empty(&osd_reqs)) { + req = list_first_entry(&osd_reqs, struct ceph_osd_request, r_unsafe_item); list_del_init(&req->r_unsafe_item); -- cgit v1.2.3 From 9a5530c63889ac928a45c4645ab0bc23b4fbfcb8 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 15 Jun 2016 16:29:18 +0800 Subject: ceph: wait unsafe sync writes for evicting inode Otherwise ceph_sync_write_unsafe() may access/modify freed inode. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 50 ++------------------------------------------------ fs/ceph/file.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ fs/ceph/inode.c | 8 ++++++++ fs/ceph/super.c | 1 + fs/ceph/super.h | 2 ++ 5 files changed, 61 insertions(+), 48 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index a08d245f16f5..1e48377f18a7 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1927,53 +1927,6 @@ static int caps_are_flushed(struct inode *inode, u64 flush_tid) return ret; } -/* - * Wait on any unsafe replies for the given inode. First wait on the - * newest request, and make that the upper bound. Then, if there are - * more requests, keep waiting on the oldest as long as it is still older - * than the original request. - */ -static void sync_write_wait(struct inode *inode) -{ - struct ceph_inode_info *ci = ceph_inode(inode); - struct list_head *head = &ci->i_unsafe_writes; - struct ceph_osd_request *req; - u64 last_tid; - - if (!S_ISREG(inode->i_mode)) - return; - - spin_lock(&ci->i_unsafe_lock); - if (list_empty(head)) - goto out; - - /* set upper bound as _last_ entry in chain */ - req = list_last_entry(head, struct ceph_osd_request, - r_unsafe_item); - last_tid = req->r_tid; - - do { - ceph_osdc_get_request(req); - spin_unlock(&ci->i_unsafe_lock); - dout("sync_write_wait on tid %llu (until %llu)\n", - req->r_tid, last_tid); - wait_for_completion(&req->r_safe_completion); - spin_lock(&ci->i_unsafe_lock); - ceph_osdc_put_request(req); - - /* - * from here on look at first entry in chain, since we - * only want to wait for anything older than last_tid - */ - if (list_empty(head)) - break; - req = list_first_entry(head, struct ceph_osd_request, - r_unsafe_item); - } while (req->r_tid < last_tid); -out: - spin_unlock(&ci->i_unsafe_lock); -} - /* * wait for any unsafe requests to complete. */ @@ -2026,7 +1979,8 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync) int dirty; dout("fsync %p%s\n", inode, datasync ? " datasync" : ""); - sync_write_wait(inode); + + ceph_sync_write_wait(inode); ret = filemap_write_and_wait_range(inode->i_mapping, start, end); if (ret < 0) diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 033e88753875..7f2ef262cdf7 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -821,6 +821,54 @@ static void ceph_sync_write_unsafe(struct ceph_osd_request *req, bool unsafe) } } +/* + * Wait on any unsafe replies for the given inode. First wait on the + * newest request, and make that the upper bound. Then, if there are + * more requests, keep waiting on the oldest as long as it is still older + * than the original request. + */ +void ceph_sync_write_wait(struct inode *inode) +{ + struct ceph_inode_info *ci = ceph_inode(inode); + struct list_head *head = &ci->i_unsafe_writes; + struct ceph_osd_request *req; + u64 last_tid; + + if (!S_ISREG(inode->i_mode)) + return; + + spin_lock(&ci->i_unsafe_lock); + if (list_empty(head)) + goto out; + + /* set upper bound as _last_ entry in chain */ + + req = list_last_entry(head, struct ceph_osd_request, + r_unsafe_item); + last_tid = req->r_tid; + + do { + ceph_osdc_get_request(req); + spin_unlock(&ci->i_unsafe_lock); + + dout("sync_write_wait on tid %llu (until %llu)\n", + req->r_tid, last_tid); + wait_for_completion(&req->r_safe_completion); + ceph_osdc_put_request(req); + + spin_lock(&ci->i_unsafe_lock); + /* + * from here on look at first entry in chain, since we + * only want to wait for anything older than last_tid + */ + if (list_empty(head)) + break; + req = list_first_entry(head, struct ceph_osd_request, + r_unsafe_item); + } while (req->r_tid < last_tid); +out: + spin_unlock(&ci->i_unsafe_lock); +} static ssize_t ceph_direct_read_write(struct kiocb *iocb, struct iov_iter *iter, diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 8ca843371d4b..6e16269277bd 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -585,6 +585,14 @@ int ceph_drop_inode(struct inode *inode) return 1; } +void ceph_evict_inode(struct inode *inode) +{ + /* wait unsafe sync writes */ + ceph_sync_write_wait(inode); + truncate_inode_pages_final(&inode->i_data); + clear_inode(inode); +} + static inline blkcnt_t calc_inode_blocks(u64 size) { return (size + (1<<9) - 1) >> 9; diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 91e02481ce06..a5b2275e1573 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -731,6 +731,7 @@ static const struct super_operations ceph_super_ops = { .destroy_inode = ceph_destroy_inode, .write_inode = ceph_write_inode, .drop_inode = ceph_drop_inode, + .evict_inode = ceph_evict_inode, .sync_fs = ceph_sync_fs, .put_super = ceph_put_super, .show_options = ceph_show_options, diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 50846e6f6a8c..d5b9077467a4 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -749,6 +749,7 @@ extern const struct inode_operations ceph_file_iops; extern struct inode *ceph_alloc_inode(struct super_block *sb); extern void ceph_destroy_inode(struct inode *inode); extern int ceph_drop_inode(struct inode *inode); +extern void ceph_evict_inode(struct inode *inode); extern struct inode *ceph_get_inode(struct super_block *sb, struct ceph_vino vino); @@ -927,6 +928,7 @@ extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry, extern int ceph_release(struct inode *inode, struct file *filp); extern void ceph_fill_inline_data(struct inode *inode, struct page *locked_page, char *data, size_t len); +extern void ceph_sync_write_wait(struct inode *inode); /* dir.c */ extern const struct file_operations ceph_dir_fops; extern const struct file_operations ceph_snapdir_fops; -- cgit v1.2.3 From fce8515741dfb6a233927262555295788ad22ca7 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 15 Jun 2016 20:51:22 +0800 Subject: ceph: fix NULL dereference in ceph_queue_cap_snap() old_snapc->seq is used in dout(...) Signed-off-by: Yan, Zheng --- fs/ceph/snap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 9caaa7ffc93f..eadf2c33edc6 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -551,7 +551,6 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) ci->i_wrbuffer_ref_head = 0; capsnap->context = old_snapc; list_add_tail(&capsnap->ci_item, &ci->i_cap_snaps); - old_snapc = NULL; if (used & CEPH_CAP_FILE_WR) { dout("queue_cap_snap %p cap_snap %p snapc %p" @@ -563,6 +562,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) __ceph_finish_cap_snap(ci, capsnap); } capsnap = NULL; + old_snapc = NULL; update_snapc: if (ci->i_head_snapc) { -- cgit v1.2.3 From 679f0b825d84f8c9a618730b00ae816976bc240f Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Thu, 23 Jun 2016 14:45:24 +0100 Subject: ceph: fix spelling mistake: "resgister" -> "register" trivial fix to spelling mistake in pr_err message Signed-off-by: Colin Ian King --- fs/ceph/cache.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/cache.c b/fs/ceph/cache.c index 238c55b01723..5bc5d37b1217 100644 --- a/fs/ceph/cache.c +++ b/fs/ceph/cache.c @@ -71,7 +71,7 @@ int ceph_fscache_register_fs(struct ceph_fs_client* fsc) &ceph_fscache_fsid_object_def, fsc, true); if (!fsc->fscache) - pr_err("Unable to resgister fsid: %p fscache cookie", fsc); + pr_err("Unable to register fsid: %p fscache cookie\n", fsc); return 0; } -- cgit v1.2.3 From 9b16f03c474d05b16cbd9eed1ec335c6e71cb57b Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 22 Jun 2016 16:35:04 +0200 Subject: ceph: don't use ->d_time Pretty simple: just use ceph_dentry_info.time instead (which was already there, unused). Signed-off-by: Miklos Szeredi --- fs/ceph/dir.c | 6 +++--- fs/ceph/inode.c | 4 ++-- fs/ceph/mds_client.c | 4 ++-- fs/ceph/super.h | 2 +- 4 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 6e0fedf6713b..8ff7bcc7fc88 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -59,7 +59,7 @@ int ceph_init_dentry(struct dentry *dentry) di->dentry = dentry; di->lease_session = NULL; - dentry->d_time = jiffies; + di->time = jiffies; /* avoid reordering d_fsdata setup so that the check above is safe */ smp_mb(); dentry->d_fsdata = di; @@ -1124,7 +1124,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry, void ceph_invalidate_dentry_lease(struct dentry *dentry) { spin_lock(&dentry->d_lock); - dentry->d_time = jiffies; + ceph_dentry(dentry)->time = jiffies; ceph_dentry(dentry)->lease_shared_gen = 0; spin_unlock(&dentry->d_lock); } @@ -1154,7 +1154,7 @@ static int dentry_lease_is_valid(struct dentry *dentry) spin_unlock(&s->s_gen_ttl_lock); if (di->lease_gen == gen && - time_before(jiffies, dentry->d_time) && + time_before(jiffies, di->time) && time_before(jiffies, ttl)) { valid = 1; if (di->lease_renew_after && diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 6e16269277bd..a38b768bc158 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1042,7 +1042,7 @@ static void update_dentry_lease(struct dentry *dentry, goto out_unlock; if (di->lease_gen == session->s_cap_gen && - time_before(ttl, dentry->d_time)) + time_before(ttl, di->time)) goto out_unlock; /* we already have a newer lease. */ if (di->lease_session && di->lease_session != session) @@ -1056,7 +1056,7 @@ static void update_dentry_lease(struct dentry *dentry, di->lease_seq = le32_to_cpu(lease->seq); di->lease_renew_after = half_ttl; di->lease_renew_from = 0; - dentry->d_time = ttl; + di->time = ttl; out_unlock: spin_unlock(&dentry->d_lock); return; diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 46641bbc8056..0d4bb24c670a 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -3226,7 +3226,7 @@ static void handle_lease(struct ceph_mds_client *mdsc, msecs_to_jiffies(le32_to_cpu(h->duration_ms)); di->lease_seq = seq; - dentry->d_time = di->lease_renew_from + duration; + di->time = di->lease_renew_from + duration; di->lease_renew_after = di->lease_renew_from + (duration >> 1); di->lease_renew_from = 0; @@ -3311,7 +3311,7 @@ void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode, if (!di || !di->lease_session || di->lease_session->s_mds < 0 || di->lease_gen != di->lease_session->s_cap_gen || - !time_before(jiffies, dentry->d_time)) { + !time_before(jiffies, di->time)) { dout("lease_release inode %p dentry %p -- " "no lease\n", inode, dentry); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index d5b9077467a4..5aa3158e8611 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -246,7 +246,7 @@ struct ceph_dentry_info { unsigned long lease_renew_after, lease_renew_from; struct list_head lru; struct dentry *dentry; - u64 time; + unsigned long time; u64 offset; }; -- cgit v1.2.3 From 8aa152c77890abd0731f119e4e6662375503e288 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 1 Jul 2016 09:39:20 -0400 Subject: ceph: remove ceph_mdsc_lease_release Nothing calls it. Signed-off-by: Jeff Layton Reviewed-by: Yan, Zheng --- fs/ceph/mds_client.c | 41 ----------------------------------------- fs/ceph/mds_client.h | 4 ---- 2 files changed, 45 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 0d4bb24c670a..78a3495a11be 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -3291,47 +3291,6 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, ceph_con_send(&session->s_con, msg); } -/* - * Preemptively release a lease we expect to invalidate anyway. - * Pass @inode always, @dentry is optional. - */ -void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode, - struct dentry *dentry) -{ - struct ceph_dentry_info *di; - struct ceph_mds_session *session; - u32 seq; - - BUG_ON(inode == NULL); - BUG_ON(dentry == NULL); - - /* is dentry lease valid? */ - spin_lock(&dentry->d_lock); - di = ceph_dentry(dentry); - if (!di || !di->lease_session || - di->lease_session->s_mds < 0 || - di->lease_gen != di->lease_session->s_cap_gen || - !time_before(jiffies, di->time)) { - dout("lease_release inode %p dentry %p -- " - "no lease\n", - inode, dentry); - spin_unlock(&dentry->d_lock); - return; - } - - /* we do have a lease on this dentry; note mds and seq */ - session = ceph_get_mds_session(di->lease_session); - seq = di->lease_seq; - __ceph_mdsc_drop_dentry_lease(dentry); - spin_unlock(&dentry->d_lock); - - dout("lease_release inode %p dentry %p to mds%d\n", - inode, dentry, session->s_mds); - ceph_mdsc_lease_send_msg(session, inode, dentry, - CEPH_MDS_LEASE_RELEASE, seq); - ceph_put_mds_session(session); -} - /* * drop all leases (and dentry refs) in preparation for umount */ diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 2ce8e9f9bfc9..9dd2c82379f8 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -385,10 +385,6 @@ extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc); extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc); -extern void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, - struct inode *inode, - struct dentry *dn); - extern void ceph_invalidate_dir_request(struct ceph_mds_request *req); extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req, struct inode *dir); -- cgit v1.2.3 From 5b484a513149f53613d376a9d1cd0391de099fb4 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 1 Jul 2016 09:39:20 -0400 Subject: ceph: clear d_fsinfo pointer under d_lock To check for a valid dentry lease, we need to get at the ceph_dentry_info. Under rcuwalk though, we may end up with a dentry that is on its way to destruction. Since we need to take the d_lock in dentry_lease_is_valid already, we can just ensure that we clear the d_fsinfo pointer out under the same lock before destroying it. Signed-off-by: Jeff Layton Reviewed-by: Yan, Zheng --- fs/ceph/dir.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 8ff7bcc7fc88..904dc671580f 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -1286,10 +1286,14 @@ static void ceph_d_release(struct dentry *dentry) dout("d_release %p\n", dentry); ceph_dentry_lru_del(dentry); + + spin_lock(&dentry->d_lock); + dentry->d_fsdata = NULL; + spin_unlock(&dentry->d_lock); + if (di->lease_session) ceph_put_mds_session(di->lease_session); kmem_cache_free(ceph_dentry_cachep, di); - dentry->d_fsdata = NULL; } static int ceph_snapdir_d_revalidate(struct dentry *dentry, -- cgit v1.2.3 From 14fb9c9efe3570459b6bddd76a990140917237ad Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 1 Jul 2016 09:39:20 -0400 Subject: ceph: allow dentry_lease_is_valid to work under RCU walk Under rcuwalk, we need to take extra care when dereferencing d_parent. We want to do that once and pass a pointer to dentry_lease_is_valid. Also, we must ensure that that function can handle the case where we're racing with d_release. Check whether "di" is NULL under the d_lock, and just return 0 if so. Finally, we still need to kick off a renewal job if the lease is getting close to expiration. If that's the case, then just drop out of rcuwalk mode since that could block. Signed-off-by: Jeff Layton Reviewed-by: Yan, Zheng --- fs/ceph/dir.c | 41 ++++++++++++++++++++++++++--------------- 1 file changed, 26 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 904dc671580f..60dcca163ba0 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -1133,7 +1133,8 @@ void ceph_invalidate_dentry_lease(struct dentry *dentry) * Check if dentry lease is valid. If not, delete the lease. Try to * renew if the least is more than half up. */ -static int dentry_lease_is_valid(struct dentry *dentry) +static int dentry_lease_is_valid(struct dentry *dentry, unsigned int flags, + struct inode *dir) { struct ceph_dentry_info *di; struct ceph_mds_session *s; @@ -1141,12 +1142,11 @@ static int dentry_lease_is_valid(struct dentry *dentry) u32 gen; unsigned long ttl; struct ceph_mds_session *session = NULL; - struct inode *dir = NULL; u32 seq = 0; spin_lock(&dentry->d_lock); di = ceph_dentry(dentry); - if (di->lease_session) { + if (di && di->lease_session) { s = di->lease_session; spin_lock(&s->s_gen_ttl_lock); gen = s->s_cap_gen; @@ -1159,12 +1159,19 @@ static int dentry_lease_is_valid(struct dentry *dentry) valid = 1; if (di->lease_renew_after && time_after(jiffies, di->lease_renew_after)) { - /* we should renew */ - dir = d_inode(dentry->d_parent); - session = ceph_get_mds_session(s); - seq = di->lease_seq; - di->lease_renew_after = 0; - di->lease_renew_from = jiffies; + /* + * We should renew. If we're in RCU walk mode + * though, we can't do that so just return + * -ECHILD. + */ + if (flags & LOOKUP_RCU) { + valid = -ECHILD; + } else { + session = ceph_get_mds_session(s); + seq = di->lease_seq; + di->lease_renew_after = 0; + di->lease_renew_from = jiffies; + } } } } @@ -1224,12 +1231,16 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) } else if (d_really_is_positive(dentry) && ceph_snap(d_inode(dentry)) == CEPH_SNAPDIR) { valid = 1; - } else if (dentry_lease_is_valid(dentry) || - dir_lease_is_valid(dir, dentry)) { - if (d_really_is_positive(dentry)) - valid = ceph_is_any_caps(d_inode(dentry)); - else - valid = 1; + } else { + valid = dentry_lease_is_valid(dentry, flags, dir); + if (valid == -ECHILD) + return valid; + if (valid || dir_lease_is_valid(dir, dentry)) { + if (d_really_is_positive(dentry)) + valid = ceph_is_any_caps(d_inode(dentry)); + else + valid = 1; + } } if (!valid) { -- cgit v1.2.3 From f49d1e058d23a5fabeb1499be739af8b3db52164 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 1 Jul 2016 09:39:21 -0400 Subject: ceph: handle LOOKUP_RCU in ceph_d_revalidate We can now handle the snapshot cases under RCU, as well as the non-snapshot case when we don't need to queue up a lease renewal allow LOOKUP_RCU walks to proceed under those conditions. Signed-off-by: Jeff Layton Reviewed-by: Yan, Zheng --- fs/ceph/dir.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 60dcca163ba0..c64a0b794d49 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -1214,15 +1214,19 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) struct dentry *parent; struct inode *dir; - if (flags & LOOKUP_RCU) - return -ECHILD; + if (flags & LOOKUP_RCU) { + parent = ACCESS_ONCE(dentry->d_parent); + dir = d_inode_rcu(parent); + if (!dir) + return -ECHILD; + } else { + parent = dget_parent(dentry); + dir = d_inode(parent); + } dout("d_revalidate %p '%pd' inode %p offset %lld\n", dentry, dentry, d_inode(dentry), ceph_dentry(dentry)->offset); - parent = dget_parent(dentry); - dir = d_inode(parent); - /* always trust cached snapped dentries, snapdir dentry */ if (ceph_snap(dir) != CEPH_NOSNAP) { dout("d_revalidate %p '%pd' inode %p is SNAPPED\n", dentry, @@ -1249,6 +1253,9 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) struct ceph_mds_request *req; int op, mask, err; + if (flags & LOOKUP_RCU) + return -ECHILD; + op = ceph_snap(dir) == CEPH_SNAPDIR ? CEPH_MDS_OP_LOOKUPSNAP : CEPH_MDS_OP_LOOKUP; req = ceph_mdsc_create_request(mdsc, op, USE_ANY_MDS); @@ -1284,7 +1291,8 @@ static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags) ceph_dir_clear_complete(dir); } - dput(parent); + if (!(flags & LOOKUP_RCU)) + dput(parent); return valid; } -- cgit v1.2.3 From 430afbadd6c885557ef2fb8c454bd5bba23a9850 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 8 Jul 2016 11:25:38 +0800 Subject: ceph: mount non-default filesystem by name To mount non-default filesytem, user currently needs to provide mds namespace ID. This is inconvenience. This patch makes user be able to mount filesystem by name. If user wants to mount non-default filesystem. Client first subscribes to fsmap.user. Subscribe to mdsmap. after getting ID of filesystem. Signed-off-by: Yan, Zheng --- fs/ceph/mds_client.c | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++-- fs/ceph/mds_client.h | 7 +++-- fs/ceph/super.c | 38 +++++++++++++++-------- fs/ceph/super.h | 2 +- 4 files changed, 117 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 78a3495a11be..e555745883da 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -2166,6 +2166,11 @@ static int __do_request(struct ceph_mds_client *mdsc, mds = __choose_mds(mdsc, req); if (mds < 0 || ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) { + if (mdsc->mdsmap_err) { + err = mdsc->mdsmap_err; + dout("do_request mdsmap err %d\n", err); + goto finish; + } dout("do_request no mds or not active, waiting for map\n"); list_add(&req->r_wait, &mdsc->waiting_for_map); goto out; @@ -3683,11 +3688,86 @@ void ceph_mdsc_destroy(struct ceph_fs_client *fsc) dout("mdsc_destroy %p done\n", mdsc); } +void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) +{ + struct ceph_fs_client *fsc = mdsc->fsc; + const char *mds_namespace = fsc->mount_options->mds_namespace; + void *p = msg->front.iov_base; + void *end = p + msg->front.iov_len; + u32 epoch; + u32 map_len; + u32 num_fs; + u32 mount_fscid = (u32)-1; + u8 struct_v, struct_cv; + int err = -EINVAL; + + ceph_decode_need(&p, end, sizeof(u32), bad); + epoch = ceph_decode_32(&p); + + dout("handle_fsmap epoch %u\n", epoch); + + ceph_decode_need(&p, end, 2 + sizeof(u32), bad); + struct_v = ceph_decode_8(&p); + struct_cv = ceph_decode_8(&p); + map_len = ceph_decode_32(&p); + + ceph_decode_need(&p, end, sizeof(u32) * 3, bad); + p += sizeof(u32) * 2; /* skip epoch and legacy_client_fscid */ + + num_fs = ceph_decode_32(&p); + while (num_fs-- > 0) { + void *info_p, *info_end; + u32 info_len; + u8 info_v, info_cv; + u32 fscid, namelen; + + ceph_decode_need(&p, end, 2 + sizeof(u32), bad); + info_v = ceph_decode_8(&p); + info_cv = ceph_decode_8(&p); + info_len = ceph_decode_32(&p); + ceph_decode_need(&p, end, info_len, bad); + info_p = p; + info_end = p + info_len; + p = info_end; + + ceph_decode_need(&info_p, info_end, sizeof(u32) * 2, bad); + fscid = ceph_decode_32(&info_p); + namelen = ceph_decode_32(&info_p); + ceph_decode_need(&info_p, info_end, namelen, bad); + + if (mds_namespace && + strlen(mds_namespace) == namelen && + !strncmp(mds_namespace, (char *)info_p, namelen)) { + mount_fscid = fscid; + break; + } + } + + ceph_monc_got_map(&fsc->client->monc, CEPH_SUB_FSMAP, epoch); + if (mount_fscid != (u32)-1) { + fsc->client->monc.fs_cluster_id = mount_fscid; + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, + 0, true); + ceph_monc_renew_subs(&fsc->client->monc); + } else { + err = -ENOENT; + goto err_out; + } + return; +bad: + pr_err("error decoding fsmap\n"); +err_out: + mutex_lock(&mdsc->mutex); + mdsc->mdsmap_err = -ENOENT; + __wake_requests(mdsc, &mdsc->waiting_for_map); + mutex_unlock(&mdsc->mutex); + return; +} /* * handle mds map update. */ -void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg) +void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg) { u32 epoch; u32 maplen; @@ -3794,7 +3874,10 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg) switch (type) { case CEPH_MSG_MDS_MAP: - ceph_mdsc_handle_map(mdsc, msg); + ceph_mdsc_handle_mdsmap(mdsc, msg); + break; + case CEPH_MSG_FS_MAP_USER: + ceph_mdsc_handle_fsmap(mdsc, msg); break; case CEPH_MSG_CLIENT_SESSION: handle_session(s, msg); diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 9dd2c82379f8..3c154b8d49bf 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -293,6 +293,7 @@ struct ceph_mds_client { struct completion safe_umount_waiters; wait_queue_head_t session_close_wq; struct list_head waiting_for_map; + int mdsmap_err; struct ceph_mds_session **sessions; /* NULL for mds if no session */ atomic_t num_sessions; @@ -419,8 +420,10 @@ extern void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session, struct dentry *dentry, char action, u32 seq); -extern void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, - struct ceph_msg *msg); +extern void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, + struct ceph_msg *msg); +extern void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, + struct ceph_msg *msg); extern struct ceph_mds_session * ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target); diff --git a/fs/ceph/super.c b/fs/ceph/super.c index a5b2275e1573..7736a931376e 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -108,7 +108,6 @@ static int ceph_sync_fs(struct super_block *sb, int wait) * mount options */ enum { - Opt_mds_namespace, Opt_wsize, Opt_rsize, Opt_rasize, @@ -121,6 +120,7 @@ enum { Opt_last_int, /* int args above */ Opt_snapdirname, + Opt_mds_namespace, Opt_last_string, /* string args above */ Opt_dirstat, @@ -144,7 +144,6 @@ enum { }; static match_table_t fsopt_tokens = { - {Opt_mds_namespace, "mds_namespace=%d"}, {Opt_wsize, "wsize=%d"}, {Opt_rsize, "rsize=%d"}, {Opt_rasize, "rasize=%d"}, @@ -156,6 +155,7 @@ static match_table_t fsopt_tokens = { {Opt_congestion_kb, "write_congestion_kb=%d"}, /* int args above */ {Opt_snapdirname, "snapdirname=%s"}, + {Opt_mds_namespace, "mds_namespace=%s"}, /* string args above */ {Opt_dirstat, "dirstat"}, {Opt_nodirstat, "nodirstat"}, @@ -212,11 +212,14 @@ static int parse_fsopt_token(char *c, void *private) if (!fsopt->snapdir_name) return -ENOMEM; break; - - /* misc */ case Opt_mds_namespace: - fsopt->mds_namespace = intval; + fsopt->mds_namespace = kstrndup(argstr[0].from, + argstr[0].to-argstr[0].from, + GFP_KERNEL); + if (!fsopt->mds_namespace) + return -ENOMEM; break; + /* misc */ case Opt_wsize: fsopt->wsize = intval; break; @@ -302,6 +305,7 @@ static void destroy_mount_options(struct ceph_mount_options *args) { dout("destroy_mount_options %p\n", args); kfree(args->snapdir_name); + kfree(args->mds_namespace); kfree(args->server_path); kfree(args); } @@ -331,6 +335,9 @@ static int compare_mount_options(struct ceph_mount_options *new_fsopt, return ret; ret = strcmp_null(fsopt1->snapdir_name, fsopt2->snapdir_name); + if (ret) + return ret; + ret = strcmp_null(fsopt1->mds_namespace, fsopt2->mds_namespace); if (ret) return ret; @@ -376,7 +383,6 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt, fsopt->max_readdir = CEPH_MAX_READDIR_DEFAULT; fsopt->max_readdir_bytes = CEPH_MAX_READDIR_BYTES_DEFAULT; fsopt->congestion_kb = default_congestion_kb(); - fsopt->mds_namespace = CEPH_FS_CLUSTER_ID_NONE; /* * Distinguish the server list from the path in "dev_name". @@ -469,8 +475,8 @@ static int ceph_show_options(struct seq_file *m, struct dentry *root) seq_puts(m, ",noacl"); #endif - if (fsopt->mds_namespace != CEPH_FS_CLUSTER_ID_NONE) - seq_printf(m, ",mds_namespace=%d", fsopt->mds_namespace); + if (fsopt->mds_namespace) + seq_printf(m, ",mds_namespace=%s", fsopt->mds_namespace); if (fsopt->wsize) seq_printf(m, ",wsize=%d", fsopt->wsize); if (fsopt->rsize != CEPH_RSIZE_DEFAULT) @@ -509,9 +515,11 @@ static int extra_mon_dispatch(struct ceph_client *client, struct ceph_msg *msg) switch (type) { case CEPH_MSG_MDS_MAP: - ceph_mdsc_handle_map(fsc->mdsc, msg); + ceph_mdsc_handle_mdsmap(fsc->mdsc, msg); + return 0; + case CEPH_MSG_FS_MAP_USER: + ceph_mdsc_handle_fsmap(fsc->mdsc, msg); return 0; - default: return -1; } @@ -543,8 +551,14 @@ static struct ceph_fs_client *create_fs_client(struct ceph_mount_options *fsopt, goto fail; } fsc->client->extra_mon_dispatch = extra_mon_dispatch; - fsc->client->monc.fs_cluster_id = fsopt->mds_namespace; - ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, 0, true); + + if (fsopt->mds_namespace == NULL) { + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP, + 0, true); + } else { + ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_FSMAP, + 0, false); + } fsc->mount_options = fsopt; diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 5aa3158e8611..9e82e29f86a1 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -62,7 +62,6 @@ struct ceph_mount_options { int cap_release_safety; int max_readdir; /* max readdir result (entires) */ int max_readdir_bytes; /* max readdir result (bytes) */ - int mds_namespace; /* * everything above this point can be memcmp'd; everything below @@ -70,6 +69,7 @@ struct ceph_mount_options { */ char *snapdir_name; /* default ".snap" */ + char *mds_namespace; /* default NULL */ char *server_path; /* default "/" */ }; -- cgit v1.2.3 From 121f22a19a30574abff68f2a22716c34cfae0b05 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 4 Jul 2016 22:05:18 +0800 Subject: ceph: update cap reconnect message to version 3 Signed-off-by: Yan, Zheng --- fs/ceph/mds_client.c | 68 ++++++++++++++++++++++++++++++++++++---------------- 1 file changed, 47 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index e555745883da..131a61f17d8a 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -48,7 +48,7 @@ struct ceph_reconnect_state { int nr_caps; struct ceph_pagelist *pagelist; - bool flock; + unsigned msg_version; }; static void __wake_requests(struct ceph_mds_client *mdsc, @@ -2791,7 +2791,6 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, struct ceph_mds_cap_reconnect v2; struct ceph_mds_cap_reconnect_v1 v1; } rec; - size_t reclen; struct ceph_inode_info *ci; struct ceph_reconnect_state *recon_state = arg; struct ceph_pagelist *pagelist = recon_state->pagelist; @@ -2820,9 +2819,6 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, path = NULL; pathlen = 0; } - err = ceph_pagelist_encode_string(pagelist, path, pathlen); - if (err) - goto out_free; spin_lock(&ci->i_ceph_lock); cap->seq = 0; /* reset cap seq */ @@ -2830,14 +2826,13 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, cap->mseq = 0; /* and migrate_seq */ cap->cap_gen = cap->session->s_cap_gen; - if (recon_state->flock) { + if (recon_state->msg_version >= 2) { rec.v2.cap_id = cpu_to_le64(cap->cap_id); rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); rec.v2.issued = cpu_to_le32(cap->issued); rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); rec.v2.pathbase = cpu_to_le64(pathbase); rec.v2.flock_len = 0; - reclen = sizeof(rec.v2); } else { rec.v1.cap_id = cpu_to_le64(cap->cap_id); rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci)); @@ -2847,13 +2842,14 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, ceph_encode_timespec(&rec.v1.atime, &inode->i_atime); rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); rec.v1.pathbase = cpu_to_le64(pathbase); - reclen = sizeof(rec.v1); } spin_unlock(&ci->i_ceph_lock); - if (recon_state->flock) { + if (recon_state->msg_version >= 2) { int num_fcntl_locks, num_flock_locks; struct ceph_filelock *flocks; + size_t struct_len, total_len = 0; + u8 struct_v = 0; encode_again: ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks); @@ -2872,20 +2868,46 @@ encode_again: goto encode_again; goto out_free; } + + if (recon_state->msg_version >= 3) { + /* version, compat_version and struct_len */ + total_len = 2 * sizeof(u8) + sizeof(u32); + struct_v = 1; + } /* * number of encoded locks is stable, so copy to pagelist */ - rec.v2.flock_len = cpu_to_le32(2*sizeof(u32) + - (num_fcntl_locks+num_flock_locks) * - sizeof(struct ceph_filelock)); - err = ceph_pagelist_append(pagelist, &rec, reclen); - if (!err) - err = ceph_locks_to_pagelist(flocks, pagelist, - num_fcntl_locks, - num_flock_locks); + struct_len = 2 * sizeof(u32) + + (num_fcntl_locks + num_flock_locks) * + sizeof(struct ceph_filelock); + rec.v2.flock_len = cpu_to_le32(struct_len); + + struct_len += sizeof(rec.v2); + struct_len += sizeof(u32) + pathlen; + + total_len += struct_len; + err = ceph_pagelist_reserve(pagelist, total_len); + + if (!err) { + if (recon_state->msg_version >= 3) { + ceph_pagelist_encode_8(pagelist, struct_v); + ceph_pagelist_encode_8(pagelist, 1); + ceph_pagelist_encode_32(pagelist, struct_len); + } + ceph_pagelist_encode_string(pagelist, path, pathlen); + ceph_pagelist_append(pagelist, &rec, sizeof(rec.v2)); + ceph_locks_to_pagelist(flocks, pagelist, + num_fcntl_locks, + num_flock_locks); + } kfree(flocks); } else { - err = ceph_pagelist_append(pagelist, &rec, reclen); + size_t size = sizeof(u32) + pathlen + sizeof(rec.v1); + err = ceph_pagelist_reserve(pagelist, size); + if (!err) { + ceph_pagelist_encode_string(pagelist, path, pathlen); + ceph_pagelist_append(pagelist, &rec, sizeof(rec.v1)); + } } recon_state->nr_caps++; @@ -2976,7 +2998,12 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, recon_state.nr_caps = 0; recon_state.pagelist = pagelist; - recon_state.flock = session->s_con.peer_features & CEPH_FEATURE_FLOCK; + if (session->s_con.peer_features & CEPH_FEATURE_MDSENC) + recon_state.msg_version = 3; + else if (session->s_con.peer_features & CEPH_FEATURE_FLOCK) + recon_state.msg_version = 2; + else + recon_state.msg_version = 1; err = iterate_session_caps(session, encode_caps_cb, &recon_state); if (err < 0) goto fail; @@ -3005,8 +3032,7 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, goto fail; } - if (recon_state.flock) - reply->hdr.version = cpu_to_le16(2); + reply->hdr.version = cpu_to_le16(recon_state.msg_version); /* raced with cap release? */ if (s_nr_caps != recon_state.nr_caps) { -- cgit v1.2.3 From 3469ed0d149ba7066e8fef72be55f67ee7de196d Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 5 Jul 2016 09:32:31 +0800 Subject: ceph: include 'follows' of pending snapflush in cap reconnect message This helps the recovering MDS to reconstruct the internal states that tracking pending snapflush. Signed-off-by: Yan, Zheng --- fs/ceph/mds_client.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 131a61f17d8a..bcf20344d904 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -2797,6 +2797,7 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, char *path; int pathlen, err; u64 pathbase; + u64 snap_follows; struct dentry *dentry; ci = cap->ci; @@ -2843,6 +2844,15 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino); rec.v1.pathbase = cpu_to_le64(pathbase); } + + if (list_empty(&ci->i_cap_snaps)) { + snap_follows = 0; + } else { + struct ceph_cap_snap *capsnap = + list_first_entry(&ci->i_cap_snaps, + struct ceph_cap_snap, ci_item); + snap_follows = capsnap->follows; + } spin_unlock(&ci->i_ceph_lock); if (recon_state->msg_version >= 2) { @@ -2872,7 +2882,7 @@ encode_again: if (recon_state->msg_version >= 3) { /* version, compat_version and struct_len */ total_len = 2 * sizeof(u8) + sizeof(u32); - struct_v = 1; + struct_v = 2; } /* * number of encoded locks is stable, so copy to pagelist @@ -2885,6 +2895,9 @@ encode_again: struct_len += sizeof(rec.v2); struct_len += sizeof(u32) + pathlen; + if (struct_v >= 2) + struct_len += sizeof(u64); /* snap_follows */ + total_len += struct_len; err = ceph_pagelist_reserve(pagelist, total_len); @@ -2899,6 +2912,8 @@ encode_again: ceph_locks_to_pagelist(flocks, pagelist, num_fcntl_locks, num_flock_locks); + if (struct_v >= 2) + ceph_pagelist_encode_64(pagelist, snap_follows); } kfree(flocks); } else { -- cgit v1.2.3 From 3609404f8c782c01fe626a03949641dddbc65c34 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 6 Jul 2016 15:37:42 +0800 Subject: ceph: update types of some local varibles Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 1e48377f18a7..698cf002d3f1 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -993,7 +993,7 @@ static int send_cap_msg(struct ceph_mds_session *session, u32 seq, u64 flush_tid, u64 oldest_flush_tid, u32 issue_seq, u32 mseq, u64 size, u64 max_size, struct timespec *mtime, struct timespec *atime, - struct timespec *ctime, u64 time_warp_seq, + struct timespec *ctime, u32 time_warp_seq, kuid_t uid, kgid_t gid, umode_t mode, u64 xattr_version, struct ceph_buffer *xattrs_buf, @@ -1118,8 +1118,8 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, struct inode *inode = &ci->vfs_inode; u64 cap_id = cap->cap_id; int held, revoking, dropping, keep; - u64 seq, issue_seq, mseq, time_warp_seq, follows; - u64 size, max_size; + u64 follows, size, max_size; + u32 seq, issue_seq, mseq, time_warp_seq; struct timespec mtime, atime, ctime; int wake = 0; umode_t mode; @@ -1585,10 +1585,11 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, int mds = -1; /* keep track of how far we've gone through i_caps list to avoid an infinite loop on retry */ struct rb_node *p; - int tried_invalidate = 0; - int delayed = 0, sent = 0, force_requeue = 0, num; - int queue_invalidate = 0; - int is_delayed = flags & CHECK_CAPS_NODELAY; + int delayed = 0, sent = 0, num; + bool is_delayed = flags & CHECK_CAPS_NODELAY; + bool queue_invalidate = false; + bool force_requeue = false; + bool tried_invalidate = false; /* if we are unmounting, flush any unused caps immediately. */ if (mdsc->stopping) @@ -1668,17 +1669,17 @@ retry_locked: if (revoking & (CEPH_CAP_FILE_CACHE| CEPH_CAP_FILE_LAZYIO)) { dout("check_caps queuing invalidate\n"); - queue_invalidate = 1; + queue_invalidate = true; ci->i_rdcache_revoking = ci->i_rdcache_gen; } else { dout("check_caps failed to invalidate pages\n"); /* we failed to invalidate pages. check these caps again later. */ - force_requeue = 1; + force_requeue = true; __cap_set_timeouts(mdsc, ci); } } - tried_invalidate = 1; + tried_invalidate = true; goto retry_locked; } @@ -1824,7 +1825,7 @@ ack: * otherwise cancel. */ if (delayed && is_delayed) - force_requeue = 1; /* __send_cap delayed release; requeue */ + force_requeue = true; /* __send_cap delayed release; requeue */ if (!delayed && !is_delayed) __cap_delay_cancel(mdsc, ci); else if (!is_delayed || force_requeue) -- cgit v1.2.3 From e4500b5e35c213e0f97be7cb69328c0877203a79 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 6 Jul 2016 11:12:56 +0800 Subject: ceph: use list instead of rbtree to track cap flushes We don't have requirement of searching cap flush by TID. In most cases, we just need to know TID of the oldest cap flush. List is ideal for this usage. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 120 +++++++++++++-------------------------------------- fs/ceph/inode.c | 2 +- fs/ceph/mds_client.c | 41 +++++++++--------- fs/ceph/mds_client.h | 2 +- fs/ceph/super.h | 9 ++-- 5 files changed, 56 insertions(+), 118 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 698cf002d3f1..e0efa75a1b98 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1413,52 +1413,6 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask, return dirty; } -static void __add_cap_flushing_to_inode(struct ceph_inode_info *ci, - struct ceph_cap_flush *cf) -{ - struct rb_node **p = &ci->i_cap_flush_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_cap_flush *other = NULL; - - while (*p) { - parent = *p; - other = rb_entry(parent, struct ceph_cap_flush, i_node); - - if (cf->tid < other->tid) - p = &(*p)->rb_left; - else if (cf->tid > other->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&cf->i_node, parent, p); - rb_insert_color(&cf->i_node, &ci->i_cap_flush_tree); -} - -static void __add_cap_flushing_to_mdsc(struct ceph_mds_client *mdsc, - struct ceph_cap_flush *cf) -{ - struct rb_node **p = &mdsc->cap_flush_tree.rb_node; - struct rb_node *parent = NULL; - struct ceph_cap_flush *other = NULL; - - while (*p) { - parent = *p; - other = rb_entry(parent, struct ceph_cap_flush, g_node); - - if (cf->tid < other->tid) - p = &(*p)->rb_left; - else if (cf->tid > other->tid) - p = &(*p)->rb_right; - else - BUG(); - } - - rb_link_node(&cf->g_node, parent, p); - rb_insert_color(&cf->g_node, &mdsc->cap_flush_tree); -} - struct ceph_cap_flush *ceph_alloc_cap_flush(void) { return kmem_cache_alloc(ceph_cap_flush_cachep, GFP_KERNEL); @@ -1472,10 +1426,10 @@ void ceph_free_cap_flush(struct ceph_cap_flush *cf) static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) { - struct rb_node *n = rb_first(&mdsc->cap_flush_tree); - if (n) { + if (!list_empty(&mdsc->cap_flush_list)) { struct ceph_cap_flush *cf = - rb_entry(n, struct ceph_cap_flush, g_node); + list_first_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); return cf->tid; } return 0; @@ -1516,7 +1470,7 @@ static int __mark_caps_flushing(struct inode *inode, list_del_init(&ci->i_dirty_item); cf->tid = ++mdsc->last_cap_flush_tid; - __add_cap_flushing_to_mdsc(mdsc, cf); + list_add_tail(&cf->g_list, &mdsc->cap_flush_list); *oldest_flush_tid = __get_oldest_flush_tid(mdsc); if (list_empty(&ci->i_flushing_item)) { @@ -1530,7 +1484,7 @@ static int __mark_caps_flushing(struct inode *inode, } spin_unlock(&mdsc->cap_dirty_lock); - __add_cap_flushing_to_inode(ci, cf); + list_add_tail(&cf->i_list, &ci->i_cap_flush_list); *flush_tid = cf->tid; return flushing; @@ -1890,10 +1844,10 @@ retry: spin_unlock(&ci->i_ceph_lock); } } else { - struct rb_node *n = rb_last(&ci->i_cap_flush_tree); - if (n) { + if (!list_empty(&ci->i_cap_flush_list)) { struct ceph_cap_flush *cf = - rb_entry(n, struct ceph_cap_flush, i_node); + list_last_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); flush_tid = cf->tid; } flushing = ci->i_flushing_caps; @@ -1913,14 +1867,13 @@ out: static int caps_are_flushed(struct inode *inode, u64 flush_tid) { struct ceph_inode_info *ci = ceph_inode(inode); - struct ceph_cap_flush *cf; - struct rb_node *n; int ret = 1; spin_lock(&ci->i_ceph_lock); - n = rb_first(&ci->i_cap_flush_tree); - if (n) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); + if (!list_empty(&ci->i_cap_flush_list)) { + struct ceph_cap_flush * cf = + list_first_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); if (cf->tid <= flush_tid) ret = 0; } @@ -2083,7 +2036,6 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc, struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; struct ceph_cap_flush *cf; - struct rb_node *n; int delayed = 0; u64 first_tid = 0; u64 oldest_flush_tid; @@ -2092,8 +2044,11 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc, oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); - while (true) { - spin_lock(&ci->i_ceph_lock); + spin_lock(&ci->i_ceph_lock); + list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { + if (cf->tid < first_tid) + continue; + cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { pr_err("%p auth cap %p not mds%d ???\n", inode, @@ -2102,18 +2057,6 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc, break; } - for (n = rb_first(&ci->i_cap_flush_tree); n; n = rb_next(n)) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); - if (cf->tid >= first_tid) - break; - } - if (!n) { - spin_unlock(&ci->i_ceph_lock); - break; - } - - cf = rb_entry(n, struct ceph_cap_flush, i_node); - first_tid = cf->tid + 1; dout("kick_flushing_caps %p cap %p tid %llu %s\n", inode, @@ -2123,7 +2066,10 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc, __ceph_caps_wanted(ci), cap->issued | cap->implemented, cf->caps, cf->tid, oldest_flush_tid); + + spin_lock(&ci->i_ceph_lock); } + spin_unlock(&ci->i_ceph_lock); return delayed; } @@ -2995,23 +2941,19 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; - struct ceph_cap_flush *cf; - struct rb_node *n; + struct ceph_cap_flush *cf, *tmp_cf; LIST_HEAD(to_remove); unsigned seq = le32_to_cpu(m->seq); int dirty = le32_to_cpu(m->dirty); int cleaned = 0; int drop = 0; - n = rb_first(&ci->i_cap_flush_tree); - while (n) { - cf = rb_entry(n, struct ceph_cap_flush, i_node); - n = rb_next(&cf->i_node); + list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { if (cf->tid == flush_tid) cleaned = cf->caps; if (cf->tid <= flush_tid) { - rb_erase(&cf->i_node, &ci->i_cap_flush_tree); - list_add_tail(&cf->list, &to_remove); + list_del(&cf->i_list); + list_add_tail(&cf->i_list, &to_remove); } else { cleaned &= ~cf->caps; if (!cleaned) @@ -3033,12 +2975,12 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, spin_lock(&mdsc->cap_dirty_lock); if (!list_empty(&to_remove)) { - list_for_each_entry(cf, &to_remove, list) - rb_erase(&cf->g_node, &mdsc->cap_flush_tree); + u64 oldest_flush_tid; + list_for_each_entry(cf, &to_remove, i_list) + list_del(&cf->g_list); - n = rb_first(&mdsc->cap_flush_tree); - cf = n ? rb_entry(n, struct ceph_cap_flush, g_node) : NULL; - if (!cf || cf->tid > flush_tid) + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + if (oldest_flush_tid == 0 || oldest_flush_tid > flush_tid) wake_up_all(&mdsc->cap_flushing_wq); } @@ -3075,8 +3017,8 @@ out: while (!list_empty(&to_remove)) { cf = list_first_entry(&to_remove, - struct ceph_cap_flush, list); - list_del(&cf->list); + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); ceph_free_cap_flush(cf); } if (drop) diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index a38b768bc158..fd85b3c58960 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -468,7 +468,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb) INIT_LIST_HEAD(&ci->i_dirty_item); INIT_LIST_HEAD(&ci->i_flushing_item); ci->i_prealloc_cap_flush = NULL; - ci->i_cap_flush_tree = RB_ROOT; + INIT_LIST_HEAD(&ci->i_cap_flush_list); init_waitqueue_head(&ci->i_cap_wq); ci->i_hold_caps_min = 0; ci->i_hold_caps_max = 0; diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index bcf20344d904..7cd6b861c2f3 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1148,19 +1148,17 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, ACCESS_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) invalidate = true; - while (true) { - struct rb_node *n = rb_first(&ci->i_cap_flush_tree); - if (!n) - break; - cf = rb_entry(n, struct ceph_cap_flush, i_node); - rb_erase(&cf->i_node, &ci->i_cap_flush_tree); - list_add(&cf->list, &to_remove); + while (!list_empty(&ci->i_cap_flush_list)) { + cf = list_first_entry(&ci->i_cap_flush_list, + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); + list_add(&cf->i_list, &to_remove); } spin_lock(&mdsc->cap_dirty_lock); - list_for_each_entry(cf, &to_remove, list) - rb_erase(&cf->g_node, &mdsc->cap_flush_tree); + list_for_each_entry(cf, &to_remove, i_list) + list_del(&cf->g_list); if (!list_empty(&ci->i_dirty_item)) { pr_warn_ratelimited( @@ -1184,7 +1182,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, spin_unlock(&mdsc->cap_dirty_lock); if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) { - list_add(&ci->i_prealloc_cap_flush->list, &to_remove); + list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove); ci->i_prealloc_cap_flush = NULL; } } @@ -1192,8 +1190,8 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap, while (!list_empty(&to_remove)) { struct ceph_cap_flush *cf; cf = list_first_entry(&to_remove, - struct ceph_cap_flush, list); - list_del(&cf->list); + struct ceph_cap_flush, i_list); + list_del(&cf->i_list); ceph_free_cap_flush(cf); } @@ -1499,17 +1497,18 @@ static int check_capsnap_flush(struct ceph_inode_info *ci, static int check_caps_flush(struct ceph_mds_client *mdsc, u64 want_flush_tid) { - struct rb_node *n; - struct ceph_cap_flush *cf; int ret = 1; spin_lock(&mdsc->cap_dirty_lock); - n = rb_first(&mdsc->cap_flush_tree); - cf = n ? rb_entry(n, struct ceph_cap_flush, g_node) : NULL; - if (cf && cf->tid <= want_flush_tid) { - dout("check_caps_flush still flushing tid %llu <= %llu\n", - cf->tid, want_flush_tid); - ret = 0; + if (!list_empty(&mdsc->cap_flush_list)) { + struct ceph_cap_flush *cf = + list_first_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); + if (cf->tid <= want_flush_tid) { + dout("check_caps_flush still flushing tid " + "%llu <= %llu\n", cf->tid, want_flush_tid); + ret = 0; + } } spin_unlock(&mdsc->cap_dirty_lock); return ret; @@ -3470,7 +3469,7 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) INIT_LIST_HEAD(&mdsc->snap_flush_list); spin_lock_init(&mdsc->snap_flush_lock); mdsc->last_cap_flush_tid = 1; - mdsc->cap_flush_tree = RB_ROOT; + INIT_LIST_HEAD(&mdsc->cap_flush_list); INIT_LIST_HEAD(&mdsc->cap_dirty); INIT_LIST_HEAD(&mdsc->cap_dirty_migrating); mdsc->num_cap_flushing = 0; diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 3c154b8d49bf..93170b4b5d75 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -325,7 +325,7 @@ struct ceph_mds_client { spinlock_t snap_flush_lock; u64 last_cap_flush_tid; - struct rb_root cap_flush_tree; + struct list_head cap_flush_list; struct list_head cap_dirty; /* inodes with dirty caps */ struct list_head cap_dirty_migrating; /* ...that are migration... */ int num_cap_flushing; /* # caps we are flushing */ diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 9e82e29f86a1..29e8b7bd9413 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -189,11 +189,8 @@ static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap) struct ceph_cap_flush { u64 tid; int caps; - struct rb_node g_node; // global - union { - struct rb_node i_node; // inode - struct list_head list; - }; + struct list_head g_list; // global + struct list_head i_list; // per inode }; /* @@ -310,7 +307,7 @@ struct ceph_inode_info { * overlapping, pipelined cap flushes to the mds. we can probably * reduce the tid to 8 bits if we're concerned about inode size. */ struct ceph_cap_flush *i_prealloc_cap_flush; - struct rb_root i_cap_flush_tree; + struct list_head i_cap_flush_list; wait_queue_head_t i_cap_wq; /* threads waiting on a capability */ unsigned long i_hold_caps_min; /* jiffies */ unsigned long i_hold_caps_max; /* jiffies */ -- cgit v1.2.3 From 0e2943878942aee7100c94d0d40c49087dac12cb Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 4 Jul 2016 18:06:41 +0800 Subject: ceph: unify cap flush and snapcap flush This patch includes following changes - Assign flush tid to snapcap flush - Remove session's s_cap_snaps_flushing list. Add inode to session's s_cap_flushing list instead. Inode is removed from the list when there is no pending snapcap flush or cap flush. - make __kick_flushing_caps() re-send both snapcap flushes and cap flushes. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 291 +++++++++++++++++++++++++++------------------------ fs/ceph/mds_client.c | 77 ++------------ fs/ceph/mds_client.h | 1 - fs/ceph/snap.c | 4 +- fs/ceph/super.h | 24 ++--- 5 files changed, 175 insertions(+), 222 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index e0efa75a1b98..0ac604719663 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -40,6 +40,7 @@ * cluster to release server state. */ +static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc); /* * Generate readable cap strings for debugging output. @@ -1217,6 +1218,22 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, return delayed; } +static inline int __send_flush_snap(struct inode *inode, + struct ceph_mds_session *session, + struct ceph_cap_snap *capsnap, + u32 mseq, u64 oldest_flush_tid) +{ + return send_cap_msg(session, ceph_vino(inode).ino, 0, + CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0, + capsnap->dirty, 0, capsnap->cap_flush.tid, + oldest_flush_tid, 0, mseq, capsnap->size, 0, + &capsnap->mtime, &capsnap->atime, + &capsnap->ctime, capsnap->time_warp_seq, + capsnap->uid, capsnap->gid, capsnap->mode, + capsnap->xattr_version, capsnap->xattr_blob, + capsnap->follows, capsnap->inline_data); +} + /* * When a snapshot is taken, clients accumulate dirty metadata on * inodes with capabilities in ceph_cap_snaps to describe the file @@ -1224,14 +1241,10 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap, * asynchronously back to the MDS once sync writes complete and dirty * data is written out. * - * Unless @kick is true, skip cap_snaps that were already sent to - * the MDS (i.e., during this session). - * * Called under i_ceph_lock. Takes s_mutex as needed. */ void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession, - int kick) + struct ceph_mds_session **psession) __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) { @@ -1242,6 +1255,7 @@ void __ceph_flush_snaps(struct ceph_inode_info *ci, struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; struct ceph_mds_session *session = NULL; /* if session != NULL, we hold session->s_mutex */ + u64 oldest_flush_tid; u64 next_follows = 0; /* keep track of how far we've gotten through the i_cap_snaps list, and skip these entries next time around to avoid an infinite loop */ @@ -1272,7 +1286,7 @@ retry: } /* only flush each capsnap once */ - if (!kick && !list_empty(&capsnap->flushing_item)) { + if (capsnap->cap_flush.tid > 0) { dout("already flushed %p, skipping\n", capsnap); continue; } @@ -1282,8 +1296,6 @@ retry: if (session && session->s_mds != mds) { dout("oops, wrong session %p mutex\n", session); - if (kick) - goto out; mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); @@ -1309,26 +1321,27 @@ retry: } spin_lock(&mdsc->cap_dirty_lock); - capsnap->flush_tid = ++mdsc->last_cap_flush_tid; + capsnap->cap_flush.tid = ++mdsc->last_cap_flush_tid; + list_add_tail(&capsnap->cap_flush.g_list, + &mdsc->cap_flush_list); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + + if (list_empty(&ci->i_flushing_item)) { + list_add_tail(&ci->i_flushing_item, + &session->s_cap_flushing); + } spin_unlock(&mdsc->cap_dirty_lock); + list_add_tail(&capsnap->cap_flush.i_list, + &ci->i_cap_flush_list); + atomic_inc(&capsnap->nref); - if (list_empty(&capsnap->flushing_item)) - list_add_tail(&capsnap->flushing_item, - &session->s_cap_snaps_flushing); spin_unlock(&ci->i_ceph_lock); dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n", - inode, capsnap, capsnap->follows, capsnap->flush_tid); - send_cap_msg(session, ceph_vino(inode).ino, 0, - CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0, - capsnap->dirty, 0, capsnap->flush_tid, 0, - 0, mseq, capsnap->size, 0, - &capsnap->mtime, &capsnap->atime, - &capsnap->ctime, capsnap->time_warp_seq, - capsnap->uid, capsnap->gid, capsnap->mode, - capsnap->xattr_version, capsnap->xattr_blob, - capsnap->follows, capsnap->inline_data); + inode, capsnap, capsnap->follows, capsnap->cap_flush.tid); + __send_flush_snap(inode, session, capsnap, mseq, + oldest_flush_tid); next_follows = capsnap->follows + 1; ceph_put_cap_snap(capsnap); @@ -1354,7 +1367,7 @@ out: static void ceph_flush_snaps(struct ceph_inode_info *ci) { spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, NULL, 0); + __ceph_flush_snaps(ci, NULL); spin_unlock(&ci->i_ceph_lock); } @@ -1476,11 +1489,6 @@ static int __mark_caps_flushing(struct inode *inode, if (list_empty(&ci->i_flushing_item)) { list_add_tail(&ci->i_flushing_item, &session->s_cap_flushing); mdsc->num_cap_flushing++; - dout(" inode %p now flushing tid %llu\n", inode, cf->tid); - } else { - list_move_tail(&ci->i_flushing_item, &session->s_cap_flushing); - dout(" inode %p now flushing (more) tid %llu\n", - inode, cf->tid); } spin_unlock(&mdsc->cap_dirty_lock); @@ -1556,7 +1564,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, /* flush snaps first time around only */ if (!list_empty(&ci->i_cap_snaps)) - __ceph_flush_snaps(ci, &session, 0); + __ceph_flush_snaps(ci, &session); goto retry_locked; retry: spin_lock(&ci->i_ceph_lock); @@ -1997,80 +2005,74 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc) return err; } -/* - * After a recovering MDS goes active, we need to resend any caps - * we were flushing. - * - * Caller holds session->s_mutex. - */ -static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session) -{ - struct ceph_cap_snap *capsnap; - - dout("kick_flushing_capsnaps mds%d\n", session->s_mds); - list_for_each_entry(capsnap, &session->s_cap_snaps_flushing, - flushing_item) { - struct ceph_inode_info *ci = capsnap->ci; - struct inode *inode = &ci->vfs_inode; - struct ceph_cap *cap; - - spin_lock(&ci->i_ceph_lock); - cap = ci->i_auth_cap; - if (cap && cap->session == session) { - dout("kick_flushing_caps %p cap %p capsnap %p\n", inode, - cap, capsnap); - __ceph_flush_snaps(ci, &session, 1); - } else { - pr_err("%p auth cap %p not mds%d ???\n", inode, - cap, session->s_mds); - } - spin_unlock(&ci->i_ceph_lock); - } -} - -static int __kick_flushing_caps(struct ceph_mds_client *mdsc, - struct ceph_mds_session *session, - struct ceph_inode_info *ci) +static void __kick_flushing_caps(struct ceph_mds_client *mdsc, + struct ceph_mds_session *session, + struct ceph_inode_info *ci, + u64 oldest_flush_tid) + __releases(ci->i_ceph_lock) + __acquires(ci->i_ceph_lock) { struct inode *inode = &ci->vfs_inode; struct ceph_cap *cap; struct ceph_cap_flush *cf; - int delayed = 0; + int ret; u64 first_tid = 0; - u64 oldest_flush_tid; - - spin_lock(&mdsc->cap_dirty_lock); - oldest_flush_tid = __get_oldest_flush_tid(mdsc); - spin_unlock(&mdsc->cap_dirty_lock); - spin_lock(&ci->i_ceph_lock); list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { if (cf->tid < first_tid) continue; cap = ci->i_auth_cap; if (!(cap && cap->session == session)) { - pr_err("%p auth cap %p not mds%d ???\n", inode, - cap, session->s_mds); - spin_unlock(&ci->i_ceph_lock); + pr_err("%p auth cap %p not mds%d ???\n", + inode, cap, session->s_mds); break; } first_tid = cf->tid + 1; - dout("kick_flushing_caps %p cap %p tid %llu %s\n", inode, - cap, cf->tid, ceph_cap_string(cf->caps)); - delayed |= __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, - __ceph_caps_used(ci), - __ceph_caps_wanted(ci), - cap->issued | cap->implemented, - cf->caps, cf->tid, oldest_flush_tid); + if (cf->caps) { + dout("kick_flushing_caps %p cap %p tid %llu %s\n", + inode, cap, cf->tid, ceph_cap_string(cf->caps)); + ci->i_ceph_flags |= CEPH_I_NODELAY; + ret = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, + __ceph_caps_used(ci), + __ceph_caps_wanted(ci), + cap->issued | cap->implemented, + cf->caps, cf->tid, oldest_flush_tid); + if (ret) { + pr_err("kick_flushing_caps: error sending " + "cap flush, ino (%llx.%llx) " + "tid %llu flushing %s\n", + ceph_vinop(inode), cf->tid, + ceph_cap_string(cf->caps)); + } + } else { + struct ceph_cap_snap *capsnap = + container_of(cf, struct ceph_cap_snap, + cap_flush); + dout("kick_flushing_caps %p capsnap %p tid %llu %s\n", + inode, capsnap, cf->tid, + ceph_cap_string(capsnap->dirty)); + + atomic_inc(&capsnap->nref); + spin_unlock(&ci->i_ceph_lock); + + ret = __send_flush_snap(inode, session, capsnap, cap->mseq, + oldest_flush_tid); + if (ret < 0) { + pr_err("kick_flushing_caps: error sending " + "cap flushsnap, ino (%llx.%llx) " + "tid %llu follows %llu\n", + ceph_vinop(inode), cf->tid, + capsnap->follows); + } + + ceph_put_cap_snap(capsnap); + } spin_lock(&ci->i_ceph_lock); } - spin_unlock(&ci->i_ceph_lock); - return delayed; } void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, @@ -2078,8 +2080,14 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, { struct ceph_inode_info *ci; struct ceph_cap *cap; + u64 oldest_flush_tid; dout("early_kick_flushing_caps mds%d\n", session->s_mds); + + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; @@ -2099,10 +2107,8 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, */ if ((cap->issued & ci->i_flushing_caps) != ci->i_flushing_caps) { - spin_unlock(&ci->i_ceph_lock); - if (!__kick_flushing_caps(mdsc, session, ci)) - continue; - spin_lock(&ci->i_ceph_lock); + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); } spin_unlock(&ci->i_ceph_lock); @@ -2113,50 +2119,43 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) { struct ceph_inode_info *ci; - - kick_flushing_capsnaps(mdsc, session); + u64 oldest_flush_tid; dout("kick_flushing_caps mds%d\n", session->s_mds); + + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { - int delayed = __kick_flushing_caps(mdsc, session, ci); - if (delayed) { - spin_lock(&ci->i_ceph_lock); - __cap_delay_requeue(mdsc, ci); - spin_unlock(&ci->i_ceph_lock); - } + spin_lock(&ci->i_ceph_lock); + __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); + spin_unlock(&ci->i_ceph_lock); } } static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session, struct inode *inode) + __releases(ci->i_ceph_lock) { struct ceph_inode_info *ci = ceph_inode(inode); struct ceph_cap *cap; - spin_lock(&ci->i_ceph_lock); cap = ci->i_auth_cap; dout("kick_flushing_inode_caps %p flushing %s\n", inode, ceph_cap_string(ci->i_flushing_caps)); - __ceph_flush_snaps(ci, &session, 1); - - if (ci->i_flushing_caps) { - int delayed; - + if (!list_empty(&ci->i_cap_flush_list)) { + u64 oldest_flush_tid; spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &cap->session->s_cap_flushing); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); + __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); spin_unlock(&ci->i_ceph_lock); - - delayed = __kick_flushing_caps(mdsc, session, ci); - if (delayed) { - spin_lock(&ci->i_ceph_lock); - __cap_delay_requeue(mdsc, ci); - spin_unlock(&ci->i_ceph_lock); - } } else { spin_unlock(&ci->i_ceph_lock); } @@ -2487,12 +2486,11 @@ static int ceph_try_drop_cap_snap(struct ceph_cap_snap *capsnap) { if (!capsnap->need_flush && !capsnap->writing && !capsnap->dirty_pages) { - dout("dropping cap_snap %p follows %llu\n", capsnap, capsnap->follows); + BUG_ON(capsnap->cap_flush.tid > 0); ceph_put_snap_context(capsnap->context); list_del(&capsnap->ci_item); - list_del(&capsnap->flushing_item); ceph_put_cap_snap(capsnap); return 1; } @@ -2891,13 +2889,13 @@ static void handle_cap_grant(struct ceph_mds_client *mdsc, fill_inline = true; } - spin_unlock(&ci->i_ceph_lock); - if (le32_to_cpu(grant->op) == CEPH_CAP_OP_IMPORT) { - kick_flushing_inode_caps(mdsc, session, inode); - up_read(&mdsc->snap_rwsem); if (newcaps & ~issued) wake = true; + kick_flushing_inode_caps(mdsc, session, inode); + up_read(&mdsc->snap_rwsem); + } else { + spin_unlock(&ci->i_ceph_lock); } if (fill_inline) @@ -2951,6 +2949,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { if (cf->tid == flush_tid) cleaned = cf->caps; + if (cf->caps == 0) /* capsnap */ + continue; if (cf->tid <= flush_tid) { list_del(&cf->i_list); list_add_tail(&cf->i_list, &to_remove); @@ -2985,13 +2985,16 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, } if (ci->i_flushing_caps == 0) { - list_del_init(&ci->i_flushing_item); - if (!list_empty(&session->s_cap_flushing)) - dout(" mds%d still flushing cap on %p\n", - session->s_mds, - &list_entry(session->s_cap_flushing.next, - struct ceph_inode_info, - i_flushing_item)->vfs_inode); + if (list_empty(&ci->i_cap_flush_list)) { + list_del_init(&ci->i_flushing_item); + if (!list_empty(&session->s_cap_flushing)) { + dout(" mds%d still flushing cap on %p\n", + session->s_mds, + &list_first_entry(&session->s_cap_flushing, + struct ceph_inode_info, + i_flushing_item)->vfs_inode); + } + } mdsc->num_cap_flushing--; dout(" inode %p now !flushing\n", inode); @@ -3039,7 +3042,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; u64 follows = le64_to_cpu(m->snap_follows); struct ceph_cap_snap *capsnap; - int drop = 0; + int flushed = 0; dout("handle_cap_flushsnap_ack inode %p ci %p mds%d follows %lld\n", inode, ci, session->s_mds, follows); @@ -3047,30 +3050,47 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, spin_lock(&ci->i_ceph_lock); list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->follows == follows) { - if (capsnap->flush_tid != flush_tid) { + if (capsnap->cap_flush.tid != flush_tid) { dout(" cap_snap %p follows %lld tid %lld !=" " %lld\n", capsnap, follows, - flush_tid, capsnap->flush_tid); + flush_tid, capsnap->cap_flush.tid); break; } - WARN_ON(capsnap->dirty_pages || capsnap->writing); - dout(" removing %p cap_snap %p follows %lld\n", - inode, capsnap, follows); - ceph_put_snap_context(capsnap->context); - list_del(&capsnap->ci_item); - list_del(&capsnap->flushing_item); - ceph_put_cap_snap(capsnap); - wake_up_all(&mdsc->cap_flushing_wq); - drop = 1; + flushed = 1; break; } else { dout(" skipping cap_snap %p follows %lld\n", capsnap, capsnap->follows); } } + if (flushed) { + u64 oldest_flush_tid; + WARN_ON(capsnap->dirty_pages || capsnap->writing); + dout(" removing %p cap_snap %p follows %lld\n", + inode, capsnap, follows); + list_del(&capsnap->ci_item); + list_del(&capsnap->cap_flush.i_list); + + spin_lock(&mdsc->cap_dirty_lock); + + if (list_empty(&ci->i_cap_flush_list)) + list_del_init(&ci->i_flushing_item); + + list_del(&capsnap->cap_flush.g_list); + + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + if (oldest_flush_tid == 0 || oldest_flush_tid > flush_tid) + wake_up_all(&mdsc->cap_flushing_wq); + + spin_unlock(&mdsc->cap_dirty_lock); + wake_up_all(&ci->i_cap_wq); + } spin_unlock(&ci->i_ceph_lock); - if (drop) + if (flushed) { + ceph_put_snap_context(capsnap->context); + ceph_put_cap_snap(capsnap); iput(inode); + } } /* @@ -3175,7 +3195,8 @@ retry: tcap->implemented |= issued; if (cap == ci->i_auth_cap) ci->i_auth_cap = tcap; - if (ci->i_flushing_caps && ci->i_auth_cap == tcap) { + if (!list_empty(&ci->i_cap_flush_list) && + ci->i_auth_cap == tcap) { spin_lock(&mdsc->cap_dirty_lock); list_move_tail(&ci->i_flushing_item, &tcap->session->s_cap_flushing); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 7cd6b861c2f3..fa9036af5445 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -472,7 +472,6 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc, s->s_cap_iterator = NULL; INIT_LIST_HEAD(&s->s_cap_releases); INIT_LIST_HEAD(&s->s_cap_flushing); - INIT_LIST_HEAD(&s->s_cap_snaps_flushing); dout("register_session mds%d\n", mds); if (mds >= mdsc->max_sessions) { @@ -1479,21 +1478,6 @@ static int trim_caps(struct ceph_mds_client *mdsc, return 0; } -static int check_capsnap_flush(struct ceph_inode_info *ci, - u64 want_snap_seq) -{ - int ret = 1; - spin_lock(&ci->i_ceph_lock); - if (want_snap_seq > 0 && !list_empty(&ci->i_cap_snaps)) { - struct ceph_cap_snap *capsnap = - list_first_entry(&ci->i_cap_snaps, - struct ceph_cap_snap, ci_item); - ret = capsnap->follows >= want_snap_seq; - } - spin_unlock(&ci->i_ceph_lock); - return ret; -} - static int check_caps_flush(struct ceph_mds_client *mdsc, u64 want_flush_tid) { @@ -1520,54 +1504,9 @@ static int check_caps_flush(struct ceph_mds_client *mdsc, * returns true if we've flushed through want_flush_tid */ static void wait_caps_flush(struct ceph_mds_client *mdsc, - u64 want_flush_tid, u64 want_snap_seq) + u64 want_flush_tid) { - int mds; - - dout("check_caps_flush want %llu snap want %llu\n", - want_flush_tid, want_snap_seq); - mutex_lock(&mdsc->mutex); - for (mds = 0; mds < mdsc->max_sessions; ) { - struct ceph_mds_session *session = mdsc->sessions[mds]; - struct inode *inode = NULL; - - if (!session) { - mds++; - continue; - } - get_session(session); - mutex_unlock(&mdsc->mutex); - - mutex_lock(&session->s_mutex); - if (!list_empty(&session->s_cap_snaps_flushing)) { - struct ceph_cap_snap *capsnap = - list_first_entry(&session->s_cap_snaps_flushing, - struct ceph_cap_snap, - flushing_item); - struct ceph_inode_info *ci = capsnap->ci; - if (!check_capsnap_flush(ci, want_snap_seq)) { - dout("check_cap_flush still flushing snap %p " - "follows %lld <= %lld to mds%d\n", - &ci->vfs_inode, capsnap->follows, - want_snap_seq, mds); - inode = igrab(&ci->vfs_inode); - } - } - mutex_unlock(&session->s_mutex); - ceph_put_mds_session(session); - - if (inode) { - wait_event(mdsc->cap_flushing_wq, - check_capsnap_flush(ceph_inode(inode), - want_snap_seq)); - iput(inode); - } else { - mds++; - } - - mutex_lock(&mdsc->mutex); - } - mutex_unlock(&mdsc->mutex); + dout("check_caps_flush want %llu\n", want_flush_tid); wait_event(mdsc->cap_flushing_wq, check_caps_flush(mdsc, want_flush_tid)); @@ -3584,7 +3523,7 @@ restart: void ceph_mdsc_sync(struct ceph_mds_client *mdsc) { - u64 want_tid, want_flush, want_snap; + u64 want_tid, want_flush; if (ACCESS_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) return; @@ -3599,15 +3538,11 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc) want_flush = mdsc->last_cap_flush_tid; spin_unlock(&mdsc->cap_dirty_lock); - down_read(&mdsc->snap_rwsem); - want_snap = mdsc->last_snap_seq; - up_read(&mdsc->snap_rwsem); - - dout("sync want tid %lld flush_seq %lld snap_seq %lld\n", - want_tid, want_flush, want_snap); + dout("sync want tid %lld flush_seq %lld\n", + want_tid, want_flush); wait_unsafe_requests(mdsc, want_tid); - wait_caps_flush(mdsc, want_flush, want_snap); + wait_caps_flush(mdsc, want_flush); } /* diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index 93170b4b5d75..6b3679737d4a 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -152,7 +152,6 @@ struct ceph_mds_session { /* protected by mutex */ struct list_head s_cap_flushing; /* inodes w/ flushing caps */ - struct list_head s_cap_snaps_flushing; unsigned long s_renew_requested; /* last time we sent a renew req */ u64 s_renew_seq; diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index eadf2c33edc6..20d5b0cdf655 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -520,9 +520,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci) ihold(inode); atomic_set(&capsnap->nref, 1); - capsnap->ci = ci; INIT_LIST_HEAD(&capsnap->ci_item); - INIT_LIST_HEAD(&capsnap->flushing_item); capsnap->follows = old_snapc->seq; capsnap->issued = __ceph_caps_issued(ci, NULL); @@ -800,7 +798,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc) ihold(inode); spin_unlock(&mdsc->snap_flush_lock); spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, &session, 0); + __ceph_flush_snaps(ci, &session); spin_unlock(&ci->i_ceph_lock); iput(inode); spin_lock(&mdsc->snap_flush_lock); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 29e8b7bd9413..08ed51299f9f 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -147,6 +147,13 @@ struct ceph_cap { #define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */ #define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */ +struct ceph_cap_flush { + u64 tid; + int caps; /* 0 means capsnap */ + struct list_head g_list; // global + struct list_head i_list; // per inode +}; + /* * Snapped cap state that is pending flush to mds. When a snapshot occurs, * we first complete any in-process sync writes and writeback any dirty @@ -154,10 +161,11 @@ struct ceph_cap { */ struct ceph_cap_snap { atomic_t nref; - struct ceph_inode_info *ci; - struct list_head ci_item, flushing_item; + struct list_head ci_item; - u64 follows, flush_tid; + struct ceph_cap_flush cap_flush; + + u64 follows; int issued, dirty; struct ceph_snap_context *context; @@ -186,13 +194,6 @@ static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap) } } -struct ceph_cap_flush { - u64 tid; - int caps; - struct list_head g_list; // global - struct list_head i_list; // per inode -}; - /* * The frag tree describes how a directory is fragmented, potentially across * multiple metadata servers. It is also used to indicate points where @@ -888,8 +889,7 @@ extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had); extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc); extern void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession, - int again); + struct ceph_mds_session **psession); extern void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session); extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc); -- cgit v1.2.3 From 13c2b57d81ec27716b9c943fd4077264b9804e55 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 5 Jul 2016 16:45:21 +0800 Subject: ceph: avoid sending duplicated cap flush message make ceph_kick_flushing_caps() ignore inodes whose cap flushes have already been re-sent by ceph_early_kick_flushing_caps() Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 18 +++++++++++++++++- fs/ceph/super.h | 1 + 2 files changed, 18 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 0ac604719663..f12d59d26a04 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2107,8 +2107,11 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc, */ if ((cap->issued & ci->i_flushing_caps) != ci->i_flushing_caps) { + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); + } else { + ci->i_ceph_flags |= CEPH_I_KICK_FLUSH; } spin_unlock(&ci->i_ceph_lock); @@ -2119,6 +2122,7 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, struct ceph_mds_session *session) { struct ceph_inode_info *ci; + struct ceph_cap *cap; u64 oldest_flush_tid; dout("kick_flushing_caps mds%d\n", session->s_mds); @@ -2129,7 +2133,18 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc, list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) { spin_lock(&ci->i_ceph_lock); - __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); + cap = ci->i_auth_cap; + if (!(cap && cap->session == session)) { + pr_err("%p auth cap %p not mds%d ???\n", + &ci->vfs_inode, cap, session->s_mds); + spin_unlock(&ci->i_ceph_lock); + continue; + } + if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) { + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); + } spin_unlock(&ci->i_ceph_lock); } } @@ -2154,6 +2169,7 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc, oldest_flush_tid = __get_oldest_flush_tid(mdsc); spin_unlock(&mdsc->cap_dirty_lock); + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; __kick_flushing_caps(mdsc, session, ci, oldest_flush_tid); spin_unlock(&ci->i_ceph_lock); } else { diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 08ed51299f9f..16068787afb4 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -468,6 +468,7 @@ static inline struct inode *ceph_find_inode(struct super_block *sb, #define CEPH_I_POOL_WR (1 << 6) /* can write to pool */ #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */ #define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */ +#define CEPH_I_KICK_FLUSH (1 << 9) /* kick flushing caps */ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci, long long release_count, -- cgit v1.2.3 From 70220ac8c220495b2a335868293be80a31dfdd4a Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Wed, 6 Jul 2016 16:21:30 +0800 Subject: ceph: introduce an inode flag to indicates if snapflush is needed Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 48 +++++++++++++++++++++++++++++++++--------------- fs/ceph/snap.c | 2 ++ fs/ceph/super.h | 1 + 3 files changed, 36 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index f12d59d26a04..45fe7a3658dc 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1368,6 +1368,7 @@ static void ceph_flush_snaps(struct ceph_inode_info *ci) { spin_lock(&ci->i_ceph_lock); __ceph_flush_snaps(ci, NULL); + ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; spin_unlock(&ci->i_ceph_lock); } @@ -1563,8 +1564,10 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, flags |= CHECK_CAPS_FLUSH; /* flush snaps first time around only */ - if (!list_empty(&ci->i_cap_snaps)) + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { __ceph_flush_snaps(ci, &session); + ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; + } goto retry_locked; retry: spin_lock(&ci->i_ceph_lock); @@ -2498,7 +2501,8 @@ void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps) * drop cap_snap that is not associated with any snapshot. * we don't need to send FLUSHSNAP message for it. */ -static int ceph_try_drop_cap_snap(struct ceph_cap_snap *capsnap) +static int ceph_try_drop_cap_snap(struct ceph_inode_info *ci, + struct ceph_cap_snap *capsnap) { if (!capsnap->need_flush && !capsnap->writing && !capsnap->dirty_pages) { @@ -2506,6 +2510,9 @@ static int ceph_try_drop_cap_snap(struct ceph_cap_snap *capsnap) capsnap, capsnap->follows); BUG_ON(capsnap->cap_flush.tid > 0); ceph_put_snap_context(capsnap->context); + if (!list_is_last(&capsnap->ci_item, &ci->i_cap_snaps)) + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; + list_del(&capsnap->ci_item); ceph_put_cap_snap(capsnap); return 1; @@ -2553,7 +2560,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) struct ceph_cap_snap, ci_item); capsnap->writing = 0; - if (ceph_try_drop_cap_snap(capsnap)) + if (ceph_try_drop_cap_snap(ci, capsnap)) put++; else if (__ceph_finish_cap_snap(ci, capsnap)) flushsnaps = 1; @@ -2596,15 +2603,19 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc) { struct inode *inode = &ci->vfs_inode; - int last = 0; - int complete_capsnap = 0; - int drop_capsnap = 0; - int found = 0; struct ceph_cap_snap *capsnap = NULL; + int put = 0; + bool last = false; + bool found = false; + bool flush_snaps = false; + bool complete_capsnap = false; spin_lock(&ci->i_ceph_lock); ci->i_wrbuffer_ref -= nr; - last = !ci->i_wrbuffer_ref; + if (ci->i_wrbuffer_ref == 0) { + last = true; + put++; + } if (ci->i_head_snapc == snapc) { ci->i_wrbuffer_ref_head -= nr; @@ -2624,15 +2635,22 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, } else { list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { if (capsnap->context == snapc) { - found = 1; + found = true; break; } } BUG_ON(!found); capsnap->dirty_pages -= nr; if (capsnap->dirty_pages == 0) { - complete_capsnap = 1; - drop_capsnap = ceph_try_drop_cap_snap(capsnap); + complete_capsnap = true; + if (!capsnap->writing) { + if (ceph_try_drop_cap_snap(ci, capsnap)) { + put++; + } else { + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; + flush_snaps = true; + } + } } dout("put_wrbuffer_cap_refs on %p cap_snap %p " " snap %lld %d/%d -> %d/%d %s%s\n", @@ -2647,12 +2665,12 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, if (last) { ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); - iput(inode); - } else if (complete_capsnap) { + } else if (flush_snaps) { ceph_flush_snaps(ci); - wake_up_all(&ci->i_cap_wq); } - if (drop_capsnap) + if (complete_capsnap) + wake_up_all(&ci->i_cap_wq); + while (put-- > 0) iput(inode); } diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 20d5b0cdf655..c3b03ae1976c 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -601,6 +601,8 @@ int __ceph_finish_cap_snap(struct ceph_inode_info *ci, capsnap->dirty_pages); return 0; } + + ci->i_ceph_flags |= CEPH_I_FLUSH_SNAPS; dout("finish_cap_snap %p cap_snap %p snapc %p %llu %s s=%llu\n", inode, capsnap, capsnap->context, capsnap->context->seq, ceph_cap_string(capsnap->dirty), diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 16068787afb4..63fdb57606fe 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -469,6 +469,7 @@ static inline struct inode *ceph_find_inode(struct super_block *sb, #define CEPH_I_SEC_INITED (1 << 7) /* security initialized */ #define CEPH_I_CAP_DROPPED (1 << 8) /* caps were forcibly dropped */ #define CEPH_I_KICK_FLUSH (1 << 9) /* kick flushing caps */ +#define CEPH_I_FLUSH_SNAPS (1 << 10) /* need flush snapss */ static inline void __ceph_dir_set_complete(struct ceph_inode_info *ci, long long release_count, -- cgit v1.2.3 From 7bc00fddb9de7f78f742bc24d95e15abde15c078 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 7 Jul 2016 18:34:45 +0800 Subject: ceph: kick cap flushes before sending other cap message If ceph_check_caps() wants to send cap message to a recovering MDS, make sure it kicks cap flushes first. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 43 ++++++++++++++++++++++++++++++++++--------- 1 file changed, 34 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 45fe7a3658dc..39e471d0aa50 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -41,6 +41,10 @@ */ static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc); +static void __kick_flushing_caps(struct ceph_mds_client *mdsc, + struct ceph_mds_session *session, + struct ceph_inode_info *ci, + u64 oldest_flush_tid); /* * Generate readable cap strings for debugging output. @@ -1563,11 +1567,6 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags, if (ci->i_ceph_flags & CEPH_I_FLUSH) flags |= CHECK_CAPS_FLUSH; - /* flush snaps first time around only */ - if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { - __ceph_flush_snaps(ci, &session); - ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; - } goto retry_locked; retry: spin_lock(&ci->i_ceph_lock); @@ -1688,10 +1687,15 @@ retry_locked: } } /* flush anything dirty? */ - if (cap == ci->i_auth_cap && (flags & CHECK_CAPS_FLUSH) && - ci->i_dirty_caps) { - dout("flushing dirty caps\n"); - goto ack; + if (cap == ci->i_auth_cap) { + if ((flags & CHECK_CAPS_FLUSH) && ci->i_dirty_caps) { + dout("flushing dirty caps\n"); + goto ack; + } + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { + dout("flushing snap caps\n"); + goto ack; + } } /* completed revocation? going down and there are no caps? */ @@ -1750,6 +1754,27 @@ ack: goto retry; } } + + /* kick flushing and flush snaps before sending normal + * cap message */ + if (cap == ci->i_auth_cap && + (ci->i_ceph_flags & + (CEPH_I_KICK_FLUSH | CEPH_I_FLUSH_SNAPS))) { + if (ci->i_ceph_flags & CEPH_I_KICK_FLUSH) { + spin_lock(&mdsc->cap_dirty_lock); + oldest_flush_tid = __get_oldest_flush_tid(mdsc); + spin_unlock(&mdsc->cap_dirty_lock); + __kick_flushing_caps(mdsc, session, ci, + oldest_flush_tid); + ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; + } + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { + __ceph_flush_snaps(ci, &session); + ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; + } + goto retry_locked; + } + /* take snap_rwsem after session mutex */ if (!took_snap_rwsem) { if (down_read_trylock(&mdsc->snap_rwsem) == 0) { -- cgit v1.2.3 From ed9b430c9ba99e70e8ddd7e08429c4c2a620ba74 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Tue, 5 Jul 2016 21:08:07 +0800 Subject: ceph: cleanup ceph_flush_snaps() This patch devide __ceph_flush_snaps() into two stags. In the first stage, __ceph_flush_snaps() assign snapcaps flush TIDs and add them to cap flush lists. __ceph_flush_snaps() keeps holding the i_ceph_lock in this stagge. So inode's auth cap can not change. In the second stage, __ceph_flush_snaps() send flushsnap cap messages. i_ceph_lock is unlocked before sending each cap message. If auth cap changes in the middle, __ceph_flush_snaps() just stops. This is OK because kick_flushing_inode_caps() will re-send flushsnap cap messages to inode's new auth MDS. Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 185 +++++++++++++++++++++++++++++++------------------------- fs/ceph/snap.c | 4 +- fs/ceph/super.h | 4 +- 3 files changed, 105 insertions(+), 88 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 39e471d0aa50..736e1c86bcf3 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1247,32 +1247,20 @@ static inline int __send_flush_snap(struct inode *inode, * * Called under i_ceph_lock. Takes s_mutex as needed. */ -void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession) +static void __ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session *session) __releases(ci->i_ceph_lock) __acquires(ci->i_ceph_lock) { struct inode *inode = &ci->vfs_inode; - int mds; + struct ceph_mds_client *mdsc = session->s_mdsc; struct ceph_cap_snap *capsnap; - u32 mseq; - struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; - struct ceph_mds_session *session = NULL; /* if session != NULL, we hold - session->s_mutex */ - u64 oldest_flush_tid; - u64 next_follows = 0; /* keep track of how far we've gotten through the - i_cap_snaps list, and skip these entries next time - around to avoid an infinite loop */ + u64 oldest_flush_tid = 0; + u64 first_tid = 1, last_tid = 0; - if (psession) - session = *psession; + dout("__flush_snaps %p session %p\n", inode, session); - dout("__flush_snaps %p\n", inode); -retry: list_for_each_entry(capsnap, &ci->i_cap_snaps, ci_item) { - /* avoid an infiniute loop after retry */ - if (capsnap->follows < next_follows) - continue; /* * we need to wait for sync writes to complete and for dirty * pages to be written out. @@ -1283,53 +1271,18 @@ retry: /* should be removed by ceph_try_drop_cap_snap() */ BUG_ON(!capsnap->need_flush); - /* pick mds, take s_mutex */ - if (ci->i_auth_cap == NULL) { - dout("no auth cap (migrating?), doing nothing\n"); - goto out; - } - /* only flush each capsnap once */ if (capsnap->cap_flush.tid > 0) { - dout("already flushed %p, skipping\n", capsnap); + dout(" already flushed %p, skipping\n", capsnap); continue; } - mds = ci->i_auth_cap->session->s_mds; - mseq = ci->i_auth_cap->mseq; - - if (session && session->s_mds != mds) { - dout("oops, wrong session %p mutex\n", session); - - mutex_unlock(&session->s_mutex); - ceph_put_mds_session(session); - session = NULL; - } - if (!session) { - spin_unlock(&ci->i_ceph_lock); - mutex_lock(&mdsc->mutex); - session = __ceph_lookup_mds_session(mdsc, mds); - mutex_unlock(&mdsc->mutex); - if (session) { - dout("inverting session/ino locks on %p\n", - session); - mutex_lock(&session->s_mutex); - } - /* - * if session == NULL, we raced against a cap - * deletion or migration. retry, and we'll - * get a better @mds value next time. - */ - spin_lock(&ci->i_ceph_lock); - goto retry; - } - spin_lock(&mdsc->cap_dirty_lock); capsnap->cap_flush.tid = ++mdsc->last_cap_flush_tid; list_add_tail(&capsnap->cap_flush.g_list, &mdsc->cap_flush_list); - oldest_flush_tid = __get_oldest_flush_tid(mdsc); - + if (oldest_flush_tid == 0) + oldest_flush_tid = __get_oldest_flush_tid(mdsc); if (list_empty(&ci->i_flushing_item)) { list_add_tail(&ci->i_flushing_item, &session->s_cap_flushing); @@ -1339,41 +1292,108 @@ retry: list_add_tail(&capsnap->cap_flush.i_list, &ci->i_cap_flush_list); + if (first_tid == 1) + first_tid = capsnap->cap_flush.tid; + last_tid = capsnap->cap_flush.tid; + } + + ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; + + while (first_tid <= last_tid) { + struct ceph_cap *cap = ci->i_auth_cap; + struct ceph_cap_flush *cf; + int ret; + + if (!(cap && cap->session == session)) { + dout("__flush_snaps %p auth cap %p not mds%d, " + "stop\n", inode, cap, session->s_mds); + break; + } + + ret = -ENOENT; + list_for_each_entry(cf, &ci->i_cap_flush_list, i_list) { + if (cf->tid >= first_tid) { + ret = 0; + break; + } + } + if (ret < 0) + break; + + first_tid = cf->tid + 1; + + capsnap = container_of(cf, struct ceph_cap_snap, cap_flush); atomic_inc(&capsnap->nref); spin_unlock(&ci->i_ceph_lock); - dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n", - inode, capsnap, capsnap->follows, capsnap->cap_flush.tid); - __send_flush_snap(inode, session, capsnap, mseq, - oldest_flush_tid); + dout("__flush_snaps %p capsnap %p tid %llu %s\n", + inode, capsnap, cf->tid, ceph_cap_string(capsnap->dirty)); - next_follows = capsnap->follows + 1; - ceph_put_cap_snap(capsnap); + ret = __send_flush_snap(inode, session, capsnap, cap->mseq, + oldest_flush_tid); + if (ret < 0) { + pr_err("__flush_snaps: error sending cap flushsnap, " + "ino (%llx.%llx) tid %llu follows %llu\n", + ceph_vinop(inode), cf->tid, capsnap->follows); + } + ceph_put_cap_snap(capsnap); spin_lock(&ci->i_ceph_lock); - goto retry; } +} - /* we flushed them all; remove this inode from the queue */ - spin_lock(&mdsc->snap_flush_lock); - list_del_init(&ci->i_snap_flush_item); - spin_unlock(&mdsc->snap_flush_lock); +void ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session **psession) +{ + struct inode *inode = &ci->vfs_inode; + struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc; + struct ceph_mds_session *session = *psession; + int mds; + dout("ceph_flush_snaps %p\n", inode); +retry: + spin_lock(&ci->i_ceph_lock); + if (!(ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS)) { + dout(" no capsnap needs flush, doing nothing\n"); + goto out; + } + if (!ci->i_auth_cap) { + dout(" no auth cap (migrating?), doing nothing\n"); + goto out; + } -out: - if (psession) - *psession = session; - else if (session) { + mds = ci->i_auth_cap->session->s_mds; + if (session && session->s_mds != mds) { + dout(" oops, wrong session %p mutex\n", session); mutex_unlock(&session->s_mutex); ceph_put_mds_session(session); + session = NULL; + } + if (!session) { + spin_unlock(&ci->i_ceph_lock); + mutex_lock(&mdsc->mutex); + session = __ceph_lookup_mds_session(mdsc, mds); + mutex_unlock(&mdsc->mutex); + if (session) { + dout(" inverting session/ino locks on %p\n", session); + mutex_lock(&session->s_mutex); + } + goto retry; } -} -static void ceph_flush_snaps(struct ceph_inode_info *ci) -{ - spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, NULL); - ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; + __ceph_flush_snaps(ci, session); +out: spin_unlock(&ci->i_ceph_lock); + + if (psession) { + *psession = session; + } else { + mutex_unlock(&session->s_mutex); + ceph_put_mds_session(session); + } + /* we flushed them all; remove this inode from the queue */ + spin_lock(&mdsc->snap_flush_lock); + list_del_init(&ci->i_snap_flush_item); + spin_unlock(&mdsc->snap_flush_lock); } /* @@ -1768,10 +1788,9 @@ ack: oldest_flush_tid); ci->i_ceph_flags &= ~CEPH_I_KICK_FLUSH; } - if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) { - __ceph_flush_snaps(ci, &session); - ci->i_ceph_flags &= ~CEPH_I_FLUSH_SNAPS; - } + if (ci->i_ceph_flags & CEPH_I_FLUSH_SNAPS) + __ceph_flush_snaps(ci, session); + goto retry_locked; } @@ -2610,7 +2629,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had) if (last && !flushsnaps) ceph_check_caps(ci, 0, NULL); else if (flushsnaps) - ceph_flush_snaps(ci); + ceph_flush_snaps(ci, NULL); if (wake) wake_up_all(&ci->i_cap_wq); while (put-- > 0) @@ -2691,7 +2710,7 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, if (last) { ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL); } else if (flush_snaps) { - ceph_flush_snaps(ci); + ceph_flush_snaps(ci, NULL); } if (complete_capsnap) wake_up_all(&ci->i_cap_wq); diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index c3b03ae1976c..9ff5219d849e 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -799,9 +799,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc) inode = &ci->vfs_inode; ihold(inode); spin_unlock(&mdsc->snap_flush_lock); - spin_lock(&ci->i_ceph_lock); - __ceph_flush_snaps(ci, &session); - spin_unlock(&ci->i_ceph_lock); + ceph_flush_snaps(ci, &session); iput(inode); spin_lock(&mdsc->snap_flush_lock); } diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 63fdb57606fe..b097d474f888 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -890,8 +890,8 @@ extern void ceph_get_cap_refs(struct ceph_inode_info *ci, int caps); extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had); extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr, struct ceph_snap_context *snapc); -extern void __ceph_flush_snaps(struct ceph_inode_info *ci, - struct ceph_mds_session **psession); +extern void ceph_flush_snaps(struct ceph_inode_info *ci, + struct ceph_mds_session **psession); extern void ceph_check_caps(struct ceph_inode_info *ci, int flags, struct ceph_mds_session *session); extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc); -- cgit v1.2.3 From c8799fc4674fe5bb9b9391f9eac202250b8370e1 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Thu, 7 Jul 2016 15:22:38 +0800 Subject: ceph: optimize cap flush waiting Add a 'wake' flag to ceph_cap_flush struct, which indicates if there is someone waiting for it to finish. When getting flush ack message, we check the 'wake' flag in corresponding ceph_cap_flush struct to decide if we should wake up waiters. One corner case is that the acked cap flush has 'wake' flags is set, but it is not the first one on the flushing list. We do not wake up waiters in this case, set 'wake' flags of preceding ceph_cap_flush struct instead Signed-off-by: Yan, Zheng --- fs/ceph/caps.c | 91 ++++++++++++++++++++++++++++++++++++---------------- fs/ceph/mds_client.c | 8 +++++ fs/ceph/super.h | 1 + 3 files changed, 73 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 736e1c86bcf3..99115cae1652 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -1473,6 +1473,37 @@ static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) return 0; } +/* + * Remove cap_flush from the mdsc's or inode's flushing cap list. + * Return true if caller needs to wake up flush waiters. + */ +static bool __finish_cap_flush(struct ceph_mds_client *mdsc, + struct ceph_inode_info *ci, + struct ceph_cap_flush *cf) +{ + struct ceph_cap_flush *prev; + bool wake = cf->wake; + if (mdsc) { + /* are there older pending cap flushes? */ + if (wake && cf->g_list.prev != &mdsc->cap_flush_list) { + prev = list_prev_entry(cf, g_list); + prev->wake = true; + wake = false; + } + list_del(&cf->g_list); + } else if (ci) { + if (wake && cf->i_list.prev != &ci->i_cap_flush_list) { + prev = list_prev_entry(cf, i_list); + prev->wake = true; + wake = false; + } + list_del(&cf->i_list); + } else { + BUG_ON(1); + } + return wake; +} + /* * Add dirty inode to the flushing list. Assigned a seq number so we * can wait for caps to flush without starving. @@ -1480,7 +1511,7 @@ static u64 __get_oldest_flush_tid(struct ceph_mds_client *mdsc) * Called under i_ceph_lock. */ static int __mark_caps_flushing(struct inode *inode, - struct ceph_mds_session *session, + struct ceph_mds_session *session, bool wake, u64 *flush_tid, u64 *oldest_flush_tid) { struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; @@ -1503,6 +1534,7 @@ static int __mark_caps_flushing(struct inode *inode, swap(cf, ci->i_prealloc_cap_flush); cf->caps = flushing; + cf->wake = wake; spin_lock(&mdsc->cap_dirty_lock); list_del_init(&ci->i_dirty_item); @@ -1808,7 +1840,7 @@ ack: } if (cap == ci->i_auth_cap && ci->i_dirty_caps) { - flushing = __mark_caps_flushing(inode, session, + flushing = __mark_caps_flushing(inode, session, false, &flush_tid, &oldest_flush_tid); } else { @@ -1885,8 +1917,8 @@ retry: if (cap->session->s_state < CEPH_MDS_SESSION_OPEN) goto out; - flushing = __mark_caps_flushing(inode, session, &flush_tid, - &oldest_flush_tid); + flushing = __mark_caps_flushing(inode, session, true, + &flush_tid, &oldest_flush_tid); /* __send_cap drops i_ceph_lock */ delayed = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, used, want, @@ -1902,7 +1934,8 @@ retry: if (!list_empty(&ci->i_cap_flush_list)) { struct ceph_cap_flush *cf = list_last_entry(&ci->i_cap_flush_list, - struct ceph_cap_flush, i_list); + struct ceph_cap_flush, i_list); + cf->wake = true; flush_tid = cf->tid; } flushing = ci->i_flushing_caps; @@ -3022,7 +3055,9 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, unsigned seq = le32_to_cpu(m->seq); int dirty = le32_to_cpu(m->dirty); int cleaned = 0; - int drop = 0; + bool drop = false; + bool wake_ci = 0; + bool wake_mdsc = 0; list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) { if (cf->tid == flush_tid) @@ -3030,7 +3065,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, if (cf->caps == 0) /* capsnap */ continue; if (cf->tid <= flush_tid) { - list_del(&cf->i_list); + if (__finish_cap_flush(NULL, ci, cf)) + wake_ci = true; list_add_tail(&cf->i_list, &to_remove); } else { cleaned &= ~cf->caps; @@ -3052,14 +3088,9 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, spin_lock(&mdsc->cap_dirty_lock); - if (!list_empty(&to_remove)) { - u64 oldest_flush_tid; - list_for_each_entry(cf, &to_remove, i_list) - list_del(&cf->g_list); - - oldest_flush_tid = __get_oldest_flush_tid(mdsc); - if (oldest_flush_tid == 0 || oldest_flush_tid > flush_tid) - wake_up_all(&mdsc->cap_flushing_wq); + list_for_each_entry(cf, &to_remove, i_list) { + if (__finish_cap_flush(mdsc, NULL, cf)) + wake_mdsc = true; } if (ci->i_flushing_caps == 0) { @@ -3079,7 +3110,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, if (ci->i_dirty_caps == 0) { dout(" inode %p now clean\n", inode); BUG_ON(!list_empty(&ci->i_dirty_item)); - drop = 1; + drop = true; if (ci->i_wr_ref == 0 && ci->i_wrbuffer_ref_head == 0) { BUG_ON(!ci->i_head_snapc); @@ -3091,7 +3122,6 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid, } } spin_unlock(&mdsc->cap_dirty_lock); - wake_up_all(&ci->i_cap_wq); out: spin_unlock(&ci->i_ceph_lock); @@ -3102,6 +3132,11 @@ out: list_del(&cf->i_list); ceph_free_cap_flush(cf); } + + if (wake_ci) + wake_up_all(&ci->i_cap_wq); + if (wake_mdsc) + wake_up_all(&mdsc->cap_flushing_wq); if (drop) iput(inode); } @@ -3120,7 +3155,9 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; u64 follows = le64_to_cpu(m->snap_follows); struct ceph_cap_snap *capsnap; - int flushed = 0; + bool flushed = false; + bool wake_ci = false; + bool wake_mdsc = false; dout("handle_cap_flushsnap_ack inode %p ci %p mds%d follows %lld\n", inode, ci, session->s_mds, follows); @@ -3134,7 +3171,7 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, flush_tid, capsnap->cap_flush.tid); break; } - flushed = 1; + flushed = true; break; } else { dout(" skipping cap_snap %p follows %lld\n", @@ -3142,31 +3179,31 @@ static void handle_cap_flushsnap_ack(struct inode *inode, u64 flush_tid, } } if (flushed) { - u64 oldest_flush_tid; WARN_ON(capsnap->dirty_pages || capsnap->writing); dout(" removing %p cap_snap %p follows %lld\n", inode, capsnap, follows); list_del(&capsnap->ci_item); - list_del(&capsnap->cap_flush.i_list); + if (__finish_cap_flush(NULL, ci, &capsnap->cap_flush)) + wake_ci = true; spin_lock(&mdsc->cap_dirty_lock); if (list_empty(&ci->i_cap_flush_list)) list_del_init(&ci->i_flushing_item); - list_del(&capsnap->cap_flush.g_list); - - oldest_flush_tid = __get_oldest_flush_tid(mdsc); - if (oldest_flush_tid == 0 || oldest_flush_tid > flush_tid) - wake_up_all(&mdsc->cap_flushing_wq); + if (__finish_cap_flush(mdsc, NULL, &capsnap->cap_flush)) + wake_mdsc = true; spin_unlock(&mdsc->cap_dirty_lock); - wake_up_all(&ci->i_cap_wq); } spin_unlock(&ci->i_ceph_lock); if (flushed) { ceph_put_snap_context(capsnap->context); ceph_put_cap_snap(capsnap); + if (wake_ci) + wake_up_all(&ci->i_cap_wq); + if (wake_mdsc) + wake_up_all(&mdsc->cap_flushing_wq); iput(inode); } } diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index fa9036af5445..cdc6a17f5867 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -1212,6 +1212,8 @@ static void remove_session_caps(struct ceph_mds_session *session) dout("remove_session_caps on %p\n", session); iterate_session_caps(session, remove_session_caps_cb, fsc); + wake_up_all(&fsc->mdsc->cap_flushing_wq); + spin_lock(&session->s_cap_lock); if (session->s_nr_caps > 0) { struct inode *inode; @@ -3536,6 +3538,12 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc) ceph_flush_dirty_caps(mdsc); spin_lock(&mdsc->cap_dirty_lock); want_flush = mdsc->last_cap_flush_tid; + if (!list_empty(&mdsc->cap_flush_list)) { + struct ceph_cap_flush *cf = + list_last_entry(&mdsc->cap_flush_list, + struct ceph_cap_flush, g_list); + cf->wake = true; + } spin_unlock(&mdsc->cap_dirty_lock); dout("sync want tid %lld flush_seq %lld\n", diff --git a/fs/ceph/super.h b/fs/ceph/super.h index b097d474f888..3e3fa9163059 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -150,6 +150,7 @@ struct ceph_cap { struct ceph_cap_flush { u64 tid; int caps; /* 0 means capsnap */ + bool wake; /* wake up flush waiters when finish ? */ struct list_head g_list; // global struct list_head i_list; // per inode }; -- cgit v1.2.3 From 6b1a9a6c54122dade800fd61e90d441e58de19eb Mon Sep 17 00:00:00 2001 From: Nikolay Borisov Date: Mon, 25 Jul 2016 20:12:13 +0300 Subject: ceph: Mark the file cache as unreclaimable Ceph creates multiple caches with the SLAB_RECLAIMABLE flag set, so that it can satisfy its internal needs. Inspecting the code shows that most of the caches are indeed reclaimable since they are directly related to the generic inode/dentry shrinkers. However, one of the cache used to satisfy struct file is not reclaimable since its entries are freed only when the last reference to the file is dropped. If a heavily loaded node opens a lot of files it can introduce non-trivial discrepancies between memory shown as reclaimable and what is actually reclaimed when drop_caches is used. Fix this by removing the reclaimable flag for the file's cache. Signed-off-by: Nikolay Borisov Signed-off-by: Yan, Zheng --- fs/ceph/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 7736a931376e..e247f6f0feb7 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -686,8 +686,8 @@ static int __init init_caches(void) if (ceph_dentry_cachep == NULL) goto bad_dentry; - ceph_file_cachep = KMEM_CACHE(ceph_file_info, - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD); + ceph_file_cachep = KMEM_CACHE(ceph_file_info, SLAB_MEM_SPREAD); + if (ceph_file_cachep == NULL) goto bad_file; -- cgit v1.2.3 From 955818cd5b6c4b58ea574ace4573e7afa4c19c1e Mon Sep 17 00:00:00 2001 From: Phil Turnbull Date: Thu, 21 Jul 2016 13:43:09 -0400 Subject: ceph: Correctly return NXIO errors from ceph_llseek ceph_llseek does not correctly return NXIO errors because the 'out' path always returns 'offset'. Fixes: 06222e491e66 ("fs: handle SEEK_HOLE/SEEK_DATA properly in all fs's that define their own llseek") Signed-off-by: Phil Turnbull Signed-off-by: Yan, Zheng --- fs/ceph/file.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 7f2ef262cdf7..0f5375d8e030 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -1499,16 +1499,14 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) { struct inode *inode = file->f_mapping->host; loff_t i_size; - int ret; + loff_t ret; inode_lock(inode); if (whence == SEEK_END || whence == SEEK_DATA || whence == SEEK_HOLE) { ret = ceph_do_getattr(inode, CEPH_STAT_CAP_SIZE, false); - if (ret < 0) { - offset = ret; + if (ret < 0) goto out; - } } i_size = i_size_read(inode); @@ -1524,7 +1522,7 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) * write() or lseek() might have altered it */ if (offset == 0) { - offset = file->f_pos; + ret = file->f_pos; goto out; } offset += file->f_pos; @@ -1544,11 +1542,11 @@ static loff_t ceph_llseek(struct file *file, loff_t offset, int whence) break; } - offset = vfs_setpos(file, offset, inode->i_sb->s_maxbytes); + ret = vfs_setpos(file, offset, inode->i_sb->s_maxbytes); out: inode_unlock(inode); - return offset; + return ret; } static inline void ceph_zero_partial_page( -- cgit v1.2.3 From 1b2ce442ea96b1e76300553963154d68cc5b4ad0 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:22 -0400 Subject: binfmt_flat: use proper user space accessors with old relocs code Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 28 ++++++++++++++++++---------- 1 file changed, 18 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 9deafb282db5..8e812d7a62e0 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -379,35 +379,38 @@ static void old_reloc(unsigned long rl) { static const char *segment[] = { "TEXT", "DATA", "BSS", "*UNKNOWN*" }; flat_v2_reloc_t r; - unsigned long *ptr; + unsigned long __user *ptr; + unsigned long val; r.value = rl; #if defined(CONFIG_COLDFIRE) - ptr = (unsigned long *) (current->mm->start_code + r.reloc.offset); + ptr = (unsigned long __user *)(current->mm->start_code + r.reloc.offset); #else - ptr = (unsigned long *) (current->mm->start_data + r.reloc.offset); + ptr = (unsigned long __user *)(current->mm->start_data + r.reloc.offset); #endif + get_user(val, ptr); pr_debug("Relocation of variable at DATASEG+%x " "(address %p, currently %lx) into segment %s\n", - r.reloc.offset, ptr, *ptr, segment[r.reloc.type]); + r.reloc.offset, ptr, val, segment[r.reloc.type]); switch (r.reloc.type) { case OLD_FLAT_RELOC_TYPE_TEXT: - *ptr += current->mm->start_code; + val += current->mm->start_code; break; case OLD_FLAT_RELOC_TYPE_DATA: - *ptr += current->mm->start_data; + val += current->mm->start_data; break; case OLD_FLAT_RELOC_TYPE_BSS: - *ptr += current->mm->end_data; + val += current->mm->end_data; break; default: pr_err("Unknown relocation type=%x\n", r.reloc.type); break; } + put_user(val, ptr); - pr_debug("Relocation became %lx\n", *ptr); + pr_debug("Relocation became %lx\n", val); } /****************************************************************************/ @@ -780,8 +783,13 @@ static int load_flat_file(struct linux_binprm *bprm, } } } else { - for (i = 0; i < relocs; i++) - old_reloc(ntohl(reloc[i])); + for (i = 0; i < relocs; i++) { + unsigned long relval; + if (get_user(relval, reloc + i)) + return -EFAULT; + relval = ntohl(relval); + old_reloc(relval); + } } flush_icache_range(start_code, end_code); -- cgit v1.2.3 From 467aa1465a7b9f0bd1a23c83bff8c38e3c3e660e Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:23 -0400 Subject: binfmt_flat: use clear_user() rather than memset() to clear .bss This is needed on systems with a MMU. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 8e812d7a62e0..eb747a266537 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -795,10 +795,11 @@ static int load_flat_file(struct linux_binprm *bprm, flush_icache_range(start_code, end_code); /* zero the BSS, BRK and stack areas */ - memset((void *)(datapos + data_len), 0, bss_len + - (memp + memp_size - stack_len - /* end brk */ - libinfo->lib_list[id].start_brk) + /* start brk */ - stack_len); + if (clear_user((void __user *)(datapos + data_len), bss_len + + (memp + memp_size - stack_len - /* end brk */ + libinfo->lib_list[id].start_brk) + /* start brk */ + stack_len)) + return -EFAULT; return 0; err: -- cgit v1.2.3 From af521f92dcf636c59330233a61cb689d0383c291 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:24 -0400 Subject: binfmt_flat: update libraries' data segment pointer with userspace accessors This is needed on systems with a MMU. This also gets rid of the strangest C code I've seen lateli i.e. an integer indexed with a pointer value within square brackets. That really looked backwards. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index eb747a266537..83e507137112 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -894,12 +894,19 @@ static int load_flat_binary(struct linux_binprm *bprm) return res; /* Update data segment pointers for all libraries */ - for (i = 0; i < MAX_SHARED_LIBS; i++) - if (libinfo.lib_list[i].loaded) - for (j = 0; j < MAX_SHARED_LIBS; j++) - (-(j+1))[(unsigned long *)(libinfo.lib_list[i].start_data)] = - (libinfo.lib_list[j].loaded) ? - libinfo.lib_list[j].start_data : UNLOADED_LIB; + for (i = 0; i < MAX_SHARED_LIBS; i++) { + if (!libinfo.lib_list[i].loaded) + continue; + for (j = 0; j < MAX_SHARED_LIBS; j++) { + unsigned long val = libinfo.lib_list[j].loaded ? + libinfo.lib_list[j].start_data : UNLOADED_LIB; + unsigned long __user *p = (unsigned long __user *) + libinfo.lib_list[i].start_data; + p -= j + 1; + if (put_user(val, p)) + return -EFAULT; + } + } install_exec_creds(bprm); -- cgit v1.2.3 From 015feacf932108429f59b511b29d7e98a629e333 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:25 -0400 Subject: binfmt_flat: add MMU-specific support Not much else to do at this point except for the different stack setups. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 83e507137112..506139bfdc9e 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -537,7 +537,7 @@ static int load_flat_file(struct linux_binprm *bprm, * case, and then the fully copied to RAM case which lumps * it all together. */ - if ((flags & (FLAT_FLAG_RAM|FLAT_FLAG_GZIP)) == 0) { + if (!IS_ENABLED(CONFIG_MMU) && !(flags & (FLAT_FLAG_RAM|FLAT_FLAG_GZIP))) { /* * this should give us a ROM ptr, but if it doesn't we don't * really care @@ -677,7 +677,9 @@ static int load_flat_file(struct linux_binprm *bprm, */ current->mm->start_brk = datapos + data_len + bss_len; current->mm->brk = (current->mm->start_brk + 3) & ~3; +#ifndef CONFIG_MMU current->mm->context.end_brk = memp + memp_size - stack_len; +#endif } if (flags & FLAT_FLAG_KTRACE) { @@ -870,7 +872,7 @@ static int load_flat_binary(struct linux_binprm *bprm) { struct lib_info libinfo; struct pt_regs *regs = current_pt_regs(); - unsigned long stack_len; + unsigned long stack_len = 0; unsigned long start_addr; int res; int i, j; @@ -884,7 +886,9 @@ static int load_flat_binary(struct linux_binprm *bprm) * pedantic and include space for the argv/envp array as it may have * a lot of entries. */ - stack_len = PAGE_SIZE * MAX_ARG_PAGES - bprm->p; /* the strings */ +#ifndef CONFIG_MMU + stack_len += PAGE_SIZE * MAX_ARG_PAGES - bprm->p; /* the strings */ +#endif stack_len += (bprm->argc + 1) * sizeof(char *); /* the argv array */ stack_len += (bprm->envc + 1) * sizeof(char *); /* the envp array */ stack_len = ALIGN(stack_len, FLAT_STACK_ALIGN); @@ -912,6 +916,11 @@ static int load_flat_binary(struct linux_binprm *bprm) set_binfmt(&flat_format); +#ifdef CONFIG_MMU + res = setup_arg_pages(bprm, STACK_TOP, EXSTACK_DEFAULT); + if (!res) + res = create_flat_tables(bprm, bprm->p); +#else /* Stash our initial stack pointer into the mm structure */ current->mm->start_stack = ((current->mm->context.end_brk + stack_len + 3) & ~3) - 4; @@ -921,6 +930,7 @@ static int load_flat_binary(struct linux_binprm *bprm) res = transfer_args_to_stack(bprm, ¤t->mm->start_stack); if (!res) res = create_flat_tables(bprm, current->mm->start_stack); +#endif if (res) return res; -- cgit v1.2.3 From 472f95f32d5aa07eda96a6b2578b85d7b27c5110 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:26 -0400 Subject: binfmt_flat: allow compressed flat binary format to work on MMU systems Let's take the simple and obvious approach by decompressing the binary into a kernel buffer and then copying it to user space. Those who are looking for top performance on an MMU system are unlikely to choose this executable format anyway. Signed-off-by: Nicolas Pitre Reviewed-by: Greg Ungerer Signed-off-by: Greg Ungerer --- fs/binfmt_flat.c | 44 ++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 42 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 506139bfdc9e..9b2917a30294 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -34,6 +34,7 @@ #include #include #include +#include #include #include @@ -628,6 +629,7 @@ static int load_flat_file(struct linux_binprm *bprm, * load it all in and treat it like a RAM load from now on */ if (flags & FLAT_FLAG_GZIP) { +#ifndef CONFIG_MMU result = decompress_exec(bprm, sizeof(struct flat_hdr), (((char *)textpos) + sizeof(struct flat_hdr)), (text_len + full_data @@ -635,13 +637,51 @@ static int load_flat_file(struct linux_binprm *bprm, 0); memmove((void *) datapos, (void *) realdatastart, full_data); +#else + /* + * This is used on MMU systems mainly for testing. + * Let's use a kernel buffer to simplify things. + */ + long unz_text_len = text_len - sizeof(struct flat_hdr); + long unz_len = unz_text_len + full_data; + char *unz_data = vmalloc(unz_len); + if (!unz_data) { + result = -ENOMEM; + } else { + result = decompress_exec(bprm, sizeof(struct flat_hdr), + unz_data, unz_len, 0); + if (result == 0 && + (copy_to_user((void __user *)textpos + sizeof(struct flat_hdr), + unz_data, unz_text_len) || + copy_to_user((void __user *)datapos, + unz_data + unz_text_len, full_data))) + result = -EFAULT; + vfree(unz_data); + } +#endif } else if (flags & FLAT_FLAG_GZDATA) { result = read_code(bprm->file, textpos, 0, text_len); - if (!IS_ERR_VALUE(result)) + if (!IS_ERR_VALUE(result)) { +#ifndef CONFIG_MMU result = decompress_exec(bprm, text_len, (char *) datapos, full_data, 0); - } else +#else + char *unz_data = vmalloc(full_data); + if (!unz_data) { + result = -ENOMEM; + } else { + result = decompress_exec(bprm, text_len, + unz_data, full_data, 0); + if (result == 0 && + copy_to_user((void __user *)datapos, + unz_data, full_data)) + result = -EFAULT; + vfree(unz_data); + } #endif + } + } else +#endif /* CONFIG_BINFMT_ZFLAT */ { result = read_code(bprm->file, textpos, 0, text_len); if (!IS_ERR_VALUE(result)) -- cgit v1.2.3 From 002d2f01f13f1671d771e1493a3405ed4986694d Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 24 Jul 2016 11:30:28 -0400 Subject: m68k: enable binfmt_flat on systems with an MMU Now that the generic changes are in place, this can be enabled on m68k with the use of proper user space accessors in the flat_get_addr_from_rp() and flat_put_addr_at_rp() handlers as rp actually holds a user space address. Signed-off-by: Nicolas Pitre Signed-off-by: Greg Ungerer --- arch/m68k/include/asm/flat.h | 5 +++-- fs/Kconfig.binfmt | 3 ++- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/arch/m68k/include/asm/flat.h b/arch/m68k/include/asm/flat.h index 8e58d47be90d..00c392b0cabd 100644 --- a/arch/m68k/include/asm/flat.h +++ b/arch/m68k/include/asm/flat.h @@ -8,8 +8,9 @@ #define flat_argvp_envp_on_stack() 1 #define flat_old_ram_flag(flags) (flags) #define flat_reloc_valid(reloc, size) ((reloc) <= (size)) -#define flat_get_addr_from_rp(rp, relval, flags, p) get_unaligned(rp) -#define flat_put_addr_at_rp(rp, val, relval) put_unaligned(val,rp) +#define flat_get_addr_from_rp(rp, relval, flags, p) \ + ({ unsigned long __val; __get_user_unaligned(__val, rp); __val; }) +#define flat_put_addr_at_rp(rp, val, relval) __put_user_unaligned(val, rp) #define flat_get_relocate_addr(rel) (rel) static inline int flat_set_persistent(unsigned long relval, diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt index 72c03354c14b..c7efddf6e038 100644 --- a/fs/Kconfig.binfmt +++ b/fs/Kconfig.binfmt @@ -89,7 +89,8 @@ config BINFMT_SCRIPT config BINFMT_FLAT bool "Kernel support for flat binaries" - depends on !MMU && (!FRV || BROKEN) + depends on !MMU || M68K + depends on !FRV || BROKEN help Support uClinux FLAT format binaries. -- cgit v1.2.3 From 944171cbf499d3445c749f7c13c46de0a564a905 Mon Sep 17 00:00:00 2001 From: Benjamin Coddington Date: Thu, 28 Jul 2016 14:41:10 -0400 Subject: pNFS: Actively set attributes as invalid if LAYOUTCOMMIT is outstanding A LAYOUTCOMMIT then subsequent GETATTR may both return the same attributes, and in that case NFS_INO_INVALID_ATTR is never set on the second pass through nfs_update_inode(). The existing check to skip the clearing of NFS_INO_INVALID_ATTR if a LAYOUTCOMMIT is outstanding does not help in this case (see commit 10b7e9ad4488: "pNFS: Don't mark the inode as revalidated if a LAYOUTCOMMIT is outstanding"). We know that if a LAYOUTCOMMIT is outstanding then attributes will need upating, so always set NFS_INO_INVALID_ATTR. Signed-off-by: Benjamin Coddington Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index f108d58101f8..bf4ec5ecc97e 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1665,7 +1665,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) unsigned long now = jiffies; unsigned long save_cache_validity; bool have_writers = nfs_file_has_buffered_writers(nfsi); - bool cache_revalidated; + bool cache_revalidated = true; dfprintk(VFS, "NFS: %s(%s/%lu fh_crc=0x%08x ct=%d info=0x%x)\n", __func__, inode->i_sb->s_id, inode->i_ino, @@ -1714,8 +1714,10 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) /* Do atomic weak cache consistency updates */ invalid |= nfs_wcc_update_inode(inode, fattr); - - cache_revalidated = !pnfs_layoutcommit_outstanding(inode); + if (pnfs_layoutcommit_outstanding(inode)) { + nfsi->cache_validity |= save_cache_validity & NFS_INO_INVALID_ATTR; + cache_revalidated = false; + } /* More cache consistency checks */ if (fattr->valid & NFS_ATTR_FATTR_CHANGE) { -- cgit v1.2.3 From 1ae92642e5900316011736072b4fa91710840620 Mon Sep 17 00:00:00 2001 From: Daniel Golle Date: Sat, 18 Jun 2016 17:53:45 +0200 Subject: ubifs: Silence error output if MS_SILENT is set This change completes commit 90bea5a3f0 ("UBIFS: respect MS_SILENT mount flag") which already implements support for MS_SILENT except for that one error message which is still being displayed despite MS_SILENT being set. Suppress that error message as well in case MS_SILENT is set. Signed-off-by: Daniel Golle [rw: massaged commit message] Signed-off-by: Richard Weinberger --- fs/ubifs/super.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 70349954e78b..736dd5842f46 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2108,8 +2108,9 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags, */ ubi = open_ubi(name, UBI_READONLY); if (IS_ERR(ubi)) { - pr_err("UBIFS error (pid: %d): cannot open \"%s\", error %d", - current->pid, name, (int)PTR_ERR(ubi)); + if (!(flags & MS_SILENT)) + pr_err("UBIFS error (pid: %d): cannot open \"%s\", error %d", + current->pid, name, (int)PTR_ERR(ubi)); return ERR_CAST(ubi); } -- cgit v1.2.3 From dfaf8d2aeca4822e104fb3c4890d7b4b42048229 Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Tue, 21 Jun 2016 17:50:43 +0100 Subject: ubifs: Make xattr structures static Fix sparse warnings from the use of "struct xattr_handler" structures that are not exported by making them static. Fixes the following sparse warnings: /fs/ubifs/xattr.c:595:28: warning: symbol 'ubifs_user_xattr_handler' was not declared. Should it be static? /fs/ubifs/xattr.c:601:28: warning: symbol 'ubifs_trusted_xattr_handler' was not declared. Should it be static? /fs/ubifs/xattr.c:607:28: warning: symbol 'ubifs_security_xattr_handler' was not declared. Should it be static? Signed-off-by: Ben Dooks Signed-off-by: Richard Weinberger --- fs/ubifs/xattr.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c index b5fc27969e9d..e237811f09ce 100644 --- a/fs/ubifs/xattr.c +++ b/fs/ubifs/xattr.c @@ -592,19 +592,19 @@ static int ubifs_xattr_set(const struct xattr_handler *handler, return __ubifs_removexattr(inode, name); } -const struct xattr_handler ubifs_user_xattr_handler = { +static const struct xattr_handler ubifs_user_xattr_handler = { .prefix = XATTR_USER_PREFIX, .get = ubifs_xattr_get, .set = ubifs_xattr_set, }; -const struct xattr_handler ubifs_trusted_xattr_handler = { +static const struct xattr_handler ubifs_trusted_xattr_handler = { .prefix = XATTR_TRUSTED_PREFIX, .get = ubifs_xattr_get, .set = ubifs_xattr_set, }; -const struct xattr_handler ubifs_security_xattr_handler = { +static const struct xattr_handler ubifs_security_xattr_handler = { .prefix = XATTR_SECURITY_PREFIX, .get = ubifs_xattr_get, .set = ubifs_xattr_set, -- cgit v1.2.3 From d614146d18159e54d4dfc289e4c1534ccc485cf2 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 28 Jul 2016 13:05:50 -0400 Subject: fold dentry_rcuwalk_invalidate() into its only remaining caller Signed-off-by: Al Viro --- fs/dcache.c | 17 ++--------------- 1 file changed, 2 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index b90cf8e09d5b..6a629f50c3cf 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -316,20 +316,6 @@ static void dentry_free(struct dentry *dentry) call_rcu(&dentry->d_u.d_rcu, __d_free); } -/** - * dentry_rcuwalk_invalidate - invalidate in-progress rcu-walk lookups - * @dentry: the target dentry - * After this call, in-progress rcu-walk path lookup will fail. This - * should be called after unhashing, and after changing d_inode (if - * the dentry has not already been unhashed). - */ -static inline void dentry_rcuwalk_invalidate(struct dentry *dentry) -{ - lockdep_assert_held(&dentry->d_lock); - /* Go through am invalidation barrier */ - write_seqcount_invalidate(&dentry->d_seq); -} - /* * Release the dentry's inode, using the filesystem * d_iput() operation if defined. @@ -468,7 +454,8 @@ void __d_drop(struct dentry *dentry) __hlist_bl_del(&dentry->d_hash); dentry->d_hash.pprev = NULL; hlist_bl_unlock(b); - dentry_rcuwalk_invalidate(dentry); + /* After this call, in-progress rcu-walk path lookup will fail. */ + write_seqcount_invalidate(&dentry->d_seq); } } EXPORT_SYMBOL(__d_drop); -- cgit v1.2.3 From 380bc8b71081a258a4cda0060803bfde47dd8c60 Mon Sep 17 00:00:00 2001 From: Daniel Golle Date: Mon, 25 Jul 2016 13:44:50 +0200 Subject: ubifs: Update comment for ubifs_errc Signed-off-by: Daniel Golle Signed-off-by: Richard Weinberger --- fs/ubifs/ubifs.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index ddf9f6b9eee2..4617d459022a 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1783,8 +1783,8 @@ void ubifs_err(const struct ubifs_info *c, const char *fmt, ...); __printf(2, 3) void ubifs_warn(const struct ubifs_info *c, const char *fmt, ...); /* - * A variant of 'ubifs_err()' which takes the UBIFS file-sytem description - * object as an argument. + * A conditional variant of 'ubifs_err()' which doesn't output anything + * if probing (ie. MS_SILENT set). */ #define ubifs_errc(c, fmt, ...) \ do { \ -- cgit v1.2.3 From dccbc9197d2c3614f2fd6811874e1d982e4415f0 Mon Sep 17 00:00:00 2001 From: Daniel Golle Date: Tue, 19 Jul 2016 00:26:55 +0200 Subject: ubifs: Silence early error messages if MS_SILENT is set Probe-mounting a volume too small for UBIFS results in kernel log polution which might irritate users. Address this by silencing errors which may happen during boot if the rootfs is e.g. squashfs (and thus rather small) stored on a UBI volume. Signed-off-by: Daniel Golle Signed-off-by: Richard Weinberger --- fs/ubifs/super.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 736dd5842f46..4ec051089186 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -520,19 +520,19 @@ static int init_constants_early(struct ubifs_info *c) c->max_write_shift = fls(c->max_write_size) - 1; if (c->leb_size < UBIFS_MIN_LEB_SZ) { - ubifs_err(c, "too small LEBs (%d bytes), min. is %d bytes", - c->leb_size, UBIFS_MIN_LEB_SZ); + ubifs_errc(c, "too small LEBs (%d bytes), min. is %d bytes", + c->leb_size, UBIFS_MIN_LEB_SZ); return -EINVAL; } if (c->leb_cnt < UBIFS_MIN_LEB_CNT) { - ubifs_err(c, "too few LEBs (%d), min. is %d", - c->leb_cnt, UBIFS_MIN_LEB_CNT); + ubifs_errc(c, "too few LEBs (%d), min. is %d", + c->leb_cnt, UBIFS_MIN_LEB_CNT); return -EINVAL; } if (!is_power_of_2(c->min_io_size)) { - ubifs_err(c, "bad min. I/O size %d", c->min_io_size); + ubifs_errc(c, "bad min. I/O size %d", c->min_io_size); return -EINVAL; } @@ -543,8 +543,8 @@ static int init_constants_early(struct ubifs_info *c) if (c->max_write_size < c->min_io_size || c->max_write_size % c->min_io_size || !is_power_of_2(c->max_write_size)) { - ubifs_err(c, "bad write buffer size %d for %d min. I/O unit", - c->max_write_size, c->min_io_size); + ubifs_errc(c, "bad write buffer size %d for %d min. I/O unit", + c->max_write_size, c->min_io_size); return -EINVAL; } -- cgit v1.2.3 From 13cd091364111508cc001aa66b5285fefe7cdd73 Mon Sep 17 00:00:00 2001 From: Sylvain Etienne Date: Thu, 23 Jun 2016 07:53:44 +0200 Subject: ubifs: switch_gc_head: Remove redondant sync of wbuf The wbuf is already sync-ed before ubifs_leb_unmap() Signed-off-by: Sylvain Etienne Acked-by: Artem Bityutskiy Signed-off-by: Richard Weinberger --- fs/ubifs/gc.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'fs') diff --git a/fs/ubifs/gc.c b/fs/ubifs/gc.c index 9718da86ad01..821b34816976 100644 --- a/fs/ubifs/gc.c +++ b/fs/ubifs/gc.c @@ -100,10 +100,6 @@ static int switch_gc_head(struct ubifs_info *c) if (err) return err; - err = ubifs_wbuf_sync_nolock(wbuf); - if (err) - return err; - err = ubifs_add_bud_to_log(c, GCHD, gc_lnum, 0); if (err) return err; -- cgit v1.2.3 From 15d3c589f6305c50a705572dbb33620c5cba416c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 29 Jul 2016 17:45:21 -0400 Subject: fold _d_rehash() and __d_rehash() together The only place where we feed to __d_rehash() something other than d_hash(dentry->d_name.hash) is __d_move(), where we give it d_hash of another dentry. Postpone rehashing until we'd switched the names and we are rid of that exception, along with the need to keep _d_rehash() and __d_rehash() separate. Signed-off-by: Al Viro --- fs/dcache.c | 34 +++++++++++----------------------- 1 file changed, 11 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 6a629f50c3cf..d9450bd496ac 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2339,19 +2339,15 @@ again: } EXPORT_SYMBOL(d_delete); -static void __d_rehash(struct dentry * entry, struct hlist_bl_head *b) +static void __d_rehash(struct dentry *entry) { + struct hlist_bl_head *b = d_hash(entry->d_name.hash); BUG_ON(!d_unhashed(entry)); hlist_bl_lock(b); hlist_bl_add_head_rcu(&entry->d_hash, b); hlist_bl_unlock(b); } -static void _d_rehash(struct dentry * entry) -{ - __d_rehash(entry, d_hash(entry->d_name.hash)); -} - /** * d_rehash - add an entry back to the hash * @entry: dentry to add to the hash @@ -2362,7 +2358,7 @@ static void _d_rehash(struct dentry * entry) void d_rehash(struct dentry * entry) { spin_lock(&entry->d_lock); - _d_rehash(entry); + __d_rehash(entry); spin_unlock(&entry->d_lock); } EXPORT_SYMBOL(d_rehash); @@ -2536,7 +2532,7 @@ static inline void __d_add(struct dentry *dentry, struct inode *inode) raw_write_seqcount_end(&dentry->d_seq); fsnotify_update_flags(dentry); } - _d_rehash(dentry); + __d_rehash(dentry); if (dir) end_dir_add(dir, n); spin_unlock(&dentry->d_lock); @@ -2598,7 +2594,7 @@ struct dentry *d_exact_alias(struct dentry *entry, struct inode *inode) alias = NULL; } else { __dget_dlock(alias); - _d_rehash(alias); + __d_rehash(alias); spin_unlock(&alias->d_lock); } spin_unlock(&inode->i_lock); @@ -2782,23 +2778,10 @@ static void __d_move(struct dentry *dentry, struct dentry *target, write_seqcount_begin(&dentry->d_seq); write_seqcount_begin_nested(&target->d_seq, DENTRY_D_LOCK_NESTED); + /* unhash both */ /* __d_drop does write_seqcount_barrier, but they're OK to nest. */ - - /* - * Move the dentry to the target hash queue. Don't bother checking - * for the same hash queue because of how unlikely it is. - */ __d_drop(dentry); - __d_rehash(dentry, d_hash(target->d_name.hash)); - - /* - * Unhash the target (d_delete() is not usable here). If exchanging - * the two dentries, then rehash onto the other's hash queue. - */ __d_drop(target); - if (exchange) { - __d_rehash(target, d_hash(dentry->d_name.hash)); - } /* Switch the names.. */ if (exchange) @@ -2806,6 +2789,11 @@ static void __d_move(struct dentry *dentry, struct dentry *target, else copy_name(dentry, target); + /* rehash in new place(s) */ + __d_rehash(dentry); + if (exchange) + __d_rehash(target); + /* ... and switch them in the tree */ if (IS_ROOT(dentry)) { /* splicing a tree */ -- cgit v1.2.3 From e0b3f595d13b3e9ce9cdf53935e7f304c04b5b2b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 29 Jul 2016 18:22:49 -0400 Subject: affs ->d_compare(): don't bother with ->d_inode Use ->d_sb directly. Signed-off-by: Al Viro --- fs/affs/amigaffs.c | 4 +--- fs/affs/namei.c | 4 ++-- 2 files changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/affs/amigaffs.c b/fs/affs/amigaffs.c index d6c7a51c93e4..d8f217c711d3 100644 --- a/fs/affs/amigaffs.c +++ b/fs/affs/amigaffs.c @@ -472,9 +472,7 @@ affs_warning(struct super_block *sb, const char *function, const char *fmt, ...) bool affs_nofilenametruncate(const struct dentry *dentry) { - struct inode *inode = d_inode(dentry); - - return affs_test_opt(AFFS_SB(inode->i_sb)->s_flags, SF_NO_TRUNCATE); + return affs_test_opt(AFFS_SB(dentry->d_sb)->s_flags, SF_NO_TRUNCATE); } /* Check if the name is valid for a affs object. */ diff --git a/fs/affs/namei.c b/fs/affs/namei.c index eb32029bc776..27ca732680d9 100644 --- a/fs/affs/namei.c +++ b/fs/affs/namei.c @@ -136,7 +136,7 @@ affs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, { return __affs_compare_dentry(len, str, name, affs_toupper, - affs_nofilenametruncate(parent)); + affs_nofilenametruncate(dentry)); } static int @@ -144,7 +144,7 @@ affs_intl_compare_dentry(const struct dentry *parent, const struct dentry *dentr unsigned int len, const char *str, const struct qstr *name) { return __affs_compare_dentry(len, str, name, affs_intl_toupper, - affs_nofilenametruncate(parent)); + affs_nofilenametruncate(dentry)); } -- cgit v1.2.3 From d3fe19852e96aabe3084c82ec2f3fb4918ab6d1e Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 29 Jul 2016 18:23:59 -0400 Subject: cifs, msdos, vfat, hfs+: don't bother with parent in ->d_compare() dentry->d_sb is just as good as parent->d_sb Signed-off-by: Al Viro --- fs/cifs/dir.c | 2 +- fs/fat/namei_msdos.c | 2 +- fs/fat/namei_vfat.c | 2 +- fs/hfsplus/unicode.c | 2 +- fs/hpfs/dentry.c | 2 +- 5 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c index 4e532536cbc6..cf394da87cd4 100644 --- a/fs/cifs/dir.c +++ b/fs/cifs/dir.c @@ -906,7 +906,7 @@ static int cifs_ci_hash(const struct dentry *dentry, struct qstr *q) static int cifs_ci_compare(const struct dentry *parent, const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { - struct nls_table *codepage = CIFS_SB(parent->d_sb)->local_nls; + struct nls_table *codepage = CIFS_SB(dentry->d_sb)->local_nls; wchar_t c1, c2; int i, l1, l2; diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c index 1337c0c7527d..0f177127e4d6 100644 --- a/fs/fat/namei_msdos.c +++ b/fs/fat/namei_msdos.c @@ -165,7 +165,7 @@ static int msdos_hash(const struct dentry *dentry, struct qstr *qstr) static int msdos_cmp(const struct dentry *parent, const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { - struct fat_mount_options *options = &MSDOS_SB(parent->d_sb)->options; + struct fat_mount_options *options = &MSDOS_SB(dentry->d_sb)->options; unsigned char a_msdos_name[MSDOS_NAME], b_msdos_name[MSDOS_NAME]; int error; diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c index 6ccdf3f34f90..e60a90841381 100644 --- a/fs/fat/namei_vfat.c +++ b/fs/fat/namei_vfat.c @@ -141,7 +141,7 @@ static int vfat_hashi(const struct dentry *dentry, struct qstr *qstr) static int vfat_cmpi(const struct dentry *parent, const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { - struct nls_table *t = MSDOS_SB(parent->d_sb)->nls_io; + struct nls_table *t = MSDOS_SB(dentry->d_sb)->nls_io; unsigned int alen, blen; /* A filename cannot end in '.' or we treat it like it has none */ diff --git a/fs/hfsplus/unicode.c b/fs/hfsplus/unicode.c index c13c8a240be3..509d7e231460 100644 --- a/fs/hfsplus/unicode.c +++ b/fs/hfsplus/unicode.c @@ -388,7 +388,7 @@ int hfsplus_hash_dentry(const struct dentry *dentry, struct qstr *str) int hfsplus_compare_dentry(const struct dentry *parent, const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { - struct super_block *sb = parent->d_sb; + struct super_block *sb = dentry->d_sb; int casefold, decompose, size; int dsize1, dsize2, len1, len2; const u16 *dstr1, *dstr2; diff --git a/fs/hpfs/dentry.c b/fs/hpfs/dentry.c index 60e6d334d79a..a804300f4dce 100644 --- a/fs/hpfs/dentry.c +++ b/fs/hpfs/dentry.c @@ -50,7 +50,7 @@ static int hpfs_compare_dentry(const struct dentry *parent, const struct dentry if (hpfs_chk_name(name->name, &bl)) return 1; - if (hpfs_compare_names(parent->d_sb, str, al, name->name, bl, 0)) + if (hpfs_compare_names(dentry->d_sb, str, al, name->name, bl, 0)) return 1; return 0; } -- cgit v1.2.3 From 13983d062f17fb69f249befeba81b5ca253ad641 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 22:34:44 -0400 Subject: qstr: constify instances in fuse Signed-off-by: Al Viro --- fs/fuse/dir.c | 4 ++-- fs/fuse/fuse_i.h | 2 +- fs/fuse/inode.c | 8 ++------ 3 files changed, 5 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 264f07c7754e..419230a00b30 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -146,7 +146,7 @@ static void fuse_invalidate_entry(struct dentry *entry) } static void fuse_lookup_init(struct fuse_conn *fc, struct fuse_args *args, - u64 nodeid, struct qstr *name, + u64 nodeid, const struct qstr *name, struct fuse_entry_out *outarg) { memset(outarg, 0, sizeof(struct fuse_entry_out)); @@ -282,7 +282,7 @@ int fuse_valid_type(int m) S_ISBLK(m) || S_ISFIFO(m) || S_ISSOCK(m); } -int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name, +int fuse_lookup_name(struct super_block *sb, u64 nodeid, const struct qstr *name, struct fuse_entry_out *outarg, struct inode **inode) { struct fuse_conn *fc = get_fuse_conn_super(sb); diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 929c383432b0..9e42c0394196 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -703,7 +703,7 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid, int generation, struct fuse_attr *attr, u64 attr_valid, u64 attr_version); -int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name, +int fuse_lookup_name(struct super_block *sb, u64 nodeid, const struct qstr *name, struct fuse_entry_out *outarg, struct inode **inode); /** diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 9961d8432ce3..cdc40537da41 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -673,13 +673,11 @@ static struct dentry *fuse_get_dentry(struct super_block *sb, inode = ilookup5(sb, handle->nodeid, fuse_inode_eq, &handle->nodeid); if (!inode) { struct fuse_entry_out outarg; - struct qstr name; + const struct qstr name = QSTR_INIT(".", 1); if (!fc->export_support) goto out_err; - name.len = 1; - name.name = "."; err = fuse_lookup_name(sb, handle->nodeid, &name, &outarg, &inode); if (err && err != -ENOENT) @@ -775,14 +773,12 @@ static struct dentry *fuse_get_parent(struct dentry *child) struct inode *inode; struct dentry *parent; struct fuse_entry_out outarg; - struct qstr name; + const struct qstr name = QSTR_INIT("..", 2); int err; if (!fc->export_support) return ERR_PTR(-ESTALE); - name.len = 2; - name.name = ".."; err = fuse_lookup_name(child_inode->i_sb, get_node_id(child_inode), &name, &outarg, &inode); if (err) { -- cgit v1.2.3 From dc12e909490dcad210ee52bc1b0776c95dcc9edf Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 22:41:44 -0400 Subject: qstr: constify instances in procfs Signed-off-by: Al Viro --- fs/proc/proc_sysctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index 5e57c3e46e1d..04f69ea4b8c7 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -474,7 +474,7 @@ static struct dentry *proc_sys_lookup(struct inode *dir, struct dentry *dentry, { struct ctl_table_header *head = grab_header(dir); struct ctl_table_header *h = NULL; - struct qstr *name = &dentry->d_name; + const struct qstr *name = &dentry->d_name; struct ctl_table *p; struct inode *inode; struct dentry *err = ERR_PTR(-ENOENT); -- cgit v1.2.3 From b59091c04ae02bc37fab86937e7aa998787d6bf1 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 22:43:14 -0400 Subject: qstr: constify instances in vfat Signed-off-by: Al Viro --- fs/fat/namei_vfat.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c index 7092584f424a..1d16ed5edabf 100644 --- a/fs/fat/namei_vfat.c +++ b/fs/fat/namei_vfat.c @@ -652,8 +652,8 @@ out_free: return err; } -static int vfat_add_entry(struct inode *dir, struct qstr *qname, int is_dir, - int cluster, struct timespec *ts, +static int vfat_add_entry(struct inode *dir, const struct qstr *qname, + int is_dir, int cluster, struct timespec *ts, struct fat_slot_info *sinfo) { struct msdos_dir_slot *slots; @@ -688,7 +688,7 @@ cleanup: return err; } -static int vfat_find(struct inode *dir, struct qstr *qname, +static int vfat_find(struct inode *dir, const struct qstr *qname, struct fat_slot_info *sinfo) { unsigned int len = vfat_striptail_len(qname); -- cgit v1.2.3 From ac3ba644bc8f9072155f16e4156dc4df845112f0 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 22:47:26 -0400 Subject: qstr: constify instances in ext2 Signed-off-by: Al Viro --- fs/ext2/dir.c | 6 +++--- fs/ext2/ext2.h | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c index 19efd1197fa5..61ad490ed67b 100644 --- a/fs/ext2/dir.c +++ b/fs/ext2/dir.c @@ -358,8 +358,8 @@ ext2_readdir(struct file *file, struct dir_context *ctx) * and the entry itself. Page is returned mapped and unlocked. * Entry is guaranteed to be valid. */ -struct ext2_dir_entry_2 *ext2_find_entry (struct inode * dir, - struct qstr *child, struct page ** res_page) +struct ext2_dir_entry_2 *ext2_find_entry (struct inode *dir, + const struct qstr *child, struct page **res_page) { const char *name = child->name; int namelen = child->len; @@ -435,7 +435,7 @@ struct ext2_dir_entry_2 * ext2_dotdot (struct inode *dir, struct page **p) return de; } -ino_t ext2_inode_by_name(struct inode *dir, struct qstr *child) +ino_t ext2_inode_by_name(struct inode *dir, const struct qstr *child) { ino_t res = 0; struct ext2_dir_entry_2 *de; diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h index 170939f379d7..24a270f26bc3 100644 --- a/fs/ext2/ext2.h +++ b/fs/ext2/ext2.h @@ -754,9 +754,9 @@ extern void ext2_rsv_window_add(struct super_block *sb, struct ext2_reserve_wind /* dir.c */ extern int ext2_add_link (struct dentry *, struct inode *); -extern ino_t ext2_inode_by_name(struct inode *, struct qstr *); +extern ino_t ext2_inode_by_name(struct inode *, const struct qstr *); extern int ext2_make_empty(struct inode *, struct inode *); -extern struct ext2_dir_entry_2 * ext2_find_entry (struct inode *,struct qstr *, struct page **); +extern struct ext2_dir_entry_2 * ext2_find_entry (struct inode *,const struct qstr *, struct page **); extern int ext2_delete_entry (struct ext2_dir_entry_2 *, struct page *); extern int ext2_empty_dir (struct inode *); extern struct ext2_dir_entry_2 * ext2_dotdot (struct inode *, struct page **); -- cgit v1.2.3 From 185de68fcb397c8c275b536fe279cbfadf8b185c Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 23:07:13 -0400 Subject: qstr: constify instances in f2fs Signed-off-by: Al Viro --- fs/f2fs/dir.c | 4 ++-- fs/f2fs/f2fs.h | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index f9313f684540..3b095a35aa47 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -214,7 +214,7 @@ static struct f2fs_dir_entry *find_in_level(struct inode *dir, * Entry is guaranteed to be valid. */ struct f2fs_dir_entry *f2fs_find_entry(struct inode *dir, - struct qstr *child, struct page **res_page) + const struct qstr *child, struct page **res_page) { unsigned long npages = dir_blocks(dir); struct f2fs_dir_entry *de = NULL; @@ -277,7 +277,7 @@ struct f2fs_dir_entry *f2fs_parent_dir(struct inode *dir, struct page **p) return de; } -ino_t f2fs_inode_by_name(struct inode *dir, struct qstr *qstr) +ino_t f2fs_inode_by_name(struct inode *dir, const struct qstr *qstr) { ino_t res = 0; struct f2fs_dir_entry *de; diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 916e7c238e3d..fd0a156d7561 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1805,10 +1805,10 @@ struct page *init_inode_metadata(struct inode *, struct inode *, void update_parent_metadata(struct inode *, struct inode *, unsigned int); int room_for_filename(const void *, int, int); void f2fs_drop_nlink(struct inode *, struct inode *, struct page *); -struct f2fs_dir_entry *f2fs_find_entry(struct inode *, struct qstr *, +struct f2fs_dir_entry *f2fs_find_entry(struct inode *, const struct qstr *, struct page **); struct f2fs_dir_entry *f2fs_parent_dir(struct inode *, struct page **); -ino_t f2fs_inode_by_name(struct inode *, struct qstr *); +ino_t f2fs_inode_by_name(struct inode *, const struct qstr *); void f2fs_set_link(struct inode *, struct f2fs_dir_entry *, struct page *, struct inode *); int update_dent_inode(struct inode *, struct inode *, const struct qstr *); -- cgit v1.2.3 From 19a6d89de2f10ee3b18c3b9fd6c7c2e249a30d71 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 20 Jul 2016 23:23:32 -0400 Subject: qstr: constify instances in adfs Signed-off-by: Al Viro --- fs/adfs/dir.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c index fd4cf2c48e48..62ee2cb8163b 100644 --- a/fs/adfs/dir.c +++ b/fs/adfs/dir.c @@ -101,7 +101,7 @@ out: } static int -adfs_match(struct qstr *name, struct object_info *obj) +adfs_match(const struct qstr *name, struct object_info *obj) { int i; @@ -126,7 +126,7 @@ adfs_match(struct qstr *name, struct object_info *obj) } static int -adfs_dir_lookup_byname(struct inode *inode, struct qstr *name, struct object_info *obj) +adfs_dir_lookup_byname(struct inode *inode, const struct qstr *name, struct object_info *obj) { struct super_block *sb = inode->i_sb; const struct adfs_dir_ops *ops = ADFS_SB(sb)->s_dir; -- cgit v1.2.3 From 6fa67e707559303e086303aeecc9e8b91ef497d5 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 31 Jul 2016 16:37:25 -0400 Subject: get rid of 'parent' argument of ->d_compare() Signed-off-by: Al Viro --- Documentation/filesystems/Locking | 2 +- Documentation/filesystems/porting | 7 +++++++ Documentation/filesystems/vfs.txt | 2 +- drivers/staging/lustre/lustre/llite/dcache.c | 2 +- fs/adfs/dir.c | 2 +- fs/affs/namei.c | 8 ++++---- fs/cifs/dir.c | 2 +- fs/dcache.c | 4 ++-- fs/efivarfs/super.c | 3 +-- fs/fat/namei_msdos.c | 2 +- fs/fat/namei_vfat.c | 4 ++-- fs/hfs/hfs_fs.h | 2 +- fs/hfs/string.c | 2 +- fs/hfsplus/hfsplus_fs.h | 3 +-- fs/hfsplus/unicode.c | 2 +- fs/hpfs/dentry.c | 2 +- fs/isofs/inode.c | 15 ++++++--------- fs/isofs/namei.c | 2 +- fs/jfs/namei.c | 2 +- fs/ncpfs/dir.c | 6 +++--- fs/proc/proc_sysctl.c | 2 +- include/linux/dcache.h | 2 +- 22 files changed, 40 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking index 1b3c39a7de62..d30fb2cb5066 100644 --- a/Documentation/filesystems/Locking +++ b/Documentation/filesystems/Locking @@ -12,7 +12,7 @@ prototypes: int (*d_revalidate)(struct dentry *, unsigned int); int (*d_weak_revalidate)(struct dentry *, unsigned int); int (*d_hash)(const struct dentry *, struct qstr *); - int (*d_compare)(const struct dentry *, const struct dentry *, + int (*d_compare)(const struct dentry *, unsigned int, const char *, const struct qstr *); int (*d_delete)(struct dentry *); int (*d_init)(struct dentry *); diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting index a5fb89cac615..b1bd05ea66b2 100644 --- a/Documentation/filesystems/porting +++ b/Documentation/filesystems/porting @@ -585,3 +585,10 @@ in your dentry operations instead. in the instances. Rationale: !@#!@# security_d_instantiate() needs to be called before we attach dentry to inode and !@#!@##!@$!$#!@#$!@$!@$ smack ->d_instantiate() uses not just ->getxattr() but ->setxattr() as well. +-- +[mandatory] + ->d_compare() doesn't get parent as a separate argument anymore. If you + used it for finding the struct super_block involved, dentry->d_sb will + work just as well; if it's something more complicated, use dentry->d_parent. + Just be careful not to assume that fetching it more than once will yield + the same value - in RCU mode it could change under you. diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt index 8a196851f01d..9ace359d6cc5 100644 --- a/Documentation/filesystems/vfs.txt +++ b/Documentation/filesystems/vfs.txt @@ -931,7 +931,7 @@ struct dentry_operations { int (*d_revalidate)(struct dentry *, unsigned int); int (*d_weak_revalidate)(struct dentry *, unsigned int); int (*d_hash)(const struct dentry *, struct qstr *); - int (*d_compare)(const struct dentry *, const struct dentry *, + int (*d_compare)(const struct dentry *, unsigned int, const char *, const struct qstr *); int (*d_delete)(const struct dentry *); int (*d_init)(struct dentry *); diff --git a/drivers/staging/lustre/lustre/llite/dcache.c b/drivers/staging/lustre/lustre/llite/dcache.c index 581a63a0a63e..463b1a360733 100644 --- a/drivers/staging/lustre/lustre/llite/dcache.c +++ b/drivers/staging/lustre/lustre/llite/dcache.c @@ -78,7 +78,7 @@ static void ll_release(struct dentry *de) * INVALID) so d_lookup() matches it, but we have no lock on it (so * lock_match() fails) and we spin around real_lookup(). */ -static int ll_dcompare(const struct dentry *parent, const struct dentry *dentry, +static int ll_dcompare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c index bec25f7017c0..c30a5a65cbb5 100644 --- a/fs/adfs/dir.c +++ b/fs/adfs/dir.c @@ -227,7 +227,7 @@ adfs_hash(const struct dentry *parent, struct qstr *qstr) * requirements of the underlying filesystem. */ static int -adfs_compare(const struct dentry *parent, const struct dentry *dentry, +adfs_compare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { int i; diff --git a/fs/affs/namei.c b/fs/affs/namei.c index 27ca732680d9..a2d68f828d53 100644 --- a/fs/affs/namei.c +++ b/fs/affs/namei.c @@ -14,11 +14,11 @@ typedef int (*toupper_t)(int); static int affs_toupper(int ch); static int affs_hash_dentry(const struct dentry *, struct qstr *); -static int affs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +static int affs_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); static int affs_intl_toupper(int ch); static int affs_intl_hash_dentry(const struct dentry *, struct qstr *); -static int affs_intl_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +static int affs_intl_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); const struct dentry_operations affs_dentry_operations = { @@ -131,7 +131,7 @@ static inline int __affs_compare_dentry(unsigned int len, } static int -affs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +affs_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { @@ -140,7 +140,7 @@ affs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, } static int -affs_intl_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +affs_intl_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { return __affs_compare_dentry(len, str, name, affs_intl_toupper, diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c index cf394da87cd4..4716c54dbfc6 100644 --- a/fs/cifs/dir.c +++ b/fs/cifs/dir.c @@ -903,7 +903,7 @@ static int cifs_ci_hash(const struct dentry *dentry, struct qstr *q) return 0; } -static int cifs_ci_compare(const struct dentry *parent, const struct dentry *dentry, +static int cifs_ci_compare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct nls_table *codepage = CIFS_SB(dentry->d_sb)->local_nls; diff --git a/fs/dcache.c b/fs/dcache.c index d9450bd496ac..32a9209c8138 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -2047,7 +2047,7 @@ static inline bool d_same_name(const struct dentry *dentry, return false; return dentry_cmp(dentry, name->name, name->len) == 0; } - return parent->d_op->d_compare(parent, dentry, + return parent->d_op->d_compare(dentry, dentry->d_name.len, dentry->d_name.name, name) == 0; } @@ -2150,7 +2150,7 @@ seqretry: cpu_relax(); goto seqretry; } - if (parent->d_op->d_compare(parent, dentry, + if (parent->d_op->d_compare(dentry, tlen, tname, name) != 0) continue; } else { diff --git a/fs/efivarfs/super.c b/fs/efivarfs/super.c index a5e607e8f056..688ccc16b702 100644 --- a/fs/efivarfs/super.c +++ b/fs/efivarfs/super.c @@ -45,8 +45,7 @@ static struct super_block *efivarfs_sb; * So we need to perform a case-sensitive match on part 1 and a * case-insensitive match on part 2. */ -static int efivarfs_d_compare(const struct dentry *parent, - const struct dentry *dentry, +static int efivarfs_d_compare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c index 0f177127e4d6..664655b2c55f 100644 --- a/fs/fat/namei_msdos.c +++ b/fs/fat/namei_msdos.c @@ -162,7 +162,7 @@ static int msdos_hash(const struct dentry *dentry, struct qstr *qstr) * Compare two msdos names. If either of the names are invalid, * we fall back to doing the standard name comparison. */ -static int msdos_cmp(const struct dentry *parent, const struct dentry *dentry, +static int msdos_cmp(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct fat_mount_options *options = &MSDOS_SB(dentry->d_sb)->options; diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c index e60a90841381..0335e504e65a 100644 --- a/fs/fat/namei_vfat.c +++ b/fs/fat/namei_vfat.c @@ -138,7 +138,7 @@ static int vfat_hashi(const struct dentry *dentry, struct qstr *qstr) /* * Case insensitive compare of two vfat names. */ -static int vfat_cmpi(const struct dentry *parent, const struct dentry *dentry, +static int vfat_cmpi(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct nls_table *t = MSDOS_SB(dentry->d_sb)->nls_io; @@ -157,7 +157,7 @@ static int vfat_cmpi(const struct dentry *parent, const struct dentry *dentry, /* * Case sensitive compare of two vfat names. */ -static int vfat_cmp(const struct dentry *parent, const struct dentry *dentry, +static int vfat_cmp(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { unsigned int alen, blen; diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h index ee2f385811c8..f28d7a259105 100644 --- a/fs/hfs/hfs_fs.h +++ b/fs/hfs/hfs_fs.h @@ -233,7 +233,7 @@ extern const struct dentry_operations hfs_dentry_operations; extern int hfs_hash_dentry(const struct dentry *, struct qstr *); extern int hfs_strcmp(const unsigned char *, unsigned int, const unsigned char *, unsigned int); -extern int hfs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +extern int hfs_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); /* trans.c */ diff --git a/fs/hfs/string.c b/fs/hfs/string.c index ec9f164c35a5..3912209153a8 100644 --- a/fs/hfs/string.c +++ b/fs/hfs/string.c @@ -92,7 +92,7 @@ int hfs_strcmp(const unsigned char *s1, unsigned int len1, * Test for equality of two strings in the HFS filename character ordering. * return 1 on failure and 0 on success */ -int hfs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +int hfs_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { const unsigned char *n1, *n2; diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h index 047245bd2cd6..e95c01f1d62e 100644 --- a/fs/hfsplus/hfsplus_fs.h +++ b/fs/hfsplus/hfsplus_fs.h @@ -520,8 +520,7 @@ int hfsplus_uni2asc(struct super_block *sb, const struct hfsplus_unistr *ustr, int hfsplus_asc2uni(struct super_block *sb, struct hfsplus_unistr *ustr, int max_unistr_len, const char *astr, int len); int hfsplus_hash_dentry(const struct dentry *dentry, struct qstr *str); -int hfsplus_compare_dentry(const struct dentry *parent, - const struct dentry *dentry, unsigned int len, +int hfsplus_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); /* wrapper.c */ diff --git a/fs/hfsplus/unicode.c b/fs/hfsplus/unicode.c index 509d7e231460..e563939882f3 100644 --- a/fs/hfsplus/unicode.c +++ b/fs/hfsplus/unicode.c @@ -385,7 +385,7 @@ int hfsplus_hash_dentry(const struct dentry *dentry, struct qstr *str) * Composed unicode characters are decomposed and case-folding is performed * if the appropriate bits are (un)set on the superblock. */ -int hfsplus_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +int hfsplus_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct super_block *sb = dentry->d_sb; diff --git a/fs/hpfs/dentry.c b/fs/hpfs/dentry.c index a804300f4dce..bb87d65f0d97 100644 --- a/fs/hpfs/dentry.c +++ b/fs/hpfs/dentry.c @@ -34,7 +34,7 @@ static int hpfs_hash_dentry(const struct dentry *dentry, struct qstr *qstr) return 0; } -static int hpfs_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +static int hpfs_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { unsigned al = len; diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 761fade7680f..ad0c745ebad7 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -29,18 +29,15 @@ #define BEQUIET static int isofs_hashi(const struct dentry *parent, struct qstr *qstr); -static int isofs_dentry_cmpi(const struct dentry *parent, - const struct dentry *dentry, +static int isofs_dentry_cmpi(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); #ifdef CONFIG_JOLIET static int isofs_hashi_ms(const struct dentry *parent, struct qstr *qstr); static int isofs_hash_ms(const struct dentry *parent, struct qstr *qstr); -static int isofs_dentry_cmpi_ms(const struct dentry *parent, - const struct dentry *dentry, +static int isofs_dentry_cmpi_ms(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); -static int isofs_dentry_cmp_ms(const struct dentry *parent, - const struct dentry *dentry, +static int isofs_dentry_cmp_ms(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name); #endif @@ -235,7 +232,7 @@ isofs_hashi(const struct dentry *dentry, struct qstr *qstr) } static int -isofs_dentry_cmpi(const struct dentry *parent, const struct dentry *dentry, +isofs_dentry_cmpi(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { return isofs_dentry_cmp_common(len, str, name, 0, 1); @@ -276,14 +273,14 @@ isofs_hashi_ms(const struct dentry *dentry, struct qstr *qstr) } static int -isofs_dentry_cmp_ms(const struct dentry *parent, const struct dentry *dentry, +isofs_dentry_cmp_ms(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { return isofs_dentry_cmp_common(len, str, name, 1, 0); } static int -isofs_dentry_cmpi_ms(const struct dentry *parent, const struct dentry *dentry, +isofs_dentry_cmpi_ms(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { return isofs_dentry_cmp_common(len, str, name, 1, 1); diff --git a/fs/isofs/namei.c b/fs/isofs/namei.c index 7b543e6b6526..aee592767f1d 100644 --- a/fs/isofs/namei.c +++ b/fs/isofs/namei.c @@ -22,7 +22,7 @@ isofs_cmp(struct dentry *dentry, const char *compare, int dlen) qstr.len = dlen; if (likely(!dentry->d_op)) return dentry->d_name.len != dlen || memcmp(dentry->d_name.name, compare, dlen); - return dentry->d_op->d_compare(NULL, NULL, dentry->d_name.len, dentry->d_name.name, &qstr); + return dentry->d_op->d_compare(NULL, dentry->d_name.len, dentry->d_name.name, &qstr); } /* diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c index 04baf0dfc40c..814b0c58016c 100644 --- a/fs/jfs/namei.c +++ b/fs/jfs/namei.c @@ -1572,7 +1572,7 @@ static int jfs_ci_hash(const struct dentry *dir, struct qstr *this) return 0; } -static int jfs_ci_compare(const struct dentry *parent, const struct dentry *dentry, +static int jfs_ci_compare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { int i, result = 1; diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c index 9add7ab747a5..17de5c13dfae 100644 --- a/fs/ncpfs/dir.c +++ b/fs/ncpfs/dir.c @@ -74,7 +74,7 @@ const struct inode_operations ncp_dir_inode_operations = */ static int ncp_lookup_validate(struct dentry *, unsigned int); static int ncp_hash_dentry(const struct dentry *, struct qstr *); -static int ncp_compare_dentry(const struct dentry *, const struct dentry *, +static int ncp_compare_dentry(const struct dentry *, unsigned int, const char *, const struct qstr *); static int ncp_delete_dentry(const struct dentry *); static void ncp_d_prune(struct dentry *dentry); @@ -154,7 +154,7 @@ ncp_hash_dentry(const struct dentry *dentry, struct qstr *this) * the callers will handle races. */ static int -ncp_compare_dentry(const struct dentry *parent, const struct dentry *dentry, +ncp_compare_dentry(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct inode *pinode; @@ -162,7 +162,7 @@ ncp_compare_dentry(const struct dentry *parent, const struct dentry *dentry, if (len != name->len) return 1; - pinode = d_inode_rcu(parent); + pinode = d_inode_rcu(dentry->d_parent); if (!pinode) return 1; diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index b59db94d2ff4..30bb00130d0f 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -834,7 +834,7 @@ static int sysctl_is_seen(struct ctl_table_header *p) return res; } -static int proc_sys_compare(const struct dentry *parent, const struct dentry *dentry, +static int proc_sys_compare(const struct dentry *dentry, unsigned int len, const char *str, const struct qstr *name) { struct ctl_table_header *head; diff --git a/include/linux/dcache.h b/include/linux/dcache.h index 98044a8d1487..107d9abe7166 100644 --- a/include/linux/dcache.h +++ b/include/linux/dcache.h @@ -130,7 +130,7 @@ struct dentry_operations { int (*d_revalidate)(struct dentry *, unsigned int); int (*d_weak_revalidate)(struct dentry *, unsigned int); int (*d_hash)(const struct dentry *, struct qstr *); - int (*d_compare)(const struct dentry *, const struct dentry *, + int (*d_compare)(const struct dentry *, unsigned int, const char *, const struct qstr *); int (*d_delete)(const struct dentry *); int (*d_init)(struct dentry *); -- cgit v1.2.3 From 31b7c1ab4ed14d3aeb658e11a114860a64290fea Mon Sep 17 00:00:00 2001 From: Martin Brandenburg Date: Mon, 8 Feb 2016 17:01:29 -0500 Subject: orangefs: Use d_time to avoid excessive lookups Signed-off-by: Martin Brandenburg --- fs/orangefs/dcache.c | 4 ++++ fs/orangefs/namei.c | 5 +++++ 2 files changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/orangefs/dcache.c b/fs/orangefs/dcache.c index 5dfc4f3cfe68..dcb55bbb0b69 100644 --- a/fs/orangefs/dcache.c +++ b/fs/orangefs/dcache.c @@ -73,6 +73,7 @@ static int orangefs_revalidate_lookup(struct dentry *dentry) } } + dentry->d_time = jiffies + HZ; ret = 1; out_release_op: op_release(new_op); @@ -94,6 +95,9 @@ static int orangefs_d_revalidate(struct dentry *dentry, unsigned int flags) { int ret; + if (dentry->d_time > jiffies) + return 1; + if (flags & LOOKUP_RCU) return -ECHILD; diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c index 7e8dfa97c44a..b37f80bf0bec 100644 --- a/fs/orangefs/namei.c +++ b/fs/orangefs/namei.c @@ -72,6 +72,7 @@ static int orangefs_create(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + HZ; gossip_debug(GOSSIP_NAME_DEBUG, "%s: dentry instantiated for %s\n", @@ -181,6 +182,8 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } + dentry->d_time = jiffies + HZ; + inode = orangefs_iget(dir->i_sb, &new_op->downcall.resp.lookup.refn); if (IS_ERR(inode)) { gossip_debug(GOSSIP_NAME_DEBUG, @@ -316,6 +319,7 @@ static int orangefs_symlink(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + HZ; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Symlink) %pU -> %s\n", @@ -378,6 +382,7 @@ static int orangefs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode d_instantiate(dentry, inode); unlock_new_inode(inode); + dentry->d_time = jiffies + HZ; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Directory) %pU -> %s\n", -- cgit v1.2.3 From 71680c18c8f22deafbaaf76d1c2d0eed2899a3d2 Mon Sep 17 00:00:00 2001 From: Martin Brandenburg Date: Thu, 9 Jun 2016 16:32:38 -0400 Subject: orangefs: Cache getattr results. The userspace component attempts to do this, but this will prevent us from even needing to go into userspace to satisfy certain getattr requests. Signed-off-by: Martin Brandenburg --- fs/orangefs/inode.c | 6 +++--- fs/orangefs/namei.c | 7 +++++++ fs/orangefs/orangefs-kernel.h | 4 +++- fs/orangefs/orangefs-utils.c | 38 +++++++++++++++++++++----------------- fs/orangefs/protocol.h | 8 -------- 5 files changed, 34 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/orangefs/inode.c b/fs/orangefs/inode.c index 2e63e6d0a68e..28a0557a69be 100644 --- a/fs/orangefs/inode.c +++ b/fs/orangefs/inode.c @@ -262,7 +262,7 @@ int orangefs_getattr(struct vfsmount *mnt, "orangefs_getattr: called on %s\n", dentry->d_name.name); - ret = orangefs_inode_getattr(inode, 0, 1); + ret = orangefs_inode_getattr(inode, 0, 0); if (ret == 0) { generic_fillattr(inode, kstat); @@ -384,7 +384,7 @@ struct inode *orangefs_iget(struct super_block *sb, struct orangefs_object_kref if (!inode || !(inode->i_state & I_NEW)) return inode; - error = orangefs_inode_getattr(inode, 1, 0); + error = orangefs_inode_getattr(inode, 1, 1); if (error) { iget_failed(inode); return ERR_PTR(error); @@ -429,7 +429,7 @@ struct inode *orangefs_new_inode(struct super_block *sb, struct inode *dir, orangefs_set_inode(inode, ref); inode->i_ino = hash; /* needed for stat etc */ - error = orangefs_inode_getattr(inode, 1, 0); + error = orangefs_inode_getattr(inode, 1, 1); if (error) goto out_iput; diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c index b37f80bf0bec..dde6c36f5a69 100644 --- a/fs/orangefs/namei.c +++ b/fs/orangefs/namei.c @@ -73,6 +73,7 @@ static int orangefs_create(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + HZ; + ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, "%s: dentry instantiated for %s\n", @@ -192,6 +193,8 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } + ORANGEFS_I(inode)->getattr_time = 0; + gossip_debug(GOSSIP_NAME_DEBUG, "%s:%s:%d " "Found good inode [%lu] with count [%d]\n", @@ -320,6 +323,7 @@ static int orangefs_symlink(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + HZ; + ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Symlink) %pU -> %s\n", @@ -383,6 +387,7 @@ static int orangefs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + HZ; + ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Directory) %pU -> %s\n", @@ -413,6 +418,8 @@ static int orangefs_rename(struct inode *old_dir, "orangefs_rename: called (%pd2 => %pd2) ct=%d\n", old_dentry, new_dentry, d_count(new_dentry)); + ORANGEFS_I(new_dentry->d_parent->d_inode)->getattr_time = 0; + new_op = op_alloc(ORANGEFS_VFS_OP_RENAME); if (!new_op) return -EINVAL; diff --git a/fs/orangefs/orangefs-kernel.h b/fs/orangefs/orangefs-kernel.h index 4b6e132d5a0f..9338fe64820d 100644 --- a/fs/orangefs/orangefs-kernel.h +++ b/fs/orangefs/orangefs-kernel.h @@ -246,6 +246,8 @@ struct orangefs_inode_s { * with this object */ unsigned long pinode_flags; + + unsigned long getattr_time; }; #define P_ATIME_FLAG 0 @@ -527,7 +529,7 @@ int orangefs_inode_setxattr(struct inode *inode, size_t size, int flags); -int orangefs_inode_getattr(struct inode *inode, int new, int size); +int orangefs_inode_getattr(struct inode *inode, int new, int bypass); int orangefs_inode_check_changed(struct inode *inode); diff --git a/fs/orangefs/orangefs-utils.c b/fs/orangefs/orangefs-utils.c index c5fbc62357c6..69b4d8af8d81 100644 --- a/fs/orangefs/orangefs-utils.c +++ b/fs/orangefs/orangefs-utils.c @@ -251,7 +251,7 @@ static int orangefs_inode_is_stale(struct inode *inode, int new, return 0; } -int orangefs_inode_getattr(struct inode *inode, int new, int size) +int orangefs_inode_getattr(struct inode *inode, int new, int bypass) { struct orangefs_inode_s *orangefs_inode = ORANGEFS_I(inode); struct orangefs_kernel_op_s *new_op; @@ -261,12 +261,16 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) gossip_debug(GOSSIP_UTILS_DEBUG, "%s: called on inode %pU\n", __func__, get_khandle_from_ino(inode)); + if (!new && !bypass) { + if (orangefs_inode->getattr_time > jiffies) + return 0; + } + new_op = op_alloc(ORANGEFS_VFS_OP_GETATTR); if (!new_op) return -ENOMEM; new_op->upcall.req.getattr.refn = orangefs_inode->refn; - new_op->upcall.req.getattr.mask = size ? - ORANGEFS_ATTR_SYS_ALL_NOHINT : ORANGEFS_ATTR_SYS_ALL_NOHINT_NOSIZE; + new_op->upcall.req.getattr.mask = ORANGEFS_ATTR_SYS_ALL_NOHINT; ret = service_operation(new_op, __func__, get_interruptible_flag(inode)); @@ -287,20 +291,18 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) case S_IFREG: inode->i_flags = orangefs_inode_flags(&new_op-> downcall.resp.getattr.attributes); - if (size) { - inode_size = (loff_t)new_op-> - downcall.resp.getattr.attributes.size; - rounded_up_size = - (inode_size + (4096 - (inode_size % 4096))); - inode->i_size = inode_size; - orangefs_inode->blksize = - new_op->downcall.resp.getattr.attributes.blksize; - spin_lock(&inode->i_lock); - inode->i_bytes = inode_size; - inode->i_blocks = - (unsigned long)(rounded_up_size / 512); - spin_unlock(&inode->i_lock); - } + inode_size = (loff_t)new_op-> + downcall.resp.getattr.attributes.size; + rounded_up_size = + (inode_size + (4096 - (inode_size % 4096))); + inode->i_size = inode_size; + orangefs_inode->blksize = + new_op->downcall.resp.getattr.attributes.blksize; + spin_lock(&inode->i_lock); + inode->i_bytes = inode_size; + inode->i_blocks = + (unsigned long)(rounded_up_size / 512); + spin_unlock(&inode->i_lock); break; case S_IFDIR: inode->i_size = PAGE_SIZE; @@ -345,6 +347,7 @@ int orangefs_inode_getattr(struct inode *inode, int new, int size) inode->i_mode = type | (is_root_handle(inode) ? S_ISVTX : 0) | orangefs_inode_perms(&new_op->downcall.resp.getattr.attributes); + orangefs_inode->getattr_time = jiffies + HZ; ret = 0; out: op_release(new_op); @@ -418,6 +421,7 @@ int orangefs_inode_setattr(struct inode *inode, struct iattr *iattr) ClearMtimeFlag(orangefs_inode); ClearCtimeFlag(orangefs_inode); ClearModeFlag(orangefs_inode); + orangefs_inode->getattr_time = 0; } return ret; diff --git a/fs/orangefs/protocol.h b/fs/orangefs/protocol.h index 1efc6f8a5224..3d7418c728f5 100644 --- a/fs/orangefs/protocol.h +++ b/fs/orangefs/protocol.h @@ -207,14 +207,6 @@ typedef __s64 ORANGEFS_offset; ORANGEFS_ATTR_SYS_DIRENT_COUNT | \ ORANGEFS_ATTR_SYS_BLKSIZE) -#define ORANGEFS_ATTR_SYS_ALL_NOHINT_NOSIZE \ - (ORANGEFS_ATTR_SYS_COMMON_ALL | \ - ORANGEFS_ATTR_SYS_LNK_TARGET | \ - ORANGEFS_ATTR_SYS_DFILE_COUNT | \ - ORANGEFS_ATTR_SYS_MIRROR_COPIES_COUNT | \ - ORANGEFS_ATTR_SYS_DIRENT_COUNT | \ - ORANGEFS_ATTR_SYS_BLKSIZE) - #define ORANGEFS_XATTR_REPLACE 0x2 #define ORANGEFS_XATTR_CREATE 0x1 #define ORANGEFS_MAX_SERVER_ADDR_LEN 256 -- cgit v1.2.3 From 4cd8f319441a323280eaa0acf390ffb18c64d86b Mon Sep 17 00:00:00 2001 From: Martin Brandenburg Date: Mon, 25 Jul 2016 13:58:24 -0400 Subject: orangefs: Allow dcache and getattr cache time to be configured. Signed-off-by: Martin Brandenburg --- fs/orangefs/dcache.c | 2 +- fs/orangefs/namei.c | 8 ++++---- fs/orangefs/orangefs-kernel.h | 2 ++ fs/orangefs/orangefs-mod.c | 2 ++ fs/orangefs/orangefs-sysfs.c | 43 ++++++++++++++++++++++++++++++++++++++++++- fs/orangefs/orangefs-utils.c | 2 +- 6 files changed, 52 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/orangefs/dcache.c b/fs/orangefs/dcache.c index dcb55bbb0b69..2521af6707f7 100644 --- a/fs/orangefs/dcache.c +++ b/fs/orangefs/dcache.c @@ -73,7 +73,7 @@ static int orangefs_revalidate_lookup(struct dentry *dentry) } } - dentry->d_time = jiffies + HZ; + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; ret = 1; out_release_op: op_release(new_op); diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c index dde6c36f5a69..77071b78816a 100644 --- a/fs/orangefs/namei.c +++ b/fs/orangefs/namei.c @@ -72,7 +72,7 @@ static int orangefs_create(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); - dentry->d_time = jiffies + HZ; + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, @@ -183,7 +183,7 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } - dentry->d_time = jiffies + HZ; + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; inode = orangefs_iget(dir->i_sb, &new_op->downcall.resp.lookup.refn); if (IS_ERR(inode)) { @@ -322,7 +322,7 @@ static int orangefs_symlink(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); - dentry->d_time = jiffies + HZ; + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, @@ -386,7 +386,7 @@ static int orangefs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode d_instantiate(dentry, inode); unlock_new_inode(inode); - dentry->d_time = jiffies + HZ; + dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; ORANGEFS_I(inode)->getattr_time = 0; gossip_debug(GOSSIP_NAME_DEBUG, diff --git a/fs/orangefs/orangefs-kernel.h b/fs/orangefs/orangefs-kernel.h index 9338fe64820d..633c07a6e3d8 100644 --- a/fs/orangefs/orangefs-kernel.h +++ b/fs/orangefs/orangefs-kernel.h @@ -548,6 +548,8 @@ extern struct mutex request_mutex; extern int debug; extern int op_timeout_secs; extern int slot_timeout_secs; +extern int dcache_timeout_msecs; +extern int getattr_timeout_msecs; extern struct list_head orangefs_superblocks; extern spinlock_t orangefs_superblocks_lock; extern struct list_head orangefs_request_list; diff --git a/fs/orangefs/orangefs-mod.c b/fs/orangefs/orangefs-mod.c index 6f072a8c0de1..1caef0aab63e 100644 --- a/fs/orangefs/orangefs-mod.c +++ b/fs/orangefs/orangefs-mod.c @@ -47,6 +47,8 @@ struct client_debug_mask client_debug_mask = { NULL, 0, 0 }; unsigned int kernel_mask_set_mod_init; /* implicitly false */ int op_timeout_secs = ORANGEFS_DEFAULT_OP_TIMEOUT_SECS; int slot_timeout_secs = ORANGEFS_DEFAULT_SLOT_TIMEOUT_SECS; +int dcache_timeout_msecs = 1000; +int getattr_timeout_msecs = 1000; MODULE_LICENSE("GPL"); MODULE_AUTHOR("ORANGEFS Development Team"); diff --git a/fs/orangefs/orangefs-sysfs.c b/fs/orangefs/orangefs-sysfs.c index 5c03113e3ad2..375708c2db87 100644 --- a/fs/orangefs/orangefs-sysfs.c +++ b/fs/orangefs/orangefs-sysfs.c @@ -61,10 +61,21 @@ * Slots are requested and waited for, * the wait times out after slot_timeout_secs. * + * What: /sys/fs/orangefs/dcache_timeout_msecs + * Date: Jul 2016 + * Contact: Martin Brandenburg + * Description: + * Time lookup is valid in milliseconds. + * + * What: /sys/fs/orangefs/getattr_timeout_msecs + * Date: Jul 2016 + * Contact: Martin Brandenburg + * Description: + * Time getattr is valid in milliseconds. * * What: /sys/fs/orangefs/acache/... * Date: Jun 2015 - * Contact: Mike Marshall + * Contact: Martin Brandenburg * Description: * Attribute cache configurable settings. * @@ -117,6 +128,8 @@ struct orangefs_obj { int perf_history_size; int perf_time_interval_secs; int slot_timeout_secs; + int dcache_timeout_msecs; + int getattr_timeout_msecs; }; struct acache_orangefs_obj { @@ -658,6 +671,20 @@ static ssize_t sysfs_int_show(char *kobj_id, char *buf, void *attr) "%d\n", slot_timeout_secs); goto out; + } else if (!strcmp(orangefs_attr->attr.name, + "dcache_timeout_msecs")) { + rc = scnprintf(buf, + PAGE_SIZE, + "%d\n", + dcache_timeout_msecs); + goto out; + } else if (!strcmp(orangefs_attr->attr.name, + "getattr_timeout_msecs")) { + rc = scnprintf(buf, + PAGE_SIZE, + "%d\n", + getattr_timeout_msecs); + goto out; } else { goto out; } @@ -734,6 +761,12 @@ static ssize_t int_store(struct orangefs_obj *orangefs_obj, } else if (!strcmp(attr->attr.name, "slot_timeout_secs")) { rc = kstrtoint(buf, 0, &slot_timeout_secs); goto out; + } else if (!strcmp(attr->attr.name, "dcache_timeout_msecs")) { + rc = kstrtoint(buf, 0, &dcache_timeout_msecs); + goto out; + } else if (!strcmp(attr->attr.name, "getattr_timeout_msecs")) { + rc = kstrtoint(buf, 0, &getattr_timeout_msecs); + goto out; } else { goto out; } @@ -1361,6 +1394,12 @@ static struct orangefs_attribute op_timeout_secs_attribute = static struct orangefs_attribute slot_timeout_secs_attribute = __ATTR(slot_timeout_secs, 0664, int_orangefs_show, int_store); +static struct orangefs_attribute dcache_timeout_msecs_attribute = + __ATTR(dcache_timeout_msecs, 0664, int_orangefs_show, int_store); + +static struct orangefs_attribute getattr_timeout_msecs_attribute = + __ATTR(getattr_timeout_msecs, 0664, int_orangefs_show, int_store); + static struct orangefs_attribute perf_counter_reset_attribute = __ATTR(perf_counter_reset, 0664, @@ -1382,6 +1421,8 @@ static struct orangefs_attribute perf_time_interval_secs_attribute = static struct attribute *orangefs_default_attrs[] = { &op_timeout_secs_attribute.attr, &slot_timeout_secs_attribute.attr, + &dcache_timeout_msecs_attribute.attr, + &getattr_timeout_msecs_attribute.attr, &perf_counter_reset_attribute.attr, &perf_history_size_attribute.attr, &perf_time_interval_secs_attribute.attr, diff --git a/fs/orangefs/orangefs-utils.c b/fs/orangefs/orangefs-utils.c index 69b4d8af8d81..cd512a3a9c52 100644 --- a/fs/orangefs/orangefs-utils.c +++ b/fs/orangefs/orangefs-utils.c @@ -347,7 +347,7 @@ int orangefs_inode_getattr(struct inode *inode, int new, int bypass) inode->i_mode = type | (is_root_handle(inode) ? S_ISVTX : 0) | orangefs_inode_perms(&new_op->downcall.resp.getattr.attributes); - orangefs_inode->getattr_time = jiffies + HZ; + orangefs_inode->getattr_time = jiffies + getattr_timeout_msecs*HZ/1000; ret = 0; out: op_release(new_op); -- cgit v1.2.3 From 957ee43718c5a226fa70ef5e680d037ca9156af6 Mon Sep 17 00:00:00 2001 From: Martin Brandenburg Date: Tue, 26 Jul 2016 13:23:04 -0400 Subject: orangefs: Change default dcache and getattr timeout to 50 msec. Signed-off-by: Martin Brandenburg --- fs/orangefs/orangefs-mod.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/orangefs/orangefs-mod.c b/fs/orangefs/orangefs-mod.c index 1caef0aab63e..e9fd5755c05f 100644 --- a/fs/orangefs/orangefs-mod.c +++ b/fs/orangefs/orangefs-mod.c @@ -47,8 +47,8 @@ struct client_debug_mask client_debug_mask = { NULL, 0, 0 }; unsigned int kernel_mask_set_mod_init; /* implicitly false */ int op_timeout_secs = ORANGEFS_DEFAULT_OP_TIMEOUT_SECS; int slot_timeout_secs = ORANGEFS_DEFAULT_SLOT_TIMEOUT_SECS; -int dcache_timeout_msecs = 1000; -int getattr_timeout_msecs = 1000; +int dcache_timeout_msecs = 50; +int getattr_timeout_msecs = 50; MODULE_LICENSE("GPL"); MODULE_AUTHOR("ORANGEFS Development Team"); -- cgit v1.2.3 From 8bbb20a863ca72dfb9025a4653f21b5abf926d20 Mon Sep 17 00:00:00 2001 From: Martin Brandenburg Date: Thu, 28 Jul 2016 14:46:36 -0400 Subject: orangefs: Account for jiffies wraparound. Signed-off-by: Martin Brandenburg --- fs/orangefs/dcache.c | 2 +- fs/orangefs/namei.c | 10 +++++----- fs/orangefs/orangefs-utils.c | 4 ++-- 3 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/orangefs/dcache.c b/fs/orangefs/dcache.c index 2521af6707f7..00235bf644dc 100644 --- a/fs/orangefs/dcache.c +++ b/fs/orangefs/dcache.c @@ -95,7 +95,7 @@ static int orangefs_d_revalidate(struct dentry *dentry, unsigned int flags) { int ret; - if (dentry->d_time > jiffies) + if (time_before(jiffies, dentry->d_time)) return 1; if (flags & LOOKUP_RCU) diff --git a/fs/orangefs/namei.c b/fs/orangefs/namei.c index 77071b78816a..62c525936ee8 100644 --- a/fs/orangefs/namei.c +++ b/fs/orangefs/namei.c @@ -73,7 +73,7 @@ static int orangefs_create(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; - ORANGEFS_I(inode)->getattr_time = 0; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "%s: dentry instantiated for %s\n", @@ -193,7 +193,7 @@ static struct dentry *orangefs_lookup(struct inode *dir, struct dentry *dentry, goto out; } - ORANGEFS_I(inode)->getattr_time = 0; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "%s:%s:%d " @@ -323,7 +323,7 @@ static int orangefs_symlink(struct inode *dir, d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; - ORANGEFS_I(inode)->getattr_time = 0; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Symlink) %pU -> %s\n", @@ -387,7 +387,7 @@ static int orangefs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode d_instantiate(dentry, inode); unlock_new_inode(inode); dentry->d_time = jiffies + dcache_timeout_msecs*HZ/1000; - ORANGEFS_I(inode)->getattr_time = 0; + ORANGEFS_I(inode)->getattr_time = jiffies - 1; gossip_debug(GOSSIP_NAME_DEBUG, "Inode (Directory) %pU -> %s\n", @@ -418,7 +418,7 @@ static int orangefs_rename(struct inode *old_dir, "orangefs_rename: called (%pd2 => %pd2) ct=%d\n", old_dentry, new_dentry, d_count(new_dentry)); - ORANGEFS_I(new_dentry->d_parent->d_inode)->getattr_time = 0; + ORANGEFS_I(new_dentry->d_parent->d_inode)->getattr_time = jiffies - 1; new_op = op_alloc(ORANGEFS_VFS_OP_RENAME); if (!new_op) diff --git a/fs/orangefs/orangefs-utils.c b/fs/orangefs/orangefs-utils.c index cd512a3a9c52..d13c7291fd05 100644 --- a/fs/orangefs/orangefs-utils.c +++ b/fs/orangefs/orangefs-utils.c @@ -262,7 +262,7 @@ int orangefs_inode_getattr(struct inode *inode, int new, int bypass) get_khandle_from_ino(inode)); if (!new && !bypass) { - if (orangefs_inode->getattr_time > jiffies) + if (time_before(jiffies, orangefs_inode->getattr_time)) return 0; } @@ -421,7 +421,7 @@ int orangefs_inode_setattr(struct inode *inode, struct iattr *iattr) ClearMtimeFlag(orangefs_inode); ClearCtimeFlag(orangefs_inode); ClearModeFlag(orangefs_inode); - orangefs_inode->getattr_time = 0; + orangefs_inode->getattr_time = jiffies - 1; } return ret; -- cgit v1.2.3 From 01a36b6758e723f919420ef20cea5fca1fc06e2b Mon Sep 17 00:00:00 2001 From: Gang He Date: Tue, 2 Aug 2016 14:02:07 -0700 Subject: ocfs2: ensure that dlm lockspace is created by kernel module We encountered a bug from the customer, the user did a fsck.ocfs2 on the file system and exited unusually, the lockspace (with LVB size = 32) was left in the kernel space, next, the user mounted this file system, the kernel module did not create a new lockspace (LVB size = 64) via calling dlm_new_lockspace() function in mounting stage, just used the existing lockspace, created by the user space tool, this would lead the user was not able to mount this file system from the other nodes, with the error message like: dlm: 032F5......: config mismatch: 64,0 nodeid 177127961: 32,0 (mount.ocfs2,26981,46):ocfs2_dlm_init:2995 ERROR: status = -71 ocfs2_mount_volume:1881 ERROR: status = -71 ocfs2_fill_super:1236 ERROR: status = -71 The user found it very difficult to find the root cause, then, we brought out this patch to relieve such problem. First, we add one more flag in calling dlm_new_lockspace() function, to make sure the lockspace is created by kernel module itself, and this change will not affect the backward compatibility. Second, the obvious error message is reported in the kernel log, let the user be more easy to find the root cause. This patch will be used to insure the dlm lockspace is created by kernel module when mounting a ocfs2 file system. There are two ways to create a lockspace, from user space and kernel space, but the same name lockspaces probably have different lvblen lengths/flags. To avoid this mix using, we add one more flag DLM_LSFL_NEWEXCL, it will make sure the dlm lockspace is created by kernel module when mounting. Secondly, if a user space program (ocfs2-tools) is running on a file system, the user tries to mount this file system in the cluster, DLM module will return a -EEXIST or -EPROTO errno, we should give the user a obvious error message, then, the user can let that user space tool exit before mounting the file system again. Link: http://lkml.kernel.org/r/1463731940-13044-2-git-send-email-ghe@suse.com Signed-off-by: Gang He Reviewed-by: Goldwyn Rodrigues Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Joseph Qi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/stack_user.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index ced70c8139f7..c9e828ec3c8e 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -1007,10 +1007,17 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn) lc->oc_type = NO_CONTROLD; rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name, - DLM_LSFL_FS, DLM_LVB_LEN, + DLM_LSFL_FS | DLM_LSFL_NEWEXCL, DLM_LVB_LEN, &ocfs2_ls_ops, conn, &ops_rv, &fsdlm); - if (rc) + if (rc) { + if (rc == -EEXIST || rc == -EPROTO) + printk(KERN_ERR "ocfs2: Unable to create the " + "lockspace %s (%d), because a ocfs2-tools " + "program is running on this file system " + "with the same name lockspace\n", + conn->cc_name, rc); goto out; + } if (ops_rv == -EOPNOTSUPP) { lc->oc_type = WITH_CONTROLD; -- cgit v1.2.3 From 2070ad1aebfff2c26190188844c38e55d2df2ae2 Mon Sep 17 00:00:00 2001 From: Eric Ren Date: Tue, 2 Aug 2016 14:02:10 -0700 Subject: ocfs2: retry on ENOSPC if sufficient space in truncate log The testcase "mmaptruncate" in ocfs2 test suite always fails with ENOSPC error on small volume (say less than 10G). This testcase repeatedly performs "extend" and "truncate" on a file. Continuously, it truncates the file to 1/2 of the size, and then extends to 100% of the size. The main bitmap will quickly run out of space because the "truncate" code prevent truncate log from being flushed by ocfs2_schedule_truncate_log_flush(osb, 1), while truncate log may have cached lots of clusters. So retry to allocate after flushing truncate log when ENOSPC is returned. And we cannot reuse the deleted blocks before the transaction committed. Fortunately, we already have a function to do this - ocfs2_try_to_free_truncate_log(). Just need to remove the "static" modifier and put it into the right place. The "unlock"/"lock" code isn't elegant, but there seems to be no better option. [zren@suse.com: locking fix] Link: http://lkml.kernel.org/r/1468031546-4797-1-git-send-email-zren@suse.com Link: http://lkml.kernel.org/r/1466586469-5541-1-git-send-email-zren@suse.com Signed-off-by: Eric Ren Reviewed-by: Gang He Reviewed-by: Joseph Qi Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/alloc.c | 37 +++++++++++++++++++++++++++++++++++++ fs/ocfs2/alloc.h | 2 ++ fs/ocfs2/aops.c | 37 ------------------------------------- fs/ocfs2/suballoc.c | 20 +++++++++++++++++++- 4 files changed, 58 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 460c0cedab3a..7dabbc31060e 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -6106,6 +6106,43 @@ void ocfs2_schedule_truncate_log_flush(struct ocfs2_super *osb, } } +/* + * Try to flush truncate logs if we can free enough clusters from it. + * As for return value, "< 0" means error, "0" no space and "1" means + * we have freed enough spaces and let the caller try to allocate again. + */ +int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, + unsigned int needed) +{ + tid_t target; + int ret = 0; + unsigned int truncated_clusters; + + inode_lock(osb->osb_tl_inode); + truncated_clusters = osb->truncated_clusters; + inode_unlock(osb->osb_tl_inode); + + /* + * Check whether we can succeed in allocating if we free + * the truncate log. + */ + if (truncated_clusters < needed) + goto out; + + ret = ocfs2_flush_truncate_log(osb); + if (ret) { + mlog_errno(ret); + goto out; + } + + if (jbd2_journal_start_commit(osb->journal->j_journal, &target)) { + jbd2_log_wait_commit(osb->journal->j_journal, target); + ret = 1; + } +out: + return ret; +} + static int ocfs2_get_truncate_log_info(struct ocfs2_super *osb, int slot_num, struct inode **tl_inode, diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h index f3dc1b0dfffc..4a5152ec88a3 100644 --- a/fs/ocfs2/alloc.h +++ b/fs/ocfs2/alloc.h @@ -188,6 +188,8 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb, u64 start_blk, unsigned int num_clusters); int __ocfs2_flush_truncate_log(struct ocfs2_super *osb); +int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, + unsigned int needed); /* * Process local structure which describes the block unlinks done diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index af2adfcb0f6f..98d36548153d 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -1645,43 +1645,6 @@ static int ocfs2_zero_tail(struct inode *inode, struct buffer_head *di_bh, return ret; } -/* - * Try to flush truncate logs if we can free enough clusters from it. - * As for return value, "< 0" means error, "0" no space and "1" means - * we have freed enough spaces and let the caller try to allocate again. - */ -static int ocfs2_try_to_free_truncate_log(struct ocfs2_super *osb, - unsigned int needed) -{ - tid_t target; - int ret = 0; - unsigned int truncated_clusters; - - inode_lock(osb->osb_tl_inode); - truncated_clusters = osb->truncated_clusters; - inode_unlock(osb->osb_tl_inode); - - /* - * Check whether we can succeed in allocating if we free - * the truncate log. - */ - if (truncated_clusters < needed) - goto out; - - ret = ocfs2_flush_truncate_log(osb); - if (ret) { - mlog_errno(ret); - goto out; - } - - if (jbd2_journal_start_commit(osb->journal->j_journal, &target)) { - jbd2_log_wait_commit(osb->journal->j_journal, target); - ret = 1; - } -out: - return ret; -} - int ocfs2_write_begin_nolock(struct address_space *mapping, loff_t pos, unsigned len, ocfs2_write_type_t type, struct page **pagep, void **fsdata, diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 2f19aeec5482..ea47120a85ff 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -1164,7 +1164,8 @@ static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb, int flags, struct ocfs2_alloc_context **ac) { - int status; + int status, ret = 0; + int retried = 0; *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL); if (!(*ac)) { @@ -1189,7 +1190,24 @@ static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb, } if (status == -ENOSPC) { +retry: status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac); + /* Retry if there is sufficient space cached in truncate log */ + if (status == -ENOSPC && !retried) { + retried = 1; + ocfs2_inode_unlock((*ac)->ac_inode, 1); + inode_unlock((*ac)->ac_inode); + + ret = ocfs2_try_to_free_truncate_log(osb, bits_wanted); + if (ret == 1) + goto retry; + + if (ret < 0) + mlog_errno(ret); + + inode_lock((*ac)->ac_inode); + ocfs2_inode_lock((*ac)->ac_inode, NULL, 1); + } if (status < 0) { if (status != -ENOSPC) mlog_errno(status); -- cgit v1.2.3 From 86b652b93adb57d8fed8edd532ed2eb8a791950d Mon Sep 17 00:00:00 2001 From: piaojun Date: Tue, 2 Aug 2016 14:02:13 -0700 Subject: ocfs2/dlm: disable BUG_ON when DLM_LOCK_RES_DROPPING_REF is cleared before dlm_deref_lockres_done_handler We found a BUG situation in which DLM_LOCK_RES_DROPPING_REF is cleared unexpected that described below. To solve the bug, we disable the BUG_ON and purge lockres in dlm_do_local_recovery_cleanup. Node 1 Node 2(master) dlm_purge_lockres dlm_deref_lockres_handler DLM_LOCK_RES_SETREF_INPROG is set response DLM_DEREF_RESPONSE_INPROG receive DLM_DEREF_RESPONSE_INPROG stop puring in dlm_purge_lockres and wait for DLM_DEREF_RESPONSE_DONE dispatch dlm_deref_lockres_worker response DLM_DEREF_RESPONSE_DONE receive DLM_DEREF_RESPONSE_DONE and prepare to purge lockres Node 2 goes down find Node2 down and do local clean up for Node2: dlm_do_local_recovery_cleanup -> clear DLM_LOCK_RES_DROPPING_REF when purging lockres, BUG_ON happens because DLM_LOCK_RES_DROPPING_REF is clear: dlm_deref_lockres_done_handler ->BUG_ON(!(res->state & DLM_LOCK_RES_DROPPING_REF)); [akpm@linux-foundation.org: fix duplicated write to `ret'] Fixes: 60d663cb5273 ("ocfs2/dlm: add DEREF_DONE message") Link: http://lkml.kernel.org/r/57845055.9080702@huawei.com Signed-off-by: Jun Piao Reviewed-by: Joseph Qi Reviewed-by: Jiufei Xue Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/dlm/dlmmaster.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 13719d3f35f8..525dc06468c4 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -2416,7 +2416,17 @@ int dlm_deref_lockres_done_handler(struct o2net_msg *msg, u32 len, void *data, } spin_lock(&res->spinlock); - BUG_ON(!(res->state & DLM_LOCK_RES_DROPPING_REF)); + if (!(res->state & DLM_LOCK_RES_DROPPING_REF)) { + spin_unlock(&res->spinlock); + spin_unlock(&dlm->spinlock); + mlog(ML_NOTICE, "%s:%.*s: node %u sends deref done " + "but it is already derefed!\n", dlm->name, + res->lockname.len, res->lockname.name, node); + dlm_lockres_put(res); + ret = 0; + goto done; + } + if (!list_empty(&res->purge)) { mlog(0, "%s: Removing res %.*s from purgelist\n", dlm->name, res->lockname.len, res->lockname.name); @@ -2456,7 +2466,6 @@ int dlm_deref_lockres_done_handler(struct o2net_msg *msg, u32 len, void *data, spin_unlock(&dlm->spinlock); ret = 0; - done: dlm_put(dlm); return ret; -- cgit v1.2.3 From 309e91911daede6adde0364f489e69909c3f6894 Mon Sep 17 00:00:00 2001 From: piaojun Date: Tue, 2 Aug 2016 14:02:16 -0700 Subject: ocfs2/dlm: solve a BUG when deref failed in dlm_drop_lockres_ref We found a BUG situation that lockres is migrated during deref described below. To solve the BUG, we could purge lockres directly when other node says I did not have a ref. Additionally, we'd better purge lockres if master goes down, as no one will response deref done. Node 1 Node 2(old master) Node3(new master) dlm_purge_lockres send deref to N2 leave domain migrate lockres to N3 finish migration send do assert master to N1 receive do assert msg form N3, but can not find lockres because DROPPING_REF is set, so the owner is still N2. receive deref from N1 and response -EINVAL because lockres is migrated BUG when receive -EINVAL in dlm_drop_lockres_ref Fixes: 842b90b62461d ("ocfs2/dlm: return in progress if master can not clear the refmap bit right now") Link: http://lkml.kernel.org/r/57845103.3070406@huawei.com Signed-off-by: Jun Piao Reviewed-by: Joseph Qi Reviewed-by: Jiufei Xue Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/dlm/dlmmaster.c | 9 ++++++--- fs/ocfs2/dlm/dlmthread.c | 13 +++++++++++-- 2 files changed, 17 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 525dc06468c4..553d220df406 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -2276,9 +2276,12 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) mlog(ML_ERROR, "%s: res %.*s, DEREF to node %u got %d\n", dlm->name, namelen, lockname, res->owner, r); dlm_print_one_lock_resource(res); - BUG(); - } - return ret ? ret : r; + if (r == -ENOMEM) + BUG(); + } else + ret = r; + + return ret; } int dlm_deref_lockres_handler(struct o2net_msg *msg, u32 len, void *data, diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index 68d239ba0c63..ce397229acc0 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -175,6 +175,15 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, res->lockname.len, res->lockname.name, master); if (!master) { + if (res->state & DLM_LOCK_RES_DROPPING_REF) { + mlog(ML_NOTICE, "%s: res %.*s already in " + "DLM_LOCK_RES_DROPPING_REF state\n", + dlm->name, res->lockname.len, + res->lockname.name); + spin_unlock(&res->spinlock); + return; + } + res->state |= DLM_LOCK_RES_DROPPING_REF; /* drop spinlock... retake below */ spin_unlock(&res->spinlock); @@ -203,8 +212,8 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, dlm->purge_count--; } - if (!master && ret != 0) { - mlog(0, "%s: deref %.*s in progress or master goes down\n", + if (!master && ret == DLM_DEREF_RESPONSE_INPROG) { + mlog(0, "%s: deref %.*s in progress\n", dlm->name, res->lockname.len, res->lockname.name); spin_unlock(&res->spinlock); return; -- cgit v1.2.3 From ee8f7fcbe638b07e8d1c3dc98e8be35e56306d05 Mon Sep 17 00:00:00 2001 From: piaojun Date: Tue, 2 Aug 2016 14:02:19 -0700 Subject: ocfs2/dlm: continue to purge recovery lockres when recovery master goes down We found a dlm-blocked situation caused by continuous breakdown of recovery masters described below. To solve this problem, we should purge recovery lock once detecting recovery master goes down. N3 N2 N1(reco master) go down pick up recovery lock and begin recoverying for N2 go down pick up recovery lock failed, then purge it: dlm_purge_lockres ->DROPPING_REF is set send deref to N1 failed, recovery lock is not purged find N1 go down, begin recoverying for N1, but blocked in dlm_do_recovery as DROPPING_REF is set: dlm_do_recovery ->dlm_pick_recovery_master ->dlmlock ->dlm_get_lock_resource ->__dlm_wait_on_lockres_flags(tmpres, DLM_LOCK_RES_DROPPING_REF); Fixes: 8c0343968163 ("ocfs2/dlm: clear DROPPING_REF flag when the master goes down") Link: http://lkml.kernel.org/r/578453AF.8030404@huawei.com Signed-off-by: Jun Piao Reviewed-by: Joseph Qi Reviewed-by: Jiufei Xue Reviewed-by: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ocfs2/dlm/dlmcommon.h | 2 ++ fs/ocfs2/dlm/dlmmaster.c | 37 +++------------------------------ fs/ocfs2/dlm/dlmrecovery.c | 29 +++++++++++++++++++------- fs/ocfs2/dlm/dlmthread.c | 52 ++++++++++++++++++++++++++++++++++++++++++---- 4 files changed, 74 insertions(+), 46 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 8107d0d0c3f6..e9f3705c4c9f 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -1004,6 +1004,8 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data, int dlm_do_master_requery(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, u8 nodenum, u8 *real_master); +void __dlm_do_purge_lockres(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res); int dlm_dispatch_assert_master(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 553d220df406..6ea06f8a7d29 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -2425,51 +2425,20 @@ int dlm_deref_lockres_done_handler(struct o2net_msg *msg, u32 len, void *data, mlog(ML_NOTICE, "%s:%.*s: node %u sends deref done " "but it is already derefed!\n", dlm->name, res->lockname.len, res->lockname.name, node); - dlm_lockres_put(res); ret = 0; goto done; } - if (!list_empty(&res->purge)) { - mlog(0, "%s: Removing res %.*s from purgelist\n", - dlm->name, res->lockname.len, res->lockname.name); - list_del_init(&res->purge); - dlm_lockres_put(res); - dlm->purge_count--; - } - - if (!__dlm_lockres_unused(res)) { - mlog(ML_ERROR, "%s: res %.*s in use after deref\n", - dlm->name, res->lockname.len, res->lockname.name); - __dlm_print_one_lock_resource(res); - BUG(); - } - - __dlm_unhash_lockres(dlm, res); - - spin_lock(&dlm->track_lock); - if (!list_empty(&res->tracking)) - list_del_init(&res->tracking); - else { - mlog(ML_ERROR, "%s: Resource %.*s not on the Tracking list\n", - dlm->name, res->lockname.len, res->lockname.name); - __dlm_print_one_lock_resource(res); - } - spin_unlock(&dlm->track_lock); - - /* lockres is not in the hash now. drop the flag and wake up - * any processes waiting in dlm_get_lock_resource. - */ - res->state &= ~DLM_LOCK_RES_DROPPING_REF; + __dlm_do_purge_lockres(dlm, res); spin_unlock(&res->spinlock); wake_up(&res->wq); - dlm_lockres_put(res); - spin_unlock(&dlm->spinlock); ret = 0; done: + if (res) + dlm_lockres_put(res); dlm_put(dlm); return ret; } diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index f6b313898763..dd5cb8bcefd1 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -2343,6 +2343,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) struct dlm_lock_resource *res; int i; struct hlist_head *bucket; + struct hlist_node *tmp; struct dlm_lock *lock; @@ -2365,7 +2366,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) */ for (i = 0; i < DLM_HASH_BUCKETS; i++) { bucket = dlm_lockres_hash(dlm, i); - hlist_for_each_entry(res, bucket, hash_node) { + hlist_for_each_entry_safe(res, tmp, bucket, hash_node) { /* always prune any $RECOVERY entries for dead nodes, * otherwise hangs can occur during later recovery */ if (dlm_is_recovery_lock(res->lockname.name, @@ -2386,8 +2387,17 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) break; } } - dlm_lockres_clear_refmap_bit(dlm, res, - dead_node); + + if ((res->owner == dead_node) && + (res->state & DLM_LOCK_RES_DROPPING_REF)) { + dlm_lockres_get(res); + __dlm_do_purge_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); + dlm_lockres_put(res); + continue; + } else if (res->owner == dlm->node_num) + dlm_lockres_clear_refmap_bit(dlm, res, dead_node); spin_unlock(&res->spinlock); continue; } @@ -2398,14 +2408,17 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) if (res->state & DLM_LOCK_RES_DROPPING_REF) { mlog(0, "%s:%.*s: owned by " "dead node %u, this node was " - "dropping its ref when it died. " - "continue, dropping the flag.\n", + "dropping its ref when master died. " + "continue, purging the lockres.\n", dlm->name, res->lockname.len, res->lockname.name, dead_node); + dlm_lockres_get(res); + __dlm_do_purge_lockres(dlm, res); + spin_unlock(&res->spinlock); + wake_up(&res->wq); + dlm_lockres_put(res); + continue; } - res->state &= ~DLM_LOCK_RES_DROPPING_REF; - dlm_move_lockres_to_recovery_list(dlm, - res); } else if (res->owner == dlm->node_num) { dlm_free_dead_locks(dlm, res, dead_node); __dlm_lockres_calc_usage(dlm, res); diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c index ce397229acc0..838a06d4066a 100644 --- a/fs/ocfs2/dlm/dlmthread.c +++ b/fs/ocfs2/dlm/dlmthread.c @@ -160,6 +160,52 @@ void dlm_lockres_calc_usage(struct dlm_ctxt *dlm, spin_unlock(&dlm->spinlock); } +/* + * Do the real purge work: + * unhash the lockres, and + * clear flag DLM_LOCK_RES_DROPPING_REF. + * It requires dlm and lockres spinlock to be taken. + */ +void __dlm_do_purge_lockres(struct dlm_ctxt *dlm, + struct dlm_lock_resource *res) +{ + assert_spin_locked(&dlm->spinlock); + assert_spin_locked(&res->spinlock); + + if (!list_empty(&res->purge)) { + mlog(0, "%s: Removing res %.*s from purgelist\n", + dlm->name, res->lockname.len, res->lockname.name); + list_del_init(&res->purge); + dlm_lockres_put(res); + dlm->purge_count--; + } + + if (!__dlm_lockres_unused(res)) { + mlog(ML_ERROR, "%s: res %.*s in use after deref\n", + dlm->name, res->lockname.len, res->lockname.name); + __dlm_print_one_lock_resource(res); + BUG(); + } + + __dlm_unhash_lockres(dlm, res); + + spin_lock(&dlm->track_lock); + if (!list_empty(&res->tracking)) + list_del_init(&res->tracking); + else { + mlog(ML_ERROR, "%s: Resource %.*s not on the Tracking list\n", + dlm->name, res->lockname.len, res->lockname.name); + __dlm_print_one_lock_resource(res); + } + spin_unlock(&dlm->track_lock); + + /* + * lockres is not in the hash now. drop the flag and wake up + * any processes waiting in dlm_get_lock_resource. + */ + res->state &= ~DLM_LOCK_RES_DROPPING_REF; +} + static void dlm_purge_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res) { @@ -176,10 +222,8 @@ static void dlm_purge_lockres(struct dlm_ctxt *dlm, if (!master) { if (res->state & DLM_LOCK_RES_DROPPING_REF) { - mlog(ML_NOTICE, "%s: res %.*s already in " - "DLM_LOCK_RES_DROPPING_REF state\n", - dlm->name, res->lockname.len, - res->lockname.name); + mlog(ML_NOTICE, "%s: res %.*s already in DLM_LOCK_RES_DROPPING_REF state\n", + dlm->name, res->lockname.len, res->lockname.name); spin_unlock(&res->spinlock); return; } -- cgit v1.2.3 From 05eb6e7263185a6bb0de9501ccf2addc52429414 Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Tue, 2 Aug 2016 14:03:01 -0700 Subject: radix-tree: account nodes to memcg only if explicitly requested Radix trees may be used not only for storing page cache pages, so unconditionally accounting radix tree nodes to the current memory cgroup is bad: if a radix tree node is used for storing data shared among different cgroups we risk pinning dead memory cgroups forever. So let's only account radix tree nodes if it was explicitly requested by passing __GFP_ACCOUNT to INIT_RADIX_TREE. Currently, we only want to account page cache entries, so mark mapping->page_tree so. Fixes: 58e698af4c63 ("radix-tree: account radix_tree_node to memory cgroup") Link: http://lkml.kernel.org/r/1470057188-7864-1-git-send-email-vdavydov@virtuozzo.com Signed-off-by: Vladimir Davydov Acked-by: Johannes Weiner Acked-by: Michal Hocko Cc: [4.6+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/inode.c | 2 +- lib/radix-tree.c | 14 ++++++++++---- 2 files changed, 11 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 9cef4e16aeda..ad445542c285 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -345,7 +345,7 @@ EXPORT_SYMBOL(inc_nlink); void address_space_init_once(struct address_space *mapping) { memset(mapping, 0, sizeof(*mapping)); - INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC); + INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC | __GFP_ACCOUNT); spin_lock_init(&mapping->tree_lock); init_rwsem(&mapping->i_mmap_rwsem); INIT_LIST_HEAD(&mapping->private_list); diff --git a/lib/radix-tree.c b/lib/radix-tree.c index 61b8fb529cef..1b7bf7314141 100644 --- a/lib/radix-tree.c +++ b/lib/radix-tree.c @@ -277,10 +277,11 @@ radix_tree_node_alloc(struct radix_tree_root *root) /* * Even if the caller has preloaded, try to allocate from the - * cache first for the new node to get accounted. + * cache first for the new node to get accounted to the memory + * cgroup. */ ret = kmem_cache_alloc(radix_tree_node_cachep, - gfp_mask | __GFP_ACCOUNT | __GFP_NOWARN); + gfp_mask | __GFP_NOWARN); if (ret) goto out; @@ -303,8 +304,7 @@ radix_tree_node_alloc(struct radix_tree_root *root) kmemleak_update_trace(ret); goto out; } - ret = kmem_cache_alloc(radix_tree_node_cachep, - gfp_mask | __GFP_ACCOUNT); + ret = kmem_cache_alloc(radix_tree_node_cachep, gfp_mask); out: BUG_ON(radix_tree_is_internal_node(ret)); return ret; @@ -351,6 +351,12 @@ static int __radix_tree_preload(gfp_t gfp_mask, int nr) struct radix_tree_node *node; int ret = -ENOMEM; + /* + * Nodes preloaded by one cgroup can be be used by another cgroup, so + * they should never be accounted to any particular memory cgroup. + */ + gfp_mask &= ~__GFP_ACCOUNT; + preempt_disable(); rtp = this_cpu_ptr(&radix_tree_preloads); while (rtp->nr < nr) { -- cgit v1.2.3 From ef419398b68925f21fd3d8463c7bf6934d2ec926 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 2 Aug 2016 14:03:19 -0700 Subject: proc_oom_score: remove tasklist_lock and pid_alive() This was needed before to ensure that ->signal != 0 and do_each_thread() is safe, see commit b95c35e76b29b ("oom: fix the unsafe usage of badness() in proc_oom_score()") for details. Today tsk->signal can't go away and for_each_thread(tsk) is always safe. Link: http://lkml.kernel.org/r/20160608211921.GA15508@redhat.com Signed-off-by: Oleg Nesterov Acked-by: David Rientjes Acked-by: Michal Hocko Cc: Tetsuo Handa Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/base.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index 31370da2ee7c..54e270262979 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -579,11 +579,8 @@ static int proc_oom_score(struct seq_file *m, struct pid_namespace *ns, unsigned long totalpages = totalram_pages + total_swap_pages; unsigned long points = 0; - read_lock(&tasklist_lock); - if (pid_alive(task)) - points = oom_badness(task, NULL, NULL, totalpages) * - 1000 / totalpages; - read_unlock(&tasklist_lock); + points = oom_badness(task, NULL, NULL, totalpages) * + 1000 / totalpages; seq_printf(m, "%lu\n", points); return 0; -- cgit v1.2.3 From 519ded5a89ec0e46e5b0867ba9f5752239b73898 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 2 Aug 2016 14:03:22 -0700 Subject: procfs: avoid 32-bit time_t in /proc/*/stat /proc/stat shows (among lots of other things) the current boottime (i.e. number of seconds since boot). While a 32-bit number is sufficient for this particular case, we want to get rid of the 'struct timespec' suffers from a 32-bit overflow in 2038. This changes the code to use a struct timespec64, which is known to be safe in all cases. Link: http://lkml.kernel.org/r/20160617201247.2292101-1-arnd@arndb.de Signed-off-by: Arnd Bergmann Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/stat.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/proc/stat.c b/fs/proc/stat.c index 510413eb25b8..7907e456ac4f 100644 --- a/fs/proc/stat.c +++ b/fs/proc/stat.c @@ -80,19 +80,17 @@ static u64 get_iowait_time(int cpu) static int show_stat(struct seq_file *p, void *v) { int i, j; - unsigned long jif; u64 user, nice, system, idle, iowait, irq, softirq, steal; u64 guest, guest_nice; u64 sum = 0; u64 sum_softirq = 0; unsigned int per_softirq_sums[NR_SOFTIRQS] = {0}; - struct timespec boottime; + struct timespec64 boottime; user = nice = system = idle = iowait = irq = softirq = steal = 0; guest = guest_nice = 0; - getboottime(&boottime); - jif = boottime.tv_sec; + getboottime64(&boottime); for_each_possible_cpu(i) { user += kcpustat_cpu(i).cpustat[CPUTIME_USER]; @@ -163,12 +161,12 @@ static int show_stat(struct seq_file *p, void *v) seq_printf(p, "\nctxt %llu\n" - "btime %lu\n" + "btime %llu\n" "processes %lu\n" "procs_running %lu\n" "procs_blocked %lu\n", nr_context_switches(), - (unsigned long)jif, + (unsigned long long)boottime.tv_sec, total_forks, nr_running(), nr_iowait()); -- cgit v1.2.3 From ca52953f5f24aff0aa8fa8de750b76ba0302142d Mon Sep 17 00:00:00 2001 From: Valdis Kletnieks Date: Tue, 2 Aug 2016 14:03:25 -0700 Subject: fs/proc/task_mmu.c: suppress compilation warnings with W=1 Suppress a bunch of warnings of the form: fs/proc/task_mmu.c: In function 'show_smap_vma_flags': fs/proc/task_mmu.c:635:22: warning: initialized field overwritten [-Wt override-init] [ilog2(VM_READ)] = "rd", ^~~~ fs/proc/task_mmu.c:635:22: note: (near initialization for 'mnemonics[0]') They happen because of the way we intentionally build the table, so silence the warning when building with 'make W=1'. Link: http://lkml.kernel.org/r/8727.1470022083@turing-police.cc.vt.edu Signed-off-by: Valdis Kletnieks Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/Makefile | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/proc/Makefile b/fs/proc/Makefile index 7151ea428041..a8c13605b434 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -4,6 +4,7 @@ obj-y += proc.o +CFLAGS_task_mmu.o += -Wno-override-init proc-y := nommu.o task_nommu.o proc-$(CONFIG_MMU) := task_mmu.o -- cgit v1.2.3 From a098ecd2fa7db8fa4fcc178a43627b29b226edb9 Mon Sep 17 00:00:00 2001 From: Stephen Boyd Date: Tue, 2 Aug 2016 14:04:28 -0700 Subject: firmware: support loading into a pre-allocated buffer Some systems are memory constrained but they need to load very large firmwares. The firmware subsystem allows drivers to request this firmware be loaded from the filesystem, but this requires that the entire firmware be loaded into kernel memory first before it's provided to the driver. This can lead to a situation where we map the firmware twice, once to load the firmware into kernel memory and once to copy the firmware into the final resting place. This creates needless memory pressure and delays loading because we have to copy from kernel memory to somewhere else. Let's add a request_firmware_into_buf() API that allows drivers to request firmware be loaded directly into a pre-allocated buffer. This skips the intermediate step of allocating a buffer in kernel memory to hold the firmware image while it's read from the filesystem. It also requires that drivers know how much memory they'll require before requesting the firmware and negates any benefits of firmware caching because the firmware layer doesn't manage the buffer lifetime. For a 16MB buffer, about half the time is spent performing a memcpy from the buffer to the final resting place. I see loading times go from 0.081171 seconds to 0.047696 seconds after applying this patch. Plus the vmalloc pressure is reduced. This is based on a patch from Vikram Mulukutla on codeaurora.org: https://www.codeaurora.org/cgit/quic/la/kernel/msm-3.18/commit/drivers/base/firmware_class.c?h=rel/msm-3.18&id=0a328c5f6cd999f5c591f172216835636f39bcb5 Link: http://lkml.kernel.org/r/20160607164741.31849-4-stephen.boyd@linaro.org Signed-off-by: Stephen Boyd Cc: Mimi Zohar Cc: Vikram Mulukutla Cc: Mark Brown Cc: Ming Lei Cc: Greg Kroah-Hartman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- drivers/base/firmware_class.c | 125 +++++++++++++++++++++++++++++++++--------- fs/exec.c | 9 ++- include/linux/firmware.h | 8 +++ include/linux/fs.h | 1 + 4 files changed, 114 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/drivers/base/firmware_class.c b/drivers/base/firmware_class.c index 45ed20cefa10..22d1760a4278 100644 --- a/drivers/base/firmware_class.c +++ b/drivers/base/firmware_class.c @@ -46,7 +46,8 @@ MODULE_LICENSE("GPL"); extern struct builtin_fw __start_builtin_fw[]; extern struct builtin_fw __end_builtin_fw[]; -static bool fw_get_builtin_firmware(struct firmware *fw, const char *name) +static bool fw_get_builtin_firmware(struct firmware *fw, const char *name, + void *buf, size_t size) { struct builtin_fw *b_fw; @@ -54,6 +55,9 @@ static bool fw_get_builtin_firmware(struct firmware *fw, const char *name) if (strcmp(name, b_fw->name) == 0) { fw->size = b_fw->size; fw->data = b_fw->data; + + if (buf && fw->size <= size) + memcpy(buf, fw->data, fw->size); return true; } } @@ -74,7 +78,9 @@ static bool fw_is_builtin_firmware(const struct firmware *fw) #else /* Module case - no builtin firmware support */ -static inline bool fw_get_builtin_firmware(struct firmware *fw, const char *name) +static inline bool fw_get_builtin_firmware(struct firmware *fw, + const char *name, void *buf, + size_t size) { return false; } @@ -144,6 +150,7 @@ struct firmware_buf { unsigned long status; void *data; size_t size; + size_t allocated_size; #ifdef CONFIG_FW_LOADER_USER_HELPER bool is_paged_buf; bool need_uevent; @@ -179,7 +186,8 @@ static DEFINE_MUTEX(fw_lock); static struct firmware_cache fw_cache; static struct firmware_buf *__allocate_fw_buf(const char *fw_name, - struct firmware_cache *fwc) + struct firmware_cache *fwc, + void *dbuf, size_t size) { struct firmware_buf *buf; @@ -195,6 +203,8 @@ static struct firmware_buf *__allocate_fw_buf(const char *fw_name, kref_init(&buf->ref); buf->fwc = fwc; + buf->data = dbuf; + buf->allocated_size = size; init_completion(&buf->completion); #ifdef CONFIG_FW_LOADER_USER_HELPER INIT_LIST_HEAD(&buf->pending_list); @@ -218,7 +228,8 @@ static struct firmware_buf *__fw_lookup_buf(const char *fw_name) static int fw_lookup_and_allocate_buf(const char *fw_name, struct firmware_cache *fwc, - struct firmware_buf **buf) + struct firmware_buf **buf, void *dbuf, + size_t size) { struct firmware_buf *tmp; @@ -230,7 +241,7 @@ static int fw_lookup_and_allocate_buf(const char *fw_name, *buf = tmp; return 1; } - tmp = __allocate_fw_buf(fw_name, fwc); + tmp = __allocate_fw_buf(fw_name, fwc, dbuf, size); if (tmp) list_add(&tmp->list, &fwc->head); spin_unlock(&fwc->lock); @@ -262,6 +273,7 @@ static void __fw_free_buf(struct kref *ref) vfree(buf->pages); } else #endif + if (!buf->allocated_size) vfree(buf->data); kfree_const(buf->fw_id); kfree(buf); @@ -302,13 +314,21 @@ static void fw_finish_direct_load(struct device *device, mutex_unlock(&fw_lock); } -static int fw_get_filesystem_firmware(struct device *device, - struct firmware_buf *buf) +static int +fw_get_filesystem_firmware(struct device *device, struct firmware_buf *buf) { loff_t size; int i, len; int rc = -ENOENT; char *path; + enum kernel_read_file_id id = READING_FIRMWARE; + size_t msize = INT_MAX; + + /* Already populated data member means we're loading into a buffer */ + if (buf->data) { + id = READING_FIRMWARE_PREALLOC_BUFFER; + msize = buf->allocated_size; + } path = __getname(); if (!path) @@ -327,8 +347,8 @@ static int fw_get_filesystem_firmware(struct device *device, } buf->size = 0; - rc = kernel_read_file_from_path(path, &buf->data, &size, - INT_MAX, READING_FIRMWARE); + rc = kernel_read_file_from_path(path, &buf->data, &size, msize, + id); if (rc) { if (rc == -ENOENT) dev_dbg(device, "loading %s failed with error %d\n", @@ -692,6 +712,15 @@ out: static DEVICE_ATTR(loading, 0644, firmware_loading_show, firmware_loading_store); +static void firmware_rw_buf(struct firmware_buf *buf, char *buffer, + loff_t offset, size_t count, bool read) +{ + if (read) + memcpy(buffer, buf->data + offset, count); + else + memcpy(buf->data + offset, buffer, count); +} + static void firmware_rw(struct firmware_buf *buf, char *buffer, loff_t offset, size_t count, bool read) { @@ -739,7 +768,10 @@ static ssize_t firmware_data_read(struct file *filp, struct kobject *kobj, ret_count = count; - firmware_rw(buf, buffer, offset, count, true); + if (buf->data) + firmware_rw_buf(buf, buffer, offset, count, true); + else + firmware_rw(buf, buffer, offset, count, true); out: mutex_unlock(&fw_lock); @@ -815,12 +847,21 @@ static ssize_t firmware_data_write(struct file *filp, struct kobject *kobj, goto out; } - retval = fw_realloc_buffer(fw_priv, offset + count); - if (retval) - goto out; + if (buf->data) { + if (offset + count > buf->allocated_size) { + retval = -ENOMEM; + goto out; + } + firmware_rw_buf(buf, buffer, offset, count, false); + retval = count; + } else { + retval = fw_realloc_buffer(fw_priv, offset + count); + if (retval) + goto out; - retval = count; - firmware_rw(buf, buffer, offset, count, false); + retval = count; + firmware_rw(buf, buffer, offset, count, false); + } buf->size = max_t(size_t, offset + count, buf->size); out: @@ -890,7 +931,8 @@ static int _request_firmware_load(struct firmware_priv *fw_priv, struct firmware_buf *buf = fw_priv->buf; /* fall back on userspace loading */ - buf->is_paged_buf = true; + if (!buf->data) + buf->is_paged_buf = true; dev_set_uevent_suppress(f_dev, true); @@ -925,7 +967,7 @@ static int _request_firmware_load(struct firmware_priv *fw_priv, if (is_fw_load_aborted(buf)) retval = -EAGAIN; - else if (!buf->data) + else if (buf->is_paged_buf && !buf->data) retval = -ENOMEM; device_del(f_dev); @@ -1008,7 +1050,7 @@ static int sync_cached_firmware_buf(struct firmware_buf *buf) */ static int _request_firmware_prepare(struct firmware **firmware_p, const char *name, - struct device *device) + struct device *device, void *dbuf, size_t size) { struct firmware *firmware; struct firmware_buf *buf; @@ -1021,12 +1063,12 @@ _request_firmware_prepare(struct firmware **firmware_p, const char *name, return -ENOMEM; } - if (fw_get_builtin_firmware(firmware, name)) { + if (fw_get_builtin_firmware(firmware, name, dbuf, size)) { dev_dbg(device, "using built-in %s\n", name); return 0; /* assigned */ } - ret = fw_lookup_and_allocate_buf(name, &fw_cache, &buf); + ret = fw_lookup_and_allocate_buf(name, &fw_cache, &buf, dbuf, size); /* * bind with 'buf' now to avoid warning in failure path @@ -1089,7 +1131,8 @@ static int assign_firmware_buf(struct firmware *fw, struct device *device, /* called from request_firmware() and request_firmware_work_func() */ static int _request_firmware(const struct firmware **firmware_p, const char *name, - struct device *device, unsigned int opt_flags) + struct device *device, void *buf, size_t size, + unsigned int opt_flags) { struct firmware *fw = NULL; long timeout; @@ -1103,7 +1146,7 @@ _request_firmware(const struct firmware **firmware_p, const char *name, goto out; } - ret = _request_firmware_prepare(&fw, name, device); + ret = _request_firmware_prepare(&fw, name, device, buf, size); if (ret <= 0) /* error or already assigned */ goto out; @@ -1182,7 +1225,7 @@ request_firmware(const struct firmware **firmware_p, const char *name, /* Need to pin this module until return */ __module_get(THIS_MODULE); - ret = _request_firmware(firmware_p, name, device, + ret = _request_firmware(firmware_p, name, device, NULL, 0, FW_OPT_UEVENT | FW_OPT_FALLBACK); module_put(THIS_MODULE); return ret; @@ -1206,13 +1249,43 @@ int request_firmware_direct(const struct firmware **firmware_p, int ret; __module_get(THIS_MODULE); - ret = _request_firmware(firmware_p, name, device, + ret = _request_firmware(firmware_p, name, device, NULL, 0, FW_OPT_UEVENT | FW_OPT_NO_WARN); module_put(THIS_MODULE); return ret; } EXPORT_SYMBOL_GPL(request_firmware_direct); +/** + * request_firmware_into_buf - load firmware into a previously allocated buffer + * @firmware_p: pointer to firmware image + * @name: name of firmware file + * @device: device for which firmware is being loaded and DMA region allocated + * @buf: address of buffer to load firmware into + * @size: size of buffer + * + * This function works pretty much like request_firmware(), but it doesn't + * allocate a buffer to hold the firmware data. Instead, the firmware + * is loaded directly into the buffer pointed to by @buf and the @firmware_p + * data member is pointed at @buf. + * + * This function doesn't cache firmware either. + */ +int +request_firmware_into_buf(const struct firmware **firmware_p, const char *name, + struct device *device, void *buf, size_t size) +{ + int ret; + + __module_get(THIS_MODULE); + ret = _request_firmware(firmware_p, name, device, buf, size, + FW_OPT_UEVENT | FW_OPT_FALLBACK | + FW_OPT_NOCACHE); + module_put(THIS_MODULE); + return ret; +} +EXPORT_SYMBOL(request_firmware_into_buf); + /** * release_firmware: - release the resource associated with a firmware image * @fw: firmware resource to release @@ -1245,7 +1318,7 @@ static void request_firmware_work_func(struct work_struct *work) fw_work = container_of(work, struct firmware_work, work); - _request_firmware(&fw, fw_work->name, fw_work->device, + _request_firmware(&fw, fw_work->name, fw_work->device, NULL, 0, fw_work->opt_flags); fw_work->cont(fw, fw_work->context); put_device(fw_work->device); /* taken in request_firmware_nowait() */ @@ -1378,7 +1451,7 @@ static int uncache_firmware(const char *fw_name) pr_debug("%s: %s\n", __func__, fw_name); - if (fw_get_builtin_firmware(&fw, fw_name)) + if (fw_get_builtin_firmware(&fw, fw_name, NULL, 0)) return 0; buf = fw_lookup_buf(fw_name); diff --git a/fs/exec.c b/fs/exec.c index ca239fc86d8d..a1789cd684bf 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -866,7 +866,8 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, goto out; } - *buf = vmalloc(i_size); + if (id != READING_FIRMWARE_PREALLOC_BUFFER) + *buf = vmalloc(i_size); if (!*buf) { ret = -ENOMEM; goto out; @@ -897,8 +898,10 @@ int kernel_read_file(struct file *file, void **buf, loff_t *size, out_free: if (ret < 0) { - vfree(*buf); - *buf = NULL; + if (id != READING_FIRMWARE_PREALLOC_BUFFER) { + vfree(*buf); + *buf = NULL; + } } out: diff --git a/include/linux/firmware.h b/include/linux/firmware.h index 5c41c5e75b5c..b1f9f0ccb8ac 100644 --- a/include/linux/firmware.h +++ b/include/linux/firmware.h @@ -47,6 +47,8 @@ int request_firmware_nowait( void (*cont)(const struct firmware *fw, void *context)); int request_firmware_direct(const struct firmware **fw, const char *name, struct device *device); +int request_firmware_into_buf(const struct firmware **firmware_p, + const char *name, struct device *device, void *buf, size_t size); void release_firmware(const struct firmware *fw); #else @@ -75,5 +77,11 @@ static inline int request_firmware_direct(const struct firmware **fw, return -EINVAL; } +static inline int request_firmware_into_buf(const struct firmware **firmware_p, + const char *name, struct device *device, void *buf, size_t size) +{ + return -EINVAL; +} + #endif #endif diff --git a/include/linux/fs.h b/include/linux/fs.h index 577365a77b47..f3f0b4c8e8ac 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2652,6 +2652,7 @@ extern int do_pipe_flags(int *, int); #define __kernel_read_file_id(id) \ id(UNKNOWN, unknown) \ id(FIRMWARE, firmware) \ + id(FIRMWARE_PREALLOC_BUFFER, firmware) \ id(MODULE, kernel-module) \ id(KEXEC_IMAGE, kexec-image) \ id(KEXEC_INITRAMFS, kexec-initramfs) \ -- cgit v1.2.3 From 0036d1f7eb95bcc52977f15507f00dd07018e7e2 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Tue, 2 Aug 2016 14:04:51 -0700 Subject: binfmt_elf: fix calculations for bss padding A double-bug exists in the bss calculation code, where an overflow can happen in the "last_bss - elf_bss" calculation, but vm_brk internally aligns the argument, underflowing it, wrapping back around safe. We shouldn't depend on these bugs staying in sync, so this cleans up the bss padding handling to avoid the overflow. This moves the bss padzero() before the last_bss > elf_bss case, since the zero-filling of the ELF_PAGE should have nothing to do with the relationship of last_bss and elf_bss: any trailing portion should be zeroed, and a zero size is already handled by padzero(). Then it handles the math on elf_bss vs last_bss correctly. These need to both be ELF_PAGE aligned to get the comparison correct, since that's the expected granularity of the mappings. Since elf_bss already had alignment-based padding happen in padzero(), the "start" of the new vm_brk() should be moved forward as done in the original code. However, since the "end" of the vm_brk() area will already become PAGE_ALIGNed in vm_brk() then last_bss should get aligned here to avoid hiding it as a side-effect. Additionally makes a cosmetic change to the initial last_bss calculation so it's easier to read in comparison to the load_addr calculation above it (i.e. the only difference is p_filesz vs p_memsz). Link: http://lkml.kernel.org/r/1468014494-25291-2-git-send-email-keescook@chromium.org Signed-off-by: Kees Cook Reported-by: Hector Marco-Gisbert Cc: Ismael Ripoll Ripoll Cc: Alexander Viro Cc: "Kirill A. Shutemov" Cc: Oleg Nesterov Cc: Chen Gang Cc: Michal Hocko Cc: Konstantin Khlebnikov Cc: Andrea Arcangeli Cc: Andrey Ryabinin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_elf.c | 34 ++++++++++++++++++---------------- 1 file changed, 18 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index a7a28110dc80..7f6aff3f72eb 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -605,28 +605,30 @@ static unsigned long load_elf_interp(struct elfhdr *interp_elf_ex, * Do the same thing for the memory mapping - between * elf_bss and last_bss is the bss section. */ - k = load_addr + eppnt->p_memsz + eppnt->p_vaddr; + k = load_addr + eppnt->p_vaddr + eppnt->p_memsz; if (k > last_bss) last_bss = k; } } + /* + * Now fill out the bss section: first pad the last page from + * the file up to the page boundary, and zero it from elf_bss + * up to the end of the page. + */ + if (padzero(elf_bss)) { + error = -EFAULT; + goto out; + } + /* + * Next, align both the file and mem bss up to the page size, + * since this is where elf_bss was just zeroed up to, and where + * last_bss will end after the vm_brk() below. + */ + elf_bss = ELF_PAGEALIGN(elf_bss); + last_bss = ELF_PAGEALIGN(last_bss); + /* Finally, if there is still more bss to allocate, do it. */ if (last_bss > elf_bss) { - /* - * Now fill out the bss section. First pad the last page up - * to the page boundary, and then perform a mmap to make sure - * that there are zero-mapped pages up to and including the - * last bss page. - */ - if (padzero(elf_bss)) { - error = -EFAULT; - goto out; - } - - /* What we have mapped so far */ - elf_bss = ELF_PAGESTART(elf_bss + ELF_MIN_ALIGN - 1); - - /* Map the last of the bss segment */ error = vm_brk(elf_bss, last_bss - elf_bss); if (error) goto out; -- cgit v1.2.3 From a310dcb7a43f971ee7d810fbbe36bd766a299717 Mon Sep 17 00:00:00 2001 From: Daniel Wagner Date: Tue, 2 Aug 2016 14:04:57 -0700 Subject: fs/binfmt_em86.c: fix incompatible pointer type Since the -Wincompatible-pointer-types is reported as error, alpha doesn't build anymore. Let's fix it in a minimal way. fs/binfmt_em86.c:73:35: error: passing argument 2 of `copy_strings_kernel' from incompatible pointer type [-Werror=incompatible-pointer-types] retval = copy_strings_kernel(1, &i_arg, bprm); ^ ^ fs/binfmt_em86.c:77:34: error: passing argument 2 of `copy_strings_kernel' from incompatible pointer type [-Werror=incompatible-pointer-types] retval = copy_strings_kernel(1, &i_name, bprm); ^ Link: http://lkml.kernel.org/r/1469525978-23359-1-git-send-email-wagi@monom.org Signed-off-by: Daniel Wagner Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_em86.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/binfmt_em86.c b/fs/binfmt_em86.c index 490538536cb4..dd2d3f0cd55d 100644 --- a/fs/binfmt_em86.c +++ b/fs/binfmt_em86.c @@ -24,7 +24,8 @@ static int load_em86(struct linux_binprm *bprm) { - char *interp, *i_name, *i_arg; + const char *i_name, *i_arg; + char *interp; struct file * file; int retval; struct elfhdr elf_ex; -- cgit v1.2.3 From cae3d4ca6fd6872d8e9c21eff0e56398c938100a Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:00 -0700 Subject: nilfs2: hide function name argument from nilfs_error() Simplify nilfs_error(), an output function used to report critical issues in file system. This renames the original nilfs_error() function to __nilfs_error() and redefines it as a macro to hide its function name argument within the macro. Every call site of nilfs_error() is changed to strip __func__ argument except nilfs_bmap_convert_error(); nilfs_bmap_convert_error() directly calls __nilfs_error() because it inherits caller's function name. Link: http://lkml.kernel.org/r/1464875891-5443-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/bmap.c | 4 ++-- fs/nilfs2/dir.c | 38 +++++++++++++++++--------------------- fs/nilfs2/ifile.c | 3 +-- fs/nilfs2/nilfs.h | 20 +++++++++++++++++++- fs/nilfs2/super.c | 22 ++++++++++++---------- 5 files changed, 51 insertions(+), 36 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index f2a7877e0c8c..01fb1831ca25 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -41,8 +41,8 @@ static int nilfs_bmap_convert_error(struct nilfs_bmap *bmap, struct inode *inode = bmap->b_inode; if (err == -EINVAL) { - nilfs_error(inode->i_sb, fname, - "broken bmap (inode number=%lu)", inode->i_ino); + __nilfs_error(inode->i_sb, fname, + "broken bmap (inode number=%lu)", inode->i_ino); err = -EIO; } return err; diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index e506f4f7120a..746956d2937a 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -140,10 +140,9 @@ out: /* Too bad, we had an error */ Ebadsize: - nilfs_error(sb, "nilfs_check_page", + nilfs_error(sb, "size of directory #%lu is not a multiple of chunk size", - dir->i_ino - ); + dir->i_ino); goto fail; Eshort: error = "rec_len is smaller than minimal"; @@ -157,19 +156,18 @@ Enamelen: Espan: error = "directory entry across blocks"; bad_entry: - nilfs_error(sb, "nilfs_check_page", "bad entry in directory #%lu: %s - " - "offset=%lu, inode=%lu, rec_len=%d, name_len=%d", - dir->i_ino, error, (page->index<inode), + nilfs_error(sb, + "bad entry in directory #%lu: %s - offset=%lu, inode=%lu, rec_len=%d, name_len=%d", + dir->i_ino, error, (page->index << PAGE_SHIFT) + offs, + (unsigned long)le64_to_cpu(p->inode), rec_len, p->name_len); goto fail; Eend: p = (struct nilfs_dir_entry *)(kaddr + offs); - nilfs_error(sb, "nilfs_check_page", - "entry in directory #%lu spans the page boundary" - "offset=%lu, inode=%lu", - dir->i_ino, (page->index<inode)); + nilfs_error(sb, + "entry in directory #%lu spans the page boundary offset=%lu, inode=%lu", + dir->i_ino, (page->index << PAGE_SHIFT) + offs, + (unsigned long)le64_to_cpu(p->inode)); fail: SetPageError(page); return false; @@ -267,8 +265,7 @@ static int nilfs_readdir(struct file *file, struct dir_context *ctx) struct page *page = nilfs_get_page(inode, n); if (IS_ERR(page)) { - nilfs_error(sb, __func__, "bad page in #%lu", - inode->i_ino); + nilfs_error(sb, "bad page in #%lu", inode->i_ino); ctx->pos += PAGE_SIZE - offset; return -EIO; } @@ -278,8 +275,7 @@ static int nilfs_readdir(struct file *file, struct dir_context *ctx) NILFS_DIR_REC_LEN(1); for ( ; (char *)de <= limit; de = nilfs_next_entry(de)) { if (de->rec_len == 0) { - nilfs_error(sb, __func__, - "zero-length directory entry"); + nilfs_error(sb, "zero-length directory entry"); nilfs_put_page(page); return -EIO; } @@ -345,7 +341,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, kaddr += nilfs_last_byte(dir, n) - reclen; while ((char *) de <= kaddr) { if (de->rec_len == 0) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "zero-length directory entry"); nilfs_put_page(page); goto out; @@ -360,7 +356,7 @@ nilfs_find_entry(struct inode *dir, const struct qstr *qstr, n = 0; /* next page is past the blocks we've got */ if (unlikely(n > (dir->i_blocks >> (PAGE_SHIFT - 9)))) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "dir %lu size %lld exceeds block count %llu", dir->i_ino, dir->i_size, (unsigned long long)dir->i_blocks); @@ -469,7 +465,7 @@ int nilfs_add_link(struct dentry *dentry, struct inode *inode) goto got_it; } if (de->rec_len == 0) { - nilfs_error(dir->i_sb, __func__, + nilfs_error(dir->i_sb, "zero-length directory entry"); err = -EIO; goto out_unlock; @@ -541,7 +537,7 @@ int nilfs_delete_entry(struct nilfs_dir_entry *dir, struct page *page) while ((char *)de < (char *)dir) { if (de->rec_len == 0) { - nilfs_error(inode->i_sb, __func__, + nilfs_error(inode->i_sb, "zero-length directory entry"); err = -EIO; goto out; @@ -628,7 +624,7 @@ int nilfs_empty_dir(struct inode *inode) while ((char *)de <= kaddr) { if (de->rec_len == 0) { - nilfs_error(inode->i_sb, __func__, + nilfs_error(inode->i_sb, "zero-length directory entry (kaddr=%p, de=%p)", kaddr, de); goto not_empty; diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index 1d2b1805327a..b1c96285aa4a 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -145,8 +145,7 @@ int nilfs_ifile_get_inode_block(struct inode *ifile, ino_t ino, int err; if (unlikely(!NILFS_VALID_INODE(sb, ino))) { - nilfs_error(sb, __func__, "bad inode number: %lu", - (unsigned long) ino); + nilfs_error(sb, "bad inode number: %lu", (unsigned long)ino); return -EINVAL; } diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index b1d48bc0532d..e482c78bcc86 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -299,10 +299,28 @@ static inline int nilfs_mark_inode_dirty_sync(struct inode *inode) /* super.c */ extern struct inode *nilfs_alloc_inode(struct super_block *); extern void nilfs_destroy_inode(struct inode *); + extern __printf(3, 4) -void nilfs_error(struct super_block *, const char *, const char *, ...); +void __nilfs_error(struct super_block *sb, const char *function, + const char *fmt, ...); extern __printf(3, 4) void nilfs_warning(struct super_block *, const char *, const char *, ...); + +#ifdef CONFIG_PRINTK + +#define nilfs_error(sb, fmt, ...) \ + __nilfs_error(sb, __func__, fmt, ##__VA_ARGS__) + +#else + +#define nilfs_error(sb, fmt, ...) \ + do { \ + no_printk(fmt, ##__VA_ARGS__); \ + __nilfs_error(sb, "", " "); \ + } while (0) + +#endif /* CONFIG_PRINTK */ + extern struct nilfs_super_block * nilfs_read_super_block(struct super_block *, u64, int, struct buffer_head **); extern int nilfs_store_magic_and_option(struct super_block *, diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 666107a18a22..7fe497eb2181 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -91,19 +91,21 @@ static void nilfs_set_error(struct super_block *sb) } /** - * nilfs_error() - report failure condition on a filesystem + * __nilfs_error() - report failure condition on a filesystem * - * nilfs_error() sets an ERROR_FS flag on the superblock as well as - * reporting an error message. It should be called when NILFS detects - * incoherences or defects of meta data on disk. As for sustainable - * errors such as a single-shot I/O error, nilfs_warning() or the printk() - * function should be used instead. + * __nilfs_error() sets an ERROR_FS flag on the superblock as well as + * reporting an error message. This function should be called when + * NILFS detects incoherences or defects of meta data on disk. * - * The segment constructor must not call this function because it can - * kill itself. + * This implements the body of nilfs_error() macro. Normally, + * nilfs_error() should be used. As for sustainable errors such as a + * single-shot I/O error, nilfs_warning() or printk() should be used + * instead. + * + * Callers should not add a trailing newline since this will do it. */ -void nilfs_error(struct super_block *sb, const char *function, - const char *fmt, ...) +void __nilfs_error(struct super_block *sb, const char *function, + const char *fmt, ...) { struct the_nilfs *nilfs = sb->s_fs_info; struct va_format vaf; -- cgit v1.2.3 From a66dfb0a91c211c77b5d4e503d3e760e2e566189 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:02 -0700 Subject: nilfs2: add nilfs_msg() message interface Define an own output routine to replace bare use of printk() function. The output routine is implemented with a macro and a helper function, which are named nilfs_msg() and __nilfs_msg(), respectively. __nilfs_msg() formats a message like "NILFS (): ", prefixing it with a given log level, and terminates the statement with a newline. The "device-name" is optional to make it available in early stages; it will be omitted if a NULL pointer is passed to super block instance argument. nilfs_msg() wraps __nilfs_msg() and is removed if CONFIG_PRINTK is not set. Link: http://lkml.kernel.org/r/1464875891-5443-3-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/nilfs.h | 7 +++++++ fs/nilfs2/super.c | 16 ++++++++++++++++ 2 files changed, 23 insertions(+) (limited to 'fs') diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index e482c78bcc86..b57ce41e8a1a 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -300,6 +300,9 @@ static inline int nilfs_mark_inode_dirty_sync(struct inode *inode) extern struct inode *nilfs_alloc_inode(struct super_block *); extern void nilfs_destroy_inode(struct inode *); +extern __printf(3, 4) +void __nilfs_msg(struct super_block *sb, const char *level, + const char *fmt, ...); extern __printf(3, 4) void __nilfs_error(struct super_block *sb, const char *function, const char *fmt, ...); @@ -308,11 +311,15 @@ void nilfs_warning(struct super_block *, const char *, const char *, ...); #ifdef CONFIG_PRINTK +#define nilfs_msg(sb, level, fmt, ...) \ + __nilfs_msg(sb, level, fmt, ##__VA_ARGS__) #define nilfs_error(sb, fmt, ...) \ __nilfs_error(sb, __func__, fmt, ##__VA_ARGS__) #else +#define nilfs_msg(sb, level, fmt, ...) \ + no_printk(fmt, ##__VA_ARGS__) #define nilfs_error(sb, fmt, ...) \ do { \ no_printk(fmt, ##__VA_ARGS__); \ diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 7fe497eb2181..86e3c00994e2 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -71,6 +71,22 @@ struct kmem_cache *nilfs_btree_path_cache; static int nilfs_setup_super(struct super_block *sb, int is_mount); static int nilfs_remount(struct super_block *sb, int *flags, char *data); +void __nilfs_msg(struct super_block *sb, const char *level, const char *fmt, + ...) +{ + struct va_format vaf; + va_list args; + + va_start(args, fmt); + vaf.fmt = fmt; + vaf.va = &args; + if (sb) + printk("%sNILFS (%s): %pV\n", level, sb->s_id, &vaf); + else + printk("%sNILFS: %pV\n", level, &vaf); + va_end(args); +} + static void nilfs_set_error(struct super_block *sb) { struct the_nilfs *nilfs = sb->s_fs_info; -- cgit v1.2.3 From 6625689e159fa1d43572ee113713ab23bec03131 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:06 -0700 Subject: nilfs2: embed a back pointer to super block instance in nilfs object Insert a back pointer to super block instance in nilfs object so that functions of nilfs2 easily refer to the super block instance. This simplifies replacement of printk() in the successive change. Link: http://lkml.kernel.org/r/1464875891-5443-4-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/super.c | 2 +- fs/nilfs2/the_nilfs.c | 7 ++++--- fs/nilfs2/the_nilfs.h | 4 +++- 3 files changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 86e3c00994e2..2d4d0bec711e 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1076,7 +1076,7 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent) __u64 cno; int err; - nilfs = alloc_nilfs(sb->s_bdev); + nilfs = alloc_nilfs(sb); if (!nilfs) return -ENOMEM; diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index e9fd241b9a0a..702115164cf3 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -56,12 +56,12 @@ void nilfs_set_last_segment(struct the_nilfs *nilfs, /** * alloc_nilfs - allocate a nilfs object - * @bdev: block device to which the_nilfs is related + * @sb: super block instance * * Return Value: On success, pointer to the_nilfs is returned. * On error, NULL is returned. */ -struct the_nilfs *alloc_nilfs(struct block_device *bdev) +struct the_nilfs *alloc_nilfs(struct super_block *sb) { struct the_nilfs *nilfs; @@ -69,7 +69,8 @@ struct the_nilfs *alloc_nilfs(struct block_device *bdev) if (!nilfs) return NULL; - nilfs->ns_bdev = bdev; + nilfs->ns_sb = sb; + nilfs->ns_bdev = sb->s_bdev; atomic_set(&nilfs->ns_ndirtyblks, 0); init_rwsem(&nilfs->ns_sem); mutex_init(&nilfs->ns_snapshot_mount_mutex); diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 79369fd6b13b..79d1421896d0 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -43,6 +43,7 @@ enum { * struct the_nilfs - struct to supervise multiple nilfs mount points * @ns_flags: flags * @ns_flushed_device: flag indicating if all volatile data was flushed + * @ns_sb: back pointer to super block instance * @ns_bdev: block device * @ns_sem: semaphore for shared states * @ns_snapshot_mount_mutex: mutex to protect snapshot mounts @@ -102,6 +103,7 @@ struct the_nilfs { unsigned long ns_flags; int ns_flushed_device; + struct super_block *ns_sb; struct block_device *ns_bdev; struct rw_semaphore ns_sem; struct mutex ns_snapshot_mount_mutex; @@ -281,7 +283,7 @@ static inline int nilfs_sb_will_flip(struct the_nilfs *nilfs) } void nilfs_set_last_segment(struct the_nilfs *, sector_t, u64, __u64); -struct the_nilfs *alloc_nilfs(struct block_device *bdev); +struct the_nilfs *alloc_nilfs(struct super_block *sb); void destroy_nilfs(struct the_nilfs *nilfs); int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data); int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb); -- cgit v1.2.3 From feee880fa58254fcc1c78bc8b6446a435cc1baf0 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:10 -0700 Subject: nilfs2: reduce bare use of printk() with nilfs_msg() Replace most use of printk() in nilfs2 implementation with nilfs_msg(), and reduce the following checkpatch.pl warning: "WARNING: Prefer [subsystem eg: netdev]_crit([subsystem]dev, ... then dev_crit(dev, ... then pr_crit(... to printk(KERN_CRIT ..." This patch also fixes a minor checkpatch warning "WARNING: quoted string split across lines" that often accompanies the prior warning, and amends message format as needed. Link: http://lkml.kernel.org/r/1464875891-5443-5-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/btree.c | 58 +++++++++++++---------- fs/nilfs2/cpfile.c | 22 ++++----- fs/nilfs2/dat.c | 19 ++++---- fs/nilfs2/direct.c | 10 ++-- fs/nilfs2/inode.c | 11 ++--- fs/nilfs2/ioctl.c | 47 +++++++++---------- fs/nilfs2/recovery.c | 72 ++++++++++++++-------------- fs/nilfs2/segbuf.c | 6 ++- fs/nilfs2/segment.c | 25 +++++----- fs/nilfs2/sufile.c | 31 ++++++------ fs/nilfs2/super.c | 100 +++++++++++++++++++-------------------- fs/nilfs2/sysfs.c | 30 ++++++------ fs/nilfs2/the_nilfs.c | 127 ++++++++++++++++++++++++++------------------------ 13 files changed, 283 insertions(+), 275 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 982d1e3df3a5..2c52693a69a4 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -339,12 +339,14 @@ static int nilfs_btree_node_lookup(const struct nilfs_btree_node *node, * nilfs_btree_node_broken - verify consistency of btree node * @node: btree node block to be examined * @size: node size (in bytes) + * @inode: host inode of btree * @blocknr: block number * * Return Value: If node is broken, 1 is returned. Otherwise, 0 is returned. */ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, - size_t size, sector_t blocknr) + size_t size, struct inode *inode, + sector_t blocknr) { int level, flags, nchildren; int ret = 0; @@ -358,9 +360,10 @@ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, (flags & NILFS_BTREE_NODE_ROOT) || nchildren < 0 || nchildren > NILFS_BTREE_NODE_NCHILDREN_MAX(size))) { - printk(KERN_CRIT "NILFS: bad btree node (blocknr=%llu): " - "level = %d, flags = 0x%x, nchildren = %d\n", - (unsigned long long)blocknr, level, flags, nchildren); + nilfs_msg(inode->i_sb, KERN_CRIT, + "bad btree node (ino=%lu, blocknr=%llu): level = %d, flags = 0x%x, nchildren = %d", + inode->i_ino, (unsigned long long)blocknr, level, + flags, nchildren); ret = 1; } return ret; @@ -369,12 +372,12 @@ static int nilfs_btree_node_broken(const struct nilfs_btree_node *node, /** * nilfs_btree_root_broken - verify consistency of btree root node * @node: btree root node to be examined - * @ino: inode number + * @inode: host inode of btree * * Return Value: If node is broken, 1 is returned. Otherwise, 0 is returned. */ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, - unsigned long ino) + struct inode *inode) { int level, flags, nchildren; int ret = 0; @@ -387,8 +390,9 @@ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, level >= NILFS_BTREE_LEVEL_MAX || nchildren < 0 || nchildren > NILFS_BTREE_ROOT_NCHILDREN_MAX)) { - pr_crit("NILFS: bad btree root (inode number=%lu): level = %d, flags = 0x%x, nchildren = %d\n", - ino, level, flags, nchildren); + nilfs_msg(inode->i_sb, KERN_CRIT, + "bad btree root (ino=%lu): level = %d, flags = 0x%x, nchildren = %d", + inode->i_ino, level, flags, nchildren); ret = 1; } return ret; @@ -396,13 +400,15 @@ static int nilfs_btree_root_broken(const struct nilfs_btree_node *node, int nilfs_btree_broken_node_block(struct buffer_head *bh) { + struct inode *inode; int ret; if (buffer_nilfs_checked(bh)) return 0; + inode = bh->b_page->mapping->host; ret = nilfs_btree_node_broken((struct nilfs_btree_node *)bh->b_data, - bh->b_size, bh->b_blocknr); + bh->b_size, inode, bh->b_blocknr); if (likely(!ret)) set_buffer_nilfs_checked(bh); return ret; @@ -448,13 +454,15 @@ nilfs_btree_get_node(const struct nilfs_bmap *btree, return node; } -static int -nilfs_btree_bad_node(struct nilfs_btree_node *node, int level) +static int nilfs_btree_bad_node(const struct nilfs_bmap *btree, + struct nilfs_btree_node *node, int level) { if (unlikely(nilfs_btree_node_get_level(node) != level)) { dump_stack(); - printk(KERN_CRIT "NILFS: btree level mismatch: %d != %d\n", - nilfs_btree_node_get_level(node), level); + nilfs_msg(btree->b_inode->i_sb, KERN_CRIT, + "btree level mismatch (ino=%lu): %d != %d", + btree->b_inode->i_ino, + nilfs_btree_node_get_level(node), level); return 1; } return 0; @@ -568,7 +576,7 @@ static int nilfs_btree_do_lookup(const struct nilfs_bmap *btree, return ret; node = nilfs_btree_get_nonroot_node(path, level); - if (nilfs_btree_bad_node(node, level)) + if (nilfs_btree_bad_node(btree, node, level)) return -EINVAL; if (!found) found = nilfs_btree_node_lookup(node, key, &index); @@ -616,7 +624,7 @@ static int nilfs_btree_do_lookup_last(const struct nilfs_bmap *btree, if (ret < 0) return ret; node = nilfs_btree_get_nonroot_node(path, level); - if (nilfs_btree_bad_node(node, level)) + if (nilfs_btree_bad_node(btree, node, level)) return -EINVAL; index = nilfs_btree_node_get_nchildren(node) - 1; ptr = nilfs_btree_node_get_ptr(node, index, ncmax); @@ -2072,8 +2080,10 @@ static int nilfs_btree_propagate(struct nilfs_bmap *btree, ret = nilfs_btree_do_lookup(btree, path, key, NULL, level + 1, 0); if (ret < 0) { if (unlikely(ret == -ENOENT)) - printk(KERN_CRIT "%s: key = %llu, level == %d\n", - __func__, (unsigned long long)key, level); + nilfs_msg(btree->b_inode->i_sb, KERN_CRIT, + "writing node/leaf block does not appear in b-tree (ino=%lu) at key=%llu, level=%d", + btree->b_inode->i_ino, + (unsigned long long)key, level); goto out; } @@ -2110,12 +2120,11 @@ static void nilfs_btree_add_dirty_buffer(struct nilfs_bmap *btree, if (level < NILFS_BTREE_LEVEL_NODE_MIN || level >= NILFS_BTREE_LEVEL_MAX) { dump_stack(); - printk(KERN_WARNING - "%s: invalid btree level: %d (key=%llu, ino=%lu, " - "blocknr=%llu)\n", - __func__, level, (unsigned long long)key, - NILFS_BMAP_I(btree)->vfs_inode.i_ino, - (unsigned long long)bh->b_blocknr); + nilfs_msg(btree->b_inode->i_sb, KERN_WARNING, + "invalid btree level: %d (key=%llu, ino=%lu, blocknr=%llu)", + level, (unsigned long long)key, + btree->b_inode->i_ino, + (unsigned long long)bh->b_blocknr); return; } @@ -2394,8 +2403,7 @@ int nilfs_btree_init(struct nilfs_bmap *bmap) __nilfs_btree_init(bmap); - if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), - bmap->b_inode->i_ino)) + if (nilfs_btree_root_broken(nilfs_btree_get_root(bmap), bmap->b_inode)) ret = -EIO; return ret; } diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 8a3d3b65af3f..19d9f4ae8347 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -332,9 +332,9 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, int ret, ncps, nicps, nss, count, i; if (unlikely(start == 0 || start > end)) { - printk(KERN_ERR "%s: invalid range of checkpoint numbers: " - "[%llu, %llu)\n", __func__, - (unsigned long long)start, (unsigned long long)end); + nilfs_msg(cpfile->i_sb, KERN_ERR, + "cannot delete checkpoints: invalid range [%llu, %llu)", + (unsigned long long)start, (unsigned long long)end); return -EINVAL; } @@ -386,9 +386,9 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, cpfile, cno); if (ret == 0) continue; - printk(KERN_ERR - "%s: cannot delete block\n", - __func__); + nilfs_msg(cpfile->i_sb, KERN_ERR, + "error %d deleting checkpoint block", + ret); break; } } @@ -991,14 +991,12 @@ int nilfs_cpfile_read(struct super_block *sb, size_t cpsize, int err; if (cpsize > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large checkpoint size: %zu bytes.\n", - cpsize); + nilfs_msg(sb, KERN_ERR, + "too large checkpoint size: %zu bytes", cpsize); return -EINVAL; } else if (cpsize < NILFS_MIN_CHECKPOINT_SIZE) { - printk(KERN_ERR - "NILFS: too small checkpoint size: %zu bytes.\n", - cpsize); + nilfs_msg(sb, KERN_ERR, + "too small checkpoint size: %zu bytes", cpsize); return -EINVAL; } diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index 7367610ea807..dffedb2f8817 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -349,10 +349,11 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) kaddr = kmap_atomic(entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); if (unlikely(entry->de_blocknr == cpu_to_le64(0))) { - printk(KERN_CRIT "%s: vbn = %llu, [%llu, %llu)\n", __func__, - (unsigned long long)vblocknr, - (unsigned long long)le64_to_cpu(entry->de_start), - (unsigned long long)le64_to_cpu(entry->de_end)); + nilfs_msg(dat->i_sb, KERN_CRIT, + "%s: invalid vblocknr = %llu, [%llu, %llu)", + __func__, (unsigned long long)vblocknr, + (unsigned long long)le64_to_cpu(entry->de_start), + (unsigned long long)le64_to_cpu(entry->de_end)); kunmap_atomic(kaddr); brelse(entry_bh); return -EINVAL; @@ -479,14 +480,12 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size, int err; if (entry_size > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large DAT entry size: %zu bytes.\n", - entry_size); + nilfs_msg(sb, KERN_ERR, "too large DAT entry size: %zu bytes", + entry_size); return -EINVAL; } else if (entry_size < NILFS_MIN_DAT_ENTRY_SIZE) { - printk(KERN_ERR - "NILFS: too small DAT entry size: %zu bytes.\n", - entry_size); + nilfs_msg(sb, KERN_ERR, "too small DAT entry size: %zu bytes", + entry_size); return -EINVAL; } diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 251a44928405..96e3ed0d9652 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -337,14 +337,16 @@ static int nilfs_direct_assign(struct nilfs_bmap *bmap, key = nilfs_bmap_data_get_key(bmap, *bh); if (unlikely(key > NILFS_DIRECT_KEY_MAX)) { - printk(KERN_CRIT "%s: invalid key: %llu\n", __func__, - (unsigned long long)key); + nilfs_msg(bmap->b_inode->i_sb, KERN_CRIT, + "%s (ino=%lu): invalid key: %llu", __func__, + bmap->b_inode->i_ino, (unsigned long long)key); return -EINVAL; } ptr = nilfs_direct_get_ptr(bmap, key); if (unlikely(ptr == NILFS_BMAP_INVALID_PTR)) { - printk(KERN_CRIT "%s: invalid pointer: %llu\n", __func__, - (unsigned long long)ptr); + nilfs_msg(bmap->b_inode->i_sb, KERN_CRIT, + "%s (ino=%lu): invalid pointer: %llu", __func__, + bmap->b_inode->i_ino, (unsigned long long)ptr); return -EINVAL; } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index a0ebdb17e912..a965fcf77955 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -112,13 +112,10 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff, * However, the page having this block must * be locked in this case. */ - printk(KERN_WARNING - "nilfs_get_block: a race condition " - "while inserting a data block. " - "(inode number=%lu, file block " - "offset=%llu)\n", - inode->i_ino, - (unsigned long long)blkoff); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): a race condition while inserting a data block at offset=%llu", + __func__, inode->i_ino, + (unsigned long long)blkoff); err = 0; } nilfs_transaction_abort(inode->i_sb); diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 358b57e2cdf9..827283fe9525 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -584,27 +584,25 @@ static int nilfs_ioctl_move_inode_block(struct inode *inode, if (unlikely(ret < 0)) { if (ret == -ENOENT) - printk(KERN_CRIT - "%s: invalid virtual block address (%s): " - "ino=%llu, cno=%llu, offset=%llu, " - "blocknr=%llu, vblocknr=%llu\n", - __func__, vdesc->vd_flags ? "node" : "data", - (unsigned long long)vdesc->vd_ino, - (unsigned long long)vdesc->vd_cno, - (unsigned long long)vdesc->vd_offset, - (unsigned long long)vdesc->vd_blocknr, - (unsigned long long)vdesc->vd_vblocknr); + nilfs_msg(inode->i_sb, KERN_CRIT, + "%s: invalid virtual block address (%s): ino=%llu, cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu", + __func__, vdesc->vd_flags ? "node" : "data", + (unsigned long long)vdesc->vd_ino, + (unsigned long long)vdesc->vd_cno, + (unsigned long long)vdesc->vd_offset, + (unsigned long long)vdesc->vd_blocknr, + (unsigned long long)vdesc->vd_vblocknr); return ret; } if (unlikely(!list_empty(&bh->b_assoc_buffers))) { - printk(KERN_CRIT "%s: conflicting %s buffer: ino=%llu, " - "cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu\n", - __func__, vdesc->vd_flags ? "node" : "data", - (unsigned long long)vdesc->vd_ino, - (unsigned long long)vdesc->vd_cno, - (unsigned long long)vdesc->vd_offset, - (unsigned long long)vdesc->vd_blocknr, - (unsigned long long)vdesc->vd_vblocknr); + nilfs_msg(inode->i_sb, KERN_CRIT, + "%s: conflicting %s buffer: ino=%llu, cno=%llu, offset=%llu, blocknr=%llu, vblocknr=%llu", + __func__, vdesc->vd_flags ? "node" : "data", + (unsigned long long)vdesc->vd_ino, + (unsigned long long)vdesc->vd_cno, + (unsigned long long)vdesc->vd_offset, + (unsigned long long)vdesc->vd_blocknr, + (unsigned long long)vdesc->vd_vblocknr); brelse(bh); return -EEXIST; } @@ -854,8 +852,8 @@ int nilfs_ioctl_prepare_clean_segments(struct the_nilfs *nilfs, return 0; failed: - printk(KERN_ERR "NILFS: GC failed during preparation: %s: err=%d\n", - msg, ret); + nilfs_msg(nilfs->ns_sb, KERN_ERR, "error %d preparing GC: %s", ret, + msg); return ret; } @@ -963,10 +961,11 @@ static int nilfs_ioctl_clean_segments(struct inode *inode, struct file *filp, } ret = nilfs_ioctl_move_blocks(inode->i_sb, &argv[0], kbufs[0]); - if (ret < 0) - printk(KERN_ERR "NILFS: GC failed during preparation: " - "cannot read source blocks: err=%d\n", ret); - else { + if (ret < 0) { + nilfs_msg(inode->i_sb, KERN_ERR, + "error %d preparing GC: cannot read source blocks", + ret); + } else { if (nilfs_sb_need_update(nilfs)) set_nilfs_discontinued(nilfs); ret = nilfs_clean_segments(inode->i_sb, argv, kbufs); diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index d893dc912b62..5139efed1888 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -54,38 +54,37 @@ struct nilfs_recovery_block { }; -static int nilfs_warn_segment_error(int err) +static int nilfs_warn_segment_error(struct super_block *sb, int err) { + const char *msg = NULL; + switch (err) { case NILFS_SEG_FAIL_IO: - printk(KERN_WARNING - "NILFS warning: I/O error on loading last segment\n"); + nilfs_msg(sb, KERN_ERR, "I/O error reading segment"); return -EIO; case NILFS_SEG_FAIL_MAGIC: - printk(KERN_WARNING - "NILFS warning: Segment magic number invalid\n"); + msg = "Magic number mismatch"; break; case NILFS_SEG_FAIL_SEQ: - printk(KERN_WARNING - "NILFS warning: Sequence number mismatch\n"); + msg = "Sequence number mismatch"; break; case NILFS_SEG_FAIL_CHECKSUM_SUPER_ROOT: - printk(KERN_WARNING - "NILFS warning: Checksum error in super root\n"); + msg = "Checksum error in super root"; break; case NILFS_SEG_FAIL_CHECKSUM_FULL: - printk(KERN_WARNING - "NILFS warning: Checksum error in segment payload\n"); + msg = "Checksum error in segment payload"; break; case NILFS_SEG_FAIL_CONSISTENCY: - printk(KERN_WARNING - "NILFS warning: Inconsistent segment\n"); + msg = "Inconsistency found"; break; case NILFS_SEG_NO_SUPER_ROOT: - printk(KERN_WARNING - "NILFS warning: No super root in the last segment\n"); + msg = "No super root in the last segment"; break; + default: + nilfs_msg(sb, KERN_ERR, "unrecognized segment error %d", err); + return -EINVAL; } + nilfs_msg(sb, KERN_WARNING, "invalid segment: %s", msg); return -EINVAL; } @@ -178,7 +177,7 @@ int nilfs_read_super_root_block(struct the_nilfs *nilfs, sector_t sr_block, brelse(bh_sr); failed: - return nilfs_warn_segment_error(ret); + return nilfs_warn_segment_error(nilfs->ns_sb, ret); } /** @@ -553,11 +552,10 @@ static int nilfs_recover_dsync_blocks(struct the_nilfs *nilfs, put_page(page); failed_inode: - printk(KERN_WARNING - "NILFS warning: error recovering data block " - "(err=%d, ino=%lu, block-offset=%llu)\n", - err, (unsigned long)rb->ino, - (unsigned long long)rb->blkoff); + nilfs_msg(sb, KERN_WARNING, + "error %d recovering data block (ino=%lu, block-offset=%llu)", + err, (unsigned long)rb->ino, + (unsigned long long)rb->blkoff); if (!err2) err2 = err; next: @@ -680,8 +678,8 @@ static int nilfs_do_roll_forward(struct the_nilfs *nilfs, } if (nsalvaged_blocks) { - printk(KERN_INFO "NILFS (device %s): salvaged %lu blocks\n", - sb->s_id, nsalvaged_blocks); + nilfs_msg(sb, KERN_INFO, "salvaged %lu blocks", + nsalvaged_blocks); ri->ri_need_recovery = NILFS_RECOVERY_ROLLFORWARD_DONE; } out: @@ -692,10 +690,9 @@ static int nilfs_do_roll_forward(struct the_nilfs *nilfs, confused: err = -EINVAL; failed: - printk(KERN_ERR - "NILFS (device %s): Error roll-forwarding " - "(err=%d, pseg block=%llu). ", - sb->s_id, err, (unsigned long long)pseg_start); + nilfs_msg(sb, KERN_ERR, + "error %d roll-forwarding partial segment at blocknr = %llu", + err, (unsigned long long)pseg_start); goto out; } @@ -715,9 +712,8 @@ static void nilfs_finish_roll_forward(struct the_nilfs *nilfs, set_buffer_dirty(bh); err = sync_dirty_buffer(bh); if (unlikely(err)) - printk(KERN_WARNING - "NILFS warning: buffer sync write failed during " - "post-cleaning of recovery.\n"); + nilfs_msg(nilfs->ns_sb, KERN_WARNING, + "buffer sync write failed during post-cleaning of recovery."); brelse(bh); } @@ -752,8 +748,8 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, err = nilfs_attach_checkpoint(sb, ri->ri_cno, true, &root); if (unlikely(err)) { - printk(KERN_ERR - "NILFS: error loading the latest checkpoint.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d loading the latest checkpoint", err); return err; } @@ -764,8 +760,9 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, if (ri->ri_need_recovery == NILFS_RECOVERY_ROLLFORWARD_DONE) { err = nilfs_prepare_segment_for_recovery(nilfs, sb, ri); if (unlikely(err)) { - printk(KERN_ERR "NILFS: Error preparing segments for " - "recovery.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d preparing segment for recovery", + err); goto failed; } @@ -778,8 +775,9 @@ int nilfs_salvage_orphan_logs(struct the_nilfs *nilfs, nilfs_detach_log_writer(sb); if (unlikely(err)) { - printk(KERN_ERR "NILFS: Oops! recovery failed. " - "(err=%d)\n", err); + nilfs_msg(sb, KERN_ERR, + "error %d writing segment for recovery", + err); goto failed; } @@ -961,5 +959,5 @@ int nilfs_search_super_root(struct the_nilfs *nilfs, failed: brelse(bh_sum); nilfs_dispose_segment_list(&segments); - return (ret < 0) ? ret : nilfs_warn_segment_error(ret); + return ret < 0 ? ret : nilfs_warn_segment_error(nilfs->ns_sb, ret); } diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index a962d7d83447..6f87b2ac1aeb 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -514,7 +514,11 @@ static int nilfs_segbuf_wait(struct nilfs_segment_buffer *segbuf) } while (--segbuf->sb_nbio > 0); if (unlikely(atomic_read(&segbuf->sb_err) > 0)) { - printk(KERN_ERR "NILFS: IO error writing segment\n"); + nilfs_msg(segbuf->sb_super, KERN_ERR, + "I/O error writing log (start-blocknr=%llu, block-count=%lu) in segment %llu", + (unsigned long long)segbuf->sb_pseg_start, + segbuf->sb_sum.nblocks, + (unsigned long long)segbuf->sb_segnum); err = -EIO; } return err; diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index e78b68a81aec..1cc968502e53 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -150,7 +150,8 @@ static void nilfs_dispose_list(struct the_nilfs *, struct list_head *, int); #define nilfs_cnt32_lt(a, b) nilfs_cnt32_gt(b, a) #define nilfs_cnt32_le(a, b) nilfs_cnt32_ge(b, a) -static int nilfs_prepare_segment_lock(struct nilfs_transaction_info *ti) +static int nilfs_prepare_segment_lock(struct super_block *sb, + struct nilfs_transaction_info *ti) { struct nilfs_transaction_info *cur_ti = current->journal_info; void *save = NULL; @@ -164,8 +165,7 @@ static int nilfs_prepare_segment_lock(struct nilfs_transaction_info *ti) * it is saved and will be restored on * nilfs_transaction_commit(). */ - printk(KERN_WARNING - "NILFS warning: journal info from a different FS\n"); + nilfs_msg(sb, KERN_WARNING, "journal info from a different FS"); save = current->journal_info; } if (!ti) { @@ -215,7 +215,7 @@ int nilfs_transaction_begin(struct super_block *sb, int vacancy_check) { struct the_nilfs *nilfs; - int ret = nilfs_prepare_segment_lock(ti); + int ret = nilfs_prepare_segment_lock(sb, ti); struct nilfs_transaction_info *trace_ti; if (unlikely(ret < 0)) @@ -2467,9 +2467,9 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv, int ret = nilfs_discard_segments(nilfs, sci->sc_freesegs, sci->sc_nfreesegs); if (ret) { - printk(KERN_WARNING - "NILFS warning: error %d on discard request, " - "turning discards off for the device\n", ret); + nilfs_msg(sb, KERN_WARNING, + "error %d on discard request, turning discards off for the device", + ret); nilfs_clear_opt(nilfs, DISCARD); } } @@ -2551,10 +2551,9 @@ static int nilfs_segctor_thread(void *arg) /* start sync. */ sci->sc_task = current; wake_up(&sci->sc_wait_task); /* for nilfs_segctor_start_thread() */ - printk(KERN_INFO - "segctord starting. Construction interval = %lu seconds, " - "CP frequency < %lu seconds\n", - sci->sc_interval / HZ, sci->sc_mjcp_freq / HZ); + nilfs_msg(sci->sc_super, KERN_INFO, + "segctord starting. Construction interval = %lu seconds, CP frequency < %lu seconds", + sci->sc_interval / HZ, sci->sc_mjcp_freq / HZ); spin_lock(&sci->sc_state_lock); loop: @@ -2628,8 +2627,8 @@ static int nilfs_segctor_start_thread(struct nilfs_sc_info *sci) if (IS_ERR(t)) { int err = PTR_ERR(t); - printk(KERN_ERR "NILFS: error %d creating segctord thread\n", - err); + nilfs_msg(sci->sc_super, KERN_ERR, + "error %d creating segctord thread", err); return err; } wait_event(sci->sc_wait_task, sci->sc_task != NULL); diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 1963595a1580..5b495c469471 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -181,9 +181,9 @@ int nilfs_sufile_updatev(struct inode *sufile, __u64 *segnumv, size_t nsegs, down_write(&NILFS_MDT(sufile)->mi_sem); for (seg = segnumv; seg < segnumv + nsegs; seg++) { if (unlikely(*seg >= nilfs_sufile_get_nsegments(sufile))) { - printk(KERN_WARNING - "%s: invalid segment number: %llu\n", __func__, - (unsigned long long)*seg); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: invalid segment number: %llu", + __func__, (unsigned long long)*seg); nerr++; } } @@ -240,8 +240,9 @@ int nilfs_sufile_update(struct inode *sufile, __u64 segnum, int create, int ret; if (unlikely(segnum >= nilfs_sufile_get_nsegments(sufile))) { - printk(KERN_WARNING "%s: invalid segment number: %llu\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: invalid segment number: %llu", + __func__, (unsigned long long)segnum); return -EINVAL; } down_write(&NILFS_MDT(sufile)->mi_sem); @@ -419,8 +420,9 @@ void nilfs_sufile_do_cancel_free(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (unlikely(!nilfs_segment_usage_clean(su))) { - printk(KERN_WARNING "%s: segment %llu must be clean\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: segment %llu must be clean", __func__, + (unsigned long long)segnum); kunmap_atomic(kaddr); return; } @@ -476,8 +478,9 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (nilfs_segment_usage_clean(su)) { - printk(KERN_WARNING "%s: segment %llu is already clean\n", - __func__, (unsigned long long)segnum); + nilfs_msg(sufile->i_sb, KERN_WARNING, + "%s: segment %llu is already clean", + __func__, (unsigned long long)segnum); kunmap_atomic(kaddr); return; } @@ -1175,14 +1178,12 @@ int nilfs_sufile_read(struct super_block *sb, size_t susize, int err; if (susize > sb->s_blocksize) { - printk(KERN_ERR - "NILFS: too large segment usage size: %zu bytes.\n", - susize); + nilfs_msg(sb, KERN_ERR, + "too large segment usage size: %zu bytes", susize); return -EINVAL; } else if (susize < NILFS_MIN_SEGMENT_USAGE_SIZE) { - printk(KERN_ERR - "NILFS: too small segment usage size: %zu bytes.\n", - susize); + nilfs_msg(sb, KERN_ERR, + "too small segment usage size: %zu bytes", susize); return -EINVAL; } diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 2d4d0bec711e..90c62b489857 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -214,8 +214,8 @@ static int nilfs_sync_super(struct super_block *sb, int flag) } if (unlikely(err)) { - printk(KERN_ERR - "NILFS: unable to write superblock (err=%d)\n", err); + nilfs_msg(sb, KERN_ERR, "unable to write superblock: err=%d", + err); if (err == -EIO && nilfs->ns_sbh[1]) { /* * sbp[0] points to newer log than sbp[1], @@ -285,8 +285,7 @@ struct nilfs_super_block **nilfs_prepare_super(struct super_block *sb, sbp[1]->s_magic == cpu_to_le16(NILFS_SUPER_MAGIC)) { memcpy(sbp[0], sbp[1], nilfs->ns_sbsize); } else { - printk(KERN_CRIT "NILFS: superblock broke on dev %s\n", - sb->s_id); + nilfs_msg(sb, KERN_CRIT, "superblock broke"); return NULL; } } else if (sbp[1] && @@ -396,9 +395,9 @@ static int nilfs_move_2nd_super(struct super_block *sb, loff_t sb2off) offset = sb2off & (nilfs->ns_blocksize - 1); nsbh = sb_getblk(sb, newblocknr); if (!nsbh) { - printk(KERN_WARNING - "NILFS warning: unable to move secondary superblock " - "to block %llu\n", (unsigned long long)newblocknr); + nilfs_msg(sb, KERN_WARNING, + "unable to move secondary superblock to block %llu", + (unsigned long long)newblocknr); ret = -EIO; goto out; } @@ -561,10 +560,9 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt, up_read(&nilfs->ns_segctor_sem); if (unlikely(err)) { if (err == -ENOENT || err == -EINVAL) { - printk(KERN_ERR - "NILFS: Invalid checkpoint " - "(checkpoint number=%llu)\n", - (unsigned long long)cno); + nilfs_msg(sb, KERN_ERR, + "Invalid checkpoint (checkpoint number=%llu)", + (unsigned long long)cno); err = -EINVAL; } goto failed; @@ -660,9 +658,8 @@ static int nilfs_statfs(struct dentry *dentry, struct kstatfs *buf) err = nilfs_ifile_count_free_inodes(root->ifile, &nmaxinodes, &nfreeinodes); if (unlikely(err)) { - printk(KERN_WARNING - "NILFS warning: fail to count free inodes: err %d.\n", - err); + nilfs_msg(sb, KERN_WARNING, + "failed to count free inodes: err=%d", err); if (err == -ERANGE) { /* * If nilfs_palloc_count_max_entries() returns @@ -794,9 +791,9 @@ static int parse_options(char *options, struct super_block *sb, int is_remount) break; case Opt_snapshot: if (is_remount) { - printk(KERN_ERR - "NILFS: \"%s\" option is invalid " - "for remount.\n", p); + nilfs_msg(sb, KERN_ERR, + "\"%s\" option is invalid for remount", + p); return 0; } break; @@ -810,8 +807,8 @@ static int parse_options(char *options, struct super_block *sb, int is_remount) nilfs_clear_opt(nilfs, DISCARD); break; default: - printk(KERN_ERR - "NILFS: Unrecognized mount option \"%s\"\n", p); + nilfs_msg(sb, KERN_ERR, + "unrecognized mount option \"%s\"", p); return 0; } } @@ -847,12 +844,10 @@ static int nilfs_setup_super(struct super_block *sb, int is_mount) mnt_count = le16_to_cpu(sbp[0]->s_mnt_count); if (nilfs->ns_mount_state & NILFS_ERROR_FS) { - printk(KERN_WARNING - "NILFS warning: mounting fs with errors\n"); + nilfs_msg(sb, KERN_WARNING, "mounting fs with errors"); #if 0 } else if (max_mnt_count >= 0 && mnt_count >= max_mnt_count) { - printk(KERN_WARNING - "NILFS warning: maximal mount count reached\n"); + nilfs_msg(sb, KERN_WARNING, "maximal mount count reached"); #endif } if (!max_mnt_count) @@ -915,17 +910,17 @@ int nilfs_check_feature_compatibility(struct super_block *sb, features = le64_to_cpu(sbp->s_feature_incompat) & ~NILFS_FEATURE_INCOMPAT_SUPP; if (features) { - printk(KERN_ERR "NILFS: couldn't mount because of unsupported " - "optional features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't mount because of unsupported optional features (%llx)", + (unsigned long long)features); return -EINVAL; } features = le64_to_cpu(sbp->s_feature_compat_ro) & ~NILFS_FEATURE_COMPAT_RO_SUPP; if (!(sb->s_flags & MS_RDONLY) && features) { - printk(KERN_ERR "NILFS: couldn't mount RDWR because of " - "unsupported optional features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't mount RDWR because of unsupported optional features (%llx)", + (unsigned long long)features); return -EINVAL; } return 0; @@ -941,13 +936,13 @@ static int nilfs_get_root_dentry(struct super_block *sb, inode = nilfs_iget(sb, root, NILFS_ROOT_INO); if (IS_ERR(inode)) { - printk(KERN_ERR "NILFS: get root inode failed\n"); ret = PTR_ERR(inode); + nilfs_msg(sb, KERN_ERR, "error %d getting root inode", ret); goto out; } if (!S_ISDIR(inode->i_mode) || !inode->i_blocks || !inode->i_size) { iput(inode); - printk(KERN_ERR "NILFS: corrupt root inode.\n"); + nilfs_msg(sb, KERN_ERR, "corrupt root inode"); ret = -EINVAL; goto out; } @@ -975,7 +970,7 @@ static int nilfs_get_root_dentry(struct super_block *sb, return ret; failed_dentry: - printk(KERN_ERR "NILFS: get root dentry failed\n"); + nilfs_msg(sb, KERN_ERR, "error %d getting root dentry", ret); goto out; } @@ -995,18 +990,18 @@ static int nilfs_attach_snapshot(struct super_block *s, __u64 cno, ret = (ret == -ENOENT) ? -EINVAL : ret; goto out; } else if (!ret) { - printk(KERN_ERR "NILFS: The specified checkpoint is " - "not a snapshot (checkpoint number=%llu).\n", - (unsigned long long)cno); + nilfs_msg(s, KERN_ERR, + "The specified checkpoint is not a snapshot (checkpoint number=%llu)", + (unsigned long long)cno); ret = -EINVAL; goto out; } ret = nilfs_attach_checkpoint(s, cno, false, &root); if (ret) { - printk(KERN_ERR "NILFS: error loading snapshot " - "(checkpoint number=%llu).\n", - (unsigned long long)cno); + nilfs_msg(s, KERN_ERR, + "error %d while loading snapshot (checkpoint number=%llu)", + ret, (unsigned long long)cno); goto out; } ret = nilfs_get_root_dentry(s, root, root_dentry); @@ -1101,8 +1096,9 @@ nilfs_fill_super(struct super_block *sb, void *data, int silent) cno = nilfs_last_cno(nilfs); err = nilfs_attach_checkpoint(sb, cno, true, &fsroot); if (err) { - printk(KERN_ERR "NILFS: error loading last checkpoint " - "(checkpoint number=%llu).\n", (unsigned long long)cno); + nilfs_msg(sb, KERN_ERR, + "error %d while loading last checkpoint (checkpoint number=%llu)", + err, (unsigned long long)cno); goto failed_unload; } @@ -1162,9 +1158,8 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) err = -EINVAL; if (!nilfs_valid_fs(nilfs)) { - printk(KERN_WARNING "NILFS (device %s): couldn't " - "remount because the filesystem is in an " - "incomplete recovery state.\n", sb->s_id); + nilfs_msg(sb, KERN_WARNING, + "couldn't remount because the filesystem is in an incomplete recovery state"); goto restore_opts; } @@ -1196,10 +1191,9 @@ static int nilfs_remount(struct super_block *sb, int *flags, char *data) ~NILFS_FEATURE_COMPAT_RO_SUPP; up_read(&nilfs->ns_sem); if (features) { - printk(KERN_WARNING "NILFS (device %s): couldn't " - "remount RDWR because of unsupported optional " - "features (%llx)\n", - sb->s_id, (unsigned long long)features); + nilfs_msg(sb, KERN_WARNING, + "couldn't remount RDWR because of unsupported optional features (%llx)", + (unsigned long long)features); err = -EROFS; goto restore_opts; } @@ -1262,8 +1256,8 @@ static int nilfs_identify(char *data, struct nilfs_super_data *sd) } } if (ret) - printk(KERN_ERR - "NILFS: invalid mount option: %s\n", p); + nilfs_msg(NULL, KERN_ERR, + "invalid mount option: %s", p); } if (!options) break; @@ -1344,10 +1338,10 @@ nilfs_mount(struct file_system_type *fs_type, int flags, } else if (!sd.cno) { if (nilfs_tree_is_busy(s->s_root)) { if ((flags ^ s->s_flags) & MS_RDONLY) { - printk(KERN_ERR "NILFS: the device already " - "has a %s mount.\n", - (s->s_flags & MS_RDONLY) ? - "read-only" : "read/write"); + nilfs_msg(s, KERN_ERR, + "the device already has a %s mount.", + (s->s_flags & MS_RDONLY) ? + "read-only" : "read/write"); err = -EBUSY; goto failed_super; } diff --git a/fs/nilfs2/sysfs.c b/fs/nilfs2/sysfs.c index 8ffa42b704d8..8e57bb91fe16 100644 --- a/fs/nilfs2/sysfs.c +++ b/fs/nilfs2/sysfs.c @@ -272,8 +272,8 @@ nilfs_checkpoints_checkpoints_number_show(struct nilfs_checkpoints_attr *attr, err = nilfs_cpfile_get_stat(nilfs->ns_cpfile, &cpstat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get checkpoint stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get checkpoint stat: err=%d", err); return err; } @@ -295,8 +295,8 @@ nilfs_checkpoints_snapshots_number_show(struct nilfs_checkpoints_attr *attr, err = nilfs_cpfile_get_stat(nilfs->ns_cpfile, &cpstat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get checkpoint stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get checkpoint stat: err=%d", err); return err; } @@ -414,8 +414,8 @@ nilfs_segments_dirty_segments_show(struct nilfs_segments_attr *attr, err = nilfs_sufile_get_stat(nilfs->ns_sufile, &sustat); up_read(&nilfs->ns_segctor_sem); if (err < 0) { - printk(KERN_ERR "NILFS: unable to get segment stat: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to get segment stat: err=%d", err); return err; } @@ -789,14 +789,15 @@ nilfs_superblock_sb_update_frequency_store(struct nilfs_superblock_attr *attr, err = kstrtouint(skip_spaces(buf), 0, &val); if (err) { - printk(KERN_ERR "NILFS: unable to convert string: err=%d\n", - err); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unable to convert string: err=%d", err); return err; } if (val < NILFS_SB_FREQ) { val = NILFS_SB_FREQ; - printk(KERN_WARNING "NILFS: superblock update frequency cannot be lesser than 10 seconds\n"); + nilfs_msg(nilfs->ns_sb, KERN_WARNING, + "superblock update frequency cannot be lesser than 10 seconds"); } down_write(&nilfs->ns_sem); @@ -999,7 +1000,8 @@ int nilfs_sysfs_create_device_group(struct super_block *sb) nilfs->ns_dev_subgroups = kzalloc(devgrp_size, GFP_KERNEL); if (unlikely(!nilfs->ns_dev_subgroups)) { err = -ENOMEM; - printk(KERN_ERR "NILFS: unable to allocate memory for device group\n"); + nilfs_msg(sb, KERN_ERR, + "unable to allocate memory for device group"); goto failed_create_device_group; } @@ -1109,15 +1111,15 @@ int __init nilfs_sysfs_init(void) nilfs_kset = kset_create_and_add(NILFS_ROOT_GROUP_NAME, NULL, fs_kobj); if (!nilfs_kset) { err = -ENOMEM; - printk(KERN_ERR "NILFS: unable to create sysfs entry: err %d\n", - err); + nilfs_msg(NULL, KERN_ERR, + "unable to create sysfs entry: err=%d", err); goto failed_sysfs_init; } err = sysfs_create_group(&nilfs_kset->kobj, &nilfs_feature_attr_group); if (unlikely(err)) { - printk(KERN_ERR "NILFS: unable to create feature group: err %d\n", - err); + nilfs_msg(NULL, KERN_ERR, + "unable to create feature group: err=%d", err); goto cleanup_sysfs_init; } diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 702115164cf3..2dd75bf619ad 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -192,7 +192,10 @@ static int nilfs_store_log_cursor(struct the_nilfs *nilfs, nilfs_get_segnum_of_block(nilfs, nilfs->ns_last_pseg); nilfs->ns_cno = nilfs->ns_last_cno + 1; if (nilfs->ns_segnum >= nilfs->ns_nsegments) { - printk(KERN_ERR "NILFS invalid last segment number.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "pointed segment number is out of range: segnum=%llu, nsegments=%lu", + (unsigned long long)nilfs->ns_segnum, + nilfs->ns_nsegments); ret = -EINVAL; } return ret; @@ -216,12 +219,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) int err; if (!valid_fs) { - printk(KERN_WARNING "NILFS warning: mounting unchecked fs\n"); + nilfs_msg(sb, KERN_WARNING, "mounting unchecked fs"); if (s_flags & MS_RDONLY) { - printk(KERN_INFO "NILFS: INFO: recovery " - "required for readonly filesystem.\n"); - printk(KERN_INFO "NILFS: write access will " - "be enabled during recovery.\n"); + nilfs_msg(sb, KERN_INFO, + "recovery required for readonly filesystem"); + nilfs_msg(sb, KERN_INFO, + "write access will be enabled during recovery"); } } @@ -236,13 +239,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) goto scan_error; if (!nilfs_valid_sb(sbp[1])) { - printk(KERN_WARNING - "NILFS warning: unable to fall back to spare" - "super block\n"); + nilfs_msg(sb, KERN_WARNING, + "unable to fall back to spare super block"); goto scan_error; } - printk(KERN_INFO - "NILFS: try rollback from an earlier position\n"); + nilfs_msg(sb, KERN_INFO, + "trying rollback from an earlier position"); /* * restore super block with its spare and reconfigure @@ -255,10 +257,9 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) /* verify consistency between two super blocks */ blocksize = BLOCK_SIZE << le32_to_cpu(sbp[0]->s_log_block_size); if (blocksize != nilfs->ns_blocksize) { - printk(KERN_WARNING - "NILFS warning: blocksize differs between " - "two super blocks (%d != %d)\n", - blocksize, nilfs->ns_blocksize); + nilfs_msg(sb, KERN_WARNING, + "blocksize differs between two super blocks (%d != %d)", + blocksize, nilfs->ns_blocksize); goto scan_error; } @@ -277,7 +278,8 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) err = nilfs_load_super_root(nilfs, sb, ri.ri_super_root); if (unlikely(err)) { - printk(KERN_ERR "NILFS: error loading super root.\n"); + nilfs_msg(sb, KERN_ERR, "error %d while loading super root", + err); goto failed; } @@ -288,30 +290,29 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) __u64 features; if (nilfs_test_opt(nilfs, NORECOVERY)) { - printk(KERN_INFO "NILFS: norecovery option specified. " - "skipping roll-forward recovery\n"); + nilfs_msg(sb, KERN_INFO, + "norecovery option specified, skipping roll-forward recovery"); goto skip_recovery; } features = le64_to_cpu(nilfs->ns_sbp[0]->s_feature_compat_ro) & ~NILFS_FEATURE_COMPAT_RO_SUPP; if (features) { - printk(KERN_ERR "NILFS: couldn't proceed with " - "recovery because of unsupported optional " - "features (%llx)\n", - (unsigned long long)features); + nilfs_msg(sb, KERN_ERR, + "couldn't proceed with recovery because of unsupported optional features (%llx)", + (unsigned long long)features); err = -EROFS; goto failed_unload; } if (really_read_only) { - printk(KERN_ERR "NILFS: write access " - "unavailable, cannot proceed.\n"); + nilfs_msg(sb, KERN_ERR, + "write access unavailable, cannot proceed"); err = -EROFS; goto failed_unload; } sb->s_flags &= ~MS_RDONLY; } else if (nilfs_test_opt(nilfs, NORECOVERY)) { - printk(KERN_ERR "NILFS: recovery cancelled because norecovery " - "option was specified for a read/write mount\n"); + nilfs_msg(sb, KERN_ERR, + "recovery cancelled because norecovery option was specified for a read/write mount"); err = -EINVAL; goto failed_unload; } @@ -326,11 +327,12 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) up_write(&nilfs->ns_sem); if (err) { - printk(KERN_ERR "NILFS: failed to update super block. " - "recovery unfinished.\n"); + nilfs_msg(sb, KERN_ERR, + "error %d updating super block. recovery unfinished.", + err); goto failed_unload; } - printk(KERN_INFO "NILFS: recovery complete.\n"); + nilfs_msg(sb, KERN_INFO, "recovery complete"); skip_recovery: nilfs_clear_recovery_info(&ri); @@ -338,7 +340,7 @@ int load_nilfs(struct the_nilfs *nilfs, struct super_block *sb) return 0; scan_error: - printk(KERN_ERR "NILFS: error searching super root.\n"); + nilfs_msg(sb, KERN_ERR, "error %d while searching super root", err); goto failed; failed_unload: @@ -385,12 +387,11 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, struct nilfs_super_block *sbp) { if (le32_to_cpu(sbp->s_rev_level) < NILFS_MIN_SUPP_REV) { - printk(KERN_ERR "NILFS: unsupported revision " - "(superblock rev.=%d.%d, current rev.=%d.%d). " - "Please check the version of mkfs.nilfs.\n", - le32_to_cpu(sbp->s_rev_level), - le16_to_cpu(sbp->s_minor_rev_level), - NILFS_CURRENT_REV, NILFS_MINOR_REV); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "unsupported revision (superblock rev.=%d.%d, current rev.=%d.%d). Please check the version of mkfs.nilfs(2).", + le32_to_cpu(sbp->s_rev_level), + le16_to_cpu(sbp->s_minor_rev_level), + NILFS_CURRENT_REV, NILFS_MINOR_REV); return -EINVAL; } nilfs->ns_sbsize = le16_to_cpu(sbp->s_bytes); @@ -399,12 +400,14 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, nilfs->ns_inode_size = le16_to_cpu(sbp->s_inode_size); if (nilfs->ns_inode_size > nilfs->ns_blocksize) { - printk(KERN_ERR "NILFS: too large inode size: %d bytes.\n", - nilfs->ns_inode_size); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too large inode size: %d bytes", + nilfs->ns_inode_size); return -EINVAL; } else if (nilfs->ns_inode_size < NILFS_MIN_INODE_SIZE) { - printk(KERN_ERR "NILFS: too small inode size: %d bytes.\n", - nilfs->ns_inode_size); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too small inode size: %d bytes", + nilfs->ns_inode_size); return -EINVAL; } @@ -412,7 +415,9 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, nilfs->ns_blocks_per_segment = le32_to_cpu(sbp->s_blocks_per_segment); if (nilfs->ns_blocks_per_segment < NILFS_SEG_MIN_BLOCKS) { - printk(KERN_ERR "NILFS: too short segment.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "too short segment: %lu blocks", + nilfs->ns_blocks_per_segment); return -EINVAL; } @@ -421,7 +426,9 @@ static int nilfs_store_disk_layout(struct the_nilfs *nilfs, le32_to_cpu(sbp->s_r_segments_percentage); if (nilfs->ns_r_segments_percentage < 1 || nilfs->ns_r_segments_percentage > 99) { - printk(KERN_ERR "NILFS: invalid reserved segments percentage.\n"); + nilfs_msg(nilfs->ns_sb, KERN_ERR, + "invalid reserved segments percentage: %lu", + nilfs->ns_r_segments_percentage); return -EINVAL; } @@ -505,16 +512,16 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs, if (!sbp[0]) { if (!sbp[1]) { - printk(KERN_ERR "NILFS: unable to read superblock\n"); + nilfs_msg(sb, KERN_ERR, "unable to read superblock"); return -EIO; } - printk(KERN_WARNING - "NILFS warning: unable to read primary superblock " - "(blocksize = %d)\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "unable to read primary superblock (blocksize = %d)", + blocksize); } else if (!sbp[1]) { - printk(KERN_WARNING - "NILFS warning: unable to read secondary superblock " - "(blocksize = %d)\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "unable to read secondary superblock (blocksize = %d)", + blocksize); } /* @@ -536,14 +543,14 @@ static int nilfs_load_super_block(struct the_nilfs *nilfs, } if (!valid[swp]) { nilfs_release_super_block(nilfs); - printk(KERN_ERR "NILFS: Can't find nilfs on dev %s.\n", - sb->s_id); + nilfs_msg(sb, KERN_ERR, "couldn't find nilfs on the device"); return -EINVAL; } if (!valid[!swp]) - printk(KERN_WARNING "NILFS warning: broken superblock. " - "using spare superblock (blocksize = %d).\n", blocksize); + nilfs_msg(sb, KERN_WARNING, + "broken superblock, retrying with spare superblock (blocksize = %d)", + blocksize); if (swp) nilfs_swap_super_block(nilfs); @@ -577,7 +584,7 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) blocksize = sb_min_blocksize(sb, NILFS_MIN_BLOCK_SIZE); if (!blocksize) { - printk(KERN_ERR "NILFS: unable to set blocksize\n"); + nilfs_msg(sb, KERN_ERR, "unable to set blocksize"); err = -EINVAL; goto out; } @@ -596,8 +603,9 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) blocksize = BLOCK_SIZE << le32_to_cpu(sbp->s_log_block_size); if (blocksize < NILFS_MIN_BLOCK_SIZE || blocksize > NILFS_MAX_BLOCK_SIZE) { - printk(KERN_ERR "NILFS: couldn't mount because of unsupported " - "filesystem blocksize %d\n", blocksize); + nilfs_msg(sb, KERN_ERR, + "couldn't mount because of unsupported filesystem blocksize %d", + blocksize); err = -EINVAL; goto failed_sbh; } @@ -605,10 +613,9 @@ int init_nilfs(struct the_nilfs *nilfs, struct super_block *sb, char *data) int hw_blocksize = bdev_logical_block_size(sb->s_bdev); if (blocksize < hw_blocksize) { - printk(KERN_ERR - "NILFS: blocksize %d too small for device " - "(sector-size = %d).\n", - blocksize, hw_blocksize); + nilfs_msg(sb, KERN_ERR, + "blocksize %d too small for device (sector-size = %d)", + blocksize, hw_blocksize); err = -EINVAL; goto failed_sbh; } -- cgit v1.2.3 From d6517deb014954d3229910e46f3b85b7ad80db3e Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:14 -0700 Subject: nilfs2: replace nilfs_warning() with nilfs_msg() Use nilfs_msg() to output warning messages and get rid of nilfs_warning() function. This also removes function names from the messages unless we embed them explicitly in format strings. Instead, some messages are revised to clarify the context. [arnd@arndb.de: avoid warning about unused variables] Link: http://lkml.kernel.org/r/20160615201945.3348205-1-arnd@arndb.de Link: http://lkml.kernel.org/r/1464875891-5443-6-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/alloc.c | 45 +++++++++++++++++++++------------------------ fs/nilfs2/ifile.c | 4 ++-- fs/nilfs2/inode.c | 21 +++++++++++---------- fs/nilfs2/namei.c | 6 +++--- fs/nilfs2/nilfs.h | 7 ++++--- fs/nilfs2/page.c | 19 +++++++++---------- fs/nilfs2/segment.c | 20 ++++++++++---------- fs/nilfs2/super.c | 21 +-------------------- 8 files changed, 61 insertions(+), 82 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 1a85d94f5b25..2c90e285d7c6 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -622,10 +622,10 @@ void nilfs_palloc_commit_free_entry(struct inode *inode, lock = nilfs_mdt_bgl_lock(inode, group); if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)req->pr_entry_nr, - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)req->pr_entry_nr); else nilfs_palloc_group_desc_add_entries(desc, lock, 1); @@ -663,10 +663,10 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode, lock = nilfs_mdt_bgl_lock(inode, group); if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)req->pr_entry_nr, - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)req->pr_entry_nr); else nilfs_palloc_group_desc_add_entries(desc, lock, 1); @@ -772,10 +772,10 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) do { if (!nilfs_clear_bit_atomic(lock, group_offset, bitmap)) { - nilfs_warning(inode->i_sb, __func__, - "entry number %llu already freed: ino=%lu", - (unsigned long long)entry_nrs[j], - (unsigned long)inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "%s (ino=%lu): entry number %llu already freed", + __func__, inode->i_ino, + (unsigned long long)entry_nrs[j]); } else { n++; } @@ -816,12 +816,11 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) for (k = 0; k < nempties; k++) { ret = nilfs_palloc_delete_entry_block(inode, last_nrs[k]); - if (ret && ret != -ENOENT) { - nilfs_warning(inode->i_sb, __func__, - "failed to delete block of entry %llu: ino=%lu, err=%d", - (unsigned long long)last_nrs[k], - (unsigned long)inode->i_ino, ret); - } + if (ret && ret != -ENOENT) + nilfs_msg(inode->i_sb, KERN_WARNING, + "error %d deleting block that object (entry=%llu, ino=%lu) belongs to", + ret, (unsigned long long)last_nrs[k], + inode->i_ino); } desc_kaddr = kmap_atomic(desc_bh->b_page); @@ -835,12 +834,10 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) if (nfree == nilfs_palloc_entries_per_group(inode)) { ret = nilfs_palloc_delete_bitmap_block(inode, group); - if (ret && ret != -ENOENT) { - nilfs_warning(inode->i_sb, __func__, - "failed to delete bitmap block of group %lu: ino=%lu, err=%d", - group, - (unsigned long)inode->i_ino, ret); - } + if (ret && ret != -ENOENT) + nilfs_msg(inode->i_sb, KERN_WARNING, + "error %d deleting bitmap block of group=%lu, ino=%lu", + ret, group, inode->i_ino); } } return 0; diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index b1c96285aa4a..b8fa45c20c63 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -151,8 +151,8 @@ int nilfs_ifile_get_inode_block(struct inode *ifile, ino_t ino, err = nilfs_palloc_get_entry_block(ifile, ino, 0, out_bh); if (unlikely(err)) - nilfs_warning(sb, __func__, "unable to read inode: %lu", - (unsigned long) ino); + nilfs_msg(sb, KERN_WARNING, "error %d reading inode: ino=%lu", + err, (unsigned long)ino); return err; } diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index a965fcf77955..b286b35174a5 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -723,9 +723,9 @@ repeat: goto repeat; failed: - nilfs_warning(ii->vfs_inode.i_sb, __func__, - "failed to truncate bmap (ino=%lu, err=%d)", - ii->vfs_inode.i_ino, ret); + nilfs_msg(ii->vfs_inode.i_sb, KERN_WARNING, + "error %d truncating bmap (ino=%lu)", ret, + ii->vfs_inode.i_ino); } void nilfs_truncate(struct inode *inode) @@ -936,9 +936,9 @@ int nilfs_set_file_dirty(struct inode *inode, unsigned int nr_dirty) * This will happen when somebody is freeing * this inode. */ - nilfs_warning(inode->i_sb, __func__, - "cannot get inode (ino=%lu)", - inode->i_ino); + nilfs_msg(inode->i_sb, KERN_WARNING, + "cannot set file dirty (ino=%lu): the file is being freed", + inode->i_ino); spin_unlock(&nilfs->ns_inode_lock); return -EINVAL; /* * NILFS_I_DIRTY may remain for @@ -959,8 +959,9 @@ int __nilfs_mark_inode_dirty(struct inode *inode, int flags) err = nilfs_load_inode_block(inode, &ibh); if (unlikely(err)) { - nilfs_warning(inode->i_sb, __func__, - "failed to reget inode block."); + nilfs_msg(inode->i_sb, KERN_WARNING, + "cannot mark inode dirty (ino=%lu): error %d loading inode block", + inode->i_ino, err); return err; } nilfs_update_inode(inode, ibh, flags); @@ -986,8 +987,8 @@ void nilfs_dirty_inode(struct inode *inode, int flags) struct nilfs_mdt_info *mdi = NILFS_MDT(inode); if (is_bad_inode(inode)) { - nilfs_warning(inode->i_sb, __func__, - "tried to mark bad_inode dirty. ignored."); + nilfs_msg(inode->i_sb, KERN_WARNING, + "tried to mark bad_inode dirty. ignored."); dump_stack(); return; } diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 1ec8ae5995a5..dbcf1dc93a51 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -283,9 +283,9 @@ static int nilfs_do_unlink(struct inode *dir, struct dentry *dentry) goto out; if (!inode->i_nlink) { - nilfs_warning(inode->i_sb, __func__, - "deleting nonexistent file (%lu), %d", - inode->i_ino, inode->i_nlink); + nilfs_msg(inode->i_sb, KERN_WARNING, + "deleting nonexistent file (ino=%lu), %d", + inode->i_ino, inode->i_nlink); set_nlink(inode, 1); } err = nilfs_delete_entry(de, page); diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index b57ce41e8a1a..46fbd4e00315 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -306,8 +306,6 @@ void __nilfs_msg(struct super_block *sb, const char *level, extern __printf(3, 4) void __nilfs_error(struct super_block *sb, const char *function, const char *fmt, ...); -extern __printf(3, 4) -void nilfs_warning(struct super_block *, const char *, const char *, ...); #ifdef CONFIG_PRINTK @@ -319,7 +317,10 @@ void nilfs_warning(struct super_block *, const char *, const char *, ...); #else #define nilfs_msg(sb, level, fmt, ...) \ - no_printk(fmt, ##__VA_ARGS__) + do { \ + no_printk(fmt, ##__VA_ARGS__); \ + (void)(sb); \ + } while (0) #define nilfs_error(sb, fmt, ...) \ do { \ no_printk(fmt, ##__VA_ARGS__); \ diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index d97ba5f11b77..eaccf12c296e 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -403,11 +403,10 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) BUG_ON(!PageLocked(page)); - if (!silent) { - nilfs_warning(sb, __func__, - "discard page: offset %lld, ino %lu", - page_offset(page), inode->i_ino); - } + if (!silent) + nilfs_msg(sb, KERN_WARNING, + "discard dirty page: offset=%lld, ino=%lu", + page_offset(page), inode->i_ino); ClearPageUptodate(page); ClearPageMappedToDisk(page); @@ -422,11 +421,11 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) bh = head = page_buffers(page); do { lock_buffer(bh); - if (!silent) { - nilfs_warning(sb, __func__, - "discard block %llu, size %zu", - (u64)bh->b_blocknr, bh->b_size); - } + if (!silent) + nilfs_msg(sb, KERN_WARNING, + "discard dirty block: blocknr=%llu, size=%zu", + (u64)bh->b_blocknr, bh->b_size); + set_mask_bits(&bh->b_state, clear_bits, 0); unlock_buffer(bh); } while (bh = bh->b_this_page, bh != head); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 1cc968502e53..7e1864c6035b 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -1951,8 +1951,9 @@ static int nilfs_segctor_collect_dirty_files(struct nilfs_sc_info *sci, err = nilfs_ifile_get_inode_block( ifile, ii->vfs_inode.i_ino, &ibh); if (unlikely(err)) { - nilfs_warning(sci->sc_super, __func__, - "failed to get inode block."); + nilfs_msg(sci->sc_super, KERN_WARNING, + "log writer: error %d getting inode block (ino=%lu)", + err, ii->vfs_inode.i_ino); return err; } mark_buffer_dirty(ibh); @@ -2458,8 +2459,7 @@ int nilfs_clean_segments(struct super_block *sb, struct nilfs_argv *argv, if (likely(!err)) break; - nilfs_warning(sb, __func__, - "segment construction failed. (err=%d)", err); + nilfs_msg(sb, KERN_WARNING, "error %d cleaning segments", err); set_current_state(TASK_INTERRUPTIBLE); schedule_timeout(sci->sc_interval); } @@ -2738,14 +2738,14 @@ static void nilfs_segctor_destroy(struct nilfs_sc_info *sci) nilfs_segctor_write_out(sci); if (!list_empty(&sci->sc_dirty_files)) { - nilfs_warning(sci->sc_super, __func__, - "dirty file(s) after the final construction"); + nilfs_msg(sci->sc_super, KERN_WARNING, + "disposed unprocessed dirty file(s) when stopping log writer"); nilfs_dispose_list(nilfs, &sci->sc_dirty_files, 1); } if (!list_empty(&sci->sc_iput_queue)) { - nilfs_warning(sci->sc_super, __func__, - "iput queue is not empty"); + nilfs_msg(sci->sc_super, KERN_WARNING, + "disposed unprocessed inode(s) in iput queue when stopping log writer"); nilfs_dispose_list(nilfs, &sci->sc_iput_queue, 1); } @@ -2821,8 +2821,8 @@ void nilfs_detach_log_writer(struct super_block *sb) spin_lock(&nilfs->ns_inode_lock); if (!list_empty(&nilfs->ns_dirty_files)) { list_splice_init(&nilfs->ns_dirty_files, &garbage_list); - nilfs_warning(sb, __func__, - "Hit dirty file after stopped log writer"); + nilfs_msg(sb, KERN_WARNING, + "disposed unprocessed dirty file(s) when detaching log writer"); } spin_unlock(&nilfs->ns_inode_lock); up_write(&nilfs->ns_segctor_sem); diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 90c62b489857..33ba6f78de69 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -115,8 +115,7 @@ static void nilfs_set_error(struct super_block *sb) * * This implements the body of nilfs_error() macro. Normally, * nilfs_error() should be used. As for sustainable errors such as a - * single-shot I/O error, nilfs_warning() or printk() should be used - * instead. + * single-shot I/O error, nilfs_msg() should be used instead. * * Callers should not add a trailing newline since this will do it. */ @@ -151,24 +150,6 @@ void __nilfs_error(struct super_block *sb, const char *function, sb->s_id); } -void nilfs_warning(struct super_block *sb, const char *function, - const char *fmt, ...) -{ - struct va_format vaf; - va_list args; - - va_start(args, fmt); - - vaf.fmt = fmt; - vaf.va = &args; - - printk(KERN_WARNING "NILFS warning (device %s): %s: %pV\n", - sb->s_id, function, &vaf); - - va_end(args); -} - - struct inode *nilfs_alloc_inode(struct super_block *sb) { struct nilfs_inode_info *ii; -- cgit v1.2.3 From 39a9dcca61a3d1375b9440676cbfc541804cd217 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:17 -0700 Subject: nilfs2: emit error message when I/O error is detected When nilfs returned -EIO as an error code, it's not always clear if it came from the underlying block device or not. This will mend the issue by having low level I/O routines of nilfs output an error message when they detected an I/O error. Link: http://lkml.kernel.org/r/1464875891-5443-7-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/btree.c | 3 +++ fs/nilfs2/gcinode.c | 9 ++++++++- fs/nilfs2/mdt.c | 6 +++++- 3 files changed, 16 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 2c52693a69a4..2e315f9f2e51 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -517,6 +517,9 @@ static int __nilfs_btree_get_block(const struct nilfs_bmap *btree, __u64 ptr, out_no_wait: if (!buffer_uptodate(bh)) { + nilfs_msg(btree->b_inode->i_sb, KERN_ERR, + "I/O error reading b-tree node block (ino=%lu, blocknr=%llu)", + btree->b_inode->i_ino, (unsigned long long)ptr); brelse(bh); return -EIO; } diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c index e9148f94d696..853a831dcde0 100644 --- a/fs/nilfs2/gcinode.c +++ b/fs/nilfs2/gcinode.c @@ -148,8 +148,15 @@ int nilfs_gccache_submit_read_node(struct inode *inode, sector_t pbn, int nilfs_gccache_wait_and_mark_dirty(struct buffer_head *bh) { wait_on_buffer(bh); - if (!buffer_uptodate(bh)) + if (!buffer_uptodate(bh)) { + struct inode *inode = bh->b_page->mapping->host; + + nilfs_msg(inode->i_sb, KERN_ERR, + "I/O error reading %s block for GC (ino=%lu, vblocknr=%llu)", + buffer_nilfs_node(bh) ? "node" : "data", + inode->i_ino, (unsigned long long)bh->b_blocknr); return -EIO; + } if (buffer_dirty(bh)) return -EEXIST; diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index 0d7b71fbeff8..d56d3a5bea88 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -207,8 +207,12 @@ static int nilfs_mdt_read_block(struct inode *inode, unsigned long block, out_no_wait: err = -EIO; - if (!buffer_uptodate(first_bh)) + if (!buffer_uptodate(first_bh)) { + nilfs_msg(inode->i_sb, KERN_ERR, + "I/O error reading meta-data file (ino=%lu, block-offset=%lu)", + inode->i_ino, block); goto failed_bh; + } out: *out_bh = first_bh; return 0; -- cgit v1.2.3 From aceb4170bb2ba88c5327cc69b9c91a708c7f7046 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:19 -0700 Subject: nilfs2: do not use yield() Use cond_resched() instead of yield() in the loop of nilfs_transaction_lock() since the usage corresponds to the "be nice for others" case that the comment of yield() says. This removes the following checkpatch.pl warning: "WARNING: Using yield() is generally wrong. See yield() kernel-doc (sched/core.c)" Link: http://lkml.kernel.org/r/1464875891-5443-8-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/segment.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 7e1864c6035b..5a97282aa074 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -373,7 +373,7 @@ static void nilfs_transaction_lock(struct super_block *sb, nilfs_segctor_do_immediate_flush(sci); up_write(&nilfs->ns_segctor_sem); - yield(); + cond_resched(); } if (gcflag) ti->ti_flags |= NILFS_TI_GC; -- cgit v1.2.3 From a7d3f104da57eecb2b9881127d6bdf9abe7fde99 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:22 -0700 Subject: nilfs2: refactor parser of snapshot mount option Move parser of snapshot mount option to a separate function nilfs_parse_snapshot_option(), replace simple_strtoull() with kstrtoull() to avoid checkpatch.pl warning "WARNING: simple_strtoull is obsolete, use kstrtoull instead", and refine the error message of the parser. Link: http://lkml.kernel.org/r/1464875891-5443-9-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/super.c | 53 +++++++++++++++++++++++++++++++++++------------------ 1 file changed, 35 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 33ba6f78de69..c95d369e90aa 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1205,6 +1205,38 @@ struct nilfs_super_data { int flags; }; +static int nilfs_parse_snapshot_option(const char *option, + const substring_t *arg, + struct nilfs_super_data *sd) +{ + unsigned long long val; + const char *msg = NULL; + int err; + + if (!(sd->flags & MS_RDONLY)) { + msg = "read-only option is not specified"; + goto parse_error; + } + + err = kstrtoull(arg->from, 0, &val); + if (err) { + if (err == -ERANGE) + msg = "too large checkpoint number"; + else + msg = "malformed argument"; + goto parse_error; + } else if (val == 0) { + msg = "invalid checkpoint number 0"; + goto parse_error; + } + sd->cno = val; + return 0; + +parse_error: + nilfs_msg(NULL, KERN_ERR, "invalid option \"%s\": %s", option, msg); + return 1; +} + /** * nilfs_identify - pre-read mount options needed to identify mount instance * @data: mount options @@ -1221,24 +1253,9 @@ static int nilfs_identify(char *data, struct nilfs_super_data *sd) p = strsep(&options, ","); if (p != NULL && *p) { token = match_token(p, tokens, args); - if (token == Opt_snapshot) { - if (!(sd->flags & MS_RDONLY)) { - ret++; - } else { - sd->cno = simple_strtoull(args[0].from, - NULL, 0); - /* - * No need to see the end pointer; - * match_token() has done syntax - * checking. - */ - if (sd->cno == 0) - ret++; - } - } - if (ret) - nilfs_msg(NULL, KERN_ERR, - "invalid mount option: %s", p); + if (token == Opt_snapshot) + ret = nilfs_parse_snapshot_option(p, &args[0], + sd); } if (!options) break; -- cgit v1.2.3 From ad980c9ab77cc4e0edc5dd3361fd69daabeb99f9 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:25 -0700 Subject: nilfs2: fix misuse of a semaphore in sysfs code Variables ns_seg_seq, ns_segnum, ns_nextnum, ns_pseg_offset, ns_cno, ns_ctime, ns_nongc_ctime, and ns_ndirtyblks, are protected by ns_segctor_sem, but ns_sem is wrongly used by the nilfs sysfs code when reading these variables. This fixes the misuse and clarifies which semaphore protects them in the comment of the_nilfs struct. Link: http://lkml.kernel.org/r/1465825507-3407-2-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/sysfs.c | 44 ++++++++++++++++++++++---------------------- fs/nilfs2/the_nilfs.h | 7 ++----- 2 files changed, 24 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/sysfs.c b/fs/nilfs2/sysfs.c index 8e57bb91fe16..490303e3d517 100644 --- a/fs/nilfs2/sysfs.c +++ b/fs/nilfs2/sysfs.c @@ -326,9 +326,9 @@ nilfs_checkpoints_next_checkpoint_show(struct nilfs_checkpoints_attr *attr, { __u64 cno; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); cno = nilfs->ns_cno; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", cno); } @@ -511,9 +511,9 @@ nilfs_segctor_current_seg_sequence_show(struct nilfs_segctor_attr *attr, { u64 seg_seq; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); seg_seq = nilfs->ns_seg_seq; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", seg_seq); } @@ -525,9 +525,9 @@ nilfs_segctor_current_last_full_seg_show(struct nilfs_segctor_attr *attr, { __u64 segnum; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); segnum = nilfs->ns_segnum; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", segnum); } @@ -539,9 +539,9 @@ nilfs_segctor_next_full_seg_show(struct nilfs_segctor_attr *attr, { __u64 nextnum; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nextnum = nilfs->ns_nextnum; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", nextnum); } @@ -553,9 +553,9 @@ nilfs_segctor_next_pseg_offset_show(struct nilfs_segctor_attr *attr, { unsigned long pseg_offset; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); pseg_offset = nilfs->ns_pseg_offset; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%lu\n", pseg_offset); } @@ -567,9 +567,9 @@ nilfs_segctor_next_checkpoint_show(struct nilfs_segctor_attr *attr, { __u64 cno; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); cno = nilfs->ns_cno; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", cno); } @@ -581,9 +581,9 @@ nilfs_segctor_last_seg_write_time_show(struct nilfs_segctor_attr *attr, { time_t ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ctime = nilfs->ns_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return NILFS_SHOW_TIME(ctime, buf); } @@ -595,9 +595,9 @@ nilfs_segctor_last_seg_write_time_secs_show(struct nilfs_segctor_attr *attr, { time_t ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ctime = nilfs->ns_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", (unsigned long long)ctime); } @@ -609,9 +609,9 @@ nilfs_segctor_last_nongc_write_time_show(struct nilfs_segctor_attr *attr, { time_t nongc_ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nongc_ctime = nilfs->ns_nongc_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return NILFS_SHOW_TIME(nongc_ctime, buf); } @@ -623,9 +623,9 @@ nilfs_segctor_last_nongc_write_time_secs_show(struct nilfs_segctor_attr *attr, { time_t nongc_ctime; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); nongc_ctime = nilfs->ns_nongc_ctime; - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%llu\n", (unsigned long long)nongc_ctime); @@ -638,9 +638,9 @@ nilfs_segctor_dirty_data_blocks_count_show(struct nilfs_segctor_attr *attr, { u32 ndirtyblks; - down_read(&nilfs->ns_sem); + down_read(&nilfs->ns_segctor_sem); ndirtyblks = atomic_read(&nilfs->ns_ndirtyblks); - up_read(&nilfs->ns_sem); + up_read(&nilfs->ns_segctor_sem); return snprintf(buf, PAGE_SIZE, "%u\n", ndirtyblks); } diff --git a/fs/nilfs2/the_nilfs.h b/fs/nilfs2/the_nilfs.h index 79d1421896d0..b305c6f033e7 100644 --- a/fs/nilfs2/the_nilfs.h +++ b/fs/nilfs2/the_nilfs.h @@ -122,11 +122,8 @@ struct the_nilfs { unsigned int ns_sb_update_freq; /* - * Following fields are dedicated to a writable FS-instance. - * Except for the period seeking checkpoint, code outside the segment - * constructor must lock a segment semaphore while accessing these - * fields. - * The writable FS-instance is sole during a lifetime of the_nilfs. + * The following fields are updated by a writable FS-instance. + * These fields are protected by ns_segctor_sem outside load_nilfs(). */ u64 ns_seg_seq; __u64 ns_segnum; -- cgit v1.2.3 From 4ce5c3426cbe9193f82345fb103e17dc3335eb4f Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:28 -0700 Subject: nilfs2: use BIT() macro Replace bit shifts by BIT macro for clarity. Link: http://lkml.kernel.org/r/1465825507-3407-3-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/nilfs2/btnode.c | 4 ++-- fs/nilfs2/inode.c | 4 ++-- fs/nilfs2/nilfs.h | 15 +++++++-------- fs/nilfs2/page.c | 26 +++++++++++++------------- fs/nilfs2/segment.c | 14 +++++++------- fs/nilfs2/sufile.c | 12 ++++++------ 6 files changed, 37 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/nilfs2/btnode.c b/fs/nilfs2/btnode.c index 4cca998ec7a0..d5c23da43513 100644 --- a/fs/nilfs2/btnode.c +++ b/fs/nilfs2/btnode.c @@ -41,7 +41,7 @@ nilfs_btnode_create_block(struct address_space *btnc, __u64 blocknr) struct inode *inode = NILFS_BTNC_I(btnc); struct buffer_head *bh; - bh = nilfs_grab_buffer(inode, btnc, blocknr, 1 << BH_NILFS_Node); + bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node)); if (unlikely(!bh)) return NULL; @@ -70,7 +70,7 @@ int nilfs_btnode_submit_block(struct address_space *btnc, __u64 blocknr, struct page *page; int err; - bh = nilfs_grab_buffer(inode, btnc, blocknr, 1 << BH_NILFS_Node); + bh = nilfs_grab_buffer(inode, btnc, blocknr, BIT(BH_NILFS_Node)); if (unlikely(!bh)) return -ENOMEM; diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index b286b35174a5..af04f553d7c9 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -356,7 +356,7 @@ struct inode *nilfs_new_inode(struct inode *dir, umode_t mode) root = NILFS_I(dir)->i_root; ii = NILFS_I(inode); - ii->i_state = 1 << NILFS_I_NEW; + ii->i_state = BIT(NILFS_I_NEW); ii->i_root = root; err = nilfs_ifile_create_inode(root->ifile, &ino, &ii->i_bh); @@ -555,7 +555,7 @@ static int nilfs_iget_set(struct inode *inode, void *opaque) inode->i_ino = args->ino; if (args->for_gc) { - NILFS_I(inode)->i_state = 1 << NILFS_I_GCINODE; + NILFS_I(inode)->i_state = BIT(NILFS_I_GCINODE); NILFS_I(inode)->i_cno = args->cno; NILFS_I(inode)->i_root = NULL; } else { diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index 46fbd4e00315..2ba8a146af1f 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -119,20 +119,19 @@ enum { /* * Macros to check inode numbers */ -#define NILFS_MDT_INO_BITS \ - ((unsigned int)(1 << NILFS_DAT_INO | 1 << NILFS_CPFILE_INO | \ - 1 << NILFS_SUFILE_INO | 1 << NILFS_IFILE_INO | \ - 1 << NILFS_ATIME_INO | 1 << NILFS_SKETCH_INO)) +#define NILFS_MDT_INO_BITS \ + (BIT(NILFS_DAT_INO) | BIT(NILFS_CPFILE_INO) | \ + BIT(NILFS_SUFILE_INO) | BIT(NILFS_IFILE_INO) | \ + BIT(NILFS_ATIME_INO) | BIT(NILFS_SKETCH_INO)) -#define NILFS_SYS_INO_BITS \ - ((unsigned int)(1 << NILFS_ROOT_INO) | NILFS_MDT_INO_BITS) +#define NILFS_SYS_INO_BITS (BIT(NILFS_ROOT_INO) | NILFS_MDT_INO_BITS) #define NILFS_FIRST_INO(sb) (((struct the_nilfs *)sb->s_fs_info)->ns_first_ino) #define NILFS_MDT_INODE(sb, ino) \ - ((ino) < NILFS_FIRST_INO(sb) && (NILFS_MDT_INO_BITS & (1 << (ino)))) + ((ino) < NILFS_FIRST_INO(sb) && (NILFS_MDT_INO_BITS & BIT(ino))) #define NILFS_VALID_INODE(sb, ino) \ - ((ino) >= NILFS_FIRST_INO(sb) || (NILFS_SYS_INO_BITS & (1 << (ino)))) + ((ino) >= NILFS_FIRST_INO(sb) || (NILFS_SYS_INO_BITS & BIT(ino))) /** * struct nilfs_transaction_info: context information for synchronization diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index eaccf12c296e..f11a3ad2df0c 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -30,9 +30,9 @@ #include "mdt.h" -#define NILFS_BUFFER_INHERENT_BITS \ - ((1UL << BH_Uptodate) | (1UL << BH_Mapped) | (1UL << BH_NILFS_Node) | \ - (1UL << BH_NILFS_Volatile) | (1UL << BH_NILFS_Checked)) +#define NILFS_BUFFER_INHERENT_BITS \ + (BIT(BH_Uptodate) | BIT(BH_Mapped) | BIT(BH_NILFS_Node) | \ + BIT(BH_NILFS_Volatile) | BIT(BH_NILFS_Checked)) static struct buffer_head * __nilfs_get_page_block(struct page *page, unsigned long block, pgoff_t index, @@ -85,9 +85,9 @@ void nilfs_forget_buffer(struct buffer_head *bh) { struct page *page = bh->b_page; const unsigned long clear_bits = - (1 << BH_Uptodate | 1 << BH_Dirty | 1 << BH_Mapped | - 1 << BH_Async_Write | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Checked | 1 << BH_NILFS_Redirected); + (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | + BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); lock_buffer(bh); set_mask_bits(&bh->b_state, clear_bits, 0); @@ -124,17 +124,17 @@ void nilfs_copy_buffer(struct buffer_head *dbh, struct buffer_head *sbh) dbh->b_bdev = sbh->b_bdev; bh = dbh; - bits = sbh->b_state & ((1UL << BH_Uptodate) | (1UL << BH_Mapped)); + bits = sbh->b_state & (BIT(BH_Uptodate) | BIT(BH_Mapped)); while ((bh = bh->b_this_page) != dbh) { lock_buffer(bh); bits &= bh->b_state; unlock_buffer(bh); } - if (bits & (1UL << BH_Uptodate)) + if (bits & BIT(BH_Uptodate)) SetPageUptodate(dpage); else ClearPageUptodate(dpage); - if (bits & (1UL << BH_Mapped)) + if (bits & BIT(BH_Mapped)) SetPageMappedToDisk(dpage); else ClearPageMappedToDisk(dpage); @@ -215,7 +215,7 @@ static void nilfs_copy_page(struct page *dst, struct page *src, int copy_dirty) create_empty_buffers(dst, sbh->b_size, 0); if (copy_dirty) - mask |= (1UL << BH_Dirty); + mask |= BIT(BH_Dirty); dbh = dbufs = page_buffers(dst); do { @@ -414,9 +414,9 @@ void nilfs_clear_dirty_page(struct page *page, bool silent) if (page_has_buffers(page)) { struct buffer_head *bh, *head; const unsigned long clear_bits = - (1 << BH_Uptodate | 1 << BH_Dirty | 1 << BH_Mapped | - 1 << BH_Async_Write | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Checked | 1 << BH_NILFS_Redirected); + (BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) | + BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected)); bh = head = page_buffers(page); do { diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 5a97282aa074..bedcae2c28e6 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -1858,11 +1858,11 @@ static void nilfs_segctor_complete_write(struct nilfs_sc_info *sci) */ list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) { - const unsigned long set_bits = (1 << BH_Uptodate); + const unsigned long set_bits = BIT(BH_Uptodate); const unsigned long clear_bits = - (1 << BH_Dirty | 1 << BH_Async_Write | - 1 << BH_Delay | 1 << BH_NILFS_Volatile | - 1 << BH_NILFS_Redirected); + (BIT(BH_Dirty) | BIT(BH_Async_Write) | + BIT(BH_Delay) | BIT(BH_NILFS_Volatile) | + BIT(BH_NILFS_Redirected)); set_mask_bits(&bh->b_state, clear_bits, set_bits); if (bh == segbuf->sb_super_root) { @@ -2132,10 +2132,10 @@ static void nilfs_segctor_start_timer(struct nilfs_sc_info *sci) static void nilfs_segctor_do_flush(struct nilfs_sc_info *sci, int bn) { spin_lock(&sci->sc_state_lock); - if (!(sci->sc_flush_request & (1 << bn))) { + if (!(sci->sc_flush_request & BIT(bn))) { unsigned long prev_req = sci->sc_flush_request; - sci->sc_flush_request |= (1 << bn); + sci->sc_flush_request |= BIT(bn); if (!prev_req) wake_up(&sci->sc_wait_daemon); } @@ -2319,7 +2319,7 @@ int nilfs_construct_dsync_segment(struct super_block *sb, struct inode *inode, } #define FLUSH_FILE_BIT (0x1) /* data file only */ -#define FLUSH_DAT_BIT (1 << NILFS_DAT_INO) /* DAT only */ +#define FLUSH_DAT_BIT BIT(NILFS_DAT_INO) /* DAT only */ /** * nilfs_segctor_accept - record accepted sequence count of log-write requests diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 5b495c469471..12d11de93602 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -446,7 +446,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, kaddr = kmap_atomic(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); - if (su->su_flags == cpu_to_le32(1UL << NILFS_SEGMENT_USAGE_DIRTY) && + if (su->su_flags == cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)) && su->su_nblocks == cpu_to_le32(0)) { kunmap_atomic(kaddr); return; @@ -457,7 +457,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, /* make the segment garbage */ su->su_lastmod = cpu_to_le64(0); su->su_nblocks = cpu_to_le32(0); - su->su_flags = cpu_to_le32(1UL << NILFS_SEGMENT_USAGE_DIRTY); + su->su_flags = cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)); kunmap_atomic(kaddr); nilfs_sufile_mod_counter(header_bh, clean ? (u64)-1 : 0, dirty ? 0 : 1); @@ -695,7 +695,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile, su2 = su; for (j = 0; j < n; j++, su = (void *)su + susz) { if ((le32_to_cpu(su->su_flags) & - ~(1UL << NILFS_SEGMENT_USAGE_ERROR)) || + ~BIT(NILFS_SEGMENT_USAGE_ERROR)) || nilfs_segment_is_active(nilfs, segnum + j)) { ret = -EBUSY; kunmap_atomic(kaddr); @@ -862,10 +862,10 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, si->sui_lastmod = le64_to_cpu(su->su_lastmod); si->sui_nblocks = le32_to_cpu(su->su_nblocks); si->sui_flags = le32_to_cpu(su->su_flags) & - ~(1UL << NILFS_SEGMENT_USAGE_ACTIVE); + ~BIT(NILFS_SEGMENT_USAGE_ACTIVE); if (nilfs_segment_is_active(nilfs, segnum + j)) si->sui_flags |= - (1UL << NILFS_SEGMENT_USAGE_ACTIVE); + BIT(NILFS_SEGMENT_USAGE_ACTIVE); } kunmap_atomic(kaddr); brelse(su_bh); @@ -953,7 +953,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf, * disk. */ sup->sup_sui.sui_flags &= - ~(1UL << NILFS_SEGMENT_USAGE_ACTIVE); + ~BIT(NILFS_SEGMENT_USAGE_ACTIVE); cleansi = nilfs_suinfo_clean(&sup->sup_sui); cleansu = nilfs_segment_usage_clean(su); -- cgit v1.2.3 From e63e88bc53bac7e4c3f592f8126c51a7569be673 Mon Sep 17 00:00:00 2001 From: Ryusuke Konishi Date: Tue, 2 Aug 2016 14:05:30 -0700 Subject: nilfs2: move ioctl interface and disk layout to uapi separately The header file "include/linux/nilfs2_fs.h" is composed of parts for ioctl and disk format, and both are intended to be shared with user space programs. This moves them to the uapi directory "include/uapi/linux" splitting the file to "nilfs2_api.h" and "nilfs2_ondisk.h". The following minor changes are accompanied by this migration: - nilfs_direct_node struct in nilfs2/direct.h is converged to nilfs2_ondisk.h because it's an on-disk structure. - inline functions nilfs_rec_len_from_disk() and nilfs_rec_len_to_disk() are moved to nilfs2/dir.c. Link: http://lkml.kernel.org/r/1465825507-3407-4-git-send-email-konishi.ryusuke@lab.ntt.co.jp Signed-off-by: Ryusuke Konishi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- Documentation/filesystems/nilfs2.txt | 3 +- Documentation/ioctl/ioctl-number.txt | 2 +- MAINTAINERS | 3 +- fs/nilfs2/bmap.h | 2 +- fs/nilfs2/btree.h | 2 +- fs/nilfs2/cpfile.c | 1 - fs/nilfs2/cpfile.h | 3 +- fs/nilfs2/dat.h | 1 + fs/nilfs2/dir.c | 22 + fs/nilfs2/direct.h | 10 - fs/nilfs2/ifile.h | 1 - fs/nilfs2/ioctl.c | 1 - fs/nilfs2/nilfs.h | 3 +- fs/nilfs2/segment.h | 1 - fs/nilfs2/sufile.c | 1 - fs/nilfs2/sufile.h | 1 - include/linux/nilfs2_fs.h | 934 ----------------------------------- include/uapi/linux/nilfs2_api.h | 292 +++++++++++ include/uapi/linux/nilfs2_ondisk.h | 650 ++++++++++++++++++++++++ 19 files changed, 976 insertions(+), 957 deletions(-) delete mode 100644 include/linux/nilfs2_fs.h create mode 100644 include/uapi/linux/nilfs2_api.h create mode 100644 include/uapi/linux/nilfs2_ondisk.h (limited to 'fs') diff --git a/Documentation/filesystems/nilfs2.txt b/Documentation/filesystems/nilfs2.txt index 5b21ef76f751..c0727dc36271 100644 --- a/Documentation/filesystems/nilfs2.txt +++ b/Documentation/filesystems/nilfs2.txt @@ -267,7 +267,8 @@ among NILFS2 files can be depicted as follows: `-- file (ino=yy) ( regular file, directory, or symlink ) -For detail on the format of each file, please see include/linux/nilfs2_fs.h. +For detail on the format of each file, please see nilfs2_ondisk.h +located at include/uapi/linux directory. There are no patents or other intellectual property that we protect with regard to the design of NILFS2. It is allowed to replicate the diff --git a/Documentation/ioctl/ioctl-number.txt b/Documentation/ioctl/ioctl-number.txt index 56af5e43e9c0..81c7f2bb7daf 100644 --- a/Documentation/ioctl/ioctl-number.txt +++ b/Documentation/ioctl/ioctl-number.txt @@ -248,7 +248,7 @@ Code Seq#(hex) Include File Comments 'm' 00 drivers/scsi/megaraid/megaraid_ioctl.h conflict! 'm' 00-1F net/irda/irmod.h conflict! 'n' 00-7F linux/ncp_fs.h and fs/ncpfs/ioctl.c -'n' 80-8F linux/nilfs2_fs.h NILFS2 +'n' 80-8F uapi/linux/nilfs2_api.h NILFS2 'n' E0-FF linux/matroxfb.h matroxfb 'o' 00-1F fs/ocfs2/ocfs2_fs.h OCFS2 'o' 00-03 mtd/ubi-user.h conflict! (OCFS2 and UBI overlaps) diff --git a/MAINTAINERS b/MAINTAINERS index bb51bbbc9e1d..e9eacacf0f08 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -8258,8 +8258,9 @@ T: git git://github.com/konis/nilfs2.git S: Supported F: Documentation/filesystems/nilfs2.txt F: fs/nilfs2/ -F: include/linux/nilfs2_fs.h F: include/trace/events/nilfs2.h +F: include/uapi/linux/nilfs2_api.h +F: include/uapi/linux/nilfs2_ondisk.h NINJA SCSI-3 / NINJA SCSI-32Bi (16bit/CardBus) PCMCIA SCSI HOST ADAPTER DRIVER M: YOKOTA Hiroshi diff --git a/fs/nilfs2/bmap.h b/fs/nilfs2/bmap.h index b6a4c8f93ac8..2b6ffbe5997a 100644 --- a/fs/nilfs2/bmap.h +++ b/fs/nilfs2/bmap.h @@ -22,7 +22,7 @@ #include #include #include -#include +#include /* nilfs_binfo, nilfs_inode, etc */ #include "alloc.h" #include "dat.h" diff --git a/fs/nilfs2/btree.h b/fs/nilfs2/btree.h index df1a25faa83b..2184e47fa4bf 100644 --- a/fs/nilfs2/btree.h +++ b/fs/nilfs2/btree.h @@ -22,7 +22,7 @@ #include #include #include -#include +#include /* nilfs_btree_node */ #include "btnode.h" #include "bmap.h" diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 19d9f4ae8347..a15a1601e931 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -21,7 +21,6 @@ #include #include #include -#include #include "mdt.h" #include "cpfile.h" diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index 0249744ae234..6eca972f9673 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -21,7 +21,8 @@ #include #include -#include +#include /* nilfs_cpstat */ +#include /* nilfs_inode, nilfs_checkpoint */ int nilfs_cpfile_get_checkpoint(struct inode *, __u64, int, diff --git a/fs/nilfs2/dat.h b/fs/nilfs2/dat.h index abbfdabcabea..57dc6cf466d0 100644 --- a/fs/nilfs2/dat.h +++ b/fs/nilfs2/dat.h @@ -22,6 +22,7 @@ #include #include #include +#include /* nilfs_inode, nilfs_checkpoint */ struct nilfs_palloc_req; diff --git a/fs/nilfs2/dir.c b/fs/nilfs2/dir.c index 746956d2937a..908ebbf0ac7e 100644 --- a/fs/nilfs2/dir.c +++ b/fs/nilfs2/dir.c @@ -42,6 +42,28 @@ #include "nilfs.h" #include "page.h" +static inline unsigned int nilfs_rec_len_from_disk(__le16 dlen) +{ + unsigned int len = le16_to_cpu(dlen); + +#if (PAGE_SIZE >= 65536) + if (len == NILFS_MAX_REC_LEN) + return 1 << 16; +#endif + return len; +} + +static inline __le16 nilfs_rec_len_to_disk(unsigned int len) +{ +#if (PAGE_SIZE >= 65536) + if (len == (1 << 16)) + return cpu_to_le16(NILFS_MAX_REC_LEN); + + BUG_ON(len > (1 << 16)); +#endif + return cpu_to_le16(len); +} + /* * nilfs uses block-sized chunks. Arguably, sector-sized ones would be * more robust, but we have what we have diff --git a/fs/nilfs2/direct.h b/fs/nilfs2/direct.h index 3015a6e78724..cfe85e848bba 100644 --- a/fs/nilfs2/direct.h +++ b/fs/nilfs2/direct.h @@ -24,16 +24,6 @@ #include "bmap.h" -/** - * struct nilfs_direct_node - direct node - * @dn_flags: flags - * @dn_pad: padding - */ -struct nilfs_direct_node { - __u8 dn_flags; - __u8 pad[7]; -}; - #define NILFS_DIRECT_NBLOCKS (NILFS_BMAP_SIZE / sizeof(__le64) - 1) #define NILFS_DIRECT_KEY_MIN 0 #define NILFS_DIRECT_KEY_MAX (NILFS_DIRECT_NBLOCKS - 1) diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index 23ad2f091e76..188b94fe0ec5 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -23,7 +23,6 @@ #include #include -#include #include "mdt.h" #include "alloc.h" diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index 827283fe9525..f1d7989459fd 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -25,7 +25,6 @@ #include /* compat_ptr() */ #include /* mnt_want_write_file(), mnt_drop_write_file() */ #include -#include #include "nilfs.h" #include "segment.h" #include "bmap.h" diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index 2ba8a146af1f..33f8c8fc96e8 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -23,7 +23,8 @@ #include #include #include -#include +#include +#include #include "the_nilfs.h" #include "bmap.h" diff --git a/fs/nilfs2/segment.h b/fs/nilfs2/segment.h index 6565c10b7b76..1060949d7dd2 100644 --- a/fs/nilfs2/segment.h +++ b/fs/nilfs2/segment.h @@ -23,7 +23,6 @@ #include #include #include -#include #include "nilfs.h" struct nilfs_root; diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 12d11de93602..1541a1e9221a 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -22,7 +22,6 @@ #include #include #include -#include #include "mdt.h" #include "sufile.h" diff --git a/fs/nilfs2/sufile.h b/fs/nilfs2/sufile.h index 46e89872294c..158a9190c8ec 100644 --- a/fs/nilfs2/sufile.h +++ b/fs/nilfs2/sufile.h @@ -21,7 +21,6 @@ #include #include -#include #include "mdt.h" diff --git a/include/linux/nilfs2_fs.h b/include/linux/nilfs2_fs.h deleted file mode 100644 index 5988dd57ba66..000000000000 --- a/include/linux/nilfs2_fs.h +++ /dev/null @@ -1,934 +0,0 @@ -/* - * nilfs2_fs.h - NILFS2 on-disk structures and common declarations. - * - * Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU Lesser General Public License as published - * by the Free Software Foundation; either version 2.1 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU Lesser General Public License for more details. - * - * Written by Koji Sato and Ryusuke Konishi. - */ -/* - * linux/include/linux/ext2_fs.h - * - * Copyright (C) 1992, 1993, 1994, 1995 - * Remy Card (card@masi.ibp.fr) - * Laboratoire MASI - Institut Blaise Pascal - * Universite Pierre et Marie Curie (Paris VI) - * - * from - * - * linux/include/linux/minix_fs.h - * - * Copyright (C) 1991, 1992 Linus Torvalds - */ - -#ifndef _LINUX_NILFS_FS_H -#define _LINUX_NILFS_FS_H - -#include -#include -#include -#include - - -#define NILFS_INODE_BMAP_SIZE 7 -/** - * struct nilfs_inode - structure of an inode on disk - * @i_blocks: blocks count - * @i_size: size in bytes - * @i_ctime: creation time (seconds) - * @i_mtime: modification time (seconds) - * @i_ctime_nsec: creation time (nano seconds) - * @i_mtime_nsec: modification time (nano seconds) - * @i_uid: user id - * @i_gid: group id - * @i_mode: file mode - * @i_links_count: links count - * @i_flags: file flags - * @i_bmap: block mapping - * @i_xattr: extended attributes - * @i_generation: file generation (for NFS) - * @i_pad: padding - */ -struct nilfs_inode { - __le64 i_blocks; - __le64 i_size; - __le64 i_ctime; - __le64 i_mtime; - __le32 i_ctime_nsec; - __le32 i_mtime_nsec; - __le32 i_uid; - __le32 i_gid; - __le16 i_mode; - __le16 i_links_count; - __le32 i_flags; - __le64 i_bmap[NILFS_INODE_BMAP_SIZE]; -#define i_device_code i_bmap[0] - __le64 i_xattr; - __le32 i_generation; - __le32 i_pad; -}; - -#define NILFS_MIN_INODE_SIZE 128 - -/** - * struct nilfs_super_root - structure of super root - * @sr_sum: check sum - * @sr_bytes: byte count of the structure - * @sr_flags: flags (reserved) - * @sr_nongc_ctime: write time of the last segment not for cleaner operation - * @sr_dat: DAT file inode - * @sr_cpfile: checkpoint file inode - * @sr_sufile: segment usage file inode - */ -struct nilfs_super_root { - __le32 sr_sum; - __le16 sr_bytes; - __le16 sr_flags; - __le64 sr_nongc_ctime; - struct nilfs_inode sr_dat; - struct nilfs_inode sr_cpfile; - struct nilfs_inode sr_sufile; -}; - -#define NILFS_SR_MDT_OFFSET(inode_size, i) \ - ((unsigned long)&((struct nilfs_super_root *)0)->sr_dat + \ - (inode_size) * (i)) -#define NILFS_SR_DAT_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 0) -#define NILFS_SR_CPFILE_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 1) -#define NILFS_SR_SUFILE_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 2) -#define NILFS_SR_BYTES(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 3) - -/* - * Maximal mount counts - */ -#define NILFS_DFL_MAX_MNT_COUNT 50 /* 50 mounts */ - -/* - * File system states (sbp->s_state, nilfs->ns_mount_state) - */ -#define NILFS_VALID_FS 0x0001 /* Unmounted cleanly */ -#define NILFS_ERROR_FS 0x0002 /* Errors detected */ -#define NILFS_RESIZE_FS 0x0004 /* Resize required */ - -/* - * Mount flags (sbi->s_mount_opt) - */ -#define NILFS_MOUNT_ERROR_MODE 0x0070 /* Error mode mask */ -#define NILFS_MOUNT_ERRORS_CONT 0x0010 /* Continue on errors */ -#define NILFS_MOUNT_ERRORS_RO 0x0020 /* Remount fs ro on errors */ -#define NILFS_MOUNT_ERRORS_PANIC 0x0040 /* Panic on errors */ -#define NILFS_MOUNT_BARRIER 0x1000 /* Use block barriers */ -#define NILFS_MOUNT_STRICT_ORDER 0x2000 /* - * Apply strict in-order - * semantics also for data - */ -#define NILFS_MOUNT_NORECOVERY 0x4000 /* - * Disable write access during - * mount-time recovery - */ -#define NILFS_MOUNT_DISCARD 0x8000 /* Issue DISCARD requests */ - - -/** - * struct nilfs_super_block - structure of super block on disk - */ -struct nilfs_super_block { -/*00*/ __le32 s_rev_level; /* Revision level */ - __le16 s_minor_rev_level; /* minor revision level */ - __le16 s_magic; /* Magic signature */ - - __le16 s_bytes; /* - * Bytes count of CRC calculation - * for this structure. s_reserved - * is excluded. - */ - __le16 s_flags; /* flags */ - __le32 s_crc_seed; /* Seed value of CRC calculation */ -/*10*/ __le32 s_sum; /* Check sum of super block */ - - __le32 s_log_block_size; /* - * Block size represented as follows - * blocksize = - * 1 << (s_log_block_size + 10) - */ - __le64 s_nsegments; /* Number of segments in filesystem */ -/*20*/ __le64 s_dev_size; /* block device size in bytes */ - __le64 s_first_data_block; /* 1st seg disk block number */ -/*30*/ __le32 s_blocks_per_segment; /* number of blocks per full segment */ - __le32 s_r_segments_percentage; /* Reserved segments percentage */ - - __le64 s_last_cno; /* Last checkpoint number */ -/*40*/ __le64 s_last_pseg; /* disk block addr pseg written last */ - __le64 s_last_seq; /* seq. number of seg written last */ -/*50*/ __le64 s_free_blocks_count; /* Free blocks count */ - - __le64 s_ctime; /* - * Creation time (execution time of - * newfs) - */ -/*60*/ __le64 s_mtime; /* Mount time */ - __le64 s_wtime; /* Write time */ -/*70*/ __le16 s_mnt_count; /* Mount count */ - __le16 s_max_mnt_count; /* Maximal mount count */ - __le16 s_state; /* File system state */ - __le16 s_errors; /* Behaviour when detecting errors */ - __le64 s_lastcheck; /* time of last check */ - -/*80*/ __le32 s_checkinterval; /* max. time between checks */ - __le32 s_creator_os; /* OS */ - __le16 s_def_resuid; /* Default uid for reserved blocks */ - __le16 s_def_resgid; /* Default gid for reserved blocks */ - __le32 s_first_ino; /* First non-reserved inode */ - -/*90*/ __le16 s_inode_size; /* Size of an inode */ - __le16 s_dat_entry_size; /* Size of a dat entry */ - __le16 s_checkpoint_size; /* Size of a checkpoint */ - __le16 s_segment_usage_size; /* Size of a segment usage */ - -/*98*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */ -/*A8*/ char s_volume_name[80]; /* volume name */ - -/*F8*/ __le32 s_c_interval; /* Commit interval of segment */ - __le32 s_c_block_max; /* - * Threshold of data amount for - * the segment construction - */ -/*100*/ __le64 s_feature_compat; /* Compatible feature set */ - __le64 s_feature_compat_ro; /* Read-only compatible feature set */ - __le64 s_feature_incompat; /* Incompatible feature set */ - __u32 s_reserved[186]; /* padding to the end of the block */ -}; - -/* - * Codes for operating systems - */ -#define NILFS_OS_LINUX 0 -/* Codes from 1 to 4 are reserved to keep compatibility with ext2 creator-OS */ - -/* - * Revision levels - */ -#define NILFS_CURRENT_REV 2 /* current major revision */ -#define NILFS_MINOR_REV 0 /* minor revision */ -#define NILFS_MIN_SUPP_REV 2 /* minimum supported revision */ - -/* - * Feature set definitions - * - * If there is a bit set in the incompatible feature set that the kernel - * doesn't know about, it should refuse to mount the filesystem. - */ -#define NILFS_FEATURE_COMPAT_RO_BLOCK_COUNT 0x00000001ULL - -#define NILFS_FEATURE_COMPAT_SUPP 0ULL -#define NILFS_FEATURE_COMPAT_RO_SUPP NILFS_FEATURE_COMPAT_RO_BLOCK_COUNT -#define NILFS_FEATURE_INCOMPAT_SUPP 0ULL - -/* - * Bytes count of super_block for CRC-calculation - */ -#define NILFS_SB_BYTES \ - ((long)&((struct nilfs_super_block *)0)->s_reserved) - -/* - * Special inode number - */ -#define NILFS_ROOT_INO 2 /* Root file inode */ -#define NILFS_DAT_INO 3 /* DAT file */ -#define NILFS_CPFILE_INO 4 /* checkpoint file */ -#define NILFS_SUFILE_INO 5 /* segment usage file */ -#define NILFS_IFILE_INO 6 /* ifile */ -#define NILFS_ATIME_INO 7 /* Atime file (reserved) */ -#define NILFS_XATTR_INO 8 /* Xattribute file (reserved) */ -#define NILFS_SKETCH_INO 10 /* Sketch file */ -#define NILFS_USER_INO 11 /* Fisrt user's file inode number */ - -#define NILFS_SB_OFFSET_BYTES 1024 /* byte offset of nilfs superblock */ - -#define NILFS_SEG_MIN_BLOCKS 16 /* - * Minimum number of blocks in - * a full segment - */ -#define NILFS_PSEG_MIN_BLOCKS 2 /* - * Minimum number of blocks in - * a partial segment - */ -#define NILFS_MIN_NRSVSEGS 8 /* - * Minimum number of reserved - * segments - */ - -/* - * We call DAT, cpfile, and sufile root metadata files. Inodes of - * these files are written in super root block instead of ifile, and - * garbage collector doesn't keep any past versions of these files. - */ -#define NILFS_ROOT_METADATA_FILE(ino) \ - ((ino) >= NILFS_DAT_INO && (ino) <= NILFS_SUFILE_INO) - -/* - * bytes offset of secondary super block - */ -#define NILFS_SB2_OFFSET_BYTES(devsize) ((((devsize) >> 12) - 1) << 12) - -/* - * Maximal count of links to a file - */ -#define NILFS_LINK_MAX 32000 - -/* - * Structure of a directory entry - * (Same as ext2) - */ - -#define NILFS_NAME_LEN 255 - -/* - * Block size limitations - */ -#define NILFS_MIN_BLOCK_SIZE 1024 -#define NILFS_MAX_BLOCK_SIZE 65536 - -/* - * The new version of the directory entry. Since V0 structures are - * stored in intel byte order, and the name_len field could never be - * bigger than 255 chars, it's safe to reclaim the extra byte for the - * file_type field. - */ -struct nilfs_dir_entry { - __le64 inode; /* Inode number */ - __le16 rec_len; /* Directory entry length */ - __u8 name_len; /* Name length */ - __u8 file_type; /* Dir entry type (file, dir, etc) */ - char name[NILFS_NAME_LEN]; /* File name */ - char pad; -}; - -/* - * NILFS directory file types. Only the low 3 bits are used. The - * other bits are reserved for now. - */ -enum { - NILFS_FT_UNKNOWN, - NILFS_FT_REG_FILE, - NILFS_FT_DIR, - NILFS_FT_CHRDEV, - NILFS_FT_BLKDEV, - NILFS_FT_FIFO, - NILFS_FT_SOCK, - NILFS_FT_SYMLINK, - NILFS_FT_MAX -}; - -/* - * NILFS_DIR_PAD defines the directory entries boundaries - * - * NOTE: It must be a multiple of 8 - */ -#define NILFS_DIR_PAD 8 -#define NILFS_DIR_ROUND (NILFS_DIR_PAD - 1) -#define NILFS_DIR_REC_LEN(name_len) (((name_len) + 12 + NILFS_DIR_ROUND) & \ - ~NILFS_DIR_ROUND) -#define NILFS_MAX_REC_LEN ((1<<16)-1) - -static inline unsigned int nilfs_rec_len_from_disk(__le16 dlen) -{ - unsigned int len = le16_to_cpu(dlen); - -#if !defined(__KERNEL__) || (PAGE_SIZE >= 65536) - if (len == NILFS_MAX_REC_LEN) - return 1 << 16; -#endif - return len; -} - -static inline __le16 nilfs_rec_len_to_disk(unsigned int len) -{ -#if !defined(__KERNEL__) || (PAGE_SIZE >= 65536) - if (len == (1 << 16)) - return cpu_to_le16(NILFS_MAX_REC_LEN); - else if (len > (1 << 16)) - BUG(); -#endif - return cpu_to_le16(len); -} - -/** - * struct nilfs_finfo - file information - * @fi_ino: inode number - * @fi_cno: checkpoint number - * @fi_nblocks: number of blocks (including intermediate blocks) - * @fi_ndatablk: number of file data blocks - */ -struct nilfs_finfo { - __le64 fi_ino; - __le64 fi_cno; - __le32 fi_nblocks; - __le32 fi_ndatablk; - /* array of virtual block numbers */ -}; - -/** - * struct nilfs_binfo_v - information for the block to which a virtual block number is assigned - * @bi_vblocknr: virtual block number - * @bi_blkoff: block offset - */ -struct nilfs_binfo_v { - __le64 bi_vblocknr; - __le64 bi_blkoff; -}; - -/** - * struct nilfs_binfo_dat - information for the block which belongs to the DAT file - * @bi_blkoff: block offset - * @bi_level: level - * @bi_pad: padding - */ -struct nilfs_binfo_dat { - __le64 bi_blkoff; - __u8 bi_level; - __u8 bi_pad[7]; -}; - -/** - * union nilfs_binfo: block information - * @bi_v: nilfs_binfo_v structure - * @bi_dat: nilfs_binfo_dat structure - */ -union nilfs_binfo { - struct nilfs_binfo_v bi_v; - struct nilfs_binfo_dat bi_dat; -}; - -/** - * struct nilfs_segment_summary - segment summary header - * @ss_datasum: checksum of data - * @ss_sumsum: checksum of segment summary - * @ss_magic: magic number - * @ss_bytes: size of this structure in bytes - * @ss_flags: flags - * @ss_seq: sequence number - * @ss_create: creation timestamp - * @ss_next: next segment - * @ss_nblocks: number of blocks - * @ss_nfinfo: number of finfo structures - * @ss_sumbytes: total size of segment summary in bytes - * @ss_pad: padding - * @ss_cno: checkpoint number - */ -struct nilfs_segment_summary { - __le32 ss_datasum; - __le32 ss_sumsum; - __le32 ss_magic; - __le16 ss_bytes; - __le16 ss_flags; - __le64 ss_seq; - __le64 ss_create; - __le64 ss_next; - __le32 ss_nblocks; - __le32 ss_nfinfo; - __le32 ss_sumbytes; - __le32 ss_pad; - __le64 ss_cno; - /* array of finfo structures */ -}; - -#define NILFS_SEGSUM_MAGIC 0x1eaffa11 /* segment summary magic number */ - -/* - * Segment summary flags - */ -#define NILFS_SS_LOGBGN 0x0001 /* begins a logical segment */ -#define NILFS_SS_LOGEND 0x0002 /* ends a logical segment */ -#define NILFS_SS_SR 0x0004 /* has super root */ -#define NILFS_SS_SYNDT 0x0008 /* includes data only updates */ -#define NILFS_SS_GC 0x0010 /* segment written for cleaner operation */ - -/** - * struct nilfs_btree_node - B-tree node - * @bn_flags: flags - * @bn_level: level - * @bn_nchildren: number of children - * @bn_pad: padding - */ -struct nilfs_btree_node { - __u8 bn_flags; - __u8 bn_level; - __le16 bn_nchildren; - __le32 bn_pad; -}; - -/* flags */ -#define NILFS_BTREE_NODE_ROOT 0x01 - -/* level */ -#define NILFS_BTREE_LEVEL_DATA 0 -#define NILFS_BTREE_LEVEL_NODE_MIN (NILFS_BTREE_LEVEL_DATA + 1) -#define NILFS_BTREE_LEVEL_MAX 14 /* Max level (exclusive) */ - -/** - * struct nilfs_palloc_group_desc - block group descriptor - * @pg_nfrees: number of free entries in block group - */ -struct nilfs_palloc_group_desc { - __le32 pg_nfrees; -}; - -/** - * struct nilfs_dat_entry - disk address translation entry - * @de_blocknr: block number - * @de_start: start checkpoint number - * @de_end: end checkpoint number - * @de_rsv: reserved for future use - */ -struct nilfs_dat_entry { - __le64 de_blocknr; - __le64 de_start; - __le64 de_end; - __le64 de_rsv; -}; - -#define NILFS_MIN_DAT_ENTRY_SIZE 32 - -/** - * struct nilfs_snapshot_list - snapshot list - * @ssl_next: next checkpoint number on snapshot list - * @ssl_prev: previous checkpoint number on snapshot list - */ -struct nilfs_snapshot_list { - __le64 ssl_next; - __le64 ssl_prev; -}; - -/** - * struct nilfs_checkpoint - checkpoint structure - * @cp_flags: flags - * @cp_checkpoints_count: checkpoints count in a block - * @cp_snapshot_list: snapshot list - * @cp_cno: checkpoint number - * @cp_create: creation timestamp - * @cp_nblk_inc: number of blocks incremented by this checkpoint - * @cp_inodes_count: inodes count - * @cp_blocks_count: blocks count - * @cp_ifile_inode: inode of ifile - */ -struct nilfs_checkpoint { - __le32 cp_flags; - __le32 cp_checkpoints_count; - struct nilfs_snapshot_list cp_snapshot_list; - __le64 cp_cno; - __le64 cp_create; - __le64 cp_nblk_inc; - __le64 cp_inodes_count; - __le64 cp_blocks_count; - - /* - * Do not change the byte offset of ifile inode. - * To keep the compatibility of the disk format, - * additional fields should be added behind cp_ifile_inode. - */ - struct nilfs_inode cp_ifile_inode; -}; - -#define NILFS_MIN_CHECKPOINT_SIZE (64 + NILFS_MIN_INODE_SIZE) - -/* checkpoint flags */ -enum { - NILFS_CHECKPOINT_SNAPSHOT, - NILFS_CHECKPOINT_INVALID, - NILFS_CHECKPOINT_SKETCH, - NILFS_CHECKPOINT_MINOR, -}; - -#define NILFS_CHECKPOINT_FNS(flag, name) \ -static inline void \ -nilfs_checkpoint_set_##name(struct nilfs_checkpoint *cp) \ -{ \ - cp->cp_flags = cpu_to_le32(le32_to_cpu(cp->cp_flags) | \ - (1UL << NILFS_CHECKPOINT_##flag)); \ -} \ -static inline void \ -nilfs_checkpoint_clear_##name(struct nilfs_checkpoint *cp) \ -{ \ - cp->cp_flags = cpu_to_le32(le32_to_cpu(cp->cp_flags) & \ - ~(1UL << NILFS_CHECKPOINT_##flag)); \ -} \ -static inline int \ -nilfs_checkpoint_##name(const struct nilfs_checkpoint *cp) \ -{ \ - return !!(le32_to_cpu(cp->cp_flags) & \ - (1UL << NILFS_CHECKPOINT_##flag)); \ -} - -NILFS_CHECKPOINT_FNS(SNAPSHOT, snapshot) -NILFS_CHECKPOINT_FNS(INVALID, invalid) -NILFS_CHECKPOINT_FNS(MINOR, minor) - -/** - * struct nilfs_cpinfo - checkpoint information - * @ci_flags: flags - * @ci_pad: padding - * @ci_cno: checkpoint number - * @ci_create: creation timestamp - * @ci_nblk_inc: number of blocks incremented by this checkpoint - * @ci_inodes_count: inodes count - * @ci_blocks_count: blocks count - * @ci_next: next checkpoint number in snapshot list - */ -struct nilfs_cpinfo { - __u32 ci_flags; - __u32 ci_pad; - __u64 ci_cno; - __u64 ci_create; - __u64 ci_nblk_inc; - __u64 ci_inodes_count; - __u64 ci_blocks_count; - __u64 ci_next; -}; - -#define NILFS_CPINFO_FNS(flag, name) \ -static inline int \ -nilfs_cpinfo_##name(const struct nilfs_cpinfo *cpinfo) \ -{ \ - return !!(cpinfo->ci_flags & (1UL << NILFS_CHECKPOINT_##flag)); \ -} - -NILFS_CPINFO_FNS(SNAPSHOT, snapshot) -NILFS_CPINFO_FNS(INVALID, invalid) -NILFS_CPINFO_FNS(MINOR, minor) - - -/** - * struct nilfs_cpfile_header - checkpoint file header - * @ch_ncheckpoints: number of checkpoints - * @ch_nsnapshots: number of snapshots - * @ch_snapshot_list: snapshot list - */ -struct nilfs_cpfile_header { - __le64 ch_ncheckpoints; - __le64 ch_nsnapshots; - struct nilfs_snapshot_list ch_snapshot_list; -}; - -#define NILFS_CPFILE_FIRST_CHECKPOINT_OFFSET \ - ((sizeof(struct nilfs_cpfile_header) + \ - sizeof(struct nilfs_checkpoint) - 1) / \ - sizeof(struct nilfs_checkpoint)) - -/** - * struct nilfs_segment_usage - segment usage - * @su_lastmod: last modified timestamp - * @su_nblocks: number of blocks in segment - * @su_flags: flags - */ -struct nilfs_segment_usage { - __le64 su_lastmod; - __le32 su_nblocks; - __le32 su_flags; -}; - -#define NILFS_MIN_SEGMENT_USAGE_SIZE 16 - -/* segment usage flag */ -enum { - NILFS_SEGMENT_USAGE_ACTIVE, - NILFS_SEGMENT_USAGE_DIRTY, - NILFS_SEGMENT_USAGE_ERROR, - - /* ... */ -}; - -#define NILFS_SEGMENT_USAGE_FNS(flag, name) \ -static inline void \ -nilfs_segment_usage_set_##name(struct nilfs_segment_usage *su) \ -{ \ - su->su_flags = cpu_to_le32(le32_to_cpu(su->su_flags) | \ - (1UL << NILFS_SEGMENT_USAGE_##flag));\ -} \ -static inline void \ -nilfs_segment_usage_clear_##name(struct nilfs_segment_usage *su) \ -{ \ - su->su_flags = \ - cpu_to_le32(le32_to_cpu(su->su_flags) & \ - ~(1UL << NILFS_SEGMENT_USAGE_##flag)); \ -} \ -static inline int \ -nilfs_segment_usage_##name(const struct nilfs_segment_usage *su) \ -{ \ - return !!(le32_to_cpu(su->su_flags) & \ - (1UL << NILFS_SEGMENT_USAGE_##flag)); \ -} - -NILFS_SEGMENT_USAGE_FNS(ACTIVE, active) -NILFS_SEGMENT_USAGE_FNS(DIRTY, dirty) -NILFS_SEGMENT_USAGE_FNS(ERROR, error) - -static inline void -nilfs_segment_usage_set_clean(struct nilfs_segment_usage *su) -{ - su->su_lastmod = cpu_to_le64(0); - su->su_nblocks = cpu_to_le32(0); - su->su_flags = cpu_to_le32(0); -} - -static inline int -nilfs_segment_usage_clean(const struct nilfs_segment_usage *su) -{ - return !le32_to_cpu(su->su_flags); -} - -/** - * struct nilfs_sufile_header - segment usage file header - * @sh_ncleansegs: number of clean segments - * @sh_ndirtysegs: number of dirty segments - * @sh_last_alloc: last allocated segment number - */ -struct nilfs_sufile_header { - __le64 sh_ncleansegs; - __le64 sh_ndirtysegs; - __le64 sh_last_alloc; - /* ... */ -}; - -#define NILFS_SUFILE_FIRST_SEGMENT_USAGE_OFFSET \ - ((sizeof(struct nilfs_sufile_header) + \ - sizeof(struct nilfs_segment_usage) - 1) / \ - sizeof(struct nilfs_segment_usage)) - -/** - * nilfs_suinfo - segment usage information - * @sui_lastmod: timestamp of last modification - * @sui_nblocks: number of written blocks in segment - * @sui_flags: segment usage flags - */ -struct nilfs_suinfo { - __u64 sui_lastmod; - __u32 sui_nblocks; - __u32 sui_flags; -}; - -#define NILFS_SUINFO_FNS(flag, name) \ -static inline int \ -nilfs_suinfo_##name(const struct nilfs_suinfo *si) \ -{ \ - return si->sui_flags & (1UL << NILFS_SEGMENT_USAGE_##flag); \ -} - -NILFS_SUINFO_FNS(ACTIVE, active) -NILFS_SUINFO_FNS(DIRTY, dirty) -NILFS_SUINFO_FNS(ERROR, error) - -static inline int nilfs_suinfo_clean(const struct nilfs_suinfo *si) -{ - return !si->sui_flags; -} - -/* ioctl */ -/** - * nilfs_suinfo_update - segment usage information update - * @sup_segnum: segment number - * @sup_flags: flags for which fields are active in sup_sui - * @sup_reserved: reserved necessary for alignment - * @sup_sui: segment usage information - */ -struct nilfs_suinfo_update { - __u64 sup_segnum; - __u32 sup_flags; - __u32 sup_reserved; - struct nilfs_suinfo sup_sui; -}; - -enum { - NILFS_SUINFO_UPDATE_LASTMOD, - NILFS_SUINFO_UPDATE_NBLOCKS, - NILFS_SUINFO_UPDATE_FLAGS, - __NR_NILFS_SUINFO_UPDATE_FIELDS, -}; - -#define NILFS_SUINFO_UPDATE_FNS(flag, name) \ -static inline void \ -nilfs_suinfo_update_set_##name(struct nilfs_suinfo_update *sup) \ -{ \ - sup->sup_flags |= 1UL << NILFS_SUINFO_UPDATE_##flag; \ -} \ -static inline void \ -nilfs_suinfo_update_clear_##name(struct nilfs_suinfo_update *sup) \ -{ \ - sup->sup_flags &= ~(1UL << NILFS_SUINFO_UPDATE_##flag); \ -} \ -static inline int \ -nilfs_suinfo_update_##name(const struct nilfs_suinfo_update *sup) \ -{ \ - return !!(sup->sup_flags & (1UL << NILFS_SUINFO_UPDATE_##flag));\ -} - -NILFS_SUINFO_UPDATE_FNS(LASTMOD, lastmod) -NILFS_SUINFO_UPDATE_FNS(NBLOCKS, nblocks) -NILFS_SUINFO_UPDATE_FNS(FLAGS, flags) - -enum { - NILFS_CHECKPOINT, - NILFS_SNAPSHOT, -}; - -/** - * struct nilfs_cpmode - change checkpoint mode structure - * @cm_cno: checkpoint number - * @cm_mode: mode of checkpoint - * @cm_pad: padding - */ -struct nilfs_cpmode { - __u64 cm_cno; - __u32 cm_mode; - __u32 cm_pad; -}; - -/** - * struct nilfs_argv - argument vector - * @v_base: pointer on data array from userspace - * @v_nmembs: number of members in data array - * @v_size: size of data array in bytes - * @v_flags: flags - * @v_index: start number of target data items - */ -struct nilfs_argv { - __u64 v_base; - __u32 v_nmembs; /* number of members */ - __u16 v_size; /* size of members */ - __u16 v_flags; - __u64 v_index; -}; - -/** - * struct nilfs_period - period of checkpoint numbers - * @p_start: start checkpoint number (inclusive) - * @p_end: end checkpoint number (exclusive) - */ -struct nilfs_period { - __u64 p_start; - __u64 p_end; -}; - -/** - * struct nilfs_cpstat - checkpoint statistics - * @cs_cno: checkpoint number - * @cs_ncps: number of checkpoints - * @cs_nsss: number of snapshots - */ -struct nilfs_cpstat { - __u64 cs_cno; - __u64 cs_ncps; - __u64 cs_nsss; -}; - -/** - * struct nilfs_sustat - segment usage statistics - * @ss_nsegs: number of segments - * @ss_ncleansegs: number of clean segments - * @ss_ndirtysegs: number of dirty segments - * @ss_ctime: creation time of the last segment - * @ss_nongc_ctime: creation time of the last segment not for GC - * @ss_prot_seq: least sequence number of segments which must not be reclaimed - */ -struct nilfs_sustat { - __u64 ss_nsegs; - __u64 ss_ncleansegs; - __u64 ss_ndirtysegs; - __u64 ss_ctime; - __u64 ss_nongc_ctime; - __u64 ss_prot_seq; -}; - -/** - * struct nilfs_vinfo - virtual block number information - * @vi_vblocknr: virtual block number - * @vi_start: start checkpoint number (inclusive) - * @vi_end: end checkpoint number (exclusive) - * @vi_blocknr: disk block number - */ -struct nilfs_vinfo { - __u64 vi_vblocknr; - __u64 vi_start; - __u64 vi_end; - __u64 vi_blocknr; -}; - -/** - * struct nilfs_vdesc - descriptor of virtual block number - * @vd_ino: inode number - * @vd_cno: checkpoint number - * @vd_vblocknr: virtual block number - * @vd_period: period of checkpoint numbers - * @vd_blocknr: disk block number - * @vd_offset: logical block offset inside a file - * @vd_flags: flags (data or node block) - * @vd_pad: padding - */ -struct nilfs_vdesc { - __u64 vd_ino; - __u64 vd_cno; - __u64 vd_vblocknr; - struct nilfs_period vd_period; - __u64 vd_blocknr; - __u64 vd_offset; - __u32 vd_flags; - __u32 vd_pad; -}; - -/** - * struct nilfs_bdesc - descriptor of disk block number - * @bd_ino: inode number - * @bd_oblocknr: disk block address (for skipping dead blocks) - * @bd_blocknr: disk block address - * @bd_offset: logical block offset inside a file - * @bd_level: level in the b-tree organization - * @bd_pad: padding - */ -struct nilfs_bdesc { - __u64 bd_ino; - __u64 bd_oblocknr; - __u64 bd_blocknr; - __u64 bd_offset; - __u32 bd_level; - __u32 bd_pad; -}; - -#define NILFS_IOCTL_IDENT 'n' - -#define NILFS_IOCTL_CHANGE_CPMODE \ - _IOW(NILFS_IOCTL_IDENT, 0x80, struct nilfs_cpmode) -#define NILFS_IOCTL_DELETE_CHECKPOINT \ - _IOW(NILFS_IOCTL_IDENT, 0x81, __u64) -#define NILFS_IOCTL_GET_CPINFO \ - _IOR(NILFS_IOCTL_IDENT, 0x82, struct nilfs_argv) -#define NILFS_IOCTL_GET_CPSTAT \ - _IOR(NILFS_IOCTL_IDENT, 0x83, struct nilfs_cpstat) -#define NILFS_IOCTL_GET_SUINFO \ - _IOR(NILFS_IOCTL_IDENT, 0x84, struct nilfs_argv) -#define NILFS_IOCTL_GET_SUSTAT \ - _IOR(NILFS_IOCTL_IDENT, 0x85, struct nilfs_sustat) -#define NILFS_IOCTL_GET_VINFO \ - _IOWR(NILFS_IOCTL_IDENT, 0x86, struct nilfs_argv) -#define NILFS_IOCTL_GET_BDESCS \ - _IOWR(NILFS_IOCTL_IDENT, 0x87, struct nilfs_argv) -#define NILFS_IOCTL_CLEAN_SEGMENTS \ - _IOW(NILFS_IOCTL_IDENT, 0x88, struct nilfs_argv[5]) -#define NILFS_IOCTL_SYNC \ - _IOR(NILFS_IOCTL_IDENT, 0x8A, __u64) -#define NILFS_IOCTL_RESIZE \ - _IOW(NILFS_IOCTL_IDENT, 0x8B, __u64) -#define NILFS_IOCTL_SET_ALLOC_RANGE \ - _IOW(NILFS_IOCTL_IDENT, 0x8C, __u64[2]) -#define NILFS_IOCTL_SET_SUINFO \ - _IOW(NILFS_IOCTL_IDENT, 0x8D, struct nilfs_argv) - -#endif /* _LINUX_NILFS_FS_H */ diff --git a/include/uapi/linux/nilfs2_api.h b/include/uapi/linux/nilfs2_api.h new file mode 100644 index 000000000000..ef4c1de89b11 --- /dev/null +++ b/include/uapi/linux/nilfs2_api.h @@ -0,0 +1,292 @@ +/* + * nilfs2_api.h - NILFS2 user space API + * + * Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License as published + * by the Free Software Foundation; either version 2.1 of the License, or + * (at your option) any later version. + */ + +#ifndef _LINUX_NILFS2_API_H +#define _LINUX_NILFS2_API_H + +#include +#include + +/** + * struct nilfs_cpinfo - checkpoint information + * @ci_flags: flags + * @ci_pad: padding + * @ci_cno: checkpoint number + * @ci_create: creation timestamp + * @ci_nblk_inc: number of blocks incremented by this checkpoint + * @ci_inodes_count: inodes count + * @ci_blocks_count: blocks count + * @ci_next: next checkpoint number in snapshot list + */ +struct nilfs_cpinfo { + __u32 ci_flags; + __u32 ci_pad; + __u64 ci_cno; + __u64 ci_create; + __u64 ci_nblk_inc; + __u64 ci_inodes_count; + __u64 ci_blocks_count; + __u64 ci_next; +}; + +/* checkpoint flags */ +enum { + NILFS_CPINFO_SNAPSHOT, + NILFS_CPINFO_INVALID, + NILFS_CPINFO_SKETCH, + NILFS_CPINFO_MINOR, +}; + +#define NILFS_CPINFO_FNS(flag, name) \ +static inline int \ +nilfs_cpinfo_##name(const struct nilfs_cpinfo *cpinfo) \ +{ \ + return !!(cpinfo->ci_flags & (1UL << NILFS_CPINFO_##flag)); \ +} + +NILFS_CPINFO_FNS(SNAPSHOT, snapshot) +NILFS_CPINFO_FNS(INVALID, invalid) +NILFS_CPINFO_FNS(MINOR, minor) + +/** + * nilfs_suinfo - segment usage information + * @sui_lastmod: timestamp of last modification + * @sui_nblocks: number of written blocks in segment + * @sui_flags: segment usage flags + */ +struct nilfs_suinfo { + __u64 sui_lastmod; + __u32 sui_nblocks; + __u32 sui_flags; +}; + +/* segment usage flags */ +enum { + NILFS_SUINFO_ACTIVE, + NILFS_SUINFO_DIRTY, + NILFS_SUINFO_ERROR, +}; + +#define NILFS_SUINFO_FNS(flag, name) \ +static inline int \ +nilfs_suinfo_##name(const struct nilfs_suinfo *si) \ +{ \ + return si->sui_flags & (1UL << NILFS_SUINFO_##flag); \ +} + +NILFS_SUINFO_FNS(ACTIVE, active) +NILFS_SUINFO_FNS(DIRTY, dirty) +NILFS_SUINFO_FNS(ERROR, error) + +static inline int nilfs_suinfo_clean(const struct nilfs_suinfo *si) +{ + return !si->sui_flags; +} + +/** + * nilfs_suinfo_update - segment usage information update + * @sup_segnum: segment number + * @sup_flags: flags for which fields are active in sup_sui + * @sup_reserved: reserved necessary for alignment + * @sup_sui: segment usage information + */ +struct nilfs_suinfo_update { + __u64 sup_segnum; + __u32 sup_flags; + __u32 sup_reserved; + struct nilfs_suinfo sup_sui; +}; + +enum { + NILFS_SUINFO_UPDATE_LASTMOD, + NILFS_SUINFO_UPDATE_NBLOCKS, + NILFS_SUINFO_UPDATE_FLAGS, + __NR_NILFS_SUINFO_UPDATE_FIELDS, +}; + +#define NILFS_SUINFO_UPDATE_FNS(flag, name) \ +static inline void \ +nilfs_suinfo_update_set_##name(struct nilfs_suinfo_update *sup) \ +{ \ + sup->sup_flags |= 1UL << NILFS_SUINFO_UPDATE_##flag; \ +} \ +static inline void \ +nilfs_suinfo_update_clear_##name(struct nilfs_suinfo_update *sup) \ +{ \ + sup->sup_flags &= ~(1UL << NILFS_SUINFO_UPDATE_##flag); \ +} \ +static inline int \ +nilfs_suinfo_update_##name(const struct nilfs_suinfo_update *sup) \ +{ \ + return !!(sup->sup_flags & (1UL << NILFS_SUINFO_UPDATE_##flag));\ +} + +NILFS_SUINFO_UPDATE_FNS(LASTMOD, lastmod) +NILFS_SUINFO_UPDATE_FNS(NBLOCKS, nblocks) +NILFS_SUINFO_UPDATE_FNS(FLAGS, flags) + +enum { + NILFS_CHECKPOINT, + NILFS_SNAPSHOT, +}; + +/** + * struct nilfs_cpmode - change checkpoint mode structure + * @cm_cno: checkpoint number + * @cm_mode: mode of checkpoint + * @cm_pad: padding + */ +struct nilfs_cpmode { + __u64 cm_cno; + __u32 cm_mode; + __u32 cm_pad; +}; + +/** + * struct nilfs_argv - argument vector + * @v_base: pointer on data array from userspace + * @v_nmembs: number of members in data array + * @v_size: size of data array in bytes + * @v_flags: flags + * @v_index: start number of target data items + */ +struct nilfs_argv { + __u64 v_base; + __u32 v_nmembs; /* number of members */ + __u16 v_size; /* size of members */ + __u16 v_flags; + __u64 v_index; +}; + +/** + * struct nilfs_period - period of checkpoint numbers + * @p_start: start checkpoint number (inclusive) + * @p_end: end checkpoint number (exclusive) + */ +struct nilfs_period { + __u64 p_start; + __u64 p_end; +}; + +/** + * struct nilfs_cpstat - checkpoint statistics + * @cs_cno: checkpoint number + * @cs_ncps: number of checkpoints + * @cs_nsss: number of snapshots + */ +struct nilfs_cpstat { + __u64 cs_cno; + __u64 cs_ncps; + __u64 cs_nsss; +}; + +/** + * struct nilfs_sustat - segment usage statistics + * @ss_nsegs: number of segments + * @ss_ncleansegs: number of clean segments + * @ss_ndirtysegs: number of dirty segments + * @ss_ctime: creation time of the last segment + * @ss_nongc_ctime: creation time of the last segment not for GC + * @ss_prot_seq: least sequence number of segments which must not be reclaimed + */ +struct nilfs_sustat { + __u64 ss_nsegs; + __u64 ss_ncleansegs; + __u64 ss_ndirtysegs; + __u64 ss_ctime; + __u64 ss_nongc_ctime; + __u64 ss_prot_seq; +}; + +/** + * struct nilfs_vinfo - virtual block number information + * @vi_vblocknr: virtual block number + * @vi_start: start checkpoint number (inclusive) + * @vi_end: end checkpoint number (exclusive) + * @vi_blocknr: disk block number + */ +struct nilfs_vinfo { + __u64 vi_vblocknr; + __u64 vi_start; + __u64 vi_end; + __u64 vi_blocknr; +}; + +/** + * struct nilfs_vdesc - descriptor of virtual block number + * @vd_ino: inode number + * @vd_cno: checkpoint number + * @vd_vblocknr: virtual block number + * @vd_period: period of checkpoint numbers + * @vd_blocknr: disk block number + * @vd_offset: logical block offset inside a file + * @vd_flags: flags (data or node block) + * @vd_pad: padding + */ +struct nilfs_vdesc { + __u64 vd_ino; + __u64 vd_cno; + __u64 vd_vblocknr; + struct nilfs_period vd_period; + __u64 vd_blocknr; + __u64 vd_offset; + __u32 vd_flags; + __u32 vd_pad; +}; + +/** + * struct nilfs_bdesc - descriptor of disk block number + * @bd_ino: inode number + * @bd_oblocknr: disk block address (for skipping dead blocks) + * @bd_blocknr: disk block address + * @bd_offset: logical block offset inside a file + * @bd_level: level in the b-tree organization + * @bd_pad: padding + */ +struct nilfs_bdesc { + __u64 bd_ino; + __u64 bd_oblocknr; + __u64 bd_blocknr; + __u64 bd_offset; + __u32 bd_level; + __u32 bd_pad; +}; + +#define NILFS_IOCTL_IDENT 'n' + +#define NILFS_IOCTL_CHANGE_CPMODE \ + _IOW(NILFS_IOCTL_IDENT, 0x80, struct nilfs_cpmode) +#define NILFS_IOCTL_DELETE_CHECKPOINT \ + _IOW(NILFS_IOCTL_IDENT, 0x81, __u64) +#define NILFS_IOCTL_GET_CPINFO \ + _IOR(NILFS_IOCTL_IDENT, 0x82, struct nilfs_argv) +#define NILFS_IOCTL_GET_CPSTAT \ + _IOR(NILFS_IOCTL_IDENT, 0x83, struct nilfs_cpstat) +#define NILFS_IOCTL_GET_SUINFO \ + _IOR(NILFS_IOCTL_IDENT, 0x84, struct nilfs_argv) +#define NILFS_IOCTL_GET_SUSTAT \ + _IOR(NILFS_IOCTL_IDENT, 0x85, struct nilfs_sustat) +#define NILFS_IOCTL_GET_VINFO \ + _IOWR(NILFS_IOCTL_IDENT, 0x86, struct nilfs_argv) +#define NILFS_IOCTL_GET_BDESCS \ + _IOWR(NILFS_IOCTL_IDENT, 0x87, struct nilfs_argv) +#define NILFS_IOCTL_CLEAN_SEGMENTS \ + _IOW(NILFS_IOCTL_IDENT, 0x88, struct nilfs_argv[5]) +#define NILFS_IOCTL_SYNC \ + _IOR(NILFS_IOCTL_IDENT, 0x8A, __u64) +#define NILFS_IOCTL_RESIZE \ + _IOW(NILFS_IOCTL_IDENT, 0x8B, __u64) +#define NILFS_IOCTL_SET_ALLOC_RANGE \ + _IOW(NILFS_IOCTL_IDENT, 0x8C, __u64[2]) +#define NILFS_IOCTL_SET_SUINFO \ + _IOW(NILFS_IOCTL_IDENT, 0x8D, struct nilfs_argv) + +#endif /* _LINUX_NILFS2_API_H */ diff --git a/include/uapi/linux/nilfs2_ondisk.h b/include/uapi/linux/nilfs2_ondisk.h new file mode 100644 index 000000000000..2a8a3addb675 --- /dev/null +++ b/include/uapi/linux/nilfs2_ondisk.h @@ -0,0 +1,650 @@ +/* + * nilfs2_ondisk.h - NILFS2 on-disk structures + * + * Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License as published + * by the Free Software Foundation; either version 2.1 of the License, or + * (at your option) any later version. + */ +/* + * linux/include/linux/ext2_fs.h + * + * Copyright (C) 1992, 1993, 1994, 1995 + * Remy Card (card@masi.ibp.fr) + * Laboratoire MASI - Institut Blaise Pascal + * Universite Pierre et Marie Curie (Paris VI) + * + * from + * + * linux/include/linux/minix_fs.h + * + * Copyright (C) 1991, 1992 Linus Torvalds + */ + +#ifndef _LINUX_NILFS2_ONDISK_H +#define _LINUX_NILFS2_ONDISK_H + +#include +#include + + +#define NILFS_INODE_BMAP_SIZE 7 + +/** + * struct nilfs_inode - structure of an inode on disk + * @i_blocks: blocks count + * @i_size: size in bytes + * @i_ctime: creation time (seconds) + * @i_mtime: modification time (seconds) + * @i_ctime_nsec: creation time (nano seconds) + * @i_mtime_nsec: modification time (nano seconds) + * @i_uid: user id + * @i_gid: group id + * @i_mode: file mode + * @i_links_count: links count + * @i_flags: file flags + * @i_bmap: block mapping + * @i_xattr: extended attributes + * @i_generation: file generation (for NFS) + * @i_pad: padding + */ +struct nilfs_inode { + __le64 i_blocks; + __le64 i_size; + __le64 i_ctime; + __le64 i_mtime; + __le32 i_ctime_nsec; + __le32 i_mtime_nsec; + __le32 i_uid; + __le32 i_gid; + __le16 i_mode; + __le16 i_links_count; + __le32 i_flags; + __le64 i_bmap[NILFS_INODE_BMAP_SIZE]; +#define i_device_code i_bmap[0] + __le64 i_xattr; + __le32 i_generation; + __le32 i_pad; +}; + +#define NILFS_MIN_INODE_SIZE 128 + +/** + * struct nilfs_super_root - structure of super root + * @sr_sum: check sum + * @sr_bytes: byte count of the structure + * @sr_flags: flags (reserved) + * @sr_nongc_ctime: write time of the last segment not for cleaner operation + * @sr_dat: DAT file inode + * @sr_cpfile: checkpoint file inode + * @sr_sufile: segment usage file inode + */ +struct nilfs_super_root { + __le32 sr_sum; + __le16 sr_bytes; + __le16 sr_flags; + __le64 sr_nongc_ctime; + struct nilfs_inode sr_dat; + struct nilfs_inode sr_cpfile; + struct nilfs_inode sr_sufile; +}; + +#define NILFS_SR_MDT_OFFSET(inode_size, i) \ + ((unsigned long)&((struct nilfs_super_root *)0)->sr_dat + \ + (inode_size) * (i)) +#define NILFS_SR_DAT_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 0) +#define NILFS_SR_CPFILE_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 1) +#define NILFS_SR_SUFILE_OFFSET(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 2) +#define NILFS_SR_BYTES(inode_size) NILFS_SR_MDT_OFFSET(inode_size, 3) + +/* + * Maximal mount counts + */ +#define NILFS_DFL_MAX_MNT_COUNT 50 /* 50 mounts */ + +/* + * File system states (sbp->s_state, nilfs->ns_mount_state) + */ +#define NILFS_VALID_FS 0x0001 /* Unmounted cleanly */ +#define NILFS_ERROR_FS 0x0002 /* Errors detected */ +#define NILFS_RESIZE_FS 0x0004 /* Resize required */ + +/* + * Mount flags (sbi->s_mount_opt) + */ +#define NILFS_MOUNT_ERROR_MODE 0x0070 /* Error mode mask */ +#define NILFS_MOUNT_ERRORS_CONT 0x0010 /* Continue on errors */ +#define NILFS_MOUNT_ERRORS_RO 0x0020 /* Remount fs ro on errors */ +#define NILFS_MOUNT_ERRORS_PANIC 0x0040 /* Panic on errors */ +#define NILFS_MOUNT_BARRIER 0x1000 /* Use block barriers */ +#define NILFS_MOUNT_STRICT_ORDER 0x2000 /* + * Apply strict in-order + * semantics also for data + */ +#define NILFS_MOUNT_NORECOVERY 0x4000 /* + * Disable write access during + * mount-time recovery + */ +#define NILFS_MOUNT_DISCARD 0x8000 /* Issue DISCARD requests */ + + +/** + * struct nilfs_super_block - structure of super block on disk + */ +struct nilfs_super_block { +/*00*/ __le32 s_rev_level; /* Revision level */ + __le16 s_minor_rev_level; /* minor revision level */ + __le16 s_magic; /* Magic signature */ + + __le16 s_bytes; /* + * Bytes count of CRC calculation + * for this structure. s_reserved + * is excluded. + */ + __le16 s_flags; /* flags */ + __le32 s_crc_seed; /* Seed value of CRC calculation */ +/*10*/ __le32 s_sum; /* Check sum of super block */ + + __le32 s_log_block_size; /* + * Block size represented as follows + * blocksize = + * 1 << (s_log_block_size + 10) + */ + __le64 s_nsegments; /* Number of segments in filesystem */ +/*20*/ __le64 s_dev_size; /* block device size in bytes */ + __le64 s_first_data_block; /* 1st seg disk block number */ +/*30*/ __le32 s_blocks_per_segment; /* number of blocks per full segment */ + __le32 s_r_segments_percentage; /* Reserved segments percentage */ + + __le64 s_last_cno; /* Last checkpoint number */ +/*40*/ __le64 s_last_pseg; /* disk block addr pseg written last */ + __le64 s_last_seq; /* seq. number of seg written last */ +/*50*/ __le64 s_free_blocks_count; /* Free blocks count */ + + __le64 s_ctime; /* + * Creation time (execution time of + * newfs) + */ +/*60*/ __le64 s_mtime; /* Mount time */ + __le64 s_wtime; /* Write time */ +/*70*/ __le16 s_mnt_count; /* Mount count */ + __le16 s_max_mnt_count; /* Maximal mount count */ + __le16 s_state; /* File system state */ + __le16 s_errors; /* Behaviour when detecting errors */ + __le64 s_lastcheck; /* time of last check */ + +/*80*/ __le32 s_checkinterval; /* max. time between checks */ + __le32 s_creator_os; /* OS */ + __le16 s_def_resuid; /* Default uid for reserved blocks */ + __le16 s_def_resgid; /* Default gid for reserved blocks */ + __le32 s_first_ino; /* First non-reserved inode */ + +/*90*/ __le16 s_inode_size; /* Size of an inode */ + __le16 s_dat_entry_size; /* Size of a dat entry */ + __le16 s_checkpoint_size; /* Size of a checkpoint */ + __le16 s_segment_usage_size; /* Size of a segment usage */ + +/*98*/ __u8 s_uuid[16]; /* 128-bit uuid for volume */ +/*A8*/ char s_volume_name[80]; /* volume name */ + +/*F8*/ __le32 s_c_interval; /* Commit interval of segment */ + __le32 s_c_block_max; /* + * Threshold of data amount for + * the segment construction + */ +/*100*/ __le64 s_feature_compat; /* Compatible feature set */ + __le64 s_feature_compat_ro; /* Read-only compatible feature set */ + __le64 s_feature_incompat; /* Incompatible feature set */ + __u32 s_reserved[186]; /* padding to the end of the block */ +}; + +/* + * Codes for operating systems + */ +#define NILFS_OS_LINUX 0 +/* Codes from 1 to 4 are reserved to keep compatibility with ext2 creator-OS */ + +/* + * Revision levels + */ +#define NILFS_CURRENT_REV 2 /* current major revision */ +#define NILFS_MINOR_REV 0 /* minor revision */ +#define NILFS_MIN_SUPP_REV 2 /* minimum supported revision */ + +/* + * Feature set definitions + * + * If there is a bit set in the incompatible feature set that the kernel + * doesn't know about, it should refuse to mount the filesystem. + */ +#define NILFS_FEATURE_COMPAT_RO_BLOCK_COUNT 0x00000001ULL + +#define NILFS_FEATURE_COMPAT_SUPP 0ULL +#define NILFS_FEATURE_COMPAT_RO_SUPP NILFS_FEATURE_COMPAT_RO_BLOCK_COUNT +#define NILFS_FEATURE_INCOMPAT_SUPP 0ULL + +/* + * Bytes count of super_block for CRC-calculation + */ +#define NILFS_SB_BYTES \ + ((long)&((struct nilfs_super_block *)0)->s_reserved) + +/* + * Special inode number + */ +#define NILFS_ROOT_INO 2 /* Root file inode */ +#define NILFS_DAT_INO 3 /* DAT file */ +#define NILFS_CPFILE_INO 4 /* checkpoint file */ +#define NILFS_SUFILE_INO 5 /* segment usage file */ +#define NILFS_IFILE_INO 6 /* ifile */ +#define NILFS_ATIME_INO 7 /* Atime file (reserved) */ +#define NILFS_XATTR_INO 8 /* Xattribute file (reserved) */ +#define NILFS_SKETCH_INO 10 /* Sketch file */ +#define NILFS_USER_INO 11 /* Fisrt user's file inode number */ + +#define NILFS_SB_OFFSET_BYTES 1024 /* byte offset of nilfs superblock */ + +#define NILFS_SEG_MIN_BLOCKS 16 /* + * Minimum number of blocks in + * a full segment + */ +#define NILFS_PSEG_MIN_BLOCKS 2 /* + * Minimum number of blocks in + * a partial segment + */ +#define NILFS_MIN_NRSVSEGS 8 /* + * Minimum number of reserved + * segments + */ + +/* + * We call DAT, cpfile, and sufile root metadata files. Inodes of + * these files are written in super root block instead of ifile, and + * garbage collector doesn't keep any past versions of these files. + */ +#define NILFS_ROOT_METADATA_FILE(ino) \ + ((ino) >= NILFS_DAT_INO && (ino) <= NILFS_SUFILE_INO) + +/* + * bytes offset of secondary super block + */ +#define NILFS_SB2_OFFSET_BYTES(devsize) ((((devsize) >> 12) - 1) << 12) + +/* + * Maximal count of links to a file + */ +#define NILFS_LINK_MAX 32000 + +/* + * Structure of a directory entry + * (Same as ext2) + */ + +#define NILFS_NAME_LEN 255 + +/* + * Block size limitations + */ +#define NILFS_MIN_BLOCK_SIZE 1024 +#define NILFS_MAX_BLOCK_SIZE 65536 + +/* + * The new version of the directory entry. Since V0 structures are + * stored in intel byte order, and the name_len field could never be + * bigger than 255 chars, it's safe to reclaim the extra byte for the + * file_type field. + */ +struct nilfs_dir_entry { + __le64 inode; /* Inode number */ + __le16 rec_len; /* Directory entry length */ + __u8 name_len; /* Name length */ + __u8 file_type; /* Dir entry type (file, dir, etc) */ + char name[NILFS_NAME_LEN]; /* File name */ + char pad; +}; + +/* + * NILFS directory file types. Only the low 3 bits are used. The + * other bits are reserved for now. + */ +enum { + NILFS_FT_UNKNOWN, + NILFS_FT_REG_FILE, + NILFS_FT_DIR, + NILFS_FT_CHRDEV, + NILFS_FT_BLKDEV, + NILFS_FT_FIFO, + NILFS_FT_SOCK, + NILFS_FT_SYMLINK, + NILFS_FT_MAX +}; + +/* + * NILFS_DIR_PAD defines the directory entries boundaries + * + * NOTE: It must be a multiple of 8 + */ +#define NILFS_DIR_PAD 8 +#define NILFS_DIR_ROUND (NILFS_DIR_PAD - 1) +#define NILFS_DIR_REC_LEN(name_len) (((name_len) + 12 + NILFS_DIR_ROUND) & \ + ~NILFS_DIR_ROUND) +#define NILFS_MAX_REC_LEN ((1 << 16) - 1) + +/** + * struct nilfs_finfo - file information + * @fi_ino: inode number + * @fi_cno: checkpoint number + * @fi_nblocks: number of blocks (including intermediate blocks) + * @fi_ndatablk: number of file data blocks + */ +struct nilfs_finfo { + __le64 fi_ino; + __le64 fi_cno; + __le32 fi_nblocks; + __le32 fi_ndatablk; +}; + +/** + * struct nilfs_binfo_v - information on a data block (except DAT) + * @bi_vblocknr: virtual block number + * @bi_blkoff: block offset + */ +struct nilfs_binfo_v { + __le64 bi_vblocknr; + __le64 bi_blkoff; +}; + +/** + * struct nilfs_binfo_dat - information on a DAT node block + * @bi_blkoff: block offset + * @bi_level: level + * @bi_pad: padding + */ +struct nilfs_binfo_dat { + __le64 bi_blkoff; + __u8 bi_level; + __u8 bi_pad[7]; +}; + +/** + * union nilfs_binfo: block information + * @bi_v: nilfs_binfo_v structure + * @bi_dat: nilfs_binfo_dat structure + */ +union nilfs_binfo { + struct nilfs_binfo_v bi_v; + struct nilfs_binfo_dat bi_dat; +}; + +/** + * struct nilfs_segment_summary - segment summary header + * @ss_datasum: checksum of data + * @ss_sumsum: checksum of segment summary + * @ss_magic: magic number + * @ss_bytes: size of this structure in bytes + * @ss_flags: flags + * @ss_seq: sequence number + * @ss_create: creation timestamp + * @ss_next: next segment + * @ss_nblocks: number of blocks + * @ss_nfinfo: number of finfo structures + * @ss_sumbytes: total size of segment summary in bytes + * @ss_pad: padding + * @ss_cno: checkpoint number + */ +struct nilfs_segment_summary { + __le32 ss_datasum; + __le32 ss_sumsum; + __le32 ss_magic; + __le16 ss_bytes; + __le16 ss_flags; + __le64 ss_seq; + __le64 ss_create; + __le64 ss_next; + __le32 ss_nblocks; + __le32 ss_nfinfo; + __le32 ss_sumbytes; + __le32 ss_pad; + __le64 ss_cno; + /* array of finfo structures */ +}; + +#define NILFS_SEGSUM_MAGIC 0x1eaffa11 /* segment summary magic number */ + +/* + * Segment summary flags + */ +#define NILFS_SS_LOGBGN 0x0001 /* begins a logical segment */ +#define NILFS_SS_LOGEND 0x0002 /* ends a logical segment */ +#define NILFS_SS_SR 0x0004 /* has super root */ +#define NILFS_SS_SYNDT 0x0008 /* includes data only updates */ +#define NILFS_SS_GC 0x0010 /* segment written for cleaner operation */ + +/** + * struct nilfs_btree_node - header of B-tree node block + * @bn_flags: flags + * @bn_level: level + * @bn_nchildren: number of children + * @bn_pad: padding + */ +struct nilfs_btree_node { + __u8 bn_flags; + __u8 bn_level; + __le16 bn_nchildren; + __le32 bn_pad; +}; + +/* flags */ +#define NILFS_BTREE_NODE_ROOT 0x01 + +/* level */ +#define NILFS_BTREE_LEVEL_DATA 0 +#define NILFS_BTREE_LEVEL_NODE_MIN (NILFS_BTREE_LEVEL_DATA + 1) +#define NILFS_BTREE_LEVEL_MAX 14 /* Max level (exclusive) */ + +/** + * struct nilfs_direct_node - header of built-in bmap array + * @dn_flags: flags + * @dn_pad: padding + */ +struct nilfs_direct_node { + __u8 dn_flags; + __u8 pad[7]; +}; + +/** + * struct nilfs_palloc_group_desc - block group descriptor + * @pg_nfrees: number of free entries in block group + */ +struct nilfs_palloc_group_desc { + __le32 pg_nfrees; +}; + +/** + * struct nilfs_dat_entry - disk address translation entry + * @de_blocknr: block number + * @de_start: start checkpoint number + * @de_end: end checkpoint number + * @de_rsv: reserved for future use + */ +struct nilfs_dat_entry { + __le64 de_blocknr; + __le64 de_start; + __le64 de_end; + __le64 de_rsv; +}; + +#define NILFS_MIN_DAT_ENTRY_SIZE 32 + +/** + * struct nilfs_snapshot_list - snapshot list + * @ssl_next: next checkpoint number on snapshot list + * @ssl_prev: previous checkpoint number on snapshot list + */ +struct nilfs_snapshot_list { + __le64 ssl_next; + __le64 ssl_prev; +}; + +/** + * struct nilfs_checkpoint - checkpoint structure + * @cp_flags: flags + * @cp_checkpoints_count: checkpoints count in a block + * @cp_snapshot_list: snapshot list + * @cp_cno: checkpoint number + * @cp_create: creation timestamp + * @cp_nblk_inc: number of blocks incremented by this checkpoint + * @cp_inodes_count: inodes count + * @cp_blocks_count: blocks count + * @cp_ifile_inode: inode of ifile + */ +struct nilfs_checkpoint { + __le32 cp_flags; + __le32 cp_checkpoints_count; + struct nilfs_snapshot_list cp_snapshot_list; + __le64 cp_cno; + __le64 cp_create; + __le64 cp_nblk_inc; + __le64 cp_inodes_count; + __le64 cp_blocks_count; + + /* + * Do not change the byte offset of ifile inode. + * To keep the compatibility of the disk format, + * additional fields should be added behind cp_ifile_inode. + */ + struct nilfs_inode cp_ifile_inode; +}; + +#define NILFS_MIN_CHECKPOINT_SIZE (64 + NILFS_MIN_INODE_SIZE) + +/* checkpoint flags */ +enum { + NILFS_CHECKPOINT_SNAPSHOT, + NILFS_CHECKPOINT_INVALID, + NILFS_CHECKPOINT_SKETCH, + NILFS_CHECKPOINT_MINOR, +}; + +#define NILFS_CHECKPOINT_FNS(flag, name) \ +static inline void \ +nilfs_checkpoint_set_##name(struct nilfs_checkpoint *cp) \ +{ \ + cp->cp_flags = cpu_to_le32(le32_to_cpu(cp->cp_flags) | \ + (1UL << NILFS_CHECKPOINT_##flag)); \ +} \ +static inline void \ +nilfs_checkpoint_clear_##name(struct nilfs_checkpoint *cp) \ +{ \ + cp->cp_flags = cpu_to_le32(le32_to_cpu(cp->cp_flags) & \ + ~(1UL << NILFS_CHECKPOINT_##flag)); \ +} \ +static inline int \ +nilfs_checkpoint_##name(const struct nilfs_checkpoint *cp) \ +{ \ + return !!(le32_to_cpu(cp->cp_flags) & \ + (1UL << NILFS_CHECKPOINT_##flag)); \ +} + +NILFS_CHECKPOINT_FNS(SNAPSHOT, snapshot) +NILFS_CHECKPOINT_FNS(INVALID, invalid) +NILFS_CHECKPOINT_FNS(MINOR, minor) + +/** + * struct nilfs_cpfile_header - checkpoint file header + * @ch_ncheckpoints: number of checkpoints + * @ch_nsnapshots: number of snapshots + * @ch_snapshot_list: snapshot list + */ +struct nilfs_cpfile_header { + __le64 ch_ncheckpoints; + __le64 ch_nsnapshots; + struct nilfs_snapshot_list ch_snapshot_list; +}; + +#define NILFS_CPFILE_FIRST_CHECKPOINT_OFFSET \ + ((sizeof(struct nilfs_cpfile_header) + \ + sizeof(struct nilfs_checkpoint) - 1) / \ + sizeof(struct nilfs_checkpoint)) + +/** + * struct nilfs_segment_usage - segment usage + * @su_lastmod: last modified timestamp + * @su_nblocks: number of blocks in segment + * @su_flags: flags + */ +struct nilfs_segment_usage { + __le64 su_lastmod; + __le32 su_nblocks; + __le32 su_flags; +}; + +#define NILFS_MIN_SEGMENT_USAGE_SIZE 16 + +/* segment usage flag */ +enum { + NILFS_SEGMENT_USAGE_ACTIVE, + NILFS_SEGMENT_USAGE_DIRTY, + NILFS_SEGMENT_USAGE_ERROR, +}; + +#define NILFS_SEGMENT_USAGE_FNS(flag, name) \ +static inline void \ +nilfs_segment_usage_set_##name(struct nilfs_segment_usage *su) \ +{ \ + su->su_flags = cpu_to_le32(le32_to_cpu(su->su_flags) | \ + (1UL << NILFS_SEGMENT_USAGE_##flag));\ +} \ +static inline void \ +nilfs_segment_usage_clear_##name(struct nilfs_segment_usage *su) \ +{ \ + su->su_flags = \ + cpu_to_le32(le32_to_cpu(su->su_flags) & \ + ~(1UL << NILFS_SEGMENT_USAGE_##flag)); \ +} \ +static inline int \ +nilfs_segment_usage_##name(const struct nilfs_segment_usage *su) \ +{ \ + return !!(le32_to_cpu(su->su_flags) & \ + (1UL << NILFS_SEGMENT_USAGE_##flag)); \ +} + +NILFS_SEGMENT_USAGE_FNS(ACTIVE, active) +NILFS_SEGMENT_USAGE_FNS(DIRTY, dirty) +NILFS_SEGMENT_USAGE_FNS(ERROR, error) + +static inline void +nilfs_segment_usage_set_clean(struct nilfs_segment_usage *su) +{ + su->su_lastmod = cpu_to_le64(0); + su->su_nblocks = cpu_to_le32(0); + su->su_flags = cpu_to_le32(0); +} + +static inline int +nilfs_segment_usage_clean(const struct nilfs_segment_usage *su) +{ + return !le32_to_cpu(su->su_flags); +} + +/** + * struct nilfs_sufile_header - segment usage file header + * @sh_ncleansegs: number of clean segments + * @sh_ndirtysegs: number of dirty segments + * @sh_last_alloc: last allocated segment number + */ +struct nilfs_sufile_header { + __le64 sh_ncleansegs; + __le64 sh_ndirtysegs; + __le64 sh_last_alloc; + /* ... */ +}; + +#define NILFS_SUFILE_FIRST_SEGMENT_USAGE_OFFSET \ + ((sizeof(struct nilfs_sufile_header) + \ + sizeof(struct nilfs_segment_usage) - 1) / \ + sizeof(struct nilfs_segment_usage)) + +#endif /* _LINUX_NILFS2_ONDISK_H */ -- cgit v1.2.3 From 0a11b9aae49adf1f952427ef1a1d9e793dd6ffb6 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 2 Aug 2016 14:05:33 -0700 Subject: reiserfs: fix "new_insert_key may be used uninitialized ..." new_insert_key only makes any sense when it's associated with a new_insert_ptr, which is initialized to NULL and changed to a buffer_head when we also initialize new_insert_key. We can key off of that to avoid the uninitialized warning. Link: http://lkml.kernel.org/r/5eca5ffb-2155-8df2-b4a2-f162f105efed@suse.com Signed-off-by: Jeff Mahoney Cc: Arnd Bergmann Cc: Jan Kara Cc: Linus Torvalds Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/ibalance.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/reiserfs/ibalance.c b/fs/reiserfs/ibalance.c index b751eea32e20..5db6f45b3fed 100644 --- a/fs/reiserfs/ibalance.c +++ b/fs/reiserfs/ibalance.c @@ -1153,8 +1153,9 @@ int balance_internal(struct tree_balance *tb, insert_ptr); } - memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE); insert_ptr[0] = new_insert_ptr; + if (new_insert_ptr) + memcpy(new_insert_key_addr, &new_insert_key, KEY_SIZE); return order; } -- cgit v1.2.3 From 0facef7fb053be4353c0a48c2f48c9dbee91cb19 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 10:58:53 +1000 Subject: xfs: in _attrlist_by_handle, copy the cursor back to userspace When we're iterating inode xattrs by handle, we have to copy the cursor back to userspace so that a subsequent invocation actually retrieves subsequent contents. Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_ioctl.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index 9a7c87809d3b..cf46658392ce 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -387,6 +387,7 @@ xfs_attrlist_by_handle( { int error = -ENOMEM; attrlist_cursor_kern_t *cursor; + struct xfs_fsop_attrlist_handlereq __user *p = arg; xfs_fsop_attrlist_handlereq_t al_hreq; struct dentry *dentry; char *kbuf; @@ -419,6 +420,11 @@ xfs_attrlist_by_handle( if (error) goto out_kfree; + if (copy_to_user(&p->pos, cursor, sizeof(attrlist_cursor_kern_t))) { + error = -EFAULT; + goto out_kfree; + } + if (copy_to_user(al_hreq.buffer, kbuf, al_hreq.buflen)) error = -EFAULT; -- cgit v1.2.3 From 3dadf901ddc0a1275b622b1a170557bd0d136862 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 10:59:42 +1000 Subject: xfs: fix attr shortform structure alignment on cris Apparently cris doesn't require structure stride to align with the largest type in the struct, so list[0] isn't at offset 4 like it is everywhere else. Fix this... insofar as existing XFSes on cris are screwed. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_da_format.h | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_da_format.h b/fs/xfs/libxfs/xfs_da_format.h index 685f23b67056..9a492a9e19bd 100644 --- a/fs/xfs/libxfs/xfs_da_format.h +++ b/fs/xfs/libxfs/xfs_da_format.h @@ -629,6 +629,7 @@ typedef struct xfs_attr_shortform { struct xfs_attr_sf_hdr { /* constant-structure header block */ __be16 totsize; /* total bytes in shortform list */ __u8 count; /* count of active entries */ + __u8 padding; } hdr; struct xfs_attr_sf_entry { __uint8_t namelen; /* actual length of name (no NULL) */ -- cgit v1.2.3 From f4a0660de34451e30f0bb8b65946b79c8bd375ca Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:00:42 +1000 Subject: xfs: fix locking of the rt bitmap/summary inodes When we're deleting realtime extents, we need to lock the summary inode in case we need to update the summary info to prevent an assert on the rsumip inode lock on a debug kernel. While we're at it, fix the locking annotations so that we avoid triggering lockdep warnings. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_bmap.c | 4 +++- fs/xfs/xfs_bmap_util.c | 4 ++-- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 2f2c85cc8117..c5981f4a44bd 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -5179,8 +5179,10 @@ xfs_bunmapi( /* * Synchronize by locking the bitmap inode. */ - xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL|XFS_ILOCK_RTBITMAP); xfs_trans_ijoin(tp, mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL|XFS_ILOCK_RTSUM); + xfs_trans_ijoin(tp, mp->m_rsumip, XFS_ILOCK_EXCL); } extno = 0; diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index cd4a850564f2..998c3e69dd33 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -214,9 +214,9 @@ xfs_bmap_rtalloc( /* * Lock out modifications to both the RT bitmap and summary inodes */ - xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL); + xfs_ilock(mp->m_rbmip, XFS_ILOCK_EXCL|XFS_ILOCK_RTBITMAP); xfs_trans_ijoin(ap->tp, mp->m_rbmip, XFS_ILOCK_EXCL); - xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL); + xfs_ilock(mp->m_rsumip, XFS_ILOCK_EXCL|XFS_ILOCK_RTSUM); xfs_trans_ijoin(ap->tp, mp->m_rsumip, XFS_ILOCK_EXCL); /* -- cgit v1.2.3 From 0d309791bdc0a92f1db5dfc171d884a6b8583702 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:01:25 +1000 Subject: xfs: set *stat=1 after iroot realloc If we make the inode root block of a btree unfull by expanding the root, we must set *stat to 1 to signal success, rather than leaving it uninitialized. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 07eeb0b4ca74..f5ff95a18e9d 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2871,6 +2871,7 @@ xfs_btree_make_block_unfull( if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) { /* A root block that can be made bigger. */ xfs_iroot_realloc(ip, 1, cur->bc_private.b.whichfork); + *stat = 1; } else { /* A root block that needs replacing */ int logflags = 0; -- cgit v1.2.3 From e5821e57af54abc36ea299bde6c101a804cfac27 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:02:39 +1000 Subject: xfs: during btree split, save new block key & ptr for future insertion When a btree block has to be split, we pass the new block's ptr from xfs_btree_split() back to xfs_btree_insert() via a pointer parameter; however, we pass the block's key through the cursor's record. It is a little weird to "initialize" a record from a key since the non-key attributes will have garbage values. When we go to add support for interval queries, we have to be able to pass the lowest and highest keys accessible via a pointer. There's no clean way to pass this back through the cursor's record field. Therefore, pass the key directly back to xfs_btree_insert() the same way that we pass the btree_ptr. As a bonus, we no longer need init_rec_from_key and can drop it from the codebase. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc_btree.c | 12 ------------ fs/xfs/libxfs/xfs_bmap_btree.c | 12 ------------ fs/xfs/libxfs/xfs_btree.c | 40 ++++++++++++++++++++-------------------- fs/xfs/libxfs/xfs_btree.h | 2 -- fs/xfs/libxfs/xfs_ialloc_btree.c | 10 ---------- 5 files changed, 20 insertions(+), 56 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c index d9b42425291e..5ba2dac5e67c 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.c +++ b/fs/xfs/libxfs/xfs_alloc_btree.c @@ -211,17 +211,6 @@ xfs_allocbt_init_key_from_rec( key->alloc.ar_blockcount = rec->alloc.ar_blockcount; } -STATIC void -xfs_allocbt_init_rec_from_key( - union xfs_btree_key *key, - union xfs_btree_rec *rec) -{ - ASSERT(key->alloc.ar_startblock != 0); - - rec->alloc.ar_startblock = key->alloc.ar_startblock; - rec->alloc.ar_blockcount = key->alloc.ar_blockcount; -} - STATIC void xfs_allocbt_init_rec_from_cur( struct xfs_btree_cur *cur, @@ -406,7 +395,6 @@ static const struct xfs_btree_ops xfs_allocbt_ops = { .get_minrecs = xfs_allocbt_get_minrecs, .get_maxrecs = xfs_allocbt_get_maxrecs, .init_key_from_rec = xfs_allocbt_init_key_from_rec, - .init_rec_from_key = xfs_allocbt_init_rec_from_key, .init_rec_from_cur = xfs_allocbt_init_rec_from_cur, .init_ptr_from_cur = xfs_allocbt_init_ptr_from_cur, .key_diff = xfs_allocbt_key_diff, diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index db0c71e470c9..714b387e11ad 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -599,17 +599,6 @@ xfs_bmbt_init_key_from_rec( cpu_to_be64(xfs_bmbt_disk_get_startoff(&rec->bmbt)); } -STATIC void -xfs_bmbt_init_rec_from_key( - union xfs_btree_key *key, - union xfs_btree_rec *rec) -{ - ASSERT(key->bmbt.br_startoff != 0); - - xfs_bmbt_disk_set_allf(&rec->bmbt, be64_to_cpu(key->bmbt.br_startoff), - 0, 0, XFS_EXT_NORM); -} - STATIC void xfs_bmbt_init_rec_from_cur( struct xfs_btree_cur *cur, @@ -760,7 +749,6 @@ static const struct xfs_btree_ops xfs_bmbt_ops = { .get_minrecs = xfs_bmbt_get_minrecs, .get_dmaxrecs = xfs_bmbt_get_dmaxrecs, .init_key_from_rec = xfs_bmbt_init_key_from_rec, - .init_rec_from_key = xfs_bmbt_init_rec_from_key, .init_rec_from_cur = xfs_bmbt_init_rec_from_cur, .init_ptr_from_cur = xfs_bmbt_init_ptr_from_cur, .key_diff = xfs_bmbt_key_diff, diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index f5ff95a18e9d..8d8e3625c365 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2858,10 +2858,9 @@ xfs_btree_make_block_unfull( int *index, /* new tree index */ union xfs_btree_ptr *nptr, /* new btree ptr */ struct xfs_btree_cur **ncur, /* new btree cursor */ - union xfs_btree_rec *nrec, /* new record */ + union xfs_btree_key *key, /* key of new block */ int *stat) { - union xfs_btree_key key; /* new btree key value */ int error = 0; if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && @@ -2907,13 +2906,12 @@ xfs_btree_make_block_unfull( * If this works we have to re-set our variables because we * could be in a different block now. */ - error = xfs_btree_split(cur, level, nptr, &key, ncur, stat); + error = xfs_btree_split(cur, level, nptr, key, ncur, stat); if (error || *stat == 0) return error; *index = cur->bc_ptrs[level]; - cur->bc_ops->init_rec_from_key(&key, nrec); return 0; } @@ -2926,16 +2924,16 @@ xfs_btree_insrec( struct xfs_btree_cur *cur, /* btree cursor */ int level, /* level to insert record at */ union xfs_btree_ptr *ptrp, /* i/o: block number inserted */ - union xfs_btree_rec *recp, /* i/o: record data inserted */ + union xfs_btree_rec *rec, /* record to insert */ + union xfs_btree_key *key, /* i/o: block key for ptrp */ struct xfs_btree_cur **curp, /* output: new cursor replacing cur */ int *stat) /* success/failure */ { struct xfs_btree_block *block; /* btree block */ struct xfs_buf *bp; /* buffer for block */ - union xfs_btree_key key; /* btree key */ union xfs_btree_ptr nptr; /* new block ptr */ struct xfs_btree_cur *ncur; /* new btree cursor */ - union xfs_btree_rec nrec; /* new record count */ + union xfs_btree_key nkey; /* new block key */ int optr; /* old key/record index */ int ptr; /* key/record index */ int numrecs;/* number of records */ @@ -2945,7 +2943,7 @@ xfs_btree_insrec( #endif XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); - XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, recp); + XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, &rec); ncur = NULL; @@ -2970,9 +2968,6 @@ xfs_btree_insrec( return 0; } - /* Make a key out of the record data to be inserted, and save it. */ - cur->bc_ops->init_key_from_rec(&key, recp); - optr = ptr; XFS_BTREE_STATS_INC(cur, insrec); @@ -2989,10 +2984,10 @@ xfs_btree_insrec( /* Check that the new entry is being inserted in the right place. */ if (ptr <= numrecs) { if (level == 0) { - ASSERT(cur->bc_ops->recs_inorder(cur, recp, + ASSERT(cur->bc_ops->recs_inorder(cur, rec, xfs_btree_rec_addr(cur, ptr, block))); } else { - ASSERT(cur->bc_ops->keys_inorder(cur, &key, + ASSERT(cur->bc_ops->keys_inorder(cur, key, xfs_btree_key_addr(cur, ptr, block))); } } @@ -3005,7 +3000,7 @@ xfs_btree_insrec( xfs_btree_set_ptr_null(cur, &nptr); if (numrecs == cur->bc_ops->get_maxrecs(cur, level)) { error = xfs_btree_make_block_unfull(cur, level, numrecs, - &optr, &ptr, &nptr, &ncur, &nrec, stat); + &optr, &ptr, &nptr, &ncur, &nkey, stat); if (error || *stat == 0) goto error0; } @@ -3055,7 +3050,7 @@ xfs_btree_insrec( #endif /* Now put the new data in, bump numrecs and log it. */ - xfs_btree_copy_keys(cur, kp, &key, 1); + xfs_btree_copy_keys(cur, kp, key, 1); xfs_btree_copy_ptrs(cur, pp, ptrp, 1); numrecs++; xfs_btree_set_numrecs(block, numrecs); @@ -3076,7 +3071,7 @@ xfs_btree_insrec( xfs_btree_shift_recs(cur, rp, 1, numrecs - ptr + 1); /* Now put the new data in, bump numrecs and log it. */ - xfs_btree_copy_recs(cur, rp, recp, 1); + xfs_btree_copy_recs(cur, rp, rec, 1); xfs_btree_set_numrecs(block, ++numrecs); xfs_btree_log_recs(cur, bp, ptr, numrecs); #ifdef DEBUG @@ -3092,7 +3087,7 @@ xfs_btree_insrec( /* If we inserted at the start of a block, update the parents' keys. */ if (optr == 1) { - error = xfs_btree_updkey(cur, &key, level + 1); + error = xfs_btree_updkey(cur, key, level + 1); if (error) goto error0; } @@ -3102,7 +3097,7 @@ xfs_btree_insrec( * we are at the far right edge of the tree, update it. */ if (xfs_btree_is_lastrec(cur, block, level)) { - cur->bc_ops->update_lastrec(cur, block, recp, + cur->bc_ops->update_lastrec(cur, block, rec, ptr, LASTREC_INSREC); } @@ -3112,7 +3107,7 @@ xfs_btree_insrec( */ *ptrp = nptr; if (!xfs_btree_ptr_is_null(cur, &nptr)) { - *recp = nrec; + *key = nkey; *curp = ncur; } @@ -3143,6 +3138,7 @@ xfs_btree_insert( union xfs_btree_ptr nptr; /* new block number (split result) */ struct xfs_btree_cur *ncur; /* new cursor (split result) */ struct xfs_btree_cur *pcur; /* previous level's cursor */ + union xfs_btree_key key; /* key of block to insert */ union xfs_btree_rec rec; /* record to insert */ level = 0; @@ -3150,7 +3146,10 @@ xfs_btree_insert( pcur = cur; xfs_btree_set_ptr_null(cur, &nptr); + + /* Make a key out of the record data to be inserted, and save it. */ cur->bc_ops->init_rec_from_cur(cur, &rec); + cur->bc_ops->init_key_from_rec(&key, &rec); /* * Loop going up the tree, starting at the leaf level. @@ -3162,7 +3161,8 @@ xfs_btree_insert( * Insert nrec/nptr into this level of the tree. * Note if we fail, nptr will be null. */ - error = xfs_btree_insrec(pcur, level, &nptr, &rec, &ncur, &i); + error = xfs_btree_insrec(pcur, level, &nptr, &rec, &key, + &ncur, &i); if (error) { if (pcur != cur) xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR); diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 785a99682159..b4f3035ae05e 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -158,8 +158,6 @@ struct xfs_btree_ops { /* init values of btree structures */ void (*init_key_from_rec)(union xfs_btree_key *key, union xfs_btree_rec *rec); - void (*init_rec_from_key)(union xfs_btree_key *key, - union xfs_btree_rec *rec); void (*init_rec_from_cur)(struct xfs_btree_cur *cur, union xfs_btree_rec *rec); void (*init_ptr_from_cur)(struct xfs_btree_cur *cur, diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index 89c21d771e35..88da2ad939d4 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -145,14 +145,6 @@ xfs_inobt_init_key_from_rec( key->inobt.ir_startino = rec->inobt.ir_startino; } -STATIC void -xfs_inobt_init_rec_from_key( - union xfs_btree_key *key, - union xfs_btree_rec *rec) -{ - rec->inobt.ir_startino = key->inobt.ir_startino; -} - STATIC void xfs_inobt_init_rec_from_cur( struct xfs_btree_cur *cur, @@ -314,7 +306,6 @@ static const struct xfs_btree_ops xfs_inobt_ops = { .get_minrecs = xfs_inobt_get_minrecs, .get_maxrecs = xfs_inobt_get_maxrecs, .init_key_from_rec = xfs_inobt_init_key_from_rec, - .init_rec_from_key = xfs_inobt_init_rec_from_key, .init_rec_from_cur = xfs_inobt_init_rec_from_cur, .init_ptr_from_cur = xfs_inobt_init_ptr_from_cur, .key_diff = xfs_inobt_key_diff, @@ -336,7 +327,6 @@ static const struct xfs_btree_ops xfs_finobt_ops = { .get_minrecs = xfs_inobt_get_minrecs, .get_maxrecs = xfs_inobt_get_maxrecs, .init_key_from_rec = xfs_inobt_init_key_from_rec, - .init_rec_from_key = xfs_inobt_init_rec_from_key, .init_rec_from_cur = xfs_inobt_init_rec_from_cur, .init_ptr_from_cur = xfs_finobt_init_ptr_from_cur, .key_diff = xfs_inobt_key_diff, -- cgit v1.2.3 From 70b2265935544c2ba64619172fd757bd0ca91800 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:03:38 +1000 Subject: xfs: add function pointers for get/update keys to the btree Add some function pointers to bc_ops to get the btree keys for leaf and node blocks, and to update parent keys of a block. Convert the _btree_updkey calls to use our new pointer, and modify the tree shape changing code to call the appropriate get_*_keys pointer instead of _btree_copy_keys because the overlapping btree has to calculate high key values. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc_btree.c | 4 + fs/xfs/libxfs/xfs_bmap_btree.c | 4 + fs/xfs/libxfs/xfs_btree.c | 164 ++++++++++++++++++++++++--------------- fs/xfs/libxfs/xfs_btree.h | 19 +++++ fs/xfs/libxfs/xfs_ialloc_btree.c | 8 ++ 5 files changed, 135 insertions(+), 64 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c index 5ba2dac5e67c..c60eeb826b9d 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.c +++ b/fs/xfs/libxfs/xfs_alloc_btree.c @@ -403,6 +403,10 @@ static const struct xfs_btree_ops xfs_allocbt_ops = { .keys_inorder = xfs_allocbt_keys_inorder, .recs_inorder = xfs_allocbt_recs_inorder, #endif + + .get_leaf_keys = xfs_btree_get_leaf_keys, + .get_node_keys = xfs_btree_get_node_keys, + .update_keys = xfs_btree_update_keys, }; /* diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 714b387e11ad..4f1a98e6fb54 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -757,6 +757,10 @@ static const struct xfs_btree_ops xfs_bmbt_ops = { .keys_inorder = xfs_bmbt_keys_inorder, .recs_inorder = xfs_bmbt_recs_inorder, #endif + + .get_leaf_keys = xfs_btree_get_leaf_keys, + .get_node_keys = xfs_btree_get_node_keys, + .update_keys = xfs_btree_update_keys, }; /* diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 8d8e3625c365..405442d8e780 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -1879,24 +1879,73 @@ error0: return error; } +/* Determine the low key of a leaf block (simple) */ +void +xfs_btree_get_leaf_keys( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key) +{ + union xfs_btree_rec *rec; + + rec = xfs_btree_rec_addr(cur, 1, block); + cur->bc_ops->init_key_from_rec(key, rec); +} + +/* Determine the low key of a node block (simple) */ +void +xfs_btree_get_node_keys( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key) +{ + memcpy(key, xfs_btree_key_addr(cur, 1, block), cur->bc_ops->key_len); +} + +/* Derive the keys for any btree block. */ +STATIC void +xfs_btree_get_keys( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key) +{ + if (be16_to_cpu(block->bb_level) == 0) + cur->bc_ops->get_leaf_keys(cur, block, key); + else + cur->bc_ops->get_node_keys(cur, block, key); +} + /* - * Update keys at all levels from here to the root along the cursor's path. + * Decide if we need to update the parent keys of a btree block. For + * a standard btree this is only necessary if we're updating the first + * record/key. */ -STATIC int -xfs_btree_updkey( +static inline bool +xfs_btree_needs_key_update( + struct xfs_btree_cur *cur, + int ptr) +{ + return ptr == 1; +} + +/* + * Update the parent keys of the given level, progressing towards the root. + */ +int +xfs_btree_update_keys( struct xfs_btree_cur *cur, - union xfs_btree_key *keyp, int level) { struct xfs_btree_block *block; struct xfs_buf *bp; union xfs_btree_key *kp; + union xfs_btree_key key; int ptr; XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); XFS_BTREE_TRACE_ARGIK(cur, level, keyp); - ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) || level >= 1); + ASSERT(level >= 0); /* * Go up the tree from this level toward the root. @@ -1904,7 +1953,9 @@ xfs_btree_updkey( * Stop when we reach a level where the cursor isn't pointing * at the first entry in the block. */ - for (ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) { + block = xfs_btree_get_block(cur, level, &bp); + xfs_btree_get_keys(cur, block, &key); + for (level++, ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) { #ifdef DEBUG int error; #endif @@ -1918,7 +1969,7 @@ xfs_btree_updkey( #endif ptr = cur->bc_ptrs[level]; kp = xfs_btree_key_addr(cur, ptr, block); - xfs_btree_copy_keys(cur, kp, keyp, 1); + xfs_btree_copy_keys(cur, kp, &key, 1); xfs_btree_log_keys(cur, bp, ptr, ptr); } @@ -1971,11 +2022,8 @@ xfs_btree_update( } /* Updating first rec in leaf. Pass new key value up to our parent. */ - if (ptr == 1) { - union xfs_btree_key key; - - cur->bc_ops->init_key_from_rec(&key, rec); - error = xfs_btree_updkey(cur, &key, 1); + if (xfs_btree_needs_key_update(cur, ptr)) { + error = cur->bc_ops->update_keys(cur, 0); if (error) goto error0; } @@ -2146,11 +2194,10 @@ xfs_btree_lshift( */ cur->bc_ops->init_key_from_rec(&key, xfs_btree_rec_addr(cur, 1, right)); - rkp = &key; } - /* Update the parent key values of right. */ - error = xfs_btree_updkey(cur, rkp, level + 1); + /* Update the parent keys of the right block. */ + error = cur->bc_ops->update_keys(cur, level); if (error) goto error0; @@ -2292,7 +2339,6 @@ xfs_btree_rshift( xfs_btree_log_recs(cur, rbp, 1, rrecs + 1); cur->bc_ops->init_key_from_rec(&key, rrp); - rkp = &key; ASSERT(cur->bc_ops->recs_inorder(cur, rrp, xfs_btree_rec_addr(cur, 2, right))); @@ -2321,7 +2367,8 @@ xfs_btree_rshift( if (error) goto error1; - error = xfs_btree_updkey(tcur, rkp, level + 1); + /* Update the parent keys of the right block. */ + error = cur->bc_ops->update_keys(tcur, level); if (error) goto error1; @@ -2422,6 +2469,11 @@ __xfs_btree_split( XFS_BTREE_STATS_ADD(cur, moves, rrecs); + /* Adjust numrecs for the later get_*_keys() calls. */ + lrecs -= rrecs; + xfs_btree_set_numrecs(left, lrecs); + xfs_btree_set_numrecs(right, xfs_btree_get_numrecs(right) + rrecs); + /* * Copy btree block entries from the left block over to the * new block, the right. Update the right block and log the @@ -2447,14 +2499,15 @@ __xfs_btree_split( } #endif + /* Copy the keys & pointers to the new block. */ xfs_btree_copy_keys(cur, rkp, lkp, rrecs); xfs_btree_copy_ptrs(cur, rpp, lpp, rrecs); xfs_btree_log_keys(cur, rbp, 1, rrecs); xfs_btree_log_ptrs(cur, rbp, 1, rrecs); - /* Grab the keys to the entries moved to the right block */ - xfs_btree_copy_keys(cur, key, rkp, 1); + /* Stash the keys of the new block for later insertion. */ + cur->bc_ops->get_node_keys(cur, right, key); } else { /* It's a leaf. Move records. */ union xfs_btree_rec *lrp; /* left record pointer */ @@ -2463,27 +2516,23 @@ __xfs_btree_split( lrp = xfs_btree_rec_addr(cur, src_index, left); rrp = xfs_btree_rec_addr(cur, 1, right); + /* Copy records to the new block. */ xfs_btree_copy_recs(cur, rrp, lrp, rrecs); xfs_btree_log_recs(cur, rbp, 1, rrecs); - cur->bc_ops->init_key_from_rec(key, - xfs_btree_rec_addr(cur, 1, right)); + /* Stash the keys of the new block for later insertion. */ + cur->bc_ops->get_leaf_keys(cur, right, key); } - /* * Find the left block number by looking in the buffer. - * Adjust numrecs, sibling pointers. + * Adjust sibling pointers. */ xfs_btree_get_sibling(cur, left, &rrptr, XFS_BB_RIGHTSIB); xfs_btree_set_sibling(cur, right, &rrptr, XFS_BB_RIGHTSIB); xfs_btree_set_sibling(cur, right, &lptr, XFS_BB_LEFTSIB); xfs_btree_set_sibling(cur, left, &rptr, XFS_BB_RIGHTSIB); - lrecs -= rrecs; - xfs_btree_set_numrecs(left, lrecs); - xfs_btree_set_numrecs(right, xfs_btree_get_numrecs(right) + rrecs); - xfs_btree_log_block(cur, rbp, XFS_BB_ALL_BITS); xfs_btree_log_block(cur, lbp, XFS_BB_NUMRECS | XFS_BB_RIGHTSIB); @@ -2802,6 +2851,7 @@ xfs_btree_new_root( bp = lbp; nptr = 2; } + /* Fill in the new block's btree header and log it. */ xfs_btree_init_block_cur(cur, nbp, cur->bc_nlevels, 2); xfs_btree_log_block(cur, nbp, XFS_BB_ALL_BITS); @@ -2810,19 +2860,24 @@ xfs_btree_new_root( /* Fill in the key data in the new root. */ if (xfs_btree_get_level(left) > 0) { - xfs_btree_copy_keys(cur, - xfs_btree_key_addr(cur, 1, new), - xfs_btree_key_addr(cur, 1, left), 1); - xfs_btree_copy_keys(cur, - xfs_btree_key_addr(cur, 2, new), - xfs_btree_key_addr(cur, 1, right), 1); + /* + * Get the keys for the left block's keys and put them directly + * in the parent block. Do the same for the right block. + */ + cur->bc_ops->get_node_keys(cur, left, + xfs_btree_key_addr(cur, 1, new)); + cur->bc_ops->get_node_keys(cur, right, + xfs_btree_key_addr(cur, 2, new)); } else { - cur->bc_ops->init_key_from_rec( - xfs_btree_key_addr(cur, 1, new), - xfs_btree_rec_addr(cur, 1, left)); - cur->bc_ops->init_key_from_rec( - xfs_btree_key_addr(cur, 2, new), - xfs_btree_rec_addr(cur, 1, right)); + /* + * Get the keys for the left block's records and put them + * directly in the parent block. Do the same for the right + * block. + */ + cur->bc_ops->get_leaf_keys(cur, left, + xfs_btree_key_addr(cur, 1, new)); + cur->bc_ops->get_leaf_keys(cur, right, + xfs_btree_key_addr(cur, 2, new)); } xfs_btree_log_keys(cur, nbp, 1, 2); @@ -2858,7 +2913,7 @@ xfs_btree_make_block_unfull( int *index, /* new tree index */ union xfs_btree_ptr *nptr, /* new btree ptr */ struct xfs_btree_cur **ncur, /* new btree cursor */ - union xfs_btree_key *key, /* key of new block */ + union xfs_btree_key *key, /* key of new block */ int *stat) { int error = 0; @@ -3086,8 +3141,8 @@ xfs_btree_insrec( xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS); /* If we inserted at the start of a block, update the parents' keys. */ - if (optr == 1) { - error = xfs_btree_updkey(cur, key, level + 1); + if (xfs_btree_needs_key_update(cur, optr)) { + error = cur->bc_ops->update_keys(cur, level); if (error) goto error0; } @@ -3107,7 +3162,7 @@ xfs_btree_insrec( */ *ptrp = nptr; if (!xfs_btree_ptr_is_null(cur, &nptr)) { - *key = nkey; + xfs_btree_copy_keys(cur, key, &nkey, 1); *curp = ncur; } @@ -3386,8 +3441,6 @@ xfs_btree_delrec( struct xfs_buf *bp; /* buffer for block */ int error; /* error return value */ int i; /* loop counter */ - union xfs_btree_key key; /* storage for keyp */ - union xfs_btree_key *keyp = &key; /* passed to the next level */ union xfs_btree_ptr lptr; /* left sibling block ptr */ struct xfs_buf *lbp; /* left buffer pointer */ struct xfs_btree_block *left; /* left btree block */ @@ -3458,13 +3511,6 @@ xfs_btree_delrec( xfs_btree_log_keys(cur, bp, ptr, numrecs - 1); xfs_btree_log_ptrs(cur, bp, ptr, numrecs - 1); } - - /* - * If it's the first record in the block, we'll need to pass a - * key up to the next level (updkey). - */ - if (ptr == 1) - keyp = xfs_btree_key_addr(cur, 1, block); } else { /* It's a leaf. operate on records */ if (ptr < numrecs) { @@ -3473,16 +3519,6 @@ xfs_btree_delrec( -1, numrecs - ptr); xfs_btree_log_recs(cur, bp, ptr, numrecs - 1); } - - /* - * If it's the first record in the block, we'll need a key - * structure to pass up to the next level (updkey). - */ - if (ptr == 1) { - cur->bc_ops->init_key_from_rec(&key, - xfs_btree_rec_addr(cur, 1, block)); - keyp = &key; - } } /* @@ -3549,8 +3585,8 @@ xfs_btree_delrec( * If we deleted the leftmost entry in the block, update the * key values above us in the tree. */ - if (ptr == 1) { - error = xfs_btree_updkey(cur, keyp, level + 1); + if (xfs_btree_needs_key_update(cur, ptr)) { + error = cur->bc_ops->update_keys(cur, level); if (error) goto error0; } diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index b4f3035ae05e..e097e60400d8 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -180,6 +180,19 @@ struct xfs_btree_ops { union xfs_btree_rec *r1, union xfs_btree_rec *r2); #endif + + /* derive the low & high keys from the records in a leaf block */ + void (*get_leaf_keys)(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key); + + /* derive the low & high keys from the keys in a node block */ + void (*get_node_keys)(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key); + + /* update the parent keys of given btree level */ + int (*update_keys)(struct xfs_btree_cur *cur, int level); }; /* @@ -475,4 +488,10 @@ bool xfs_btree_sblock_verify(struct xfs_buf *bp, unsigned int max_recs); uint xfs_btree_compute_maxlevels(struct xfs_mount *mp, uint *limits, unsigned long len); +void xfs_btree_get_leaf_keys(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, union xfs_btree_key *key); +void xfs_btree_get_node_keys(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, union xfs_btree_key *key); +int xfs_btree_update_keys(struct xfs_btree_cur *cur, int level); + #endif /* __XFS_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index 88da2ad939d4..a48f4482004c 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -314,6 +314,10 @@ static const struct xfs_btree_ops xfs_inobt_ops = { .keys_inorder = xfs_inobt_keys_inorder, .recs_inorder = xfs_inobt_recs_inorder, #endif + + .get_leaf_keys = xfs_btree_get_leaf_keys, + .get_node_keys = xfs_btree_get_node_keys, + .update_keys = xfs_btree_update_keys, }; static const struct xfs_btree_ops xfs_finobt_ops = { @@ -335,6 +339,10 @@ static const struct xfs_btree_ops xfs_finobt_ops = { .keys_inorder = xfs_inobt_keys_inorder, .recs_inorder = xfs_inobt_recs_inorder, #endif + + .get_leaf_keys = xfs_btree_get_leaf_keys, + .get_node_keys = xfs_btree_get_node_keys, + .update_keys = xfs_btree_update_keys, }; /* -- cgit v1.2.3 From 2c813ad66a7218a64db68f0a4bfa8d2d9caef4c0 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:08:36 +1000 Subject: xfs: support btrees with overlapping intervals for keys On a filesystem with both reflink and reverse mapping enabled, it's possible to have multiple rmap records referring to the same blocks on disk. When overlapping intervals are possible, querying a classic btree to find all records intersecting a given interval is inefficient because we cannot use the left side of the search interval to filter out non-matching records the same way that we can use the existing btree key to filter out records coming after the right side of the search interval. This will become important once we want to use the rmap btree to rebuild BMBTs, or implement the (future) fsmap ioctl. (For the non-overlapping case, we can perform such queries trivially by starting at the left side of the interval and walking the tree until we pass the right side.) Therefore, extend the btree code to come closer to supporting intervals as a first-class record attribute. This involves widening the btree node's key space to store both the lowest key reachable via the node pointer (as the btree does now) and the highest key reachable via the same pointer and teaching the btree modifying functions to keep the highest-key records up to date. This behavior can be turned on via a new btree ops flag so that btrees that cannot store overlapping intervals don't pay the overhead costs in terms of extra code and disk format changes. When we're deleting a record in a btree that supports overlapped interval records and the deletion results in two btree blocks being joined, we defer updating the high/low keys until after all possible joining (at higher levels in the tree) have finished. At this point, the btree pointers at all levels have been updated to remove the empty blocks and we can update the low and high keys. When we're doing this, we must be careful to update the keys of all node pointers up to the root instead of stopping at the first set of keys that don't need updating. This is because it's possible for a single deletion to cause joining of multiple levels of tree, and so we need to update everything going back to the root. The diff_two_keys functions return < 0, 0, or > 0 if key1 is less than, equal to, or greater than key2, respectively. This is consistent with the rest of the kernel and the C library. In btree_updkeys(), we need to evaluate the force_all parameter before running the key diff to avoid reading uninitialized memory when we're forcing a key update. This happens when we've allocated an empty slot at level N + 1 to point to a new block at level N and we're in the process of filling out the new keys. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 339 ++++++++++++++++++++++++++++++++++++++++++++-- fs/xfs/libxfs/xfs_btree.h | 30 ++++ fs/xfs/xfs_trace.h | 36 +++++ 3 files changed, 392 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 405442d8e780..5c1b35a23a3d 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -51,7 +51,6 @@ static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = { #define xfs_btree_magic(cur) \ xfs_magics[!!((cur)->bc_flags & XFS_BTREE_CRC_BLOCKS)][cur->bc_btnum] - STATIC int /* error (0 or EFSCORRUPTED) */ xfs_btree_check_lblock( struct xfs_btree_cur *cur, /* btree cursor */ @@ -428,6 +427,50 @@ xfs_btree_dup_cursor( * into a btree block (xfs_btree_*_offset) or return a pointer to the given * record, key or pointer (xfs_btree_*_addr). Note that all addressing * inside the btree block is done using indices starting at one, not zero! + * + * If XFS_BTREE_OVERLAPPING is set, then this btree supports keys containing + * overlapping intervals. In such a tree, records are still sorted lowest to + * highest and indexed by the smallest key value that refers to the record. + * However, nodes are different: each pointer has two associated keys -- one + * indexing the lowest key available in the block(s) below (the same behavior + * as the key in a regular btree) and another indexing the highest key + * available in the block(s) below. Because records are /not/ sorted by the + * highest key, all leaf block updates require us to compute the highest key + * that matches any record in the leaf and to recursively update the high keys + * in the nodes going further up in the tree, if necessary. Nodes look like + * this: + * + * +--------+-----+-----+-----+-----+-----+-------+-------+-----+ + * Non-Leaf: | header | lo1 | hi1 | lo2 | hi2 | ... | ptr 1 | ptr 2 | ... | + * +--------+-----+-----+-----+-----+-----+-------+-------+-----+ + * + * To perform an interval query on an overlapped tree, perform the usual + * depth-first search and use the low and high keys to decide if we can skip + * that particular node. If a leaf node is reached, return the records that + * intersect the interval. Note that an interval query may return numerous + * entries. For a non-overlapped tree, simply search for the record associated + * with the lowest key and iterate forward until a non-matching record is + * found. Section 14.3 ("Interval Trees") of _Introduction to Algorithms_ by + * Cormen, Leiserson, Rivest, and Stein (2nd or 3rd ed. only) discuss this in + * more detail. + * + * Why do we care about overlapping intervals? Let's say you have a bunch of + * reverse mapping records on a reflink filesystem: + * + * 1: +- file A startblock B offset C length D -----------+ + * 2: +- file E startblock F offset G length H --------------+ + * 3: +- file I startblock F offset J length K --+ + * 4: +- file L... --+ + * + * Now say we want to map block (B+D) into file A at offset (C+D). Ideally, + * we'd simply increment the length of record 1. But how do we find the record + * that ends at (B+D-1) (i.e. record 1)? A LE lookup of (B+D-1) would return + * record 3 because the keys are ordered first by startblock. An interval + * query would return records 1 and 2 because they both overlap (B+D-1), and + * from that we can pick out record 1 as the appropriate left neighbor. + * + * In the non-overlapped case you can do a LE lookup and decrement the cursor + * because a record's interval must end before the next record. */ /* @@ -478,6 +521,18 @@ xfs_btree_key_offset( (n - 1) * cur->bc_ops->key_len; } +/* + * Calculate offset of the n-th high key in a btree block. + */ +STATIC size_t +xfs_btree_high_key_offset( + struct xfs_btree_cur *cur, + int n) +{ + return xfs_btree_block_len(cur) + + (n - 1) * cur->bc_ops->key_len + (cur->bc_ops->key_len / 2); +} + /* * Calculate offset of the n-th block pointer in a btree block. */ @@ -518,6 +573,19 @@ xfs_btree_key_addr( ((char *)block + xfs_btree_key_offset(cur, n)); } +/* + * Return a pointer to the n-th high key in the btree block. + */ +STATIC union xfs_btree_key * +xfs_btree_high_key_addr( + struct xfs_btree_cur *cur, + int n, + struct xfs_btree_block *block) +{ + return (union xfs_btree_key *) + ((char *)block + xfs_btree_high_key_offset(cur, n)); +} + /* * Return a pointer to the n-th block pointer in the btree block. */ @@ -1902,6 +1970,73 @@ xfs_btree_get_node_keys( memcpy(key, xfs_btree_key_addr(cur, 1, block), cur->bc_ops->key_len); } +/* Find the high key storage area from a regular key. */ +STATIC union xfs_btree_key * +xfs_btree_high_key_from_key( + struct xfs_btree_cur *cur, + union xfs_btree_key *key) +{ + ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + return (union xfs_btree_key *)((char *)key + + (cur->bc_ops->key_len / 2)); +} + +/* Determine the low and high keys of a leaf block (overlapped) */ +void +xfs_btree_get_leaf_keys_overlapped( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key) +{ + int n; + union xfs_btree_rec *rec; + union xfs_btree_key max_hkey; + union xfs_btree_key hkey; + union xfs_btree_key *high; + + ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + rec = xfs_btree_rec_addr(cur, 1, block); + cur->bc_ops->init_key_from_rec(key, rec); + + cur->bc_ops->init_high_key_from_rec(&max_hkey, rec); + for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { + rec = xfs_btree_rec_addr(cur, n, block); + cur->bc_ops->init_high_key_from_rec(&hkey, rec); + if (cur->bc_ops->diff_two_keys(cur, &hkey, &max_hkey) > 0) + max_hkey = hkey; + } + + high = xfs_btree_high_key_from_key(cur, key); + memcpy(high, &max_hkey, cur->bc_ops->key_len / 2); +} + +/* Determine the low and high keys of a node block (overlapped) */ +void +xfs_btree_get_node_keys_overlapped( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + union xfs_btree_key *key) +{ + int n; + union xfs_btree_key *hkey; + union xfs_btree_key *max_hkey; + union xfs_btree_key *high; + + ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + memcpy(key, xfs_btree_key_addr(cur, 1, block), + cur->bc_ops->key_len / 2); + + max_hkey = xfs_btree_high_key_addr(cur, 1, block); + for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { + hkey = xfs_btree_high_key_addr(cur, n, block); + if (cur->bc_ops->diff_two_keys(cur, hkey, max_hkey) > 0) + max_hkey = hkey; + } + + high = xfs_btree_high_key_from_key(cur, key); + memcpy(high, max_hkey, cur->bc_ops->key_len / 2); +} + /* Derive the keys for any btree block. */ STATIC void xfs_btree_get_keys( @@ -1918,14 +2053,107 @@ xfs_btree_get_keys( /* * Decide if we need to update the parent keys of a btree block. For * a standard btree this is only necessary if we're updating the first - * record/key. + * record/key. For an overlapping btree, we must always update the + * keys because the highest key can be in any of the records or keys + * in the block. */ static inline bool xfs_btree_needs_key_update( struct xfs_btree_cur *cur, int ptr) { - return ptr == 1; + return (cur->bc_flags & XFS_BTREE_OVERLAPPING) || ptr == 1; +} + +/* + * Update the low and high parent keys of the given level, progressing + * towards the root. If force_all is false, stop if the keys for a given + * level do not need updating. + */ +STATIC int +__xfs_btree_updkeys( + struct xfs_btree_cur *cur, + int level, + struct xfs_btree_block *block, + struct xfs_buf *bp0, + bool force_all) +{ + union xfs_btree_bigkey key; /* keys from current level */ + union xfs_btree_key *lkey; /* keys from the next level up */ + union xfs_btree_key *hkey; + union xfs_btree_key *nlkey; /* keys from the next level up */ + union xfs_btree_key *nhkey; + struct xfs_buf *bp; + int ptr; + + ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + + /* Exit if there aren't any parent levels to update. */ + if (level + 1 >= cur->bc_nlevels) + return 0; + + trace_xfs_btree_updkeys(cur, level, bp0); + + lkey = (union xfs_btree_key *)&key; + hkey = xfs_btree_high_key_from_key(cur, lkey); + xfs_btree_get_keys(cur, block, lkey); + for (level++; level < cur->bc_nlevels; level++) { +#ifdef DEBUG + int error; +#endif + block = xfs_btree_get_block(cur, level, &bp); + trace_xfs_btree_updkeys(cur, level, bp); +#ifdef DEBUG + error = xfs_btree_check_block(cur, block, level, bp); + if (error) { + XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR); + return error; + } +#endif + ptr = cur->bc_ptrs[level]; + nlkey = xfs_btree_key_addr(cur, ptr, block); + nhkey = xfs_btree_high_key_addr(cur, ptr, block); + if (!force_all && + !(cur->bc_ops->diff_two_keys(cur, nlkey, lkey) != 0 || + cur->bc_ops->diff_two_keys(cur, nhkey, hkey) != 0)) + break; + xfs_btree_copy_keys(cur, nlkey, lkey, 1); + xfs_btree_log_keys(cur, bp, ptr, ptr); + if (level + 1 >= cur->bc_nlevels) + break; + cur->bc_ops->get_node_keys(cur, block, lkey); + } + + return 0; +} + +/* + * Update all the keys from some level in cursor back to the root, stopping + * when we find a key pair that don't need updating. + */ +int +xfs_btree_update_keys_overlapped( + struct xfs_btree_cur *cur, + int level) +{ + struct xfs_buf *bp; + struct xfs_btree_block *block; + + block = xfs_btree_get_block(cur, level, &bp); + return __xfs_btree_updkeys(cur, level, block, bp, false); +} + +/* Update all the keys from some level in cursor back to the root. */ +STATIC int +xfs_btree_updkeys_force( + struct xfs_btree_cur *cur, + int level) +{ + struct xfs_buf *bp; + struct xfs_btree_block *block; + + block = xfs_btree_get_block(cur, level, &bp); + return __xfs_btree_updkeys(cur, level, block, bp, true); } /* @@ -1942,6 +2170,8 @@ xfs_btree_update_keys( union xfs_btree_key key; int ptr; + ASSERT(!(cur->bc_flags & XFS_BTREE_OVERLAPPING)); + XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); XFS_BTREE_TRACE_ARGIK(cur, level, keyp); @@ -2021,7 +2251,7 @@ xfs_btree_update( ptr, LASTREC_UPDATE); } - /* Updating first rec in leaf. Pass new key value up to our parent. */ + /* Pass new key value up to our parent. */ if (xfs_btree_needs_key_update(cur, ptr)) { error = cur->bc_ops->update_keys(cur, 0); if (error) @@ -2052,12 +2282,14 @@ xfs_btree_lshift( int lrecs; /* left record count */ struct xfs_buf *rbp; /* right buffer pointer */ struct xfs_btree_block *right; /* right btree block */ + struct xfs_btree_cur *tcur; /* temporary btree cursor */ int rrecs; /* right record count */ union xfs_btree_ptr lptr; /* left btree pointer */ union xfs_btree_key *rkp = NULL; /* right btree key */ union xfs_btree_ptr *rpp = NULL; /* right address pointer */ union xfs_btree_rec *rrp = NULL; /* right record pointer */ int error; /* error return value */ + int i; XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); XFS_BTREE_TRACE_ARGI(cur, level); @@ -2196,10 +2428,33 @@ xfs_btree_lshift( xfs_btree_rec_addr(cur, 1, right)); } + /* + * Using a temporary cursor, update the parent key values of the + * block on the left. + */ + error = xfs_btree_dup_cursor(cur, &tcur); + if (error) + goto error0; + i = xfs_btree_firstrec(tcur, level); + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0); + + error = xfs_btree_decrement(tcur, level, &i); + if (error) + goto error1; + /* Update the parent keys of the right block. */ error = cur->bc_ops->update_keys(cur, level); if (error) - goto error0; + goto error1; + + /* Update the parent high keys of the left block, if needed. */ + if (tcur->bc_flags & XFS_BTREE_OVERLAPPING) { + error = tcur->bc_ops->update_keys(tcur, level); + if (error) + goto error1; + } + + xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR); /* Slide the cursor value left one. */ cur->bc_ptrs[level]--; @@ -2216,6 +2471,11 @@ out0: error0: XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR); return error; + +error1: + XFS_BTREE_TRACE_CURSOR(tcur, XBT_ERROR); + xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR); + return error; } /* @@ -2367,6 +2627,13 @@ xfs_btree_rshift( if (error) goto error1; + /* Update the parent high keys of the left block, if needed. */ + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + error = cur->bc_ops->update_keys(cur, level); + if (error) + goto error1; + } + /* Update the parent keys of the right block. */ error = cur->bc_ops->update_keys(tcur, level); if (error) @@ -2548,6 +2815,14 @@ __xfs_btree_split( xfs_btree_set_sibling(cur, rrblock, &rptr, XFS_BB_LEFTSIB); xfs_btree_log_block(cur, rrbp, XFS_BB_LEFTSIB); } + + /* Update the parent high keys of the left block, if needed. */ + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + error = cur->bc_ops->update_keys(cur, level); + if (error) + goto error0; + } + /* * If the cursor is really in the right block, move it there. * If it's just pointing past the last entry in left, then we'll @@ -2988,7 +3263,8 @@ xfs_btree_insrec( struct xfs_buf *bp; /* buffer for block */ union xfs_btree_ptr nptr; /* new block ptr */ struct xfs_btree_cur *ncur; /* new btree cursor */ - union xfs_btree_key nkey; /* new block key */ + union xfs_btree_bigkey nkey; /* new block key */ + union xfs_btree_key *lkey; int optr; /* old key/record index */ int ptr; /* key/record index */ int numrecs;/* number of records */ @@ -2996,11 +3272,13 @@ xfs_btree_insrec( #ifdef DEBUG int i; #endif + xfs_daddr_t old_bn; XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, &rec); ncur = NULL; + lkey = (union xfs_btree_key *)&nkey; /* * If we have an external root pointer, and we've made it to the @@ -3029,6 +3307,7 @@ xfs_btree_insrec( /* Get pointers to the btree buffer and block. */ block = xfs_btree_get_block(cur, level, &bp); + old_bn = bp ? bp->b_bn : XFS_BUF_DADDR_NULL; numrecs = xfs_btree_get_numrecs(block); #ifdef DEBUG @@ -3055,7 +3334,7 @@ xfs_btree_insrec( xfs_btree_set_ptr_null(cur, &nptr); if (numrecs == cur->bc_ops->get_maxrecs(cur, level)) { error = xfs_btree_make_block_unfull(cur, level, numrecs, - &optr, &ptr, &nptr, &ncur, &nkey, stat); + &optr, &ptr, &nptr, &ncur, lkey, stat); if (error || *stat == 0) goto error0; } @@ -3140,8 +3419,17 @@ xfs_btree_insrec( /* Log the new number of records in the btree header. */ xfs_btree_log_block(cur, bp, XFS_BB_NUMRECS); - /* If we inserted at the start of a block, update the parents' keys. */ - if (xfs_btree_needs_key_update(cur, optr)) { + /* + * If we just inserted into a new tree block, we have to + * recalculate nkey here because nkey is out of date. + * + * Otherwise we're just updating an existing block (having shoved + * some records into the new tree block), so use the regular key + * update mechanism. + */ + if (bp && bp->b_bn != old_bn) { + xfs_btree_get_keys(cur, block, lkey); + } else if (xfs_btree_needs_key_update(cur, optr)) { error = cur->bc_ops->update_keys(cur, level); if (error) goto error0; @@ -3162,7 +3450,7 @@ xfs_btree_insrec( */ *ptrp = nptr; if (!xfs_btree_ptr_is_null(cur, &nptr)) { - xfs_btree_copy_keys(cur, key, &nkey, 1); + xfs_btree_copy_keys(cur, key, lkey, 1); *curp = ncur; } @@ -3193,18 +3481,20 @@ xfs_btree_insert( union xfs_btree_ptr nptr; /* new block number (split result) */ struct xfs_btree_cur *ncur; /* new cursor (split result) */ struct xfs_btree_cur *pcur; /* previous level's cursor */ - union xfs_btree_key key; /* key of block to insert */ + union xfs_btree_bigkey bkey; /* key of block to insert */ + union xfs_btree_key *key; union xfs_btree_rec rec; /* record to insert */ level = 0; ncur = NULL; pcur = cur; + key = (union xfs_btree_key *)&bkey; xfs_btree_set_ptr_null(cur, &nptr); /* Make a key out of the record data to be inserted, and save it. */ cur->bc_ops->init_rec_from_cur(cur, &rec); - cur->bc_ops->init_key_from_rec(&key, &rec); + cur->bc_ops->init_key_from_rec(key, &rec); /* * Loop going up the tree, starting at the leaf level. @@ -3216,7 +3506,7 @@ xfs_btree_insert( * Insert nrec/nptr into this level of the tree. * Note if we fail, nptr will be null. */ - error = xfs_btree_insrec(pcur, level, &nptr, &rec, &key, + error = xfs_btree_insrec(pcur, level, &nptr, &rec, key, &ncur, &i); if (error) { if (pcur != cur) @@ -3915,6 +4205,16 @@ xfs_btree_delrec( if (level > 0) cur->bc_ptrs[level]--; + /* + * We combined blocks, so we have to update the parent keys if the + * btree supports overlapped intervals. However, bc_ptrs[level + 1] + * points to the old block so that the caller knows which record to + * delete. Therefore, the caller must be savvy enough to call updkeys + * for us if we return stat == 2. The other exit points from this + * function don't require deletions further up the tree, so they can + * call updkeys directly. + */ + XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); /* Return value means the next level up has something to do. */ *stat = 2; @@ -3940,6 +4240,7 @@ xfs_btree_delete( int error; /* error return value */ int level; int i; + bool joined = false; XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); @@ -3953,6 +4254,18 @@ xfs_btree_delete( error = xfs_btree_delrec(cur, level, &i); if (error) goto error0; + if (i == 2) + joined = true; + } + + /* + * If we combined blocks as part of deleting the record, delrec won't + * have updated the parent high keys so we have to do that here. + */ + if (joined && (cur->bc_flags & XFS_BTREE_OVERLAPPING)) { + error = xfs_btree_updkeys_force(cur, 0); + if (error) + goto error0; } if (i == 0) { diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index e097e60400d8..bce6daacb1f7 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -44,6 +44,20 @@ union xfs_btree_key { xfs_inobt_key_t inobt; }; +/* + * In-core key that holds both low and high keys for overlapped btrees. + * The two keys are packed next to each other on disk, so do the same + * in memory. Preserve the existing xfs_btree_key as a single key to + * avoid the mental model breakage that would happen if we passed a + * bigkey into a function that operates on a single key. + */ +union xfs_btree_bigkey { + struct xfs_bmbt_key bmbt; + xfs_bmdr_key_t bmbr; /* bmbt root block */ + xfs_alloc_key_t alloc; + struct xfs_inobt_key inobt; +}; + union xfs_btree_rec { xfs_bmbt_rec_t bmbt; xfs_bmdr_rec_t bmbr; /* bmbt root block */ @@ -162,11 +176,21 @@ struct xfs_btree_ops { union xfs_btree_rec *rec); void (*init_ptr_from_cur)(struct xfs_btree_cur *cur, union xfs_btree_ptr *ptr); + void (*init_high_key_from_rec)(union xfs_btree_key *key, + union xfs_btree_rec *rec); /* difference between key value and cursor value */ __int64_t (*key_diff)(struct xfs_btree_cur *cur, union xfs_btree_key *key); + /* + * Difference between key2 and key1 -- positive if key1 > key2, + * negative if key1 < key2, and zero if equal. + */ + __int64_t (*diff_two_keys)(struct xfs_btree_cur *cur, + union xfs_btree_key *key1, + union xfs_btree_key *key2); + const struct xfs_buf_ops *buf_ops; #if defined(DEBUG) || defined(XFS_WARN) @@ -249,6 +273,7 @@ typedef struct xfs_btree_cur #define XFS_BTREE_ROOT_IN_INODE (1<<1) /* root may be variable size */ #define XFS_BTREE_LASTREC_UPDATE (1<<2) /* track last rec externally */ #define XFS_BTREE_CRC_BLOCKS (1<<3) /* uses extended btree blocks */ +#define XFS_BTREE_OVERLAPPING (1<<4) /* overlapping intervals */ #define XFS_BTREE_NOERROR 0 @@ -493,5 +518,10 @@ void xfs_btree_get_leaf_keys(struct xfs_btree_cur *cur, void xfs_btree_get_node_keys(struct xfs_btree_cur *cur, struct xfs_btree_block *block, union xfs_btree_key *key); int xfs_btree_update_keys(struct xfs_btree_cur *cur, int level); +void xfs_btree_get_leaf_keys_overlapped(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, union xfs_btree_key *key); +void xfs_btree_get_node_keys_overlapped(struct xfs_btree_cur *cur, + struct xfs_btree_block *block, union xfs_btree_key *key); +int xfs_btree_update_keys_overlapped(struct xfs_btree_cur *cur, int level); #endif /* __XFS_BTREE_H__ */ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 145169093fe0..8fb59e644f80 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -38,6 +38,7 @@ struct xlog_recover_item; struct xfs_buf_log_format; struct xfs_inode_log_format; struct xfs_bmbt_irec; +struct xfs_btree_cur; DECLARE_EVENT_CLASS(xfs_attr_list_class, TP_PROTO(struct xfs_attr_list_context *ctx), @@ -2185,6 +2186,41 @@ DEFINE_DISCARD_EVENT(xfs_discard_toosmall); DEFINE_DISCARD_EVENT(xfs_discard_exclude); DEFINE_DISCARD_EVENT(xfs_discard_busy); +/* btree cursor events */ +DECLARE_EVENT_CLASS(xfs_btree_cur_class, + TP_PROTO(struct xfs_btree_cur *cur, int level, struct xfs_buf *bp), + TP_ARGS(cur, level, bp), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_btnum_t, btnum) + __field(int, level) + __field(int, nlevels) + __field(int, ptr) + __field(xfs_daddr_t, daddr) + ), + TP_fast_assign( + __entry->dev = cur->bc_mp->m_super->s_dev; + __entry->btnum = cur->bc_btnum; + __entry->level = level; + __entry->nlevels = cur->bc_nlevels; + __entry->ptr = cur->bc_ptrs[level]; + __entry->daddr = bp ? bp->b_bn : -1; + ), + TP_printk("dev %d:%d btnum %d level %d/%d ptr %d daddr 0x%llx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->btnum, + __entry->level, + __entry->nlevels, + __entry->ptr, + (unsigned long long)__entry->daddr) +) + +#define DEFINE_BTREE_CUR_EVENT(name) \ +DEFINE_EVENT(xfs_btree_cur_class, name, \ + TP_PROTO(struct xfs_btree_cur *cur, int level, struct xfs_buf *bp), \ + TP_ARGS(cur, level, bp)) +DEFINE_BTREE_CUR_EVENT(xfs_btree_updkeys); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH -- cgit v1.2.3 From 105f7d83db4f82ce170893eaaca946754e38541f Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:10:21 +1000 Subject: xfs: introduce interval queries on btrees Create a function to enable querying of btree records mapping to a range of keys. This will be used in subsequent patches to allow querying the reverse mapping btree to find the extents mapped to a range of physical blocks, though the generic code can be used for any range query. The overlapped query range function needs to use the btree get_block helper because the root block could be an inode, in which case bc_bufs[nlevels-1] will be NULL. Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 264 ++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_btree.h | 22 +++- fs/xfs/xfs_trace.h | 1 + 3 files changed, 282 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 5c1b35a23a3d..d79dcb08ed18 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -4521,3 +4521,267 @@ xfs_btree_compute_maxlevels( maxblocks = (maxblocks + limits[1] - 1) / limits[1]; return level; } + +/* + * Query a regular btree for all records overlapping a given interval. + * Start with a LE lookup of the key of low_rec and return all records + * until we find a record with a key greater than the key of high_rec. + */ +STATIC int +xfs_btree_simple_query_range( + struct xfs_btree_cur *cur, + union xfs_btree_key *low_key, + union xfs_btree_key *high_key, + xfs_btree_query_range_fn fn, + void *priv) +{ + union xfs_btree_rec *recp; + union xfs_btree_key rec_key; + __int64_t diff; + int stat; + bool firstrec = true; + int error; + + ASSERT(cur->bc_ops->init_high_key_from_rec); + ASSERT(cur->bc_ops->diff_two_keys); + + /* + * Find the leftmost record. The btree cursor must be set + * to the low record used to generate low_key. + */ + stat = 0; + error = xfs_btree_lookup(cur, XFS_LOOKUP_LE, &stat); + if (error) + goto out; + + while (stat) { + /* Find the record. */ + error = xfs_btree_get_rec(cur, &recp, &stat); + if (error || !stat) + break; + cur->bc_ops->init_high_key_from_rec(&rec_key, recp); + + /* Skip if high_key(rec) < low_key. */ + if (firstrec) { + firstrec = false; + diff = cur->bc_ops->diff_two_keys(cur, low_key, + &rec_key); + if (diff > 0) + goto advloop; + } + + /* Stop if high_key < low_key(rec). */ + diff = cur->bc_ops->diff_two_keys(cur, &rec_key, high_key); + if (diff > 0) + break; + + /* Callback */ + error = fn(cur, recp, priv); + if (error < 0 || error == XFS_BTREE_QUERY_RANGE_ABORT) + break; + +advloop: + /* Move on to the next record. */ + error = xfs_btree_increment(cur, 0, &stat); + if (error) + break; + } + +out: + return error; +} + +/* + * Query an overlapped interval btree for all records overlapping a given + * interval. This function roughly follows the algorithm given in + * "Interval Trees" of _Introduction to Algorithms_, which is section + * 14.3 in the 2nd and 3rd editions. + * + * First, generate keys for the low and high records passed in. + * + * For any leaf node, generate the high and low keys for the record. + * If the record keys overlap with the query low/high keys, pass the + * record to the function iterator. + * + * For any internal node, compare the low and high keys of each + * pointer against the query low/high keys. If there's an overlap, + * follow the pointer. + * + * As an optimization, we stop scanning a block when we find a low key + * that is greater than the query's high key. + */ +STATIC int +xfs_btree_overlapped_query_range( + struct xfs_btree_cur *cur, + union xfs_btree_key *low_key, + union xfs_btree_key *high_key, + xfs_btree_query_range_fn fn, + void *priv) +{ + union xfs_btree_ptr ptr; + union xfs_btree_ptr *pp; + union xfs_btree_key rec_key; + union xfs_btree_key rec_hkey; + union xfs_btree_key *lkp; + union xfs_btree_key *hkp; + union xfs_btree_rec *recp; + struct xfs_btree_block *block; + __int64_t ldiff; + __int64_t hdiff; + int level; + struct xfs_buf *bp; + int i; + int error; + + /* Load the root of the btree. */ + level = cur->bc_nlevels - 1; + cur->bc_ops->init_ptr_from_cur(cur, &ptr); + error = xfs_btree_lookup_get_block(cur, level, &ptr, &block); + if (error) + return error; + xfs_btree_get_block(cur, level, &bp); + trace_xfs_btree_overlapped_query_range(cur, level, bp); +#ifdef DEBUG + error = xfs_btree_check_block(cur, block, level, bp); + if (error) + goto out; +#endif + cur->bc_ptrs[level] = 1; + + while (level < cur->bc_nlevels) { + block = xfs_btree_get_block(cur, level, &bp); + + /* End of node, pop back towards the root. */ + if (cur->bc_ptrs[level] > be16_to_cpu(block->bb_numrecs)) { +pop_up: + if (level < cur->bc_nlevels - 1) + cur->bc_ptrs[level + 1]++; + level++; + continue; + } + + if (level == 0) { + /* Handle a leaf node. */ + recp = xfs_btree_rec_addr(cur, cur->bc_ptrs[0], block); + + cur->bc_ops->init_high_key_from_rec(&rec_hkey, recp); + ldiff = cur->bc_ops->diff_two_keys(cur, &rec_hkey, + low_key); + + cur->bc_ops->init_key_from_rec(&rec_key, recp); + hdiff = cur->bc_ops->diff_two_keys(cur, high_key, + &rec_key); + + /* + * If (record's high key >= query's low key) and + * (query's high key >= record's low key), then + * this record overlaps the query range; callback. + */ + if (ldiff >= 0 && hdiff >= 0) { + error = fn(cur, recp, priv); + if (error < 0 || + error == XFS_BTREE_QUERY_RANGE_ABORT) + break; + } else if (hdiff < 0) { + /* Record is larger than high key; pop. */ + goto pop_up; + } + cur->bc_ptrs[level]++; + continue; + } + + /* Handle an internal node. */ + lkp = xfs_btree_key_addr(cur, cur->bc_ptrs[level], block); + hkp = xfs_btree_high_key_addr(cur, cur->bc_ptrs[level], block); + pp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[level], block); + + ldiff = cur->bc_ops->diff_two_keys(cur, hkp, low_key); + hdiff = cur->bc_ops->diff_two_keys(cur, high_key, lkp); + + /* + * If (pointer's high key >= query's low key) and + * (query's high key >= pointer's low key), then + * this record overlaps the query range; follow pointer. + */ + if (ldiff >= 0 && hdiff >= 0) { + level--; + error = xfs_btree_lookup_get_block(cur, level, pp, + &block); + if (error) + goto out; + xfs_btree_get_block(cur, level, &bp); + trace_xfs_btree_overlapped_query_range(cur, level, bp); +#ifdef DEBUG + error = xfs_btree_check_block(cur, block, level, bp); + if (error) + goto out; +#endif + cur->bc_ptrs[level] = 1; + continue; + } else if (hdiff < 0) { + /* The low key is larger than the upper range; pop. */ + goto pop_up; + } + cur->bc_ptrs[level]++; + } + +out: + /* + * If we don't end this function with the cursor pointing at a record + * block, a subsequent non-error cursor deletion will not release + * node-level buffers, causing a buffer leak. This is quite possible + * with a zero-results range query, so release the buffers if we + * failed to return any results. + */ + if (cur->bc_bufs[0] == NULL) { + for (i = 0; i < cur->bc_nlevels; i++) { + if (cur->bc_bufs[i]) { + xfs_trans_brelse(cur->bc_tp, cur->bc_bufs[i]); + cur->bc_bufs[i] = NULL; + cur->bc_ptrs[i] = 0; + cur->bc_ra[i] = 0; + } + } + } + + return error; +} + +/* + * Query a btree for all records overlapping a given interval of keys. The + * supplied function will be called with each record found; return one of the + * XFS_BTREE_QUERY_RANGE_{CONTINUE,ABORT} values or the usual negative error + * code. This function returns XFS_BTREE_QUERY_RANGE_ABORT, zero, or a + * negative error code. + */ +int +xfs_btree_query_range( + struct xfs_btree_cur *cur, + union xfs_btree_irec *low_rec, + union xfs_btree_irec *high_rec, + xfs_btree_query_range_fn fn, + void *priv) +{ + union xfs_btree_rec rec; + union xfs_btree_key low_key; + union xfs_btree_key high_key; + + /* Find the keys of both ends of the interval. */ + cur->bc_rec = *high_rec; + cur->bc_ops->init_rec_from_cur(cur, &rec); + cur->bc_ops->init_key_from_rec(&high_key, &rec); + + cur->bc_rec = *low_rec; + cur->bc_ops->init_rec_from_cur(cur, &rec); + cur->bc_ops->init_key_from_rec(&low_key, &rec); + + /* Enforce low key < high key. */ + if (cur->bc_ops->diff_two_keys(cur, &low_key, &high_key) > 0) + return -EINVAL; + + if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + return xfs_btree_simple_query_range(cur, &low_key, + &high_key, fn, priv); + return xfs_btree_overlapped_query_range(cur, &low_key, &high_key, + fn, priv); +} diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index bce6daacb1f7..34c13616d255 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -227,6 +227,12 @@ struct xfs_btree_ops { #define LASTREC_DELREC 2 +union xfs_btree_irec { + struct xfs_alloc_rec_incore a; + struct xfs_bmbt_irec b; + struct xfs_inobt_rec_incore i; +}; + /* * Btree cursor structure. * This collects all information needed by the btree code in one place. @@ -237,11 +243,7 @@ typedef struct xfs_btree_cur struct xfs_mount *bc_mp; /* file system mount struct */ const struct xfs_btree_ops *bc_ops; uint bc_flags; /* btree features - below */ - union { - xfs_alloc_rec_incore_t a; - xfs_bmbt_irec_t b; - xfs_inobt_rec_incore_t i; - } bc_rec; /* current insert/search record value */ + union xfs_btree_irec bc_rec; /* current insert/search record value */ struct xfs_buf *bc_bufs[XFS_BTREE_MAXLEVELS]; /* buf ptr per level */ int bc_ptrs[XFS_BTREE_MAXLEVELS]; /* key/record # */ __uint8_t bc_ra[XFS_BTREE_MAXLEVELS]; /* readahead bits */ @@ -524,4 +526,14 @@ void xfs_btree_get_node_keys_overlapped(struct xfs_btree_cur *cur, struct xfs_btree_block *block, union xfs_btree_key *key); int xfs_btree_update_keys_overlapped(struct xfs_btree_cur *cur, int level); +/* return codes */ +#define XFS_BTREE_QUERY_RANGE_CONTINUE 0 /* keep iterating */ +#define XFS_BTREE_QUERY_RANGE_ABORT 1 /* stop iterating */ +typedef int (*xfs_btree_query_range_fn)(struct xfs_btree_cur *cur, + union xfs_btree_rec *rec, void *priv); + +int xfs_btree_query_range(struct xfs_btree_cur *cur, + union xfs_btree_irec *low_rec, union xfs_btree_irec *high_rec, + xfs_btree_query_range_fn fn, void *priv); + #endif /* __XFS_BTREE_H__ */ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 8fb59e644f80..a5862680db31 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2220,6 +2220,7 @@ DEFINE_EVENT(xfs_btree_cur_class, name, \ TP_PROTO(struct xfs_btree_cur *cur, int level, struct xfs_buf *bp), \ TP_ARGS(cur, level, bp)) DEFINE_BTREE_CUR_EVENT(xfs_btree_updkeys); +DEFINE_BTREE_CUR_EVENT(xfs_btree_overlapped_query_range); #endif /* _TRACE_XFS_H */ -- cgit v1.2.3 From 28a89567b8bd95f42c17822d276cccb5b085810d Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:10:55 +1000 Subject: xfs: refactor btree owner change into a separate visit-blocks function Refactor the btree_change_owner function into a more generic apparatus which visits all blocks in a btree. We'll use this in a subsequent patch for counting btree blocks for AG reservations. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 141 ++++++++++++++++++++++++++++++---------------- fs/xfs/libxfs/xfs_btree.h | 5 ++ 2 files changed, 96 insertions(+), 50 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index d79dcb08ed18..b5be7f41c856 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -4328,6 +4328,81 @@ xfs_btree_get_rec( return 0; } +/* Visit a block in a btree. */ +STATIC int +xfs_btree_visit_block( + struct xfs_btree_cur *cur, + int level, + xfs_btree_visit_blocks_fn fn, + void *data) +{ + struct xfs_btree_block *block; + struct xfs_buf *bp; + union xfs_btree_ptr rptr; + int error; + + /* do right sibling readahead */ + xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA); + block = xfs_btree_get_block(cur, level, &bp); + + /* process the block */ + error = fn(cur, level, data); + if (error) + return error; + + /* now read rh sibling block for next iteration */ + xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB); + if (xfs_btree_ptr_is_null(cur, &rptr)) + return -ENOENT; + + return xfs_btree_lookup_get_block(cur, level, &rptr, &block); +} + + +/* Visit every block in a btree. */ +int +xfs_btree_visit_blocks( + struct xfs_btree_cur *cur, + xfs_btree_visit_blocks_fn fn, + void *data) +{ + union xfs_btree_ptr lptr; + int level; + struct xfs_btree_block *block = NULL; + int error = 0; + + cur->bc_ops->init_ptr_from_cur(cur, &lptr); + + /* for each level */ + for (level = cur->bc_nlevels - 1; level >= 0; level--) { + /* grab the left hand block */ + error = xfs_btree_lookup_get_block(cur, level, &lptr, &block); + if (error) + return error; + + /* readahead the left most block for the next level down */ + if (level > 0) { + union xfs_btree_ptr *ptr; + + ptr = xfs_btree_ptr_addr(cur, 1, block); + xfs_btree_readahead_ptr(cur, ptr, 1); + + /* save for the next iteration of the loop */ + lptr = *ptr; + } + + /* for each buffer in the level */ + do { + error = xfs_btree_visit_block(cur, level, fn, data); + } while (!error); + + if (error != -ENOENT) + return error; + } + + return 0; +} + /* * Change the owner of a btree. * @@ -4352,26 +4427,27 @@ xfs_btree_get_rec( * just queue the modified buffer as delayed write buffer so the transaction * recovery completion writes the changes to disk. */ +struct xfs_btree_block_change_owner_info { + __uint64_t new_owner; + struct list_head *buffer_list; +}; + static int xfs_btree_block_change_owner( struct xfs_btree_cur *cur, int level, - __uint64_t new_owner, - struct list_head *buffer_list) + void *data) { + struct xfs_btree_block_change_owner_info *bbcoi = data; struct xfs_btree_block *block; struct xfs_buf *bp; - union xfs_btree_ptr rptr; - - /* do right sibling readahead */ - xfs_btree_readahead(cur, level, XFS_BTCUR_RIGHTRA); /* modify the owner */ block = xfs_btree_get_block(cur, level, &bp); if (cur->bc_flags & XFS_BTREE_LONG_PTRS) - block->bb_u.l.bb_owner = cpu_to_be64(new_owner); + block->bb_u.l.bb_owner = cpu_to_be64(bbcoi->new_owner); else - block->bb_u.s.bb_owner = cpu_to_be32(new_owner); + block->bb_u.s.bb_owner = cpu_to_be32(bbcoi->new_owner); /* * If the block is a root block hosted in an inode, we might not have a @@ -4385,19 +4461,14 @@ xfs_btree_block_change_owner( xfs_trans_ordered_buf(cur->bc_tp, bp); xfs_btree_log_block(cur, bp, XFS_BB_OWNER); } else { - xfs_buf_delwri_queue(bp, buffer_list); + xfs_buf_delwri_queue(bp, bbcoi->buffer_list); } } else { ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); ASSERT(level == cur->bc_nlevels - 1); } - /* now read rh sibling block for next iteration */ - xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB); - if (xfs_btree_ptr_is_null(cur, &rptr)) - return -ENOENT; - - return xfs_btree_lookup_get_block(cur, level, &rptr, &block); + return 0; } int @@ -4406,43 +4477,13 @@ xfs_btree_change_owner( __uint64_t new_owner, struct list_head *buffer_list) { - union xfs_btree_ptr lptr; - int level; - struct xfs_btree_block *block = NULL; - int error = 0; + struct xfs_btree_block_change_owner_info bbcoi; - cur->bc_ops->init_ptr_from_cur(cur, &lptr); + bbcoi.new_owner = new_owner; + bbcoi.buffer_list = buffer_list; - /* for each level */ - for (level = cur->bc_nlevels - 1; level >= 0; level--) { - /* grab the left hand block */ - error = xfs_btree_lookup_get_block(cur, level, &lptr, &block); - if (error) - return error; - - /* readahead the left most block for the next level down */ - if (level > 0) { - union xfs_btree_ptr *ptr; - - ptr = xfs_btree_ptr_addr(cur, 1, block); - xfs_btree_readahead_ptr(cur, ptr, 1); - - /* save for the next iteration of the loop */ - lptr = *ptr; - } - - /* for each buffer in the level */ - do { - error = xfs_btree_block_change_owner(cur, level, - new_owner, - buffer_list); - } while (!error); - - if (error != -ENOENT) - return error; - } - - return 0; + return xfs_btree_visit_blocks(cur, xfs_btree_block_change_owner, + &bbcoi); } /** diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 34c13616d255..69f564029f56 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -536,4 +536,9 @@ int xfs_btree_query_range(struct xfs_btree_cur *cur, union xfs_btree_irec *low_rec, union xfs_btree_irec *high_rec, xfs_btree_query_range_fn fn, void *priv); +typedef int (*xfs_btree_visit_blocks_fn)(struct xfs_btree_cur *cur, int level, + void *data); +int xfs_btree_visit_blocks(struct xfs_btree_cur *cur, + xfs_btree_visit_blocks_fn fn, void *data); + #endif /* __XFS_BTREE_H__ */ -- cgit v1.2.3 From 4e0cc29b91a8056f902f0974b49084b07491905f Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:12:25 +1000 Subject: xfs: move deferred operations into a separate file All the code around struct xfs_bmap_free basically implements a deferred operation framework through which we can roll transactions (to unlock buffers and avoid violating lock order rules) while managing all the necessary log redo items. Previously we only used this code to free extents after some sort of mapping operation, but with the advent of rmap and reflink, we suddenly need to do more than that. With that in mind, xfs_bmap_free really becomes a deferred ops control structure. Rename the structure and move the deferred ops into their own file to avoid further bloating of the bmap code. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/Makefile | 1 + fs/xfs/libxfs/xfs_defer.c | 444 ++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_defer.h | 95 ++++++++++ 3 files changed, 540 insertions(+) create mode 100644 fs/xfs/libxfs/xfs_defer.c create mode 100644 fs/xfs/libxfs/xfs_defer.h (limited to 'fs') diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 3542d94fddce..01857b096cae 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -39,6 +39,7 @@ xfs-y += $(addprefix libxfs/, \ xfs_btree.o \ xfs_da_btree.o \ xfs_da_format.o \ + xfs_defer.o \ xfs_dir2.o \ xfs_dir2_block.o \ xfs_dir2_data.o \ diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c new file mode 100644 index 000000000000..2809db75e0d3 --- /dev/null +++ b/fs/xfs/libxfs/xfs_defer.c @@ -0,0 +1,444 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_bit.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_trans.h" +#include "xfs_trace.h" + +/* + * Deferred Operations in XFS + * + * Due to the way locking rules work in XFS, certain transactions (block + * mapping and unmapping, typically) have permanent reservations so that + * we can roll the transaction to adhere to AG locking order rules and + * to unlock buffers between metadata updates. Prior to rmap/reflink, + * the mapping code had a mechanism to perform these deferrals for + * extents that were going to be freed; this code makes that facility + * more generic. + * + * When adding the reverse mapping and reflink features, it became + * necessary to perform complex remapping multi-transactions to comply + * with AG locking order rules, and to be able to spread a single + * refcount update operation (an operation on an n-block extent can + * update as many as n records!) among multiple transactions. XFS can + * roll a transaction to facilitate this, but using this facility + * requires us to log "intent" items in case log recovery needs to + * redo the operation, and to log "done" items to indicate that redo + * is not necessary. + * + * Deferred work is tracked in xfs_defer_pending items. Each pending + * item tracks one type of deferred work. Incoming work items (which + * have not yet had an intent logged) are attached to a pending item + * on the dop_intake list, where they wait for the caller to finish + * the deferred operations. + * + * Finishing a set of deferred operations is an involved process. To + * start, we define "rolling a deferred-op transaction" as follows: + * + * > For each xfs_defer_pending item on the dop_intake list, + * - Sort the work items in AG order. XFS locking + * order rules require us to lock buffers in AG order. + * - Create a log intent item for that type. + * - Attach it to the pending item. + * - Move the pending item from the dop_intake list to the + * dop_pending list. + * > Roll the transaction. + * + * NOTE: To avoid exceeding the transaction reservation, we limit the + * number of items that we attach to a given xfs_defer_pending. + * + * The actual finishing process looks like this: + * + * > For each xfs_defer_pending in the dop_pending list, + * - Roll the deferred-op transaction as above. + * - Create a log done item for that type, and attach it to the + * log intent item. + * - For each work item attached to the log intent item, + * * Perform the described action. + * * Attach the work item to the log done item. + * + * The key here is that we must log an intent item for all pending + * work items every time we roll the transaction, and that we must log + * a done item as soon as the work is completed. With this mechanism + * we can perform complex remapping operations, chaining intent items + * as needed. + * + * This is an example of remapping the extent (E, E+B) into file X at + * offset A and dealing with the extent (C, C+B) already being mapped + * there: + * +-------------------------------------------------+ + * | Unmap file X startblock C offset A length B | t0 + * | Intent to reduce refcount for extent (C, B) | + * | Intent to remove rmap (X, C, A, B) | + * | Intent to free extent (D, 1) (bmbt block) | + * | Intent to map (X, A, B) at startblock E | + * +-------------------------------------------------+ + * | Map file X startblock E offset A length B | t1 + * | Done mapping (X, E, A, B) | + * | Intent to increase refcount for extent (E, B) | + * | Intent to add rmap (X, E, A, B) | + * +-------------------------------------------------+ + * | Reduce refcount for extent (C, B) | t2 + * | Done reducing refcount for extent (C, B) | + * | Increase refcount for extent (E, B) | + * | Done increasing refcount for extent (E, B) | + * | Intent to free extent (C, B) | + * | Intent to free extent (F, 1) (refcountbt block) | + * | Intent to remove rmap (F, 1, REFC) | + * +-------------------------------------------------+ + * | Remove rmap (X, C, A, B) | t3 + * | Done removing rmap (X, C, A, B) | + * | Add rmap (X, E, A, B) | + * | Done adding rmap (X, E, A, B) | + * | Remove rmap (F, 1, REFC) | + * | Done removing rmap (F, 1, REFC) | + * +-------------------------------------------------+ + * | Free extent (C, B) | t4 + * | Done freeing extent (C, B) | + * | Free extent (D, 1) | + * | Done freeing extent (D, 1) | + * | Free extent (F, 1) | + * | Done freeing extent (F, 1) | + * +-------------------------------------------------+ + * + * If we should crash before t2 commits, log recovery replays + * the following intent items: + * + * - Intent to reduce refcount for extent (C, B) + * - Intent to remove rmap (X, C, A, B) + * - Intent to free extent (D, 1) (bmbt block) + * - Intent to increase refcount for extent (E, B) + * - Intent to add rmap (X, E, A, B) + * + * In the process of recovering, it should also generate and take care + * of these intent items: + * + * - Intent to free extent (C, B) + * - Intent to free extent (F, 1) (refcountbt block) + * - Intent to remove rmap (F, 1, REFC) + */ + +static const struct xfs_defer_op_type *defer_op_types[XFS_DEFER_OPS_TYPE_MAX]; + +/* + * For each pending item in the intake list, log its intent item and the + * associated extents, then add the entire intake list to the end of + * the pending list. + */ +STATIC void +xfs_defer_intake_work( + struct xfs_trans *tp, + struct xfs_defer_ops *dop) +{ + struct list_head *li; + struct xfs_defer_pending *dfp; + + list_for_each_entry(dfp, &dop->dop_intake, dfp_list) { + dfp->dfp_intent = dfp->dfp_type->create_intent(tp, + dfp->dfp_count); + list_sort(tp->t_mountp, &dfp->dfp_work, + dfp->dfp_type->diff_items); + list_for_each(li, &dfp->dfp_work) + dfp->dfp_type->log_item(tp, dfp->dfp_intent, li); + } + + list_splice_tail_init(&dop->dop_intake, &dop->dop_pending); +} + +/* Abort all the intents that were committed. */ +STATIC void +xfs_defer_trans_abort( + struct xfs_trans *tp, + struct xfs_defer_ops *dop, + int error) +{ + struct xfs_defer_pending *dfp; + + /* + * If the transaction was committed, drop the intent reference + * since we're bailing out of here. The other reference is + * dropped when the intent hits the AIL. If the transaction + * was not committed, the intent is freed by the intent item + * unlock handler on abort. + */ + if (!dop->dop_committed) + return; + + /* Abort intent items. */ + list_for_each_entry(dfp, &dop->dop_pending, dfp_list) { + if (dfp->dfp_committed) + dfp->dfp_type->abort_intent(dfp->dfp_intent); + } + + /* Shut down FS. */ + xfs_force_shutdown(tp->t_mountp, (error == -EFSCORRUPTED) ? + SHUTDOWN_CORRUPT_INCORE : SHUTDOWN_META_IO_ERROR); +} + +/* Roll a transaction so we can do some deferred op processing. */ +STATIC int +xfs_defer_trans_roll( + struct xfs_trans **tp, + struct xfs_defer_ops *dop, + struct xfs_inode *ip) +{ + int i; + int error; + + /* Log all the joined inodes except the one we passed in. */ + for (i = 0; i < XFS_DEFER_OPS_NR_INODES && dop->dop_inodes[i]; i++) { + if (dop->dop_inodes[i] == ip) + continue; + xfs_trans_log_inode(*tp, dop->dop_inodes[i], XFS_ILOG_CORE); + } + + /* Roll the transaction. */ + error = xfs_trans_roll(tp, ip); + if (error) { + xfs_defer_trans_abort(*tp, dop, error); + return error; + } + dop->dop_committed = true; + + /* Rejoin the joined inodes except the one we passed in. */ + for (i = 0; i < XFS_DEFER_OPS_NR_INODES && dop->dop_inodes[i]; i++) { + if (dop->dop_inodes[i] == ip) + continue; + xfs_trans_ijoin(*tp, dop->dop_inodes[i], 0); + } + + return error; +} + +/* Do we have any work items to finish? */ +bool +xfs_defer_has_unfinished_work( + struct xfs_defer_ops *dop) +{ + return !list_empty(&dop->dop_pending) || !list_empty(&dop->dop_intake); +} + +/* + * Add this inode to the deferred op. Each joined inode is relogged + * each time we roll the transaction, in addition to any inode passed + * to xfs_defer_finish(). + */ +int +xfs_defer_join( + struct xfs_defer_ops *dop, + struct xfs_inode *ip) +{ + int i; + + for (i = 0; i < XFS_DEFER_OPS_NR_INODES; i++) { + if (dop->dop_inodes[i] == ip) + return 0; + else if (dop->dop_inodes[i] == NULL) { + dop->dop_inodes[i] = ip; + return 0; + } + } + + return -EFSCORRUPTED; +} + +/* + * Finish all the pending work. This involves logging intent items for + * any work items that wandered in since the last transaction roll (if + * one has even happened), rolling the transaction, and finishing the + * work items in the first item on the logged-and-pending list. + * + * If an inode is provided, relog it to the new transaction. + */ +int +xfs_defer_finish( + struct xfs_trans **tp, + struct xfs_defer_ops *dop, + struct xfs_inode *ip) +{ + struct xfs_defer_pending *dfp; + struct list_head *li; + struct list_head *n; + void *done_item = NULL; + void *state; + int error = 0; + void (*cleanup_fn)(struct xfs_trans *, void *, int); + + ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES); + + /* Until we run out of pending work to finish... */ + while (xfs_defer_has_unfinished_work(dop)) { + /* Log intents for work items sitting in the intake. */ + xfs_defer_intake_work(*tp, dop); + + /* Roll the transaction. */ + error = xfs_defer_trans_roll(tp, dop, ip); + if (error) + goto out; + + /* Mark all pending intents as committed. */ + list_for_each_entry_reverse(dfp, &dop->dop_pending, dfp_list) { + if (dfp->dfp_committed) + break; + dfp->dfp_committed = true; + } + + /* Log an intent-done item for the first pending item. */ + dfp = list_first_entry(&dop->dop_pending, + struct xfs_defer_pending, dfp_list); + done_item = dfp->dfp_type->create_done(*tp, dfp->dfp_intent, + dfp->dfp_count); + cleanup_fn = dfp->dfp_type->finish_cleanup; + + /* Finish the work items. */ + state = NULL; + list_for_each_safe(li, n, &dfp->dfp_work) { + list_del(li); + dfp->dfp_count--; + error = dfp->dfp_type->finish_item(*tp, dop, li, + done_item, &state); + if (error) { + /* + * Clean up after ourselves and jump out. + * xfs_defer_cancel will take care of freeing + * all these lists and stuff. + */ + if (cleanup_fn) + cleanup_fn(*tp, state, error); + xfs_defer_trans_abort(*tp, dop, error); + goto out; + } + } + /* Done with the dfp, free it. */ + list_del(&dfp->dfp_list); + kmem_free(dfp); + + if (cleanup_fn) + cleanup_fn(*tp, state, error); + } + +out: + return error; +} + +/* + * Free up any items left in the list. + */ +void +xfs_defer_cancel( + struct xfs_defer_ops *dop) +{ + struct xfs_defer_pending *dfp; + struct xfs_defer_pending *pli; + struct list_head *pwi; + struct list_head *n; + + /* + * Free the pending items. Caller should already have arranged + * for the intent items to be released. + */ + list_for_each_entry_safe(dfp, pli, &dop->dop_intake, dfp_list) { + list_del(&dfp->dfp_list); + list_for_each_safe(pwi, n, &dfp->dfp_work) { + list_del(pwi); + dfp->dfp_count--; + dfp->dfp_type->cancel_item(pwi); + } + ASSERT(dfp->dfp_count == 0); + kmem_free(dfp); + } + list_for_each_entry_safe(dfp, pli, &dop->dop_pending, dfp_list) { + list_del(&dfp->dfp_list); + list_for_each_safe(pwi, n, &dfp->dfp_work) { + list_del(pwi); + dfp->dfp_count--; + dfp->dfp_type->cancel_item(pwi); + } + ASSERT(dfp->dfp_count == 0); + kmem_free(dfp); + } +} + +/* Add an item for later deferred processing. */ +void +xfs_defer_add( + struct xfs_defer_ops *dop, + enum xfs_defer_ops_type type, + struct list_head *li) +{ + struct xfs_defer_pending *dfp = NULL; + + /* + * Add the item to a pending item at the end of the intake list. + * If the last pending item has the same type, reuse it. Else, + * create a new pending item at the end of the intake list. + */ + if (!list_empty(&dop->dop_intake)) { + dfp = list_last_entry(&dop->dop_intake, + struct xfs_defer_pending, dfp_list); + if (dfp->dfp_type->type != type || + (dfp->dfp_type->max_items && + dfp->dfp_count >= dfp->dfp_type->max_items)) + dfp = NULL; + } + if (!dfp) { + dfp = kmem_alloc(sizeof(struct xfs_defer_pending), + KM_SLEEP | KM_NOFS); + dfp->dfp_type = defer_op_types[type]; + dfp->dfp_committed = false; + dfp->dfp_intent = NULL; + dfp->dfp_count = 0; + INIT_LIST_HEAD(&dfp->dfp_work); + list_add_tail(&dfp->dfp_list, &dop->dop_intake); + } + + list_add_tail(li, &dfp->dfp_work); + dfp->dfp_count++; +} + +/* Initialize a deferred operation list. */ +void +xfs_defer_init_op_type( + const struct xfs_defer_op_type *type) +{ + defer_op_types[type->type] = type; +} + +/* Initialize a deferred operation. */ +void +xfs_defer_init( + struct xfs_defer_ops *dop, + xfs_fsblock_t *fbp) +{ + dop->dop_committed = false; + dop->dop_low = false; + memset(&dop->dop_inodes, 0, sizeof(dop->dop_inodes)); + *fbp = NULLFSBLOCK; + INIT_LIST_HEAD(&dop->dop_intake); + INIT_LIST_HEAD(&dop->dop_pending); +} diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h new file mode 100644 index 000000000000..a227bd2cece8 --- /dev/null +++ b/fs/xfs/libxfs/xfs_defer.h @@ -0,0 +1,95 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_DEFER_H__ +#define __XFS_DEFER_H__ + +struct xfs_defer_op_type; + +/* + * Save a log intent item and a list of extents, so that we can replay + * whatever action had to happen to the extent list and file the log done + * item. + */ +struct xfs_defer_pending { + const struct xfs_defer_op_type *dfp_type; /* function pointers */ + struct list_head dfp_list; /* pending items */ + bool dfp_committed; /* committed trans? */ + void *dfp_intent; /* log intent item */ + struct list_head dfp_work; /* work items */ + unsigned int dfp_count; /* # extent items */ +}; + +/* + * Header for deferred operation list. + * + * dop_low is used by the allocator to activate the lowspace algorithm - + * when free space is running low the extent allocator may choose to + * allocate an extent from an AG without leaving sufficient space for + * a btree split when inserting the new extent. In this case the allocator + * will enable the lowspace algorithm which is supposed to allow further + * allocations (such as btree splits and newroots) to allocate from + * sequential AGs. In order to avoid locking AGs out of order the lowspace + * algorithm will start searching for free space from AG 0. If the correct + * transaction reservations have been made then this algorithm will eventually + * find all the space it needs. + */ +enum xfs_defer_ops_type { + XFS_DEFER_OPS_TYPE_MAX, +}; + +#define XFS_DEFER_OPS_NR_INODES 2 /* join up to two inodes */ + +struct xfs_defer_ops { + bool dop_committed; /* did any trans commit? */ + bool dop_low; /* alloc in low mode */ + struct list_head dop_intake; /* unlogged pending work */ + struct list_head dop_pending; /* logged pending work */ + + /* relog these inodes with each roll */ + struct xfs_inode *dop_inodes[XFS_DEFER_OPS_NR_INODES]; +}; + +void xfs_defer_add(struct xfs_defer_ops *dop, enum xfs_defer_ops_type type, + struct list_head *h); +int xfs_defer_finish(struct xfs_trans **tp, struct xfs_defer_ops *dop, + struct xfs_inode *ip); +void xfs_defer_cancel(struct xfs_defer_ops *dop); +void xfs_defer_init(struct xfs_defer_ops *dop, xfs_fsblock_t *fbp); +bool xfs_defer_has_unfinished_work(struct xfs_defer_ops *dop); +int xfs_defer_join(struct xfs_defer_ops *dop, struct xfs_inode *ip); + +/* Description of a deferred type. */ +struct xfs_defer_op_type { + enum xfs_defer_ops_type type; + unsigned int max_items; + void (*abort_intent)(void *); + void *(*create_done)(struct xfs_trans *, void *, unsigned int); + int (*finish_item)(struct xfs_trans *, struct xfs_defer_ops *, + struct list_head *, void *, void **); + void (*finish_cleanup)(struct xfs_trans *, void *, int); + void (*cancel_item)(struct list_head *); + int (*diff_items)(void *, struct list_head *, struct list_head *); + void *(*create_intent)(struct xfs_trans *, uint); + void (*log_item)(struct xfs_trans *, void *, struct list_head *); +}; + +void xfs_defer_init_op_type(const struct xfs_defer_op_type *type); + +#endif /* __XFS_DEFER_H__ */ -- cgit v1.2.3 From 3cd48abcc1f76d6cd5ce61f3540801849a6c82e0 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:13:02 +1000 Subject: xfs: add tracepoints for the deferred ops mechanism Add tracepoints for the internals of the deferred ops mechanism and tracepoint classes for clients of the dops, to make debugging easier. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_defer.c | 19 +++++ fs/xfs/xfs_trace.c | 1 + fs/xfs/xfs_trace.h | 198 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 218 insertions(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c index 2809db75e0d3..054a2032fdb3 100644 --- a/fs/xfs/libxfs/xfs_defer.c +++ b/fs/xfs/libxfs/xfs_defer.c @@ -159,6 +159,7 @@ xfs_defer_intake_work( struct xfs_defer_pending *dfp; list_for_each_entry(dfp, &dop->dop_intake, dfp_list) { + trace_xfs_defer_intake_work(tp->t_mountp, dfp); dfp->dfp_intent = dfp->dfp_type->create_intent(tp, dfp->dfp_count); list_sort(tp->t_mountp, &dfp->dfp_work, @@ -179,6 +180,7 @@ xfs_defer_trans_abort( { struct xfs_defer_pending *dfp; + trace_xfs_defer_trans_abort(tp->t_mountp, dop); /* * If the transaction was committed, drop the intent reference * since we're bailing out of here. The other reference is @@ -191,6 +193,7 @@ xfs_defer_trans_abort( /* Abort intent items. */ list_for_each_entry(dfp, &dop->dop_pending, dfp_list) { + trace_xfs_defer_pending_abort(tp->t_mountp, dfp); if (dfp->dfp_committed) dfp->dfp_type->abort_intent(dfp->dfp_intent); } @@ -217,9 +220,12 @@ xfs_defer_trans_roll( xfs_trans_log_inode(*tp, dop->dop_inodes[i], XFS_ILOG_CORE); } + trace_xfs_defer_trans_roll((*tp)->t_mountp, dop); + /* Roll the transaction. */ error = xfs_trans_roll(tp, ip); if (error) { + trace_xfs_defer_trans_roll_error((*tp)->t_mountp, dop, error); xfs_defer_trans_abort(*tp, dop, error); return error; } @@ -291,6 +297,8 @@ xfs_defer_finish( ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES); + trace_xfs_defer_finish((*tp)->t_mountp, dop); + /* Until we run out of pending work to finish... */ while (xfs_defer_has_unfinished_work(dop)) { /* Log intents for work items sitting in the intake. */ @@ -305,12 +313,14 @@ xfs_defer_finish( list_for_each_entry_reverse(dfp, &dop->dop_pending, dfp_list) { if (dfp->dfp_committed) break; + trace_xfs_defer_pending_commit((*tp)->t_mountp, dfp); dfp->dfp_committed = true; } /* Log an intent-done item for the first pending item. */ dfp = list_first_entry(&dop->dop_pending, struct xfs_defer_pending, dfp_list); + trace_xfs_defer_pending_finish((*tp)->t_mountp, dfp); done_item = dfp->dfp_type->create_done(*tp, dfp->dfp_intent, dfp->dfp_count); cleanup_fn = dfp->dfp_type->finish_cleanup; @@ -343,6 +353,10 @@ xfs_defer_finish( } out: + if (error) + trace_xfs_defer_finish_error((*tp)->t_mountp, dop, error); + else + trace_xfs_defer_finish_done((*tp)->t_mountp, dop); return error; } @@ -358,11 +372,14 @@ xfs_defer_cancel( struct list_head *pwi; struct list_head *n; + trace_xfs_defer_cancel(NULL, dop); + /* * Free the pending items. Caller should already have arranged * for the intent items to be released. */ list_for_each_entry_safe(dfp, pli, &dop->dop_intake, dfp_list) { + trace_xfs_defer_intake_cancel(NULL, dfp); list_del(&dfp->dfp_list); list_for_each_safe(pwi, n, &dfp->dfp_work) { list_del(pwi); @@ -373,6 +390,7 @@ xfs_defer_cancel( kmem_free(dfp); } list_for_each_entry_safe(dfp, pli, &dop->dop_pending, dfp_list) { + trace_xfs_defer_pending_cancel(NULL, dfp); list_del(&dfp->dfp_list); list_for_each_safe(pwi, n, &dfp->dfp_work) { list_del(pwi); @@ -441,4 +459,5 @@ xfs_defer_init( *fbp = NULLFSBLOCK; INIT_LIST_HEAD(&dop->dop_intake); INIT_LIST_HEAD(&dop->dop_pending); + trace_xfs_defer_init(NULL, dop); } diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c index 13a029806805..3971527c9a07 100644 --- a/fs/xfs/xfs_trace.c +++ b/fs/xfs/xfs_trace.c @@ -22,6 +22,7 @@ #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_inode.h" #include "xfs_btree.h" diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index a5862680db31..a45b030d4592 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2222,6 +2222,204 @@ DEFINE_EVENT(xfs_btree_cur_class, name, \ DEFINE_BTREE_CUR_EVENT(xfs_btree_updkeys); DEFINE_BTREE_CUR_EVENT(xfs_btree_overlapped_query_range); +/* deferred ops */ +struct xfs_defer_pending; +struct xfs_defer_intake; +struct xfs_defer_ops; + +DECLARE_EVENT_CLASS(xfs_defer_class, + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop), + TP_ARGS(mp, dop), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(void *, dop) + __field(bool, committed) + __field(bool, low) + ), + TP_fast_assign( + __entry->dev = mp ? mp->m_super->s_dev : 0; + __entry->dop = dop; + __entry->committed = dop->dop_committed; + __entry->low = dop->dop_low; + ), + TP_printk("dev %d:%d ops %p committed %d low %d\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->dop, + __entry->committed, + __entry->low) +) +#define DEFINE_DEFER_EVENT(name) \ +DEFINE_EVENT(xfs_defer_class, name, \ + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop), \ + TP_ARGS(mp, dop)) + +DECLARE_EVENT_CLASS(xfs_defer_error_class, + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop, int error), + TP_ARGS(mp, dop, error), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(void *, dop) + __field(bool, committed) + __field(bool, low) + __field(int, error) + ), + TP_fast_assign( + __entry->dev = mp ? mp->m_super->s_dev : 0; + __entry->dop = dop; + __entry->committed = dop->dop_committed; + __entry->low = dop->dop_low; + __entry->error = error; + ), + TP_printk("dev %d:%d ops %p committed %d low %d err %d\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->dop, + __entry->committed, + __entry->low, + __entry->error) +) +#define DEFINE_DEFER_ERROR_EVENT(name) \ +DEFINE_EVENT(xfs_defer_error_class, name, \ + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_ops *dop, int error), \ + TP_ARGS(mp, dop, error)) + +DECLARE_EVENT_CLASS(xfs_defer_pending_class, + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp), + TP_ARGS(mp, dfp), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(int, type) + __field(void *, intent) + __field(bool, committed) + __field(int, nr) + ), + TP_fast_assign( + __entry->dev = mp ? mp->m_super->s_dev : 0; + __entry->type = dfp->dfp_type->type; + __entry->intent = dfp->dfp_intent; + __entry->committed = dfp->dfp_committed; + __entry->nr = dfp->dfp_count; + ), + TP_printk("dev %d:%d optype %d intent %p committed %d nr %d\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->type, + __entry->intent, + __entry->committed, + __entry->nr) +) +#define DEFINE_DEFER_PENDING_EVENT(name) \ +DEFINE_EVENT(xfs_defer_pending_class, name, \ + TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp), \ + TP_ARGS(mp, dfp)) + +DECLARE_EVENT_CLASS(xfs_phys_extent_deferred_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + int type, xfs_agblock_t agbno, xfs_extlen_t len), + TP_ARGS(mp, agno, type, agbno, len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(int, type) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->type = type; + __entry->agbno = agbno; + __entry->len = len; + ), + TP_printk("dev %d:%d op %d agno %u agbno %u len %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->type, + __entry->agno, + __entry->agbno, + __entry->len) +); +#define DEFINE_PHYS_EXTENT_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_phys_extent_deferred_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + int type, \ + xfs_agblock_t bno, \ + xfs_extlen_t len), \ + TP_ARGS(mp, agno, type, bno, len)) + +DECLARE_EVENT_CLASS(xfs_map_extent_deferred_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + int op, + xfs_agblock_t agbno, + xfs_ino_t ino, + int whichfork, + xfs_fileoff_t offset, + xfs_filblks_t len, + xfs_exntst_t state), + TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_ino_t, ino) + __field(xfs_agblock_t, agbno) + __field(int, whichfork) + __field(xfs_fileoff_t, l_loff) + __field(xfs_filblks_t, l_len) + __field(xfs_exntst_t, l_state) + __field(int, op) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->ino = ino; + __entry->agbno = agbno; + __entry->whichfork = whichfork; + __entry->l_loff = offset; + __entry->l_len = len; + __entry->l_state = state; + __entry->op = op; + ), + TP_printk("dev %d:%d op %d agno %u agbno %u owner %lld %s offset %llu len %llu state %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->op, + __entry->agno, + __entry->agbno, + __entry->ino, + __entry->whichfork == XFS_ATTR_FORK ? "attr" : "data", + __entry->l_loff, + __entry->l_len, + __entry->l_state) +); +#define DEFINE_MAP_EXTENT_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_map_extent_deferred_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + int op, \ + xfs_agblock_t agbno, \ + xfs_ino_t ino, \ + int whichfork, \ + xfs_fileoff_t offset, \ + xfs_filblks_t len, \ + xfs_exntst_t state), \ + TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state)) + +DEFINE_DEFER_EVENT(xfs_defer_init); +DEFINE_DEFER_EVENT(xfs_defer_cancel); +DEFINE_DEFER_EVENT(xfs_defer_trans_roll); +DEFINE_DEFER_EVENT(xfs_defer_trans_abort); +DEFINE_DEFER_EVENT(xfs_defer_finish); +DEFINE_DEFER_EVENT(xfs_defer_finish_done); + +DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error); +DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error); +DEFINE_DEFER_ERROR_EVENT(xfs_defer_op_finish_error); + +DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_work); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_cancel); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_commit); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_cancel); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); + +DEFINE_PHYS_EXTENT_DEFERRED_EVENT(xfs_defer_phys_extent); +DEFINE_MAP_EXTENT_DEFERRED_EVENT(xfs_defer_map_extent); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH -- cgit v1.2.3 From bba61cbf30bebc39c5227e131625de1521d27b18 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:13:47 +1000 Subject: xfs: clean up typedef usage in the EFI/EFD handling code Replace structure typedefs with struct xfs_foo_* in the EFI/EFD handling code in preparation to move it over to deferred ops. Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig Signed-off-by: Dave Chinner --- fs/xfs/xfs_trans.h | 6 +++--- fs/xfs/xfs_trans_extfree.c | 30 +++++++++++++++--------------- 2 files changed, 18 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 9b2b9fa89331..abf9aa31b2fa 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -210,12 +210,12 @@ void xfs_trans_ichgtime(struct xfs_trans *, struct xfs_inode *, int); void xfs_trans_ijoin(struct xfs_trans *, struct xfs_inode *, uint); void xfs_trans_log_buf(xfs_trans_t *, struct xfs_buf *, uint, uint); void xfs_trans_log_inode(xfs_trans_t *, struct xfs_inode *, uint); -struct xfs_efi_log_item *xfs_trans_get_efi(xfs_trans_t *, uint); -void xfs_trans_log_efi_extent(xfs_trans_t *, +struct xfs_efi_log_item *xfs_trans_get_efi(struct xfs_trans *, uint); +void xfs_trans_log_efi_extent(struct xfs_trans *, struct xfs_efi_log_item *, xfs_fsblock_t, xfs_extlen_t); -struct xfs_efd_log_item *xfs_trans_get_efd(xfs_trans_t *, +struct xfs_efd_log_item *xfs_trans_get_efd(struct xfs_trans *, struct xfs_efi_log_item *, uint); int xfs_trans_free_extent(struct xfs_trans *, diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index a96ae540eb62..380cc4631119 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -33,11 +33,11 @@ * caller must use all nextents extents, because we are not * flexible about this at all. */ -xfs_efi_log_item_t * -xfs_trans_get_efi(xfs_trans_t *tp, - uint nextents) +struct xfs_efi_log_item * +xfs_trans_get_efi(struct xfs_trans *tp, + uint nextents) { - xfs_efi_log_item_t *efip; + struct xfs_efi_log_item *efip; ASSERT(tp != NULL); ASSERT(nextents > 0); @@ -58,13 +58,13 @@ xfs_trans_get_efi(xfs_trans_t *tp, * be called once for each extent to be freed. */ void -xfs_trans_log_efi_extent(xfs_trans_t *tp, - xfs_efi_log_item_t *efip, - xfs_fsblock_t start_block, - xfs_extlen_t ext_len) +xfs_trans_log_efi_extent(struct xfs_trans *tp, + struct xfs_efi_log_item *efip, + xfs_fsblock_t start_block, + xfs_extlen_t ext_len) { - uint next_extent; - xfs_extent_t *extp; + uint next_extent; + struct xfs_extent *extp; tp->t_flags |= XFS_TRANS_DIRTY; efip->efi_item.li_desc->lid_flags |= XFS_LID_DIRTY; @@ -88,12 +88,12 @@ xfs_trans_log_efi_extent(xfs_trans_t *tp, * caller must use all nextents extents, because we are not * flexible about this at all. */ -xfs_efd_log_item_t * -xfs_trans_get_efd(xfs_trans_t *tp, - xfs_efi_log_item_t *efip, - uint nextents) +struct xfs_efd_log_item * +xfs_trans_get_efd(struct xfs_trans *tp, + struct xfs_efi_log_item *efip, + uint nextents) { - xfs_efd_log_item_t *efdp; + struct xfs_efd_log_item *efdp; ASSERT(tp != NULL); ASSERT(nextents > 0); -- cgit v1.2.3 From 9749fee83f38fca8dbe67161a033db22e3c4a2dd Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:14:35 +1000 Subject: xfs: enable the xfs_defer mechanism to process extents to free Connect the xfs_defer mechanism with the pieces that we'll need to handle deferred extent freeing. We'll wire up the existing code to our new deferred mechanism later. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_defer.h | 1 + fs/xfs/xfs_super.c | 2 + fs/xfs/xfs_trans.h | 2 + fs/xfs/xfs_trans_extfree.c | 110 +++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 115 insertions(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h index a227bd2cece8..b9b5a92ffa6c 100644 --- a/fs/xfs/libxfs/xfs_defer.h +++ b/fs/xfs/libxfs/xfs_defer.h @@ -51,6 +51,7 @@ struct xfs_defer_pending { * find all the space it needs. */ enum xfs_defer_ops_type { + XFS_DEFER_OPS_TYPE_FREE, XFS_DEFER_OPS_TYPE_MAX, }; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 0303f1005f88..7fc508fe8d40 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1854,6 +1854,8 @@ init_xfs_fs(void) printk(KERN_INFO XFS_VERSION_STRING " with " XFS_BUILD_OPTIONS " enabled\n"); + xfs_extent_free_init_defer_op(); + xfs_dir_startup(); error = xfs_init_zones(); diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index abf9aa31b2fa..746cf0832642 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -210,6 +210,8 @@ void xfs_trans_ichgtime(struct xfs_trans *, struct xfs_inode *, int); void xfs_trans_ijoin(struct xfs_trans *, struct xfs_inode *, uint); void xfs_trans_log_buf(xfs_trans_t *, struct xfs_buf *, uint, uint); void xfs_trans_log_inode(xfs_trans_t *, struct xfs_inode *, uint); + +void xfs_extent_free_init_defer_op(void); struct xfs_efi_log_item *xfs_trans_get_efi(struct xfs_trans *, uint); void xfs_trans_log_efi_extent(struct xfs_trans *, struct xfs_efi_log_item *, diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index 380cc4631119..a9f34ee52ab4 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -22,10 +22,12 @@ #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_trans.h" #include "xfs_trans_priv.h" #include "xfs_extfree_item.h" #include "xfs_alloc.h" +#include "xfs_bmap.h" /* * This routine is called to allocate an "extent free intention" @@ -145,3 +147,111 @@ xfs_trans_free_extent( return error; } + +/* Sort bmap items by AG. */ +static int +xfs_extent_free_diff_items( + void *priv, + struct list_head *a, + struct list_head *b) +{ + struct xfs_mount *mp = priv; + struct xfs_bmap_free_item *ra; + struct xfs_bmap_free_item *rb; + + ra = container_of(a, struct xfs_bmap_free_item, xbfi_list); + rb = container_of(b, struct xfs_bmap_free_item, xbfi_list); + return XFS_FSB_TO_AGNO(mp, ra->xbfi_startblock) - + XFS_FSB_TO_AGNO(mp, rb->xbfi_startblock); +} + +/* Get an EFI. */ +STATIC void * +xfs_extent_free_create_intent( + struct xfs_trans *tp, + unsigned int count) +{ + return xfs_trans_get_efi(tp, count); +} + +/* Log a free extent to the intent item. */ +STATIC void +xfs_extent_free_log_item( + struct xfs_trans *tp, + void *intent, + struct list_head *item) +{ + struct xfs_bmap_free_item *free; + + free = container_of(item, struct xfs_bmap_free_item, xbfi_list); + xfs_trans_log_efi_extent(tp, intent, free->xbfi_startblock, + free->xbfi_blockcount); +} + +/* Get an EFD so we can process all the free extents. */ +STATIC void * +xfs_extent_free_create_done( + struct xfs_trans *tp, + void *intent, + unsigned int count) +{ + return xfs_trans_get_efd(tp, intent, count); +} + +/* Process a free extent. */ +STATIC int +xfs_extent_free_finish_item( + struct xfs_trans *tp, + struct xfs_defer_ops *dop, + struct list_head *item, + void *done_item, + void **state) +{ + struct xfs_bmap_free_item *free; + int error; + + free = container_of(item, struct xfs_bmap_free_item, xbfi_list); + error = xfs_trans_free_extent(tp, done_item, + free->xbfi_startblock, + free->xbfi_blockcount); + kmem_free(free); + return error; +} + +/* Abort all pending EFIs. */ +STATIC void +xfs_extent_free_abort_intent( + void *intent) +{ + xfs_efi_release(intent); +} + +/* Cancel a free extent. */ +STATIC void +xfs_extent_free_cancel_item( + struct list_head *item) +{ + struct xfs_bmap_free_item *free; + + free = container_of(item, struct xfs_bmap_free_item, xbfi_list); + kmem_free(free); +} + +static const struct xfs_defer_op_type xfs_extent_free_defer_type = { + .type = XFS_DEFER_OPS_TYPE_FREE, + .max_items = XFS_EFI_MAX_FAST_EXTENTS, + .diff_items = xfs_extent_free_diff_items, + .create_intent = xfs_extent_free_create_intent, + .abort_intent = xfs_extent_free_abort_intent, + .log_item = xfs_extent_free_log_item, + .create_done = xfs_extent_free_create_done, + .finish_item = xfs_extent_free_finish_item, + .cancel_item = xfs_extent_free_cancel_item, +}; + +/* Register the deferred op type. */ +void +xfs_extent_free_init_defer_op(void) +{ + xfs_defer_init_op_type(&xfs_extent_free_defer_type); +} -- cgit v1.2.3 From 3ab78df2a59a485f479d26852a060acfd8c4ecd7 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:15:38 +1000 Subject: xfs: rework xfs_bmap_free callers to use xfs_defer_ops Restructure everything that used xfs_bmap_free to use xfs_defer_ops instead. For now we'll just remove the old symbols and play some cpp magic to make it work; in the next patch we'll actually rename everything. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 1 + fs/xfs/libxfs/xfs_attr.c | 1 + fs/xfs/libxfs/xfs_attr_remote.c | 1 + fs/xfs/libxfs/xfs_bmap.c | 55 +++++------------------- fs/xfs/libxfs/xfs_bmap.h | 32 -------------- fs/xfs/libxfs/xfs_bmap_btree.c | 5 ++- fs/xfs/libxfs/xfs_btree.c | 1 + fs/xfs/libxfs/xfs_defer.h | 7 ++++ fs/xfs/libxfs/xfs_dir2.c | 1 + fs/xfs/libxfs/xfs_ialloc.c | 1 + fs/xfs/libxfs/xfs_inode_buf.c | 1 + fs/xfs/libxfs/xfs_sb.c | 1 + fs/xfs/xfs_bmap_util.c | 92 +---------------------------------------- fs/xfs/xfs_bmap_util.h | 2 - fs/xfs/xfs_dquot.c | 1 + fs/xfs/xfs_filestream.c | 3 +- fs/xfs/xfs_fsops.c | 1 + fs/xfs/xfs_inode.c | 1 + fs/xfs/xfs_iomap.c | 1 + fs/xfs/xfs_log_recover.c | 1 + fs/xfs/xfs_mount.c | 1 + fs/xfs/xfs_rtalloc.c | 1 + fs/xfs/xfs_symlink.c | 1 + fs/xfs/xfs_trace.c | 1 + 24 files changed, 42 insertions(+), 171 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 88c26b827a2d..5993f87eee3e 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" #include "xfs_alloc_btree.h" diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index 4e126f41a0aa..79d3a309b50f 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_attr_sf.h" diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c index a572532a55cd..93a9ce1da359 100644 --- a/fs/xfs/libxfs/xfs_attr_remote.c +++ b/fs/xfs/libxfs/xfs_attr_remote.c @@ -24,6 +24,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index c5981f4a44bd..e090d342c21b 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_dir2.h" @@ -595,41 +596,7 @@ xfs_bmap_add_free( new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP); new->xbfi_startblock = bno; new->xbfi_blockcount = (xfs_extlen_t)len; - list_add(&new->xbfi_list, &flist->xbf_flist); - flist->xbf_count++; -} - -/* - * Remove the entry "free" from the free item list. Prev points to the - * previous entry, unless "free" is the head of the list. - */ -void -xfs_bmap_del_free( - struct xfs_bmap_free *flist, /* free item list header */ - struct xfs_bmap_free_item *free) /* list item to be freed */ -{ - list_del(&free->xbfi_list); - flist->xbf_count--; - kmem_zone_free(xfs_bmap_free_item_zone, free); -} - -/* - * Free up any items left in the list. - */ -void -xfs_bmap_cancel( - struct xfs_bmap_free *flist) /* list of bmap_free_items */ -{ - struct xfs_bmap_free_item *free; /* free list item */ - - if (flist->xbf_count == 0) - return; - while (!list_empty(&flist->xbf_flist)) { - free = list_first_entry(&flist->xbf_flist, - struct xfs_bmap_free_item, xbfi_list); - xfs_bmap_del_free(flist, free); - } - ASSERT(flist->xbf_count == 0); + xfs_defer_add(flist, XFS_DEFER_OPS_TYPE_FREE, &new->xbfi_list); } /* @@ -767,7 +734,7 @@ xfs_bmap_extents_to_btree( if (*firstblock == NULLFSBLOCK) { args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino); - } else if (flist->xbf_low) { + } else if (flist->dop_low) { args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = *firstblock; } else { @@ -788,7 +755,7 @@ xfs_bmap_extents_to_btree( ASSERT(args.fsbno != NULLFSBLOCK); ASSERT(*firstblock == NULLFSBLOCK || args.agno == XFS_FSB_TO_AGNO(mp, *firstblock) || - (flist->xbf_low && + (flist->dop_low && args.agno > XFS_FSB_TO_AGNO(mp, *firstblock))); *firstblock = cur->bc_private.b.firstblock = args.fsbno; cur->bc_private.b.allocated++; @@ -3708,7 +3675,7 @@ xfs_bmap_btalloc( error = xfs_bmap_btalloc_nullfb(ap, &args, &blen); if (error) return error; - } else if (ap->flist->xbf_low) { + } else if (ap->flist->dop_low) { if (xfs_inode_is_filestream(ap->ip)) args.type = XFS_ALLOCTYPE_FIRST_AG; else @@ -3741,7 +3708,7 @@ xfs_bmap_btalloc( * is >= the stripe unit and the allocation offset is * at the end of file. */ - if (!ap->flist->xbf_low && ap->aeof) { + if (!ap->flist->dop_low && ap->aeof) { if (!ap->offset) { args.alignment = stripe_align; atype = args.type; @@ -3834,7 +3801,7 @@ xfs_bmap_btalloc( args.minleft = 0; if ((error = xfs_alloc_vextent(&args))) return error; - ap->flist->xbf_low = 1; + ap->flist->dop_low = true; } if (args.fsbno != NULLFSBLOCK) { /* @@ -3844,7 +3811,7 @@ xfs_bmap_btalloc( ASSERT(*ap->firstblock == NULLFSBLOCK || XFS_FSB_TO_AGNO(mp, *ap->firstblock) == XFS_FSB_TO_AGNO(mp, args.fsbno) || - (ap->flist->xbf_low && + (ap->flist->dop_low && XFS_FSB_TO_AGNO(mp, *ap->firstblock) < XFS_FSB_TO_AGNO(mp, args.fsbno))); @@ -3852,7 +3819,7 @@ xfs_bmap_btalloc( if (*ap->firstblock == NULLFSBLOCK) *ap->firstblock = args.fsbno; ASSERT(nullfb || fb_agno == args.agno || - (ap->flist->xbf_low && fb_agno < args.agno)); + (ap->flist->dop_low && fb_agno < args.agno)); ap->length = args.len; ap->ip->i_d.di_nblocks += args.len; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); @@ -4319,7 +4286,7 @@ xfs_bmapi_allocate( if (error) return error; - if (bma->flist->xbf_low) + if (bma->flist->dop_low) bma->minleft = 0; if (bma->cur) bma->cur->bc_private.b.firstblock = *bma->firstblock; @@ -4684,7 +4651,7 @@ error0: XFS_FSB_TO_AGNO(mp, *firstblock) == XFS_FSB_TO_AGNO(mp, bma.cur->bc_private.b.firstblock) || - (flist->xbf_low && + (flist->dop_low && XFS_FSB_TO_AGNO(mp, *firstblock) < XFS_FSB_TO_AGNO(mp, bma.cur->bc_private.b.firstblock))); diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index f1f3ae6c0a3f..891b7285765e 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -69,27 +69,6 @@ struct xfs_bmap_free_item struct list_head xbfi_list; }; -/* - * Header for free extent list. - * - * xbf_low is used by the allocator to activate the lowspace algorithm - - * when free space is running low the extent allocator may choose to - * allocate an extent from an AG without leaving sufficient space for - * a btree split when inserting the new extent. In this case the allocator - * will enable the lowspace algorithm which is supposed to allow further - * allocations (such as btree splits and newroots) to allocate from - * sequential AGs. In order to avoid locking AGs out of order the lowspace - * algorithm will start searching for free space from AG 0. If the correct - * transaction reservations have been made then this algorithm will eventually - * find all the space it needs. - */ -typedef struct xfs_bmap_free -{ - struct list_head xbf_flist; /* list of to-be-free extents */ - int xbf_count; /* count of items on list */ - int xbf_low; /* alloc in low mode */ -} xfs_bmap_free_t; - #define XFS_BMAP_MAX_NMAP 4 /* @@ -139,14 +118,6 @@ static inline int xfs_bmapi_aflag(int w) #define DELAYSTARTBLOCK ((xfs_fsblock_t)-1LL) #define HOLESTARTBLOCK ((xfs_fsblock_t)-2LL) -static inline void xfs_bmap_init(xfs_bmap_free_t *flp, xfs_fsblock_t *fbp) -{ - INIT_LIST_HEAD(&flp->xbf_flist); - flp->xbf_count = 0; - flp->xbf_low = 0; - *fbp = NULLFSBLOCK; -} - /* * Flags for xfs_bmap_add_extent*. */ @@ -195,9 +166,6 @@ int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); void xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork); void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_bmap_free *flist, xfs_fsblock_t bno, xfs_filblks_t len); -void xfs_bmap_cancel(struct xfs_bmap_free *flist); -int xfs_bmap_finish(struct xfs_trans **tp, struct xfs_bmap_free *flist, - struct xfs_inode *ip); void xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork); int xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip, xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork); diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 4f1a98e6fb54..1b7ba73428b4 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_trans.h" #include "xfs_inode_item.h" @@ -462,7 +463,7 @@ xfs_bmbt_alloc_block( * block allocation here and corrupt the filesystem. */ args.minleft = args.tp->t_blk_res; - } else if (cur->bc_private.b.flist->xbf_low) { + } else if (cur->bc_private.b.flist->dop_low) { args.type = XFS_ALLOCTYPE_START_BNO; } else { args.type = XFS_ALLOCTYPE_NEAR_BNO; @@ -490,7 +491,7 @@ xfs_bmbt_alloc_block( error = xfs_alloc_vextent(&args); if (error) goto error0; - cur->bc_private.b.flist->xbf_low = 1; + cur->bc_private.b.flist->dop_low = true; } if (args.fsbno == NULLFSBLOCK) { XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index b5be7f41c856..4fc46529753d 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_trans.h" #include "xfs_inode_item.h" diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h index b9b5a92ffa6c..8d2508640121 100644 --- a/fs/xfs/libxfs/xfs_defer.h +++ b/fs/xfs/libxfs/xfs_defer.h @@ -93,4 +93,11 @@ struct xfs_defer_op_type { void xfs_defer_init_op_type(const struct xfs_defer_op_type *type); +/* XXX: compatibility shims, will go away in the next patch */ +#define xfs_bmap_finish xfs_defer_finish +#define xfs_bmap_cancel xfs_defer_cancel +#define xfs_bmap_init xfs_defer_init +#define xfs_bmap_free xfs_defer_ops +typedef struct xfs_defer_ops xfs_bmap_free_t; + #endif /* __XFS_DEFER_H__ */ diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c index af0f9d171f8a..945c03459ddd 100644 --- a/fs/xfs/libxfs/xfs_dir2.c +++ b/fs/xfs/libxfs/xfs_dir2.c @@ -21,6 +21,7 @@ #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index 4b1e408169a8..c654310983f0 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" #include "xfs_ialloc.h" diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c index 9d9559eb2835..4b9769e23c83 100644 --- a/fs/xfs/libxfs/xfs_inode_buf.c +++ b/fs/xfs/libxfs/xfs_inode_buf.c @@ -22,6 +22,7 @@ #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_error.h" #include "xfs_cksum.h" diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 12ca86778e02..2194e5d4eea8 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_ialloc.h" #include "xfs_alloc.h" diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 998c3e69dd33..d3c3d46245b2 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -25,6 +25,7 @@ #include "xfs_bit.h" #include "xfs_mount.h" #include "xfs_da_format.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" #include "xfs_trans.h" @@ -79,95 +80,6 @@ xfs_zero_extent( GFP_NOFS, true); } -/* Sort bmap items by AG. */ -static int -xfs_bmap_free_list_cmp( - void *priv, - struct list_head *a, - struct list_head *b) -{ - struct xfs_mount *mp = priv; - struct xfs_bmap_free_item *ra; - struct xfs_bmap_free_item *rb; - - ra = container_of(a, struct xfs_bmap_free_item, xbfi_list); - rb = container_of(b, struct xfs_bmap_free_item, xbfi_list); - return XFS_FSB_TO_AGNO(mp, ra->xbfi_startblock) - - XFS_FSB_TO_AGNO(mp, rb->xbfi_startblock); -} - -/* - * Routine to be called at transaction's end by xfs_bmapi, xfs_bunmapi - * caller. Frees all the extents that need freeing, which must be done - * last due to locking considerations. We never free any extents in - * the first transaction. - * - * If an inode *ip is provided, rejoin it to the transaction if - * the transaction was committed. - */ -int /* error */ -xfs_bmap_finish( - struct xfs_trans **tp, /* transaction pointer addr */ - struct xfs_bmap_free *flist, /* i/o: list extents to free */ - struct xfs_inode *ip) -{ - struct xfs_efd_log_item *efd; /* extent free data */ - struct xfs_efi_log_item *efi; /* extent free intention */ - int error; /* error return value */ - int committed;/* xact committed or not */ - struct xfs_bmap_free_item *free; /* free extent item */ - - ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES); - if (flist->xbf_count == 0) - return 0; - - list_sort((*tp)->t_mountp, &flist->xbf_flist, xfs_bmap_free_list_cmp); - - efi = xfs_trans_get_efi(*tp, flist->xbf_count); - list_for_each_entry(free, &flist->xbf_flist, xbfi_list) - xfs_trans_log_efi_extent(*tp, efi, free->xbfi_startblock, - free->xbfi_blockcount); - - error = __xfs_trans_roll(tp, ip, &committed); - if (error) { - /* - * If the transaction was committed, drop the EFD reference - * since we're bailing out of here. The other reference is - * dropped when the EFI hits the AIL. - * - * If the transaction was not committed, the EFI is freed by the - * EFI item unlock handler on abort. Also, we have a new - * transaction so we should return committed=1 even though we're - * returning an error. - */ - if (committed) { - xfs_efi_release(efi); - xfs_force_shutdown((*tp)->t_mountp, - SHUTDOWN_META_IO_ERROR); - } - return error; - } - - /* - * Get an EFD and free each extent in the list, logging to the EFD in - * the process. The remaining bmap free list is cleaned up by the caller - * on error. - */ - efd = xfs_trans_get_efd(*tp, efi, flist->xbf_count); - while (!list_empty(&flist->xbf_flist)) { - free = list_first_entry(&flist->xbf_flist, - struct xfs_bmap_free_item, xbfi_list); - error = xfs_trans_free_extent(*tp, efd, free->xbfi_startblock, - free->xbfi_blockcount); - if (error) - return error; - - xfs_bmap_del_free(flist, free); - } - - return 0; -} - int xfs_bmap_rtalloc( struct xfs_bmalloca *ap) /* bmap alloc argument struct */ @@ -815,7 +727,7 @@ xfs_bmap_punch_delalloc_range( if (error) break; - ASSERT(!flist.xbf_count && list_empty(&flist.xbf_flist)); + ASSERT(!xfs_defer_has_unfinished_work(&flist)); next_block: start_fsb++; remaining--; diff --git a/fs/xfs/xfs_bmap_util.h b/fs/xfs/xfs_bmap_util.h index f20071432ca6..51aaddee7b83 100644 --- a/fs/xfs/xfs_bmap_util.h +++ b/fs/xfs/xfs_bmap_util.h @@ -40,8 +40,6 @@ int xfs_getbmap(struct xfs_inode *ip, struct getbmapx *bmv, xfs_bmap_format_t formatter, void *arg); /* functions in xfs_bmap.c that are only needed by xfs_bmap_util.c */ -void xfs_bmap_del_free(struct xfs_bmap_free *flist, - struct xfs_bmap_free_item *free); int xfs_bmap_extsize_align(struct xfs_mount *mp, struct xfs_bmbt_irec *gotp, struct xfs_bmbt_irec *prevp, xfs_extlen_t extsz, int rt, int eof, int delay, int convert, diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index ccb0811963b2..3615f806cec9 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_bmap.h" #include "xfs_bmap_util.h" diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index a51353a1f87f..3e990fb0cb53 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -22,6 +22,7 @@ #include "xfs_trans_resv.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_bmap.h" #include "xfs_bmap_util.h" @@ -385,7 +386,7 @@ xfs_filestream_new_ag( } flags = (ap->userdata ? XFS_PICK_USERDATA : 0) | - (ap->flist->xbf_low ? XFS_PICK_LOWSPACE : 0); + (ap->flist->dop_low ? XFS_PICK_LOWSPACE : 0); err = xfs_filestream_pick_ag(pip, startag, agp, flags, minlen); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 7191c3878b4a..4c191d2194e8 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 8825bcfd314c..d2389bb994b7 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -25,6 +25,7 @@ #include "xfs_trans_resv.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 620fc9120444..023a4b41dc18 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -23,6 +23,7 @@ #include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" #include "xfs_bmap_btree.h" diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 835997843846..080b54b13154 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 970c19ba2f56..5f3ba88985d5 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -24,6 +24,7 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 3938b37d1043..627f7e69ae59 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -23,6 +23,7 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_bmap.h" #include "xfs_bmap_util.h" diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 08a46c6181fd..20af47b80264 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -26,6 +26,7 @@ #include "xfs_mount.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" +#include "xfs_defer.h" #include "xfs_dir2.h" #include "xfs_inode.h" #include "xfs_ialloc.h" diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c index 3971527c9a07..7f17ae6d709a 100644 --- a/fs/xfs/xfs_trace.c +++ b/fs/xfs/xfs_trace.c @@ -24,6 +24,7 @@ #include "xfs_mount.h" #include "xfs_defer.h" #include "xfs_da_format.h" +#include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" #include "xfs_da_btree.h" -- cgit v1.2.3 From 310a75a3c6c747857ad53dd25f2ede3de13612c9 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:18:10 +1000 Subject: xfs: change xfs_bmap_{finish,cancel,init,free} -> xfs_defer_* Drop the compatibility shims that we were using to integrate the new deferred operation mechanism into the existing code. No new code. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_attr.c | 58 +++++++++++++++++++------------------- fs/xfs/libxfs/xfs_attr_remote.c | 14 +++++----- fs/xfs/libxfs/xfs_bmap.c | 46 +++++++++++++++--------------- fs/xfs/libxfs/xfs_bmap.h | 18 ++++++------ fs/xfs/libxfs/xfs_btree.h | 5 ++-- fs/xfs/libxfs/xfs_da_btree.h | 4 +-- fs/xfs/libxfs/xfs_defer.h | 7 ----- fs/xfs/libxfs/xfs_dir2.c | 6 ++-- fs/xfs/libxfs/xfs_dir2.h | 8 +++--- fs/xfs/libxfs/xfs_ialloc.c | 6 ++-- fs/xfs/libxfs/xfs_ialloc.h | 2 +- fs/xfs/libxfs/xfs_trans_resv.c | 4 +-- fs/xfs/xfs_bmap_util.c | 28 +++++++++---------- fs/xfs/xfs_bmap_util.h | 2 +- fs/xfs/xfs_dquot.c | 10 +++---- fs/xfs/xfs_inode.c | 62 ++++++++++++++++++++--------------------- fs/xfs/xfs_inode.h | 4 +-- fs/xfs/xfs_iomap.c | 24 ++++++++-------- fs/xfs/xfs_rtalloc.c | 8 +++--- fs/xfs/xfs_super.c | 2 +- fs/xfs/xfs_symlink.c | 16 +++++------ fs/xfs/xfs_trans.h | 5 ---- fs/xfs/xfs_trans_extfree.c | 36 ++++++++++++------------ 23 files changed, 182 insertions(+), 193 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index 79d3a309b50f..66baf9747221 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -204,7 +204,7 @@ xfs_attr_set( { struct xfs_mount *mp = dp->i_mount; struct xfs_da_args args; - struct xfs_bmap_free flist; + struct xfs_defer_ops flist; struct xfs_trans_res tres; xfs_fsblock_t firstblock; int rsvd = (flags & ATTR_ROOT) != 0; @@ -317,13 +317,13 @@ xfs_attr_set( * It won't fit in the shortform, transform to a leaf block. * GROT: another possible req'mt for a double-split btree op. */ - xfs_bmap_init(args.flist, args.firstblock); + xfs_defer_init(args.flist, args.firstblock); error = xfs_attr_shortform_to_leaf(&args); if (!error) - error = xfs_bmap_finish(&args.trans, args.flist, dp); + error = xfs_defer_finish(&args.trans, args.flist, dp); if (error) { args.trans = NULL; - xfs_bmap_cancel(&flist); + xfs_defer_cancel(&flist); goto out; } @@ -383,7 +383,7 @@ xfs_attr_remove( { struct xfs_mount *mp = dp->i_mount; struct xfs_da_args args; - struct xfs_bmap_free flist; + struct xfs_defer_ops flist; xfs_fsblock_t firstblock; int error; @@ -585,13 +585,13 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * Commit that transaction so that the node_addname() call * can manage its own transactions. */ - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_attr3_leaf_to_node(args); if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); return error; } @@ -675,15 +675,15 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * If the result is small enough, shrink it all into the inode. */ if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) - error = xfs_bmap_finish(&args->trans, + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); return error; } } @@ -738,14 +738,14 @@ xfs_attr_leaf_removename(xfs_da_args_t *args) * If the result is small enough, shrink it all into the inode. */ if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); return error; } } @@ -864,14 +864,14 @@ restart: */ xfs_da_state_free(state); state = NULL; - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_attr3_leaf_to_node(args); if (!error) - error = xfs_bmap_finish(&args->trans, + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); goto out; } @@ -892,13 +892,13 @@ restart: * in the index/blkno/rmtblkno/rmtblkcnt fields and * in the index2/blkno2/rmtblkno2/rmtblkcnt2 fields. */ - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_da3_split(state); if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); goto out; } } else { @@ -991,14 +991,14 @@ restart: * Check to see if the tree needs to be collapsed. */ if (retval && (state->path.active > 1)) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_da3_join(state); if (!error) - error = xfs_bmap_finish(&args->trans, + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); goto out; } } @@ -1114,13 +1114,13 @@ xfs_attr_node_removename(xfs_da_args_t *args) * Check to see if the tree needs to be collapsed. */ if (retval && (state->path.active > 1)) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_da3_join(state); if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); goto out; } /* @@ -1147,15 +1147,15 @@ xfs_attr_node_removename(xfs_da_args_t *args) goto out; if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) - error = xfs_bmap_finish(&args->trans, + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); goto out; } } else diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c index 93a9ce1da359..aabb51690b6f 100644 --- a/fs/xfs/libxfs/xfs_attr_remote.c +++ b/fs/xfs/libxfs/xfs_attr_remote.c @@ -461,16 +461,16 @@ xfs_attr_rmtval_set( * extent and then crash then the block may not contain the * correct metadata after log recovery occurs. */ - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); nmap = 1; error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno, blkcnt, XFS_BMAPI_ATTRFORK, args->firstblock, args->total, &map, &nmap, args->flist); if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->flist, dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); return error; } @@ -504,7 +504,7 @@ xfs_attr_rmtval_set( ASSERT(blkcnt > 0); - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); nmap = 1; error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno, blkcnt, &map, &nmap, @@ -604,16 +604,16 @@ xfs_attr_rmtval_remove( blkcnt = args->rmtblkcnt; done = 0; while (!done) { - xfs_bmap_init(args->flist, args->firstblock); + xfs_defer_init(args->flist, args->firstblock); error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt, XFS_BMAPI_ATTRFORK, 1, args->firstblock, args->flist, &done); if (!error) - error = xfs_bmap_finish(&args->trans, args->flist, + error = xfs_defer_finish(&args->trans, args->flist, args->dp); if (error) { args->trans = NULL; - xfs_bmap_cancel(args->flist); + xfs_defer_cancel(args->flist); return error; } diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index e090d342c21b..71bf553d685e 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -572,11 +572,11 @@ xfs_bmap_validate_ret( void xfs_bmap_add_free( struct xfs_mount *mp, /* mount point structure */ - struct xfs_bmap_free *flist, /* list of extents */ + struct xfs_defer_ops *flist, /* list of extents */ xfs_fsblock_t bno, /* fs block number of extent */ xfs_filblks_t len) /* length of extent */ { - struct xfs_bmap_free_item *new; /* new element */ + struct xfs_extent_free_item *new; /* new element */ #ifdef DEBUG xfs_agnumber_t agno; xfs_agblock_t agbno; @@ -594,9 +594,9 @@ xfs_bmap_add_free( #endif ASSERT(xfs_bmap_free_item_zone != NULL); new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP); - new->xbfi_startblock = bno; - new->xbfi_blockcount = (xfs_extlen_t)len; - xfs_defer_add(flist, XFS_DEFER_OPS_TYPE_FREE, &new->xbfi_list); + new->xefi_startblock = bno; + new->xefi_blockcount = (xfs_extlen_t)len; + xfs_defer_add(flist, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list); } /* @@ -672,7 +672,7 @@ xfs_bmap_extents_to_btree( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first-block-allocated */ - xfs_bmap_free_t *flist, /* blocks freed in xaction */ + struct xfs_defer_ops *flist, /* blocks freed in xaction */ xfs_btree_cur_t **curp, /* cursor returned to caller */ int wasdel, /* converting a delayed alloc */ int *logflagsp, /* inode logging flags */ @@ -940,7 +940,7 @@ xfs_bmap_add_attrfork_btree( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - xfs_bmap_free_t *flist, /* blocks to free at commit */ + struct xfs_defer_ops *flist, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_btree_cur_t *cur; /* btree cursor */ @@ -983,7 +983,7 @@ xfs_bmap_add_attrfork_extents( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - xfs_bmap_free_t *flist, /* blocks to free at commit */ + struct xfs_defer_ops *flist, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_btree_cur_t *cur; /* bmap btree cursor */ @@ -1018,7 +1018,7 @@ xfs_bmap_add_attrfork_local( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - xfs_bmap_free_t *flist, /* blocks to free at commit */ + struct xfs_defer_ops *flist, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_da_args_t dargs; /* args for dir/attr code */ @@ -1059,7 +1059,7 @@ xfs_bmap_add_attrfork( int rsvd) /* xact may use reserved blks */ { xfs_fsblock_t firstblock; /* 1st block/ag allocated */ - xfs_bmap_free_t flist; /* freed extent records */ + struct xfs_defer_ops flist; /* freed extent records */ xfs_mount_t *mp; /* mount structure */ xfs_trans_t *tp; /* transaction pointer */ int blks; /* space reservation */ @@ -1125,7 +1125,7 @@ xfs_bmap_add_attrfork( ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP); ip->i_afp->if_flags = XFS_IFEXTENTS; logflags = 0; - xfs_bmap_init(&flist, &firstblock); + xfs_defer_init(&flist, &firstblock); switch (ip->i_d.di_format) { case XFS_DINODE_FMT_LOCAL: error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &flist, @@ -1165,7 +1165,7 @@ xfs_bmap_add_attrfork( xfs_log_sb(tp); } - error = xfs_bmap_finish(&tp, &flist, NULL); + error = xfs_defer_finish(&tp, &flist, NULL); if (error) goto bmap_cancel; error = xfs_trans_commit(tp); @@ -1173,7 +1173,7 @@ xfs_bmap_add_attrfork( return error; bmap_cancel: - xfs_bmap_cancel(&flist); + xfs_defer_cancel(&flist); trans_cancel: xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_ILOCK_EXCL); @@ -2214,7 +2214,7 @@ xfs_bmap_add_extent_unwritten_real( xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ xfs_bmbt_irec_t *new, /* new data to add to file extents */ xfs_fsblock_t *first, /* pointer to firstblock variable */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ + struct xfs_defer_ops *flist, /* list of extents to be freed */ int *logflagsp) /* inode logging flags */ { xfs_btree_cur_t *cur; /* btree cursor */ @@ -4447,7 +4447,7 @@ xfs_bmapi_write( xfs_extlen_t total, /* total blocks needed */ struct xfs_bmbt_irec *mval, /* output: map values */ int *nmap, /* i/o: mval size/count */ - struct xfs_bmap_free *flist) /* i/o: list extents to free */ + struct xfs_defer_ops *flist) /* i/o: list extents to free */ { struct xfs_mount *mp = ip->i_mount; struct xfs_ifork *ifp; @@ -4735,7 +4735,7 @@ xfs_bmap_del_extent( xfs_inode_t *ip, /* incore inode pointer */ xfs_trans_t *tp, /* current transaction pointer */ xfs_extnum_t *idx, /* extent number to update/delete */ - xfs_bmap_free_t *flist, /* list of extents to be freed */ + struct xfs_defer_ops *flist, /* list of extents to be freed */ xfs_btree_cur_t *cur, /* if null, not a btree */ xfs_bmbt_irec_t *del, /* data to remove from extents */ int *logflagsp, /* inode logging flags */ @@ -5064,7 +5064,7 @@ xfs_bunmapi( xfs_extnum_t nexts, /* number of extents max */ xfs_fsblock_t *firstblock, /* first allocated block controls a.g. for allocs */ - xfs_bmap_free_t *flist, /* i/o: list extents to free */ + struct xfs_defer_ops *flist, /* i/o: list extents to free */ int *done) /* set if not done yet */ { xfs_btree_cur_t *cur; /* bmap btree cursor */ @@ -5680,7 +5680,7 @@ xfs_bmap_shift_extents( int *done, xfs_fileoff_t stop_fsb, xfs_fsblock_t *firstblock, - struct xfs_bmap_free *flist, + struct xfs_defer_ops *flist, enum shift_direction direction, int num_exts) { @@ -5834,7 +5834,7 @@ xfs_bmap_split_extent_at( struct xfs_inode *ip, xfs_fileoff_t split_fsb, xfs_fsblock_t *firstfsb, - struct xfs_bmap_free *free_list) + struct xfs_defer_ops *free_list) { int whichfork = XFS_DATA_FORK; struct xfs_btree_cur *cur = NULL; @@ -5973,7 +5973,7 @@ xfs_bmap_split_extent( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; - struct xfs_bmap_free free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t firstfsb; int error; @@ -5985,21 +5985,21 @@ xfs_bmap_split_extent( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_bmap_init(&free_list, &firstfsb); + xfs_defer_init(&free_list, &firstfsb); error = xfs_bmap_split_extent_at(tp, ip, split_fsb, &firstfsb, &free_list); if (error) goto out; - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out; return xfs_trans_commit(tp); out: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); xfs_trans_cancel(tp); return error; } diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index 891b7285765e..3dc839383690 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -32,7 +32,7 @@ extern kmem_zone_t *xfs_bmap_free_item_zone; */ struct xfs_bmalloca { xfs_fsblock_t *firstblock; /* i/o first block allocated */ - struct xfs_bmap_free *flist; /* bmap freelist */ + struct xfs_defer_ops *flist; /* bmap freelist */ struct xfs_trans *tp; /* transaction pointer */ struct xfs_inode *ip; /* incore inode pointer */ struct xfs_bmbt_irec prev; /* extent before the new one */ @@ -62,11 +62,11 @@ struct xfs_bmalloca { * List of extents to be free "later". * The list is kept sorted on xbf_startblock. */ -struct xfs_bmap_free_item +struct xfs_extent_free_item { - xfs_fsblock_t xbfi_startblock;/* starting fs block number */ - xfs_extlen_t xbfi_blockcount;/* number of blocks in extent */ - struct list_head xbfi_list; + xfs_fsblock_t xefi_startblock;/* starting fs block number */ + xfs_extlen_t xefi_blockcount;/* number of blocks in extent */ + struct list_head xefi_list; }; #define XFS_BMAP_MAX_NMAP 4 @@ -164,7 +164,7 @@ void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); void xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork); -void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_bmap_free *flist, +void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *flist, xfs_fsblock_t bno, xfs_filblks_t len); void xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork); int xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip, @@ -186,18 +186,18 @@ int xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_fsblock_t *firstblock, xfs_extlen_t total, struct xfs_bmbt_irec *mval, int *nmap, - struct xfs_bmap_free *flist); + struct xfs_defer_ops *flist); int xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_extnum_t nexts, xfs_fsblock_t *firstblock, - struct xfs_bmap_free *flist, int *done); + struct xfs_defer_ops *flist, int *done); int xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx, xfs_extnum_t num); uint xfs_default_attroffset(struct xfs_inode *ip); int xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t *next_fsb, xfs_fileoff_t offset_shift_fsb, int *done, xfs_fileoff_t stop_fsb, xfs_fsblock_t *firstblock, - struct xfs_bmap_free *flist, enum shift_direction direction, + struct xfs_defer_ops *flist, enum shift_direction direction, int num_exts); int xfs_bmap_split_extent(struct xfs_inode *ip, xfs_fileoff_t split_offset); diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 69f564029f56..4ee6c2b98def 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -19,7 +19,7 @@ #define __XFS_BTREE_H__ struct xfs_buf; -struct xfs_bmap_free; +struct xfs_defer_ops; struct xfs_inode; struct xfs_mount; struct xfs_trans; @@ -255,11 +255,12 @@ typedef struct xfs_btree_cur union { struct { /* needed for BNO, CNT, INO */ struct xfs_buf *agbp; /* agf/agi buffer pointer */ + struct xfs_defer_ops *flist; /* deferred updates */ xfs_agnumber_t agno; /* ag number */ } a; struct { /* needed for BMAP */ struct xfs_inode *ip; /* pointer to our inode */ - struct xfs_bmap_free *flist; /* list to free after */ + struct xfs_defer_ops *flist; /* deferred updates */ xfs_fsblock_t firstblock; /* 1st blk allocated */ int allocated; /* count of alloced */ short forksize; /* fork's inode space */ diff --git a/fs/xfs/libxfs/xfs_da_btree.h b/fs/xfs/libxfs/xfs_da_btree.h index 6e153e399a77..249813a92ea8 100644 --- a/fs/xfs/libxfs/xfs_da_btree.h +++ b/fs/xfs/libxfs/xfs_da_btree.h @@ -19,7 +19,7 @@ #ifndef __XFS_DA_BTREE_H__ #define __XFS_DA_BTREE_H__ -struct xfs_bmap_free; +struct xfs_defer_ops; struct xfs_inode; struct xfs_trans; struct zone; @@ -70,7 +70,7 @@ typedef struct xfs_da_args { xfs_ino_t inumber; /* input/output inode number */ struct xfs_inode *dp; /* directory inode to manipulate */ xfs_fsblock_t *firstblock; /* ptr to firstblock for bmap calls */ - struct xfs_bmap_free *flist; /* ptr to freelist for bmap_finish */ + struct xfs_defer_ops *flist; /* ptr to freelist for bmap_finish */ struct xfs_trans *trans; /* current trans (changes over time) */ xfs_extlen_t total; /* total blocks needed, for 1st bmap */ int whichfork; /* data or attribute fork */ diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h index 8d2508640121..b9b5a92ffa6c 100644 --- a/fs/xfs/libxfs/xfs_defer.h +++ b/fs/xfs/libxfs/xfs_defer.h @@ -93,11 +93,4 @@ struct xfs_defer_op_type { void xfs_defer_init_op_type(const struct xfs_defer_op_type *type); -/* XXX: compatibility shims, will go away in the next patch */ -#define xfs_bmap_finish xfs_defer_finish -#define xfs_bmap_cancel xfs_defer_cancel -#define xfs_bmap_init xfs_defer_init -#define xfs_bmap_free xfs_defer_ops -typedef struct xfs_defer_ops xfs_bmap_free_t; - #endif /* __XFS_DEFER_H__ */ diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c index 945c03459ddd..0523100bf89f 100644 --- a/fs/xfs/libxfs/xfs_dir2.c +++ b/fs/xfs/libxfs/xfs_dir2.c @@ -260,7 +260,7 @@ xfs_dir_createname( struct xfs_name *name, xfs_ino_t inum, /* new entry inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ - xfs_bmap_free_t *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; @@ -437,7 +437,7 @@ xfs_dir_removename( struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, /* bmap's firstblock */ - xfs_bmap_free_t *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; @@ -499,7 +499,7 @@ xfs_dir_replace( struct xfs_name *name, /* name of entry to replace */ xfs_ino_t inum, /* new inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ - xfs_bmap_free_t *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *flist, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h index e55353651f5b..c272f1a4e280 100644 --- a/fs/xfs/libxfs/xfs_dir2.h +++ b/fs/xfs/libxfs/xfs_dir2.h @@ -18,7 +18,7 @@ #ifndef __XFS_DIR2_H__ #define __XFS_DIR2_H__ -struct xfs_bmap_free; +struct xfs_defer_ops; struct xfs_da_args; struct xfs_inode; struct xfs_mount; @@ -129,18 +129,18 @@ extern int xfs_dir_init(struct xfs_trans *tp, struct xfs_inode *dp, extern int xfs_dir_createname(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, - struct xfs_bmap_free *flist, xfs_extlen_t tot); + struct xfs_defer_ops *flist, xfs_extlen_t tot); extern int xfs_dir_lookup(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t *inum, struct xfs_name *ci_name); extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, - struct xfs_bmap_free *flist, xfs_extlen_t tot); + struct xfs_defer_ops *flist, xfs_extlen_t tot); extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, - struct xfs_bmap_free *flist, xfs_extlen_t tot); + struct xfs_defer_ops *flist, xfs_extlen_t tot); extern int xfs_dir_canenter(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name); diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index c654310983f0..972b2ea6a075 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -1818,7 +1818,7 @@ xfs_difree_inode_chunk( struct xfs_mount *mp, xfs_agnumber_t agno, struct xfs_inobt_rec_incore *rec, - struct xfs_bmap_free *flist) + struct xfs_defer_ops *flist) { xfs_agblock_t sagbno = XFS_AGINO_TO_AGBNO(mp, rec->ir_startino); int startidx, endidx; @@ -1890,7 +1890,7 @@ xfs_difree_inobt( struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agino_t agino, - struct xfs_bmap_free *flist, + struct xfs_defer_ops *flist, struct xfs_icluster *xic, struct xfs_inobt_rec_incore *orec) { @@ -2122,7 +2122,7 @@ int xfs_difree( struct xfs_trans *tp, /* transaction pointer */ xfs_ino_t inode, /* inode to be freed */ - struct xfs_bmap_free *flist, /* extents to free */ + struct xfs_defer_ops *flist, /* extents to free */ struct xfs_icluster *xic) /* cluster info if deleted */ { /* REFERENCED */ diff --git a/fs/xfs/libxfs/xfs_ialloc.h b/fs/xfs/libxfs/xfs_ialloc.h index 6e450df2979b..2e06b675d8bb 100644 --- a/fs/xfs/libxfs/xfs_ialloc.h +++ b/fs/xfs/libxfs/xfs_ialloc.h @@ -95,7 +95,7 @@ int /* error */ xfs_difree( struct xfs_trans *tp, /* transaction pointer */ xfs_ino_t inode, /* inode to be freed */ - struct xfs_bmap_free *flist, /* extents to free */ + struct xfs_defer_ops *flist, /* extents to free */ struct xfs_icluster *ifree); /* cluster info if deleted */ /* diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c index 68cb1e7bf2bb..4c7eb9d7c0c7 100644 --- a/fs/xfs/libxfs/xfs_trans_resv.c +++ b/fs/xfs/libxfs/xfs_trans_resv.c @@ -153,9 +153,9 @@ xfs_calc_finobt_res( * item logged to try to account for the overhead of the transaction mechanism. * * Note: Most of the reservations underestimate the number of allocation - * groups into which they could free extents in the xfs_bmap_finish() call. + * groups into which they could free extents in the xfs_defer_finish() call. * This is because the number in the worst case is quite high and quite - * unusual. In order to fix this we need to change xfs_bmap_finish() to free + * unusual. In order to fix this we need to change xfs_defer_finish() to free * extents in only a single AG at a time. This will require changes to the * EFI code as well, however, so that the EFI for the extents not freed is * logged again in each transaction. See SGI PV #261917. diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index d3c3d46245b2..07954e8f803b 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -685,7 +685,7 @@ xfs_bmap_punch_delalloc_range( xfs_bmbt_irec_t imap; int nimaps = 1; xfs_fsblock_t firstblock; - xfs_bmap_free_t flist; + struct xfs_defer_ops flist; /* * Map the range first and check that it is a delalloc extent @@ -721,7 +721,7 @@ xfs_bmap_punch_delalloc_range( * allocated or freed for a delalloc extent and hence we need * don't cancel or finish them after the xfs_bunmapi() call. */ - xfs_bmap_init(&flist, &firstblock); + xfs_defer_init(&flist, &firstblock); error = xfs_bunmapi(NULL, ip, start_fsb, 1, 0, 1, &firstblock, &flist, &done); if (error) @@ -884,7 +884,7 @@ xfs_alloc_file_space( int rt; xfs_trans_t *tp; xfs_bmbt_irec_t imaps[1], *imapp; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; uint qblocks, resblks, resrtextents; int error; @@ -975,7 +975,7 @@ xfs_alloc_file_space( xfs_trans_ijoin(tp, ip, 0); - xfs_bmap_init(&free_list, &firstfsb); + xfs_defer_init(&free_list, &firstfsb); error = xfs_bmapi_write(tp, ip, startoffset_fsb, allocatesize_fsb, alloc_type, &firstfsb, resblks, imapp, &nimaps, &free_list); @@ -985,7 +985,7 @@ xfs_alloc_file_space( /* * Complete the transaction */ - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto error0; @@ -1008,7 +1008,7 @@ xfs_alloc_file_space( return error; error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */ - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); xfs_trans_unreserve_quota_nblks(tp, ip, (long)qblocks, 0, quota_flag); error1: /* Just cancel transaction */ @@ -1026,7 +1026,7 @@ xfs_unmap_extent( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; - struct xfs_bmap_free free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t firstfsb; uint resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0); int error; @@ -1045,13 +1045,13 @@ xfs_unmap_extent( xfs_trans_ijoin(tp, ip, 0); - xfs_bmap_init(&free_list, &firstfsb); + xfs_defer_init(&free_list, &firstfsb); error = xfs_bunmapi(tp, ip, startoffset_fsb, len_fsb, 0, 2, &firstfsb, &free_list, done); if (error) goto out_bmap_cancel; - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, ip); if (error) goto out_bmap_cancel; @@ -1061,7 +1061,7 @@ out_unlock: return error; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); goto out_unlock; @@ -1250,7 +1250,7 @@ xfs_shift_file_space( struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; int error; - struct xfs_bmap_free free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; xfs_fileoff_t stop_fsb; xfs_fileoff_t next_fsb; @@ -1328,7 +1328,7 @@ xfs_shift_file_space( xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* * We are using the write transaction in which max 2 bmbt @@ -1340,7 +1340,7 @@ xfs_shift_file_space( if (error) goto out_bmap_cancel; - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out_bmap_cancel; @@ -1350,7 +1350,7 @@ xfs_shift_file_space( return error; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); return error; diff --git a/fs/xfs/xfs_bmap_util.h b/fs/xfs/xfs_bmap_util.h index 51aaddee7b83..68a621a8e0c0 100644 --- a/fs/xfs/xfs_bmap_util.h +++ b/fs/xfs/xfs_bmap_util.h @@ -21,7 +21,7 @@ /* Kernel only BMAP related definitions and functions */ struct xfs_bmbt_irec; -struct xfs_bmap_free_item; +struct xfs_extent_free_item; struct xfs_ifork; struct xfs_inode; struct xfs_mount; diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index 3615f806cec9..d9447f1bf483 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -308,7 +308,7 @@ xfs_qm_dqalloc( xfs_buf_t **O_bpp) { xfs_fsblock_t firstblock; - xfs_bmap_free_t flist; + struct xfs_defer_ops flist; xfs_bmbt_irec_t map; int nmaps, error; xfs_buf_t *bp; @@ -321,7 +321,7 @@ xfs_qm_dqalloc( /* * Initialize the bmap freelist prior to calling bmapi code. */ - xfs_bmap_init(&flist, &firstblock); + xfs_defer_init(&flist, &firstblock); xfs_ilock(quotip, XFS_ILOCK_EXCL); /* * Return if this type of quotas is turned off while we didn't @@ -369,7 +369,7 @@ xfs_qm_dqalloc( dqp->dq_flags & XFS_DQ_ALLTYPES, bp); /* - * xfs_bmap_finish() may commit the current transaction and + * xfs_defer_finish() may commit the current transaction and * start a second transaction if the freelist is not empty. * * Since we still want to modify this buffer, we need to @@ -383,7 +383,7 @@ xfs_qm_dqalloc( xfs_trans_bhold(tp, bp); - error = xfs_bmap_finish(tpp, &flist, NULL); + error = xfs_defer_finish(tpp, &flist, NULL); if (error) goto error1; @@ -399,7 +399,7 @@ xfs_qm_dqalloc( return 0; error1: - xfs_bmap_cancel(&flist); + xfs_defer_cancel(&flist); error0: xfs_iunlock(quotip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d2389bb994b7..3ce50da885ac 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1123,7 +1123,7 @@ xfs_create( struct xfs_inode *ip = NULL; struct xfs_trans *tp = NULL; int error; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; bool unlock_dp_on_error = false; prid_t prid; @@ -1183,7 +1183,7 @@ xfs_create( XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); unlock_dp_on_error = true; - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* * Reserve disk quota and the inode. @@ -1254,7 +1254,7 @@ xfs_create( */ xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp); - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out_bmap_cancel; @@ -1270,7 +1270,7 @@ xfs_create( return 0; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); out_release_inode: @@ -1402,7 +1402,7 @@ xfs_link( xfs_mount_t *mp = tdp->i_mount; xfs_trans_t *tp; int error; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; int resblks; @@ -1453,7 +1453,7 @@ xfs_link( goto error_return; } - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* * Handle initial link state of O_TMPFILE inode @@ -1483,9 +1483,9 @@ xfs_link( if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) { - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); goto error_return; } @@ -1527,7 +1527,7 @@ xfs_itruncate_extents( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp = *tpp; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; xfs_fileoff_t first_unmap_block; xfs_fileoff_t last_block; @@ -1563,7 +1563,7 @@ xfs_itruncate_extents( ASSERT(first_unmap_block < last_block); unmap_len = last_block - first_unmap_block + 1; while (!done) { - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); error = xfs_bunmapi(tp, ip, first_unmap_block, unmap_len, xfs_bmapi_aflag(whichfork), @@ -1577,7 +1577,7 @@ xfs_itruncate_extents( * Duplicate the transaction that has the permanent * reservation and commit the old transaction. */ - error = xfs_bmap_finish(&tp, &free_list, ip); + error = xfs_defer_finish(&tp, &free_list, ip); if (error) goto out_bmap_cancel; @@ -1603,7 +1603,7 @@ out_bmap_cancel: * the transaction can be properly aborted. We just need to make sure * we're not holding any resources that we were not when we came in. */ - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); goto out; } @@ -1744,7 +1744,7 @@ STATIC int xfs_inactive_ifree( struct xfs_inode *ip) { - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; @@ -1781,7 +1781,7 @@ xfs_inactive_ifree( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); error = xfs_ifree(tp, ip, &free_list); if (error) { /* @@ -1808,11 +1808,11 @@ xfs_inactive_ifree( * Just ignore errors at this point. There is nothing we can do except * to try to keep going. Make sure it's not a silent error. */ - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) { - xfs_notice(mp, "%s: xfs_bmap_finish returned error %d", + xfs_notice(mp, "%s: xfs_defer_finish returned error %d", __func__, error); - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); } error = xfs_trans_commit(tp); if (error) @@ -2368,7 +2368,7 @@ int xfs_ifree( xfs_trans_t *tp, xfs_inode_t *ip, - xfs_bmap_free_t *flist) + struct xfs_defer_ops *flist) { int error; struct xfs_icluster xic = { 0 }; @@ -2475,7 +2475,7 @@ xfs_iunpin_wait( * directory entry. * * This is still safe from a transactional point of view - it is not until we - * get to xfs_bmap_finish() that we have the possibility of multiple + * get to xfs_defer_finish() that we have the possibility of multiple * transactions in this operation. Hence as long as we remove the directory * entry and drop the link count in the first transaction of the remove * operation, there are no transactional constraints on the ordering here. @@ -2490,7 +2490,7 @@ xfs_remove( xfs_trans_t *tp = NULL; int is_dir = S_ISDIR(VFS_I(ip)->i_mode); int error = 0; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; uint resblks; @@ -2572,7 +2572,7 @@ xfs_remove( if (error) goto out_trans_cancel; - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); error = xfs_dir_removename(tp, dp, name, ip->i_ino, &first_block, &free_list, resblks); if (error) { @@ -2588,7 +2588,7 @@ xfs_remove( if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out_bmap_cancel; @@ -2602,7 +2602,7 @@ xfs_remove( return 0; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); std_return: @@ -2663,7 +2663,7 @@ xfs_sort_for_rename( static int xfs_finish_rename( struct xfs_trans *tp, - struct xfs_bmap_free *free_list) + struct xfs_defer_ops *free_list) { int error; @@ -2674,9 +2674,9 @@ xfs_finish_rename( if (tp->t_mountp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_bmap_finish(&tp, free_list, NULL); + error = xfs_defer_finish(&tp, free_list, NULL); if (error) { - xfs_bmap_cancel(free_list); + xfs_defer_cancel(free_list); xfs_trans_cancel(tp); return error; } @@ -2698,7 +2698,7 @@ xfs_cross_rename( struct xfs_inode *dp2, struct xfs_name *name2, struct xfs_inode *ip2, - struct xfs_bmap_free *free_list, + struct xfs_defer_ops *free_list, xfs_fsblock_t *first_block, int spaceres) { @@ -2801,7 +2801,7 @@ xfs_cross_rename( return xfs_finish_rename(tp, free_list); out_trans_abort: - xfs_bmap_cancel(free_list); + xfs_defer_cancel(free_list); xfs_trans_cancel(tp); return error; } @@ -2856,7 +2856,7 @@ xfs_rename( { struct xfs_mount *mp = src_dp->i_mount; struct xfs_trans *tp; - struct xfs_bmap_free free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; struct xfs_inode *wip = NULL; /* whiteout inode */ struct xfs_inode *inodes[__XFS_SORT_INODES]; @@ -2945,7 +2945,7 @@ xfs_rename( goto out_trans_cancel; } - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* RENAME_EXCHANGE is unique from here on. */ if (flags & RENAME_EXCHANGE) @@ -3131,7 +3131,7 @@ xfs_rename( return error; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); out_release_wip: diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 8eb78ec4a6e2..e1a411e08f00 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -27,7 +27,7 @@ struct xfs_dinode; struct xfs_inode; struct xfs_buf; -struct xfs_bmap_free; +struct xfs_defer_ops; struct xfs_bmbt_irec; struct xfs_inode_log_item; struct xfs_mount; @@ -398,7 +398,7 @@ uint xfs_ilock_attr_map_shared(struct xfs_inode *); uint xfs_ip2xflags(struct xfs_inode *); int xfs_ifree(struct xfs_trans *, xfs_inode_t *, - struct xfs_bmap_free *); + struct xfs_defer_ops *); int xfs_itruncate_extents(struct xfs_trans **, struct xfs_inode *, int, xfs_fsize_t); void xfs_iext_realloc(xfs_inode_t *, int, int); diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 023a4b41dc18..dc93c20fb9ad 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -129,7 +129,7 @@ xfs_iomap_write_direct( int quota_flag; int rt; xfs_trans_t *tp; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; uint qblocks, resblks, resrtextents; int error; int lockmode; @@ -232,7 +232,7 @@ xfs_iomap_write_direct( * From this point onwards we overwrite the imap pointer that the * caller gave to us. */ - xfs_bmap_init(&free_list, &firstfsb); + xfs_defer_init(&free_list, &firstfsb); nimaps = 1; error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, bmapi_flags, &firstfsb, resblks, imap, @@ -243,7 +243,7 @@ xfs_iomap_write_direct( /* * Complete the transaction */ - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out_bmap_cancel; @@ -267,7 +267,7 @@ out_unlock: return error; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); xfs_trans_unreserve_quota_nblks(tp, ip, (long)qblocks, 0, quota_flag); out_trans_cancel: xfs_trans_cancel(tp); @@ -686,7 +686,7 @@ xfs_iomap_write_allocate( xfs_fileoff_t offset_fsb, last_block; xfs_fileoff_t end_fsb, map_start_fsb; xfs_fsblock_t first_block; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_filblks_t count_fsb; xfs_trans_t *tp; int nimaps; @@ -728,7 +728,7 @@ xfs_iomap_write_allocate( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* * it is possible that the extents have changed since @@ -788,7 +788,7 @@ xfs_iomap_write_allocate( if (error) goto trans_cancel; - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto trans_cancel; @@ -822,7 +822,7 @@ xfs_iomap_write_allocate( } trans_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); xfs_trans_cancel(tp); error0: xfs_iunlock(ip, XFS_ILOCK_EXCL); @@ -843,7 +843,7 @@ xfs_iomap_write_unwritten( int nimaps; xfs_trans_t *tp; xfs_bmbt_irec_t imap; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; xfs_fsize_t i_size; uint resblks; int error; @@ -887,7 +887,7 @@ xfs_iomap_write_unwritten( /* * Modify the unwritten extent state of the buffer. */ - xfs_bmap_init(&free_list, &firstfsb); + xfs_defer_init(&free_list, &firstfsb); nimaps = 1; error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, XFS_BMAPI_CONVERT, &firstfsb, resblks, @@ -910,7 +910,7 @@ xfs_iomap_write_unwritten( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); } - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto error_on_bmapi_transaction; @@ -937,7 +937,7 @@ xfs_iomap_write_unwritten( return 0; error_on_bmapi_transaction: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_ILOCK_EXCL); return error; diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 627f7e69ae59..c761a6a75904 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -770,7 +770,7 @@ xfs_growfs_rt_alloc( xfs_daddr_t d; /* disk block address */ int error; /* error return value */ xfs_fsblock_t firstblock;/* first block allocated in xaction */ - struct xfs_bmap_free flist; /* list of freed blocks */ + struct xfs_defer_ops flist; /* list of freed blocks */ xfs_fsblock_t fsbno; /* filesystem block for bno */ struct xfs_bmbt_irec map; /* block map output */ int nmap; /* number of block maps */ @@ -795,7 +795,7 @@ xfs_growfs_rt_alloc( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_bmap_init(&flist, &firstblock); + xfs_defer_init(&flist, &firstblock); /* * Allocate blocks to the bitmap file. */ @@ -810,7 +810,7 @@ xfs_growfs_rt_alloc( /* * Free any blocks freed up in the transaction, then commit. */ - error = xfs_bmap_finish(&tp, &flist, NULL); + error = xfs_defer_finish(&tp, &flist, NULL); if (error) goto out_bmap_cancel; error = xfs_trans_commit(tp); @@ -863,7 +863,7 @@ xfs_growfs_rt_alloc( return 0; out_bmap_cancel: - xfs_bmap_cancel(&flist); + xfs_defer_cancel(&flist); out_trans_cancel: xfs_trans_cancel(tp); return error; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 7fc508fe8d40..95a11beb48c6 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1697,7 +1697,7 @@ xfs_init_zones(void) goto out_free_ioend_bioset; xfs_bmap_free_item_zone = kmem_zone_init( - sizeof(struct xfs_bmap_free_item), + sizeof(struct xfs_extent_free_item), "xfs_bmap_free_item"); if (!xfs_bmap_free_item_zone) goto out_destroy_log_ticket_zone; diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 20af47b80264..3b005ec33886 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -173,7 +173,7 @@ xfs_symlink( struct xfs_inode *ip = NULL; int error = 0; int pathlen; - struct xfs_bmap_free free_list; + struct xfs_defer_ops free_list; xfs_fsblock_t first_block; bool unlock_dp_on_error = false; xfs_fileoff_t first_fsb; @@ -270,7 +270,7 @@ xfs_symlink( * Initialize the bmap freelist prior to calling either * bmapi or the directory create code. */ - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); /* * Allocate an inode for the symlink. @@ -377,7 +377,7 @@ xfs_symlink( xfs_trans_set_sync(tp); } - error = xfs_bmap_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &free_list, NULL); if (error) goto out_bmap_cancel; @@ -393,7 +393,7 @@ xfs_symlink( return 0; out_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); out_trans_cancel: xfs_trans_cancel(tp); out_release_inode: @@ -427,7 +427,7 @@ xfs_inactive_symlink_rmt( int done; int error; xfs_fsblock_t first_block; - xfs_bmap_free_t free_list; + struct xfs_defer_ops free_list; int i; xfs_mount_t *mp; xfs_bmbt_irec_t mval[XFS_SYMLINK_MAPS]; @@ -466,7 +466,7 @@ xfs_inactive_symlink_rmt( * Find the block(s) so we can inval and unmap them. */ done = 0; - xfs_bmap_init(&free_list, &first_block); + xfs_defer_init(&free_list, &first_block); nmaps = ARRAY_SIZE(mval); error = xfs_bmapi_read(ip, 0, xfs_symlink_blocks(mp, size), mval, &nmaps, 0); @@ -496,7 +496,7 @@ xfs_inactive_symlink_rmt( /* * Commit the first transaction. This logs the EFI and the inode. */ - error = xfs_bmap_finish(&tp, &free_list, ip); + error = xfs_defer_finish(&tp, &free_list, ip); if (error) goto error_bmap_cancel; /* @@ -526,7 +526,7 @@ xfs_inactive_symlink_rmt( return 0; error_bmap_cancel: - xfs_bmap_cancel(&free_list); + xfs_defer_cancel(&free_list); error_trans_cancel: xfs_trans_cancel(tp); error_unlock: diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 746cf0832642..a5fe1874f02f 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -212,11 +212,6 @@ void xfs_trans_log_buf(xfs_trans_t *, struct xfs_buf *, uint, uint); void xfs_trans_log_inode(xfs_trans_t *, struct xfs_inode *, uint); void xfs_extent_free_init_defer_op(void); -struct xfs_efi_log_item *xfs_trans_get_efi(struct xfs_trans *, uint); -void xfs_trans_log_efi_extent(struct xfs_trans *, - struct xfs_efi_log_item *, - xfs_fsblock_t, - xfs_extlen_t); struct xfs_efd_log_item *xfs_trans_get_efd(struct xfs_trans *, struct xfs_efi_log_item *, uint); diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index a9f34ee52ab4..b48490290b27 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -35,7 +35,7 @@ * caller must use all nextents extents, because we are not * flexible about this at all. */ -struct xfs_efi_log_item * +STATIC struct xfs_efi_log_item * xfs_trans_get_efi(struct xfs_trans *tp, uint nextents) { @@ -59,7 +59,7 @@ xfs_trans_get_efi(struct xfs_trans *tp, * extent is to be logged as needing to be freed. It should * be called once for each extent to be freed. */ -void +STATIC void xfs_trans_log_efi_extent(struct xfs_trans *tp, struct xfs_efi_log_item *efip, xfs_fsblock_t start_block, @@ -156,13 +156,13 @@ xfs_extent_free_diff_items( struct list_head *b) { struct xfs_mount *mp = priv; - struct xfs_bmap_free_item *ra; - struct xfs_bmap_free_item *rb; + struct xfs_extent_free_item *ra; + struct xfs_extent_free_item *rb; - ra = container_of(a, struct xfs_bmap_free_item, xbfi_list); - rb = container_of(b, struct xfs_bmap_free_item, xbfi_list); - return XFS_FSB_TO_AGNO(mp, ra->xbfi_startblock) - - XFS_FSB_TO_AGNO(mp, rb->xbfi_startblock); + ra = container_of(a, struct xfs_extent_free_item, xefi_list); + rb = container_of(b, struct xfs_extent_free_item, xefi_list); + return XFS_FSB_TO_AGNO(mp, ra->xefi_startblock) - + XFS_FSB_TO_AGNO(mp, rb->xefi_startblock); } /* Get an EFI. */ @@ -181,11 +181,11 @@ xfs_extent_free_log_item( void *intent, struct list_head *item) { - struct xfs_bmap_free_item *free; + struct xfs_extent_free_item *free; - free = container_of(item, struct xfs_bmap_free_item, xbfi_list); - xfs_trans_log_efi_extent(tp, intent, free->xbfi_startblock, - free->xbfi_blockcount); + free = container_of(item, struct xfs_extent_free_item, xefi_list); + xfs_trans_log_efi_extent(tp, intent, free->xefi_startblock, + free->xefi_blockcount); } /* Get an EFD so we can process all the free extents. */ @@ -207,13 +207,13 @@ xfs_extent_free_finish_item( void *done_item, void **state) { - struct xfs_bmap_free_item *free; + struct xfs_extent_free_item *free; int error; - free = container_of(item, struct xfs_bmap_free_item, xbfi_list); + free = container_of(item, struct xfs_extent_free_item, xefi_list); error = xfs_trans_free_extent(tp, done_item, - free->xbfi_startblock, - free->xbfi_blockcount); + free->xefi_startblock, + free->xefi_blockcount); kmem_free(free); return error; } @@ -231,9 +231,9 @@ STATIC void xfs_extent_free_cancel_item( struct list_head *item) { - struct xfs_bmap_free_item *free; + struct xfs_extent_free_item *free; - free = container_of(item, struct xfs_bmap_free_item, xbfi_list); + free = container_of(item, struct xfs_extent_free_item, xefi_list); kmem_free(free); } -- cgit v1.2.3 From 2c3234d1ef53030ff6a79d55ba1fb291098467c2 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:19:29 +1000 Subject: xfs: rename flist/free_list to dfops Mechanical change of flist/free_list to dfops, since they're now deferred ops, not just a freeing list. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_attr.c | 62 ++++++++++----------- fs/xfs/libxfs/xfs_attr_leaf.c | 4 +- fs/xfs/libxfs/xfs_attr_remote.c | 18 +++---- fs/xfs/libxfs/xfs_bmap.c | 116 ++++++++++++++++++++-------------------- fs/xfs/libxfs/xfs_bmap.h | 10 ++-- fs/xfs/libxfs/xfs_bmap_btree.c | 14 ++--- fs/xfs/libxfs/xfs_btree.h | 4 +- fs/xfs/libxfs/xfs_da_btree.c | 6 +-- fs/xfs/libxfs/xfs_da_btree.h | 2 +- fs/xfs/libxfs/xfs_dir2.c | 14 ++--- fs/xfs/libxfs/xfs_dir2.h | 6 +-- fs/xfs/libxfs/xfs_ialloc.c | 14 ++--- fs/xfs/libxfs/xfs_ialloc.h | 2 +- fs/xfs/xfs_bmap_util.c | 40 +++++++------- fs/xfs/xfs_dquot.c | 10 ++-- fs/xfs/xfs_filestream.c | 2 +- fs/xfs/xfs_inode.c | 94 ++++++++++++++++---------------- fs/xfs/xfs_iomap.c | 30 +++++------ fs/xfs/xfs_rtalloc.c | 10 ++-- fs/xfs/xfs_symlink.c | 24 ++++----- 20 files changed, 241 insertions(+), 241 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index 66baf9747221..af1ecb19121e 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -204,7 +204,7 @@ xfs_attr_set( { struct xfs_mount *mp = dp->i_mount; struct xfs_da_args args; - struct xfs_defer_ops flist; + struct xfs_defer_ops dfops; struct xfs_trans_res tres; xfs_fsblock_t firstblock; int rsvd = (flags & ATTR_ROOT) != 0; @@ -222,7 +222,7 @@ xfs_attr_set( args.value = value; args.valuelen = valuelen; args.firstblock = &firstblock; - args.flist = &flist; + args.dfops = &dfops; args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT; args.total = xfs_attr_calc_size(&args, &local); @@ -317,13 +317,13 @@ xfs_attr_set( * It won't fit in the shortform, transform to a leaf block. * GROT: another possible req'mt for a double-split btree op. */ - xfs_defer_init(args.flist, args.firstblock); + xfs_defer_init(args.dfops, args.firstblock); error = xfs_attr_shortform_to_leaf(&args); if (!error) - error = xfs_defer_finish(&args.trans, args.flist, dp); + error = xfs_defer_finish(&args.trans, args.dfops, dp); if (error) { args.trans = NULL; - xfs_defer_cancel(&flist); + xfs_defer_cancel(&dfops); goto out; } @@ -383,7 +383,7 @@ xfs_attr_remove( { struct xfs_mount *mp = dp->i_mount; struct xfs_da_args args; - struct xfs_defer_ops flist; + struct xfs_defer_ops dfops; xfs_fsblock_t firstblock; int error; @@ -400,7 +400,7 @@ xfs_attr_remove( return error; args.firstblock = &firstblock; - args.flist = &flist; + args.dfops = &dfops; /* * we have no control over the attribute names that userspace passes us @@ -585,13 +585,13 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * Commit that transaction so that the node_addname() call * can manage its own transactions. */ - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_attr3_leaf_to_node(args); if (!error) - error = xfs_defer_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); return error; } @@ -675,15 +675,15 @@ xfs_attr_leaf_addname(xfs_da_args_t *args) * If the result is small enough, shrink it all into the inode. */ if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) error = xfs_defer_finish(&args->trans, - args->flist, dp); + args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); return error; } } @@ -738,14 +738,14 @@ xfs_attr_leaf_removename(xfs_da_args_t *args) * If the result is small enough, shrink it all into the inode. */ if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) - error = xfs_defer_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); return error; } } @@ -864,14 +864,14 @@ restart: */ xfs_da_state_free(state); state = NULL; - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_attr3_leaf_to_node(args); if (!error) error = xfs_defer_finish(&args->trans, - args->flist, dp); + args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); goto out; } @@ -892,13 +892,13 @@ restart: * in the index/blkno/rmtblkno/rmtblkcnt fields and * in the index2/blkno2/rmtblkno2/rmtblkcnt2 fields. */ - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_da3_split(state); if (!error) - error = xfs_defer_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); goto out; } } else { @@ -991,14 +991,14 @@ restart: * Check to see if the tree needs to be collapsed. */ if (retval && (state->path.active > 1)) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_da3_join(state); if (!error) error = xfs_defer_finish(&args->trans, - args->flist, dp); + args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); goto out; } } @@ -1114,13 +1114,13 @@ xfs_attr_node_removename(xfs_da_args_t *args) * Check to see if the tree needs to be collapsed. */ if (retval && (state->path.active > 1)) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_da3_join(state); if (!error) - error = xfs_defer_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); goto out; } /* @@ -1147,15 +1147,15 @@ xfs_attr_node_removename(xfs_da_args_t *args) goto out; if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_attr3_leaf_to_shortform(bp, args, forkoff); /* bp is gone due to xfs_da_shrink_inode */ if (!error) error = xfs_defer_finish(&args->trans, - args->flist, dp); + args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); goto out; } } else diff --git a/fs/xfs/libxfs/xfs_attr_leaf.c b/fs/xfs/libxfs/xfs_attr_leaf.c index 01a5ecfedfcf..8ea91f363093 100644 --- a/fs/xfs/libxfs/xfs_attr_leaf.c +++ b/fs/xfs/libxfs/xfs_attr_leaf.c @@ -792,7 +792,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args) nargs.dp = dp; nargs.geo = args->geo; nargs.firstblock = args->firstblock; - nargs.flist = args->flist; + nargs.dfops = args->dfops; nargs.total = args->total; nargs.whichfork = XFS_ATTR_FORK; nargs.trans = args->trans; @@ -922,7 +922,7 @@ xfs_attr3_leaf_to_shortform( nargs.geo = args->geo; nargs.dp = dp; nargs.firstblock = args->firstblock; - nargs.flist = args->flist; + nargs.dfops = args->dfops; nargs.total = args->total; nargs.whichfork = XFS_ATTR_FORK; nargs.trans = args->trans; diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c index aabb51690b6f..d52f525f5b2d 100644 --- a/fs/xfs/libxfs/xfs_attr_remote.c +++ b/fs/xfs/libxfs/xfs_attr_remote.c @@ -461,16 +461,16 @@ xfs_attr_rmtval_set( * extent and then crash then the block may not contain the * correct metadata after log recovery occurs. */ - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); nmap = 1; error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno, blkcnt, XFS_BMAPI_ATTRFORK, args->firstblock, - args->total, &map, &nmap, args->flist); + args->total, &map, &nmap, args->dfops); if (!error) - error = xfs_defer_finish(&args->trans, args->flist, dp); + error = xfs_defer_finish(&args->trans, args->dfops, dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); return error; } @@ -504,7 +504,7 @@ xfs_attr_rmtval_set( ASSERT(blkcnt > 0); - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); nmap = 1; error = xfs_bmapi_read(dp, (xfs_fileoff_t)lblkno, blkcnt, &map, &nmap, @@ -604,16 +604,16 @@ xfs_attr_rmtval_remove( blkcnt = args->rmtblkcnt; done = 0; while (!done) { - xfs_defer_init(args->flist, args->firstblock); + xfs_defer_init(args->dfops, args->firstblock); error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt, XFS_BMAPI_ATTRFORK, 1, args->firstblock, - args->flist, &done); + args->dfops, &done); if (!error) - error = xfs_defer_finish(&args->trans, args->flist, + error = xfs_defer_finish(&args->trans, args->dfops, args->dp); if (error) { args->trans = NULL; - xfs_defer_cancel(args->flist); + xfs_defer_cancel(args->dfops); return error; } diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 71bf553d685e..540a6b7928af 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -572,7 +572,7 @@ xfs_bmap_validate_ret( void xfs_bmap_add_free( struct xfs_mount *mp, /* mount point structure */ - struct xfs_defer_ops *flist, /* list of extents */ + struct xfs_defer_ops *dfops, /* list of extents */ xfs_fsblock_t bno, /* fs block number of extent */ xfs_filblks_t len) /* length of extent */ { @@ -596,7 +596,7 @@ xfs_bmap_add_free( new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP); new->xefi_startblock = bno; new->xefi_blockcount = (xfs_extlen_t)len; - xfs_defer_add(flist, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list); + xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list); } /* @@ -649,7 +649,7 @@ xfs_bmap_btree_to_extents( cblock = XFS_BUF_TO_BLOCK(cbp); if ((error = xfs_btree_check_block(cur, cblock, 0, cbp))) return error; - xfs_bmap_add_free(mp, cur->bc_private.b.flist, cbno, 1); + xfs_bmap_add_free(mp, cur->bc_private.b.dfops, cbno, 1); ip->i_d.di_nblocks--; xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L); xfs_trans_binval(tp, cbp); @@ -672,7 +672,7 @@ xfs_bmap_extents_to_btree( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first-block-allocated */ - struct xfs_defer_ops *flist, /* blocks freed in xaction */ + struct xfs_defer_ops *dfops, /* blocks freed in xaction */ xfs_btree_cur_t **curp, /* cursor returned to caller */ int wasdel, /* converting a delayed alloc */ int *logflagsp, /* inode logging flags */ @@ -721,7 +721,7 @@ xfs_bmap_extents_to_btree( */ cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); cur->bc_private.b.firstblock = *firstblock; - cur->bc_private.b.flist = flist; + cur->bc_private.b.dfops = dfops; cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0; /* * Convert to a btree with two levels, one record in root. @@ -734,7 +734,7 @@ xfs_bmap_extents_to_btree( if (*firstblock == NULLFSBLOCK) { args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino); - } else if (flist->dop_low) { + } else if (dfops->dop_low) { args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = *firstblock; } else { @@ -755,7 +755,7 @@ xfs_bmap_extents_to_btree( ASSERT(args.fsbno != NULLFSBLOCK); ASSERT(*firstblock == NULLFSBLOCK || args.agno == XFS_FSB_TO_AGNO(mp, *firstblock) || - (flist->dop_low && + (dfops->dop_low && args.agno > XFS_FSB_TO_AGNO(mp, *firstblock))); *firstblock = cur->bc_private.b.firstblock = args.fsbno; cur->bc_private.b.allocated++; @@ -940,7 +940,7 @@ xfs_bmap_add_attrfork_btree( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - struct xfs_defer_ops *flist, /* blocks to free at commit */ + struct xfs_defer_ops *dfops, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_btree_cur_t *cur; /* btree cursor */ @@ -953,7 +953,7 @@ xfs_bmap_add_attrfork_btree( *flags |= XFS_ILOG_DBROOT; else { cur = xfs_bmbt_init_cursor(mp, tp, ip, XFS_DATA_FORK); - cur->bc_private.b.flist = flist; + cur->bc_private.b.dfops = dfops; cur->bc_private.b.firstblock = *firstblock; if ((error = xfs_bmbt_lookup_ge(cur, 0, 0, 0, &stat))) goto error0; @@ -983,7 +983,7 @@ xfs_bmap_add_attrfork_extents( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - struct xfs_defer_ops *flist, /* blocks to free at commit */ + struct xfs_defer_ops *dfops, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_btree_cur_t *cur; /* bmap btree cursor */ @@ -992,7 +992,7 @@ xfs_bmap_add_attrfork_extents( if (ip->i_d.di_nextents * sizeof(xfs_bmbt_rec_t) <= XFS_IFORK_DSIZE(ip)) return 0; cur = NULL; - error = xfs_bmap_extents_to_btree(tp, ip, firstblock, flist, &cur, 0, + error = xfs_bmap_extents_to_btree(tp, ip, firstblock, dfops, &cur, 0, flags, XFS_DATA_FORK); if (cur) { cur->bc_private.b.allocated = 0; @@ -1018,7 +1018,7 @@ xfs_bmap_add_attrfork_local( xfs_trans_t *tp, /* transaction pointer */ xfs_inode_t *ip, /* incore inode pointer */ xfs_fsblock_t *firstblock, /* first block allocated */ - struct xfs_defer_ops *flist, /* blocks to free at commit */ + struct xfs_defer_ops *dfops, /* blocks to free at commit */ int *flags) /* inode logging flags */ { xfs_da_args_t dargs; /* args for dir/attr code */ @@ -1031,7 +1031,7 @@ xfs_bmap_add_attrfork_local( dargs.geo = ip->i_mount->m_dir_geo; dargs.dp = ip; dargs.firstblock = firstblock; - dargs.flist = flist; + dargs.dfops = dfops; dargs.total = dargs.geo->fsbcount; dargs.whichfork = XFS_DATA_FORK; dargs.trans = tp; @@ -1059,7 +1059,7 @@ xfs_bmap_add_attrfork( int rsvd) /* xact may use reserved blks */ { xfs_fsblock_t firstblock; /* 1st block/ag allocated */ - struct xfs_defer_ops flist; /* freed extent records */ + struct xfs_defer_ops dfops; /* freed extent records */ xfs_mount_t *mp; /* mount structure */ xfs_trans_t *tp; /* transaction pointer */ int blks; /* space reservation */ @@ -1125,18 +1125,18 @@ xfs_bmap_add_attrfork( ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP); ip->i_afp->if_flags = XFS_IFEXTENTS; logflags = 0; - xfs_defer_init(&flist, &firstblock); + xfs_defer_init(&dfops, &firstblock); switch (ip->i_d.di_format) { case XFS_DINODE_FMT_LOCAL: - error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &flist, + error = xfs_bmap_add_attrfork_local(tp, ip, &firstblock, &dfops, &logflags); break; case XFS_DINODE_FMT_EXTENTS: error = xfs_bmap_add_attrfork_extents(tp, ip, &firstblock, - &flist, &logflags); + &dfops, &logflags); break; case XFS_DINODE_FMT_BTREE: - error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &flist, + error = xfs_bmap_add_attrfork_btree(tp, ip, &firstblock, &dfops, &logflags); break; default: @@ -1165,7 +1165,7 @@ xfs_bmap_add_attrfork( xfs_log_sb(tp); } - error = xfs_defer_finish(&tp, &flist, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto bmap_cancel; error = xfs_trans_commit(tp); @@ -1173,7 +1173,7 @@ xfs_bmap_add_attrfork( return error; bmap_cancel: - xfs_defer_cancel(&flist); + xfs_defer_cancel(&dfops); trans_cancel: xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_ILOCK_EXCL); @@ -1970,7 +1970,7 @@ xfs_bmap_add_extent_delay_real( if (xfs_bmap_needs_btree(bma->ip, whichfork)) { error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, - bma->firstblock, bma->flist, + bma->firstblock, bma->dfops, &bma->cur, 1, &tmp_rval, whichfork); rval |= tmp_rval; if (error) @@ -2054,7 +2054,7 @@ xfs_bmap_add_extent_delay_real( if (xfs_bmap_needs_btree(bma->ip, whichfork)) { error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, - bma->firstblock, bma->flist, &bma->cur, 1, + bma->firstblock, bma->dfops, &bma->cur, 1, &tmp_rval, whichfork); rval |= tmp_rval; if (error) @@ -2123,7 +2123,7 @@ xfs_bmap_add_extent_delay_real( if (xfs_bmap_needs_btree(bma->ip, whichfork)) { error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, - bma->firstblock, bma->flist, &bma->cur, + bma->firstblock, bma->dfops, &bma->cur, 1, &tmp_rval, whichfork); rval |= tmp_rval; if (error) @@ -2172,7 +2172,7 @@ xfs_bmap_add_extent_delay_real( ASSERT(bma->cur == NULL); error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, - bma->firstblock, bma->flist, &bma->cur, + bma->firstblock, bma->dfops, &bma->cur, da_old > 0, &tmp_logflags, whichfork); bma->logflags |= tmp_logflags; if (error) @@ -2214,7 +2214,7 @@ xfs_bmap_add_extent_unwritten_real( xfs_btree_cur_t **curp, /* if *curp is null, not a btree */ xfs_bmbt_irec_t *new, /* new data to add to file extents */ xfs_fsblock_t *first, /* pointer to firstblock variable */ - struct xfs_defer_ops *flist, /* list of extents to be freed */ + struct xfs_defer_ops *dfops, /* list of extents to be freed */ int *logflagsp) /* inode logging flags */ { xfs_btree_cur_t *cur; /* btree cursor */ @@ -2707,7 +2707,7 @@ xfs_bmap_add_extent_unwritten_real( int tmp_logflags; /* partial log flag return val */ ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, first, flist, &cur, + error = xfs_bmap_extents_to_btree(tp, ip, first, dfops, &cur, 0, &tmp_logflags, XFS_DATA_FORK); *logflagsp |= tmp_logflags; if (error) @@ -3100,7 +3100,7 @@ xfs_bmap_add_extent_hole_real( ASSERT(bma->cur == NULL); error = xfs_bmap_extents_to_btree(bma->tp, bma->ip, - bma->firstblock, bma->flist, &bma->cur, + bma->firstblock, bma->dfops, &bma->cur, 0, &tmp_logflags, whichfork); bma->logflags |= tmp_logflags; if (error) @@ -3675,7 +3675,7 @@ xfs_bmap_btalloc( error = xfs_bmap_btalloc_nullfb(ap, &args, &blen); if (error) return error; - } else if (ap->flist->dop_low) { + } else if (ap->dfops->dop_low) { if (xfs_inode_is_filestream(ap->ip)) args.type = XFS_ALLOCTYPE_FIRST_AG; else @@ -3708,7 +3708,7 @@ xfs_bmap_btalloc( * is >= the stripe unit and the allocation offset is * at the end of file. */ - if (!ap->flist->dop_low && ap->aeof) { + if (!ap->dfops->dop_low && ap->aeof) { if (!ap->offset) { args.alignment = stripe_align; atype = args.type; @@ -3801,7 +3801,7 @@ xfs_bmap_btalloc( args.minleft = 0; if ((error = xfs_alloc_vextent(&args))) return error; - ap->flist->dop_low = true; + ap->dfops->dop_low = true; } if (args.fsbno != NULLFSBLOCK) { /* @@ -3811,7 +3811,7 @@ xfs_bmap_btalloc( ASSERT(*ap->firstblock == NULLFSBLOCK || XFS_FSB_TO_AGNO(mp, *ap->firstblock) == XFS_FSB_TO_AGNO(mp, args.fsbno) || - (ap->flist->dop_low && + (ap->dfops->dop_low && XFS_FSB_TO_AGNO(mp, *ap->firstblock) < XFS_FSB_TO_AGNO(mp, args.fsbno))); @@ -3819,7 +3819,7 @@ xfs_bmap_btalloc( if (*ap->firstblock == NULLFSBLOCK) *ap->firstblock = args.fsbno; ASSERT(nullfb || fb_agno == args.agno || - (ap->flist->dop_low && fb_agno < args.agno)); + (ap->dfops->dop_low && fb_agno < args.agno)); ap->length = args.len; ap->ip->i_d.di_nblocks += args.len; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); @@ -4286,7 +4286,7 @@ xfs_bmapi_allocate( if (error) return error; - if (bma->flist->dop_low) + if (bma->dfops->dop_low) bma->minleft = 0; if (bma->cur) bma->cur->bc_private.b.firstblock = *bma->firstblock; @@ -4295,7 +4295,7 @@ xfs_bmapi_allocate( if ((ifp->if_flags & XFS_IFBROOT) && !bma->cur) { bma->cur = xfs_bmbt_init_cursor(mp, bma->tp, bma->ip, whichfork); bma->cur->bc_private.b.firstblock = *bma->firstblock; - bma->cur->bc_private.b.flist = bma->flist; + bma->cur->bc_private.b.dfops = bma->dfops; } /* * Bump the number of extents we've allocated @@ -4376,7 +4376,7 @@ xfs_bmapi_convert_unwritten( bma->cur = xfs_bmbt_init_cursor(bma->ip->i_mount, bma->tp, bma->ip, whichfork); bma->cur->bc_private.b.firstblock = *bma->firstblock; - bma->cur->bc_private.b.flist = bma->flist; + bma->cur->bc_private.b.dfops = bma->dfops; } mval->br_state = (mval->br_state == XFS_EXT_UNWRITTEN) ? XFS_EXT_NORM : XFS_EXT_UNWRITTEN; @@ -4393,7 +4393,7 @@ xfs_bmapi_convert_unwritten( } error = xfs_bmap_add_extent_unwritten_real(bma->tp, bma->ip, &bma->idx, - &bma->cur, mval, bma->firstblock, bma->flist, + &bma->cur, mval, bma->firstblock, bma->dfops, &tmp_logflags); /* * Log the inode core unconditionally in the unwritten extent conversion @@ -4447,7 +4447,7 @@ xfs_bmapi_write( xfs_extlen_t total, /* total blocks needed */ struct xfs_bmbt_irec *mval, /* output: map values */ int *nmap, /* i/o: mval size/count */ - struct xfs_defer_ops *flist) /* i/o: list extents to free */ + struct xfs_defer_ops *dfops) /* i/o: list extents to free */ { struct xfs_mount *mp = ip->i_mount; struct xfs_ifork *ifp; @@ -4537,7 +4537,7 @@ xfs_bmapi_write( bma.ip = ip; bma.total = total; bma.userdata = 0; - bma.flist = flist; + bma.dfops = dfops; bma.firstblock = firstblock; while (bno < end && n < *nmap) { @@ -4651,7 +4651,7 @@ error0: XFS_FSB_TO_AGNO(mp, *firstblock) == XFS_FSB_TO_AGNO(mp, bma.cur->bc_private.b.firstblock) || - (flist->dop_low && + (dfops->dop_low && XFS_FSB_TO_AGNO(mp, *firstblock) < XFS_FSB_TO_AGNO(mp, bma.cur->bc_private.b.firstblock))); @@ -4735,7 +4735,7 @@ xfs_bmap_del_extent( xfs_inode_t *ip, /* incore inode pointer */ xfs_trans_t *tp, /* current transaction pointer */ xfs_extnum_t *idx, /* extent number to update/delete */ - struct xfs_defer_ops *flist, /* list of extents to be freed */ + struct xfs_defer_ops *dfops, /* list of extents to be freed */ xfs_btree_cur_t *cur, /* if null, not a btree */ xfs_bmbt_irec_t *del, /* data to remove from extents */ int *logflagsp, /* inode logging flags */ @@ -5023,7 +5023,7 @@ xfs_bmap_del_extent( * If we need to, add to list of extents to delete. */ if (do_fx) - xfs_bmap_add_free(mp, flist, del->br_startblock, + xfs_bmap_add_free(mp, dfops, del->br_startblock, del->br_blockcount); /* * Adjust inode # blocks in the file. @@ -5064,7 +5064,7 @@ xfs_bunmapi( xfs_extnum_t nexts, /* number of extents max */ xfs_fsblock_t *firstblock, /* first allocated block controls a.g. for allocs */ - struct xfs_defer_ops *flist, /* i/o: list extents to free */ + struct xfs_defer_ops *dfops, /* i/o: list extents to free */ int *done) /* set if not done yet */ { xfs_btree_cur_t *cur; /* bmap btree cursor */ @@ -5137,7 +5137,7 @@ xfs_bunmapi( ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE); cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); cur->bc_private.b.firstblock = *firstblock; - cur->bc_private.b.flist = flist; + cur->bc_private.b.dfops = dfops; cur->bc_private.b.flags = 0; } else cur = NULL; @@ -5231,7 +5231,7 @@ xfs_bunmapi( } del.br_state = XFS_EXT_UNWRITTEN; error = xfs_bmap_add_extent_unwritten_real(tp, ip, - &lastx, &cur, &del, firstblock, flist, + &lastx, &cur, &del, firstblock, dfops, &logflags); if (error) goto error0; @@ -5290,7 +5290,7 @@ xfs_bunmapi( lastx--; error = xfs_bmap_add_extent_unwritten_real(tp, ip, &lastx, &cur, &prev, - firstblock, flist, &logflags); + firstblock, dfops, &logflags); if (error) goto error0; goto nodelete; @@ -5299,7 +5299,7 @@ xfs_bunmapi( del.br_state = XFS_EXT_UNWRITTEN; error = xfs_bmap_add_extent_unwritten_real(tp, ip, &lastx, &cur, &del, - firstblock, flist, &logflags); + firstblock, dfops, &logflags); if (error) goto error0; goto nodelete; @@ -5357,7 +5357,7 @@ xfs_bunmapi( } else if (cur) cur->bc_private.b.flags &= ~XFS_BTCUR_BPRV_WASDEL; - error = xfs_bmap_del_extent(ip, tp, &lastx, flist, cur, &del, + error = xfs_bmap_del_extent(ip, tp, &lastx, dfops, cur, &del, &tmp_logflags, whichfork); logflags |= tmp_logflags; if (error) @@ -5391,7 +5391,7 @@ nodelete: */ if (xfs_bmap_needs_btree(ip, whichfork)) { ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, firstblock, flist, + error = xfs_bmap_extents_to_btree(tp, ip, firstblock, dfops, &cur, 0, &tmp_logflags, whichfork); logflags |= tmp_logflags; if (error) @@ -5680,7 +5680,7 @@ xfs_bmap_shift_extents( int *done, xfs_fileoff_t stop_fsb, xfs_fsblock_t *firstblock, - struct xfs_defer_ops *flist, + struct xfs_defer_ops *dfops, enum shift_direction direction, int num_exts) { @@ -5725,7 +5725,7 @@ xfs_bmap_shift_extents( if (ifp->if_flags & XFS_IFBROOT) { cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); cur->bc_private.b.firstblock = *firstblock; - cur->bc_private.b.flist = flist; + cur->bc_private.b.dfops = dfops; cur->bc_private.b.flags = 0; } @@ -5834,7 +5834,7 @@ xfs_bmap_split_extent_at( struct xfs_inode *ip, xfs_fileoff_t split_fsb, xfs_fsblock_t *firstfsb, - struct xfs_defer_ops *free_list) + struct xfs_defer_ops *dfops) { int whichfork = XFS_DATA_FORK; struct xfs_btree_cur *cur = NULL; @@ -5896,7 +5896,7 @@ xfs_bmap_split_extent_at( if (ifp->if_flags & XFS_IFBROOT) { cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); cur->bc_private.b.firstblock = *firstfsb; - cur->bc_private.b.flist = free_list; + cur->bc_private.b.dfops = dfops; cur->bc_private.b.flags = 0; error = xfs_bmbt_lookup_eq(cur, got.br_startoff, got.br_startblock, @@ -5949,7 +5949,7 @@ xfs_bmap_split_extent_at( int tmp_logflags; /* partial log flag return val */ ASSERT(cur == NULL); - error = xfs_bmap_extents_to_btree(tp, ip, firstfsb, free_list, + error = xfs_bmap_extents_to_btree(tp, ip, firstfsb, dfops, &cur, 0, &tmp_logflags, whichfork); logflags |= tmp_logflags; } @@ -5973,7 +5973,7 @@ xfs_bmap_split_extent( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t firstfsb; int error; @@ -5985,21 +5985,21 @@ xfs_bmap_split_extent( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_defer_init(&free_list, &firstfsb); + xfs_defer_init(&dfops, &firstfsb); error = xfs_bmap_split_extent_at(tp, ip, split_fsb, - &firstfsb, &free_list); + &firstfsb, &dfops); if (error) goto out; - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out; return xfs_trans_commit(tp); out: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); xfs_trans_cancel(tp); return error; } diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index 3dc839383690..dc01bb85a6e8 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -32,7 +32,7 @@ extern kmem_zone_t *xfs_bmap_free_item_zone; */ struct xfs_bmalloca { xfs_fsblock_t *firstblock; /* i/o first block allocated */ - struct xfs_defer_ops *flist; /* bmap freelist */ + struct xfs_defer_ops *dfops; /* bmap freelist */ struct xfs_trans *tp; /* transaction pointer */ struct xfs_inode *ip; /* incore inode pointer */ struct xfs_bmbt_irec prev; /* extent before the new one */ @@ -164,7 +164,7 @@ void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); void xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork); -void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *flist, +void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *dfops, xfs_fsblock_t bno, xfs_filblks_t len); void xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork); int xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip, @@ -186,18 +186,18 @@ int xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_fsblock_t *firstblock, xfs_extlen_t total, struct xfs_bmbt_irec *mval, int *nmap, - struct xfs_defer_ops *flist); + struct xfs_defer_ops *dfops); int xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_extnum_t nexts, xfs_fsblock_t *firstblock, - struct xfs_defer_ops *flist, int *done); + struct xfs_defer_ops *dfops, int *done); int xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx, xfs_extnum_t num); uint xfs_default_attroffset(struct xfs_inode *ip); int xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t *next_fsb, xfs_fileoff_t offset_shift_fsb, int *done, xfs_fileoff_t stop_fsb, xfs_fsblock_t *firstblock, - struct xfs_defer_ops *flist, enum shift_direction direction, + struct xfs_defer_ops *dfops, enum shift_direction direction, int num_exts); int xfs_bmap_split_extent(struct xfs_inode *ip, xfs_fileoff_t split_offset); diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 1b7ba73428b4..0df5318cefff 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -407,11 +407,11 @@ xfs_bmbt_dup_cursor( cur->bc_private.b.ip, cur->bc_private.b.whichfork); /* - * Copy the firstblock, flist, and flags values, + * Copy the firstblock, dfops, and flags values, * since init cursor doesn't get them. */ new->bc_private.b.firstblock = cur->bc_private.b.firstblock; - new->bc_private.b.flist = cur->bc_private.b.flist; + new->bc_private.b.dfops = cur->bc_private.b.dfops; new->bc_private.b.flags = cur->bc_private.b.flags; return new; @@ -424,7 +424,7 @@ xfs_bmbt_update_cursor( { ASSERT((dst->bc_private.b.firstblock != NULLFSBLOCK) || (dst->bc_private.b.ip->i_d.di_flags & XFS_DIFLAG_REALTIME)); - ASSERT(dst->bc_private.b.flist == src->bc_private.b.flist); + ASSERT(dst->bc_private.b.dfops == src->bc_private.b.dfops); dst->bc_private.b.allocated += src->bc_private.b.allocated; dst->bc_private.b.firstblock = src->bc_private.b.firstblock; @@ -463,7 +463,7 @@ xfs_bmbt_alloc_block( * block allocation here and corrupt the filesystem. */ args.minleft = args.tp->t_blk_res; - } else if (cur->bc_private.b.flist->dop_low) { + } else if (cur->bc_private.b.dfops->dop_low) { args.type = XFS_ALLOCTYPE_START_BNO; } else { args.type = XFS_ALLOCTYPE_NEAR_BNO; @@ -491,7 +491,7 @@ xfs_bmbt_alloc_block( error = xfs_alloc_vextent(&args); if (error) goto error0; - cur->bc_private.b.flist->dop_low = true; + cur->bc_private.b.dfops->dop_low = true; } if (args.fsbno == NULLFSBLOCK) { XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); @@ -527,7 +527,7 @@ xfs_bmbt_free_block( struct xfs_trans *tp = cur->bc_tp; xfs_fsblock_t fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp)); - xfs_bmap_add_free(mp, cur->bc_private.b.flist, fsbno, 1); + xfs_bmap_add_free(mp, cur->bc_private.b.dfops, fsbno, 1); ip->i_d.di_nblocks--; xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); @@ -793,7 +793,7 @@ xfs_bmbt_init_cursor( cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork); cur->bc_private.b.ip = ip; cur->bc_private.b.firstblock = NULLFSBLOCK; - cur->bc_private.b.flist = NULL; + cur->bc_private.b.dfops = NULL; cur->bc_private.b.allocated = 0; cur->bc_private.b.flags = 0; cur->bc_private.b.whichfork = whichfork; diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 4ee6c2b98def..154eefdcb730 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -255,12 +255,12 @@ typedef struct xfs_btree_cur union { struct { /* needed for BNO, CNT, INO */ struct xfs_buf *agbp; /* agf/agi buffer pointer */ - struct xfs_defer_ops *flist; /* deferred updates */ + struct xfs_defer_ops *dfops; /* deferred updates */ xfs_agnumber_t agno; /* ag number */ } a; struct { /* needed for BMAP */ struct xfs_inode *ip; /* pointer to our inode */ - struct xfs_defer_ops *flist; /* deferred updates */ + struct xfs_defer_ops *dfops; /* deferred updates */ xfs_fsblock_t firstblock; /* 1st blk allocated */ int allocated; /* count of alloced */ short forksize; /* fork's inode space */ diff --git a/fs/xfs/libxfs/xfs_da_btree.c b/fs/xfs/libxfs/xfs_da_btree.c index 0f1f165f4048..f2dc1a950c85 100644 --- a/fs/xfs/libxfs/xfs_da_btree.c +++ b/fs/xfs/libxfs/xfs_da_btree.c @@ -2029,7 +2029,7 @@ xfs_da_grow_inode_int( error = xfs_bmapi_write(tp, dp, *bno, count, xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA|XFS_BMAPI_CONTIG, args->firstblock, args->total, &map, &nmap, - args->flist); + args->dfops); if (error) return error; @@ -2052,7 +2052,7 @@ xfs_da_grow_inode_int( error = xfs_bmapi_write(tp, dp, b, c, xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA, args->firstblock, args->total, - &mapp[mapi], &nmap, args->flist); + &mapp[mapi], &nmap, args->dfops); if (error) goto out_free_map; if (nmap < 1) @@ -2362,7 +2362,7 @@ xfs_da_shrink_inode( */ error = xfs_bunmapi(tp, dp, dead_blkno, count, xfs_bmapi_aflag(w), 0, args->firstblock, - args->flist, &done); + args->dfops, &done); if (error == -ENOSPC) { if (w != XFS_DATA_FORK) break; diff --git a/fs/xfs/libxfs/xfs_da_btree.h b/fs/xfs/libxfs/xfs_da_btree.h index 249813a92ea8..98c75cbe6ac2 100644 --- a/fs/xfs/libxfs/xfs_da_btree.h +++ b/fs/xfs/libxfs/xfs_da_btree.h @@ -70,7 +70,7 @@ typedef struct xfs_da_args { xfs_ino_t inumber; /* input/output inode number */ struct xfs_inode *dp; /* directory inode to manipulate */ xfs_fsblock_t *firstblock; /* ptr to firstblock for bmap calls */ - struct xfs_defer_ops *flist; /* ptr to freelist for bmap_finish */ + struct xfs_defer_ops *dfops; /* ptr to freelist for bmap_finish */ struct xfs_trans *trans; /* current trans (changes over time) */ xfs_extlen_t total; /* total blocks needed, for 1st bmap */ int whichfork; /* data or attribute fork */ diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c index 0523100bf89f..20a96dd5af7e 100644 --- a/fs/xfs/libxfs/xfs_dir2.c +++ b/fs/xfs/libxfs/xfs_dir2.c @@ -260,7 +260,7 @@ xfs_dir_createname( struct xfs_name *name, xfs_ino_t inum, /* new entry inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ - struct xfs_defer_ops *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *dfops, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; @@ -287,7 +287,7 @@ xfs_dir_createname( args->inumber = inum; args->dp = dp; args->firstblock = first; - args->flist = flist; + args->dfops = dfops; args->total = total; args->whichfork = XFS_DATA_FORK; args->trans = tp; @@ -437,7 +437,7 @@ xfs_dir_removename( struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, /* bmap's firstblock */ - struct xfs_defer_ops *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *dfops, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; @@ -459,7 +459,7 @@ xfs_dir_removename( args->inumber = ino; args->dp = dp; args->firstblock = first; - args->flist = flist; + args->dfops = dfops; args->total = total; args->whichfork = XFS_DATA_FORK; args->trans = tp; @@ -499,7 +499,7 @@ xfs_dir_replace( struct xfs_name *name, /* name of entry to replace */ xfs_ino_t inum, /* new inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ - struct xfs_defer_ops *flist, /* bmap's freeblock list */ + struct xfs_defer_ops *dfops, /* bmap's freeblock list */ xfs_extlen_t total) /* bmap's total block count */ { struct xfs_da_args *args; @@ -524,7 +524,7 @@ xfs_dir_replace( args->inumber = inum; args->dp = dp; args->firstblock = first; - args->flist = flist; + args->dfops = dfops; args->total = total; args->whichfork = XFS_DATA_FORK; args->trans = tp; @@ -681,7 +681,7 @@ xfs_dir2_shrink_inode( /* Unmap the fsblock(s). */ error = xfs_bunmapi(tp, dp, da, args->geo->fsbcount, 0, 0, - args->firstblock, args->flist, &done); + args->firstblock, args->dfops, &done); if (error) { /* * ENOSPC actually can happen if we're in a removename with no diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h index c272f1a4e280..becc926c3e3d 100644 --- a/fs/xfs/libxfs/xfs_dir2.h +++ b/fs/xfs/libxfs/xfs_dir2.h @@ -129,18 +129,18 @@ extern int xfs_dir_init(struct xfs_trans *tp, struct xfs_inode *dp, extern int xfs_dir_createname(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, - struct xfs_defer_ops *flist, xfs_extlen_t tot); + struct xfs_defer_ops *dfops, xfs_extlen_t tot); extern int xfs_dir_lookup(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t *inum, struct xfs_name *ci_name); extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, - struct xfs_defer_ops *flist, xfs_extlen_t tot); + struct xfs_defer_ops *dfops, xfs_extlen_t tot); extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, - struct xfs_defer_ops *flist, xfs_extlen_t tot); + struct xfs_defer_ops *dfops, xfs_extlen_t tot); extern int xfs_dir_canenter(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_name *name); diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index 972b2ea6a075..f0e1f713b0be 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -1818,7 +1818,7 @@ xfs_difree_inode_chunk( struct xfs_mount *mp, xfs_agnumber_t agno, struct xfs_inobt_rec_incore *rec, - struct xfs_defer_ops *flist) + struct xfs_defer_ops *dfops) { xfs_agblock_t sagbno = XFS_AGINO_TO_AGBNO(mp, rec->ir_startino); int startidx, endidx; @@ -1829,7 +1829,7 @@ xfs_difree_inode_chunk( if (!xfs_inobt_issparse(rec->ir_holemask)) { /* not sparse, calculate extent info directly */ - xfs_bmap_add_free(mp, flist, XFS_AGB_TO_FSB(mp, agno, sagbno), + xfs_bmap_add_free(mp, dfops, XFS_AGB_TO_FSB(mp, agno, sagbno), mp->m_ialloc_blks); return; } @@ -1873,7 +1873,7 @@ xfs_difree_inode_chunk( ASSERT(agbno % mp->m_sb.sb_spino_align == 0); ASSERT(contigblk % mp->m_sb.sb_spino_align == 0); - xfs_bmap_add_free(mp, flist, XFS_AGB_TO_FSB(mp, agno, agbno), + xfs_bmap_add_free(mp, dfops, XFS_AGB_TO_FSB(mp, agno, agbno), contigblk); /* reset range to current bit and carry on... */ @@ -1890,7 +1890,7 @@ xfs_difree_inobt( struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agino_t agino, - struct xfs_defer_ops *flist, + struct xfs_defer_ops *dfops, struct xfs_icluster *xic, struct xfs_inobt_rec_incore *orec) { @@ -1977,7 +1977,7 @@ xfs_difree_inobt( goto error0; } - xfs_difree_inode_chunk(mp, agno, &rec, flist); + xfs_difree_inode_chunk(mp, agno, &rec, dfops); } else { xic->deleted = 0; @@ -2122,7 +2122,7 @@ int xfs_difree( struct xfs_trans *tp, /* transaction pointer */ xfs_ino_t inode, /* inode to be freed */ - struct xfs_defer_ops *flist, /* extents to free */ + struct xfs_defer_ops *dfops, /* extents to free */ struct xfs_icluster *xic) /* cluster info if deleted */ { /* REFERENCED */ @@ -2174,7 +2174,7 @@ xfs_difree( /* * Fix up the inode allocation btree. */ - error = xfs_difree_inobt(mp, tp, agbp, agino, flist, xic, &rec); + error = xfs_difree_inobt(mp, tp, agbp, agino, dfops, xic, &rec); if (error) goto error0; diff --git a/fs/xfs/libxfs/xfs_ialloc.h b/fs/xfs/libxfs/xfs_ialloc.h index 2e06b675d8bb..0bb89669fc07 100644 --- a/fs/xfs/libxfs/xfs_ialloc.h +++ b/fs/xfs/libxfs/xfs_ialloc.h @@ -95,7 +95,7 @@ int /* error */ xfs_difree( struct xfs_trans *tp, /* transaction pointer */ xfs_ino_t inode, /* inode to be freed */ - struct xfs_defer_ops *flist, /* extents to free */ + struct xfs_defer_ops *dfops, /* extents to free */ struct xfs_icluster *ifree); /* cluster info if deleted */ /* diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 07954e8f803b..3bbe46bdbc6e 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -685,7 +685,7 @@ xfs_bmap_punch_delalloc_range( xfs_bmbt_irec_t imap; int nimaps = 1; xfs_fsblock_t firstblock; - struct xfs_defer_ops flist; + struct xfs_defer_ops dfops; /* * Map the range first and check that it is a delalloc extent @@ -716,18 +716,18 @@ xfs_bmap_punch_delalloc_range( WARN_ON(imap.br_blockcount == 0); /* - * Note: while we initialise the firstblock/flist pair, they + * Note: while we initialise the firstblock/dfops pair, they * should never be used because blocks should never be * allocated or freed for a delalloc extent and hence we need * don't cancel or finish them after the xfs_bunmapi() call. */ - xfs_defer_init(&flist, &firstblock); + xfs_defer_init(&dfops, &firstblock); error = xfs_bunmapi(NULL, ip, start_fsb, 1, 0, 1, &firstblock, - &flist, &done); + &dfops, &done); if (error) break; - ASSERT(!xfs_defer_has_unfinished_work(&flist)); + ASSERT(!xfs_defer_has_unfinished_work(&dfops)); next_block: start_fsb++; remaining--; @@ -884,7 +884,7 @@ xfs_alloc_file_space( int rt; xfs_trans_t *tp; xfs_bmbt_irec_t imaps[1], *imapp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; uint qblocks, resblks, resrtextents; int error; @@ -975,17 +975,17 @@ xfs_alloc_file_space( xfs_trans_ijoin(tp, ip, 0); - xfs_defer_init(&free_list, &firstfsb); + xfs_defer_init(&dfops, &firstfsb); error = xfs_bmapi_write(tp, ip, startoffset_fsb, allocatesize_fsb, alloc_type, &firstfsb, - resblks, imapp, &nimaps, &free_list); + resblks, imapp, &nimaps, &dfops); if (error) goto error0; /* * Complete the transaction */ - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto error0; @@ -1008,7 +1008,7 @@ xfs_alloc_file_space( return error; error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */ - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); xfs_trans_unreserve_quota_nblks(tp, ip, (long)qblocks, 0, quota_flag); error1: /* Just cancel transaction */ @@ -1026,7 +1026,7 @@ xfs_unmap_extent( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t firstfsb; uint resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0); int error; @@ -1045,13 +1045,13 @@ xfs_unmap_extent( xfs_trans_ijoin(tp, ip, 0); - xfs_defer_init(&free_list, &firstfsb); + xfs_defer_init(&dfops, &firstfsb); error = xfs_bunmapi(tp, ip, startoffset_fsb, len_fsb, 0, 2, &firstfsb, - &free_list, done); + &dfops, done); if (error) goto out_bmap_cancel; - error = xfs_defer_finish(&tp, &free_list, ip); + error = xfs_defer_finish(&tp, &dfops, ip); if (error) goto out_bmap_cancel; @@ -1061,7 +1061,7 @@ out_unlock: return error; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); goto out_unlock; @@ -1250,7 +1250,7 @@ xfs_shift_file_space( struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; int error; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; xfs_fileoff_t stop_fsb; xfs_fileoff_t next_fsb; @@ -1328,19 +1328,19 @@ xfs_shift_file_space( xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* * We are using the write transaction in which max 2 bmbt * updates are allowed */ error = xfs_bmap_shift_extents(tp, ip, &next_fsb, shift_fsb, - &done, stop_fsb, &first_block, &free_list, + &done, stop_fsb, &first_block, &dfops, direction, XFS_BMAP_MAX_SHIFT_EXTENTS); if (error) goto out_bmap_cancel; - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; @@ -1350,7 +1350,7 @@ xfs_shift_file_space( return error; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); return error; diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index d9447f1bf483..7a30b8f11db7 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -308,7 +308,7 @@ xfs_qm_dqalloc( xfs_buf_t **O_bpp) { xfs_fsblock_t firstblock; - struct xfs_defer_ops flist; + struct xfs_defer_ops dfops; xfs_bmbt_irec_t map; int nmaps, error; xfs_buf_t *bp; @@ -321,7 +321,7 @@ xfs_qm_dqalloc( /* * Initialize the bmap freelist prior to calling bmapi code. */ - xfs_defer_init(&flist, &firstblock); + xfs_defer_init(&dfops, &firstblock); xfs_ilock(quotip, XFS_ILOCK_EXCL); /* * Return if this type of quotas is turned off while we didn't @@ -337,7 +337,7 @@ xfs_qm_dqalloc( error = xfs_bmapi_write(tp, quotip, offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB, XFS_BMAPI_METADATA, &firstblock, XFS_QM_DQALLOC_SPACE_RES(mp), - &map, &nmaps, &flist); + &map, &nmaps, &dfops); if (error) goto error0; ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB); @@ -383,7 +383,7 @@ xfs_qm_dqalloc( xfs_trans_bhold(tp, bp); - error = xfs_defer_finish(tpp, &flist, NULL); + error = xfs_defer_finish(tpp, &dfops, NULL); if (error) goto error1; @@ -399,7 +399,7 @@ xfs_qm_dqalloc( return 0; error1: - xfs_defer_cancel(&flist); + xfs_defer_cancel(&dfops); error0: xfs_iunlock(quotip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 3e990fb0cb53..4a33a3304369 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -386,7 +386,7 @@ xfs_filestream_new_ag( } flags = (ap->userdata ? XFS_PICK_USERDATA : 0) | - (ap->flist->dop_low ? XFS_PICK_LOWSPACE : 0); + (ap->dfops->dop_low ? XFS_PICK_LOWSPACE : 0); err = xfs_filestream_pick_ag(pip, startag, agp, flags, minlen); diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 3ce50da885ac..e08eaea6327b 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1123,7 +1123,7 @@ xfs_create( struct xfs_inode *ip = NULL; struct xfs_trans *tp = NULL; int error; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; bool unlock_dp_on_error = false; prid_t prid; @@ -1183,7 +1183,7 @@ xfs_create( XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); unlock_dp_on_error = true; - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* * Reserve disk quota and the inode. @@ -1220,7 +1220,7 @@ xfs_create( unlock_dp_on_error = false; error = xfs_dir_createname(tp, dp, name, ip->i_ino, - &first_block, &free_list, resblks ? + &first_block, &dfops, resblks ? resblks - XFS_IALLOC_SPACE_RES(mp) : 0); if (error) { ASSERT(error != -ENOSPC); @@ -1254,7 +1254,7 @@ xfs_create( */ xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp); - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; @@ -1270,7 +1270,7 @@ xfs_create( return 0; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); out_release_inode: @@ -1402,7 +1402,7 @@ xfs_link( xfs_mount_t *mp = tdp->i_mount; xfs_trans_t *tp; int error; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; int resblks; @@ -1453,7 +1453,7 @@ xfs_link( goto error_return; } - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* * Handle initial link state of O_TMPFILE inode @@ -1465,7 +1465,7 @@ xfs_link( } error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino, - &first_block, &free_list, resblks); + &first_block, &dfops, resblks); if (error) goto error_return; xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); @@ -1483,9 +1483,9 @@ xfs_link( if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) { - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); goto error_return; } @@ -1527,7 +1527,7 @@ xfs_itruncate_extents( { struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp = *tpp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; xfs_fileoff_t first_unmap_block; xfs_fileoff_t last_block; @@ -1563,12 +1563,12 @@ xfs_itruncate_extents( ASSERT(first_unmap_block < last_block); unmap_len = last_block - first_unmap_block + 1; while (!done) { - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); error = xfs_bunmapi(tp, ip, first_unmap_block, unmap_len, xfs_bmapi_aflag(whichfork), XFS_ITRUNC_MAX_EXTENTS, - &first_block, &free_list, + &first_block, &dfops, &done); if (error) goto out_bmap_cancel; @@ -1577,7 +1577,7 @@ xfs_itruncate_extents( * Duplicate the transaction that has the permanent * reservation and commit the old transaction. */ - error = xfs_defer_finish(&tp, &free_list, ip); + error = xfs_defer_finish(&tp, &dfops, ip); if (error) goto out_bmap_cancel; @@ -1603,7 +1603,7 @@ out_bmap_cancel: * the transaction can be properly aborted. We just need to make sure * we're not holding any resources that we were not when we came in. */ - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); goto out; } @@ -1744,7 +1744,7 @@ STATIC int xfs_inactive_ifree( struct xfs_inode *ip) { - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; @@ -1781,8 +1781,8 @@ xfs_inactive_ifree( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); - xfs_defer_init(&free_list, &first_block); - error = xfs_ifree(tp, ip, &free_list); + xfs_defer_init(&dfops, &first_block); + error = xfs_ifree(tp, ip, &dfops); if (error) { /* * If we fail to free the inode, shut down. The cancel @@ -1808,11 +1808,11 @@ xfs_inactive_ifree( * Just ignore errors at this point. There is nothing we can do except * to try to keep going. Make sure it's not a silent error. */ - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) { xfs_notice(mp, "%s: xfs_defer_finish returned error %d", __func__, error); - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); } error = xfs_trans_commit(tp); if (error) @@ -2368,7 +2368,7 @@ int xfs_ifree( xfs_trans_t *tp, xfs_inode_t *ip, - struct xfs_defer_ops *flist) + struct xfs_defer_ops *dfops) { int error; struct xfs_icluster xic = { 0 }; @@ -2387,7 +2387,7 @@ xfs_ifree( if (error) return error; - error = xfs_difree(tp, ip->i_ino, flist, &xic); + error = xfs_difree(tp, ip->i_ino, dfops, &xic); if (error) return error; @@ -2490,7 +2490,7 @@ xfs_remove( xfs_trans_t *tp = NULL; int is_dir = S_ISDIR(VFS_I(ip)->i_mode); int error = 0; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; uint resblks; @@ -2572,9 +2572,9 @@ xfs_remove( if (error) goto out_trans_cancel; - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); error = xfs_dir_removename(tp, dp, name, ip->i_ino, - &first_block, &free_list, resblks); + &first_block, &dfops, resblks); if (error) { ASSERT(error != -ENOENT); goto out_bmap_cancel; @@ -2588,7 +2588,7 @@ xfs_remove( if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; @@ -2602,7 +2602,7 @@ xfs_remove( return 0; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); std_return: @@ -2663,7 +2663,7 @@ xfs_sort_for_rename( static int xfs_finish_rename( struct xfs_trans *tp, - struct xfs_defer_ops *free_list) + struct xfs_defer_ops *dfops) { int error; @@ -2674,9 +2674,9 @@ xfs_finish_rename( if (tp->t_mountp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) xfs_trans_set_sync(tp); - error = xfs_defer_finish(&tp, free_list, NULL); + error = xfs_defer_finish(&tp, dfops, NULL); if (error) { - xfs_defer_cancel(free_list); + xfs_defer_cancel(dfops); xfs_trans_cancel(tp); return error; } @@ -2698,7 +2698,7 @@ xfs_cross_rename( struct xfs_inode *dp2, struct xfs_name *name2, struct xfs_inode *ip2, - struct xfs_defer_ops *free_list, + struct xfs_defer_ops *dfops, xfs_fsblock_t *first_block, int spaceres) { @@ -2710,14 +2710,14 @@ xfs_cross_rename( /* Swap inode number for dirent in first parent */ error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino, - first_block, free_list, spaceres); + first_block, dfops, spaceres); if (error) goto out_trans_abort; /* Swap inode number for dirent in second parent */ error = xfs_dir_replace(tp, dp2, name2, ip1->i_ino, - first_block, free_list, spaceres); + first_block, dfops, spaceres); if (error) goto out_trans_abort; @@ -2732,7 +2732,7 @@ xfs_cross_rename( if (S_ISDIR(VFS_I(ip2)->i_mode)) { error = xfs_dir_replace(tp, ip2, &xfs_name_dotdot, dp1->i_ino, first_block, - free_list, spaceres); + dfops, spaceres); if (error) goto out_trans_abort; @@ -2759,7 +2759,7 @@ xfs_cross_rename( if (S_ISDIR(VFS_I(ip1)->i_mode)) { error = xfs_dir_replace(tp, ip1, &xfs_name_dotdot, dp2->i_ino, first_block, - free_list, spaceres); + dfops, spaceres); if (error) goto out_trans_abort; @@ -2798,10 +2798,10 @@ xfs_cross_rename( } xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE); - return xfs_finish_rename(tp, free_list); + return xfs_finish_rename(tp, dfops); out_trans_abort: - xfs_defer_cancel(free_list); + xfs_defer_cancel(dfops); xfs_trans_cancel(tp); return error; } @@ -2856,7 +2856,7 @@ xfs_rename( { struct xfs_mount *mp = src_dp->i_mount; struct xfs_trans *tp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; struct xfs_inode *wip = NULL; /* whiteout inode */ struct xfs_inode *inodes[__XFS_SORT_INODES]; @@ -2945,13 +2945,13 @@ xfs_rename( goto out_trans_cancel; } - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* RENAME_EXCHANGE is unique from here on. */ if (flags & RENAME_EXCHANGE) return xfs_cross_rename(tp, src_dp, src_name, src_ip, target_dp, target_name, target_ip, - &free_list, &first_block, spaceres); + &dfops, &first_block, spaceres); /* * Set up the target. @@ -2973,7 +2973,7 @@ xfs_rename( */ error = xfs_dir_createname(tp, target_dp, target_name, src_ip->i_ino, &first_block, - &free_list, spaceres); + &dfops, spaceres); if (error) goto out_bmap_cancel; @@ -3013,7 +3013,7 @@ xfs_rename( */ error = xfs_dir_replace(tp, target_dp, target_name, src_ip->i_ino, - &first_block, &free_list, spaceres); + &first_block, &dfops, spaceres); if (error) goto out_bmap_cancel; @@ -3048,7 +3048,7 @@ xfs_rename( */ error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot, target_dp->i_ino, - &first_block, &free_list, spaceres); + &first_block, &dfops, spaceres); ASSERT(error != -EEXIST); if (error) goto out_bmap_cancel; @@ -3087,10 +3087,10 @@ xfs_rename( */ if (wip) { error = xfs_dir_replace(tp, src_dp, src_name, wip->i_ino, - &first_block, &free_list, spaceres); + &first_block, &dfops, spaceres); } else error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino, - &first_block, &free_list, spaceres); + &first_block, &dfops, spaceres); if (error) goto out_bmap_cancel; @@ -3125,13 +3125,13 @@ xfs_rename( if (new_parent) xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE); - error = xfs_finish_rename(tp, &free_list); + error = xfs_finish_rename(tp, &dfops); if (wip) IRELE(wip); return error; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); out_release_wip: diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index dc93c20fb9ad..2114d53df433 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -129,7 +129,7 @@ xfs_iomap_write_direct( int quota_flag; int rt; xfs_trans_t *tp; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; uint qblocks, resblks, resrtextents; int error; int lockmode; @@ -232,18 +232,18 @@ xfs_iomap_write_direct( * From this point onwards we overwrite the imap pointer that the * caller gave to us. */ - xfs_defer_init(&free_list, &firstfsb); + xfs_defer_init(&dfops, &firstfsb); nimaps = 1; error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, bmapi_flags, &firstfsb, resblks, imap, - &nimaps, &free_list); + &nimaps, &dfops); if (error) goto out_bmap_cancel; /* * Complete the transaction */ - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; @@ -267,7 +267,7 @@ out_unlock: return error; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); xfs_trans_unreserve_quota_nblks(tp, ip, (long)qblocks, 0, quota_flag); out_trans_cancel: xfs_trans_cancel(tp); @@ -686,7 +686,7 @@ xfs_iomap_write_allocate( xfs_fileoff_t offset_fsb, last_block; xfs_fileoff_t end_fsb, map_start_fsb; xfs_fsblock_t first_block; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_filblks_t count_fsb; xfs_trans_t *tp; int nimaps; @@ -728,7 +728,7 @@ xfs_iomap_write_allocate( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* * it is possible that the extents have changed since @@ -784,11 +784,11 @@ xfs_iomap_write_allocate( error = xfs_bmapi_write(tp, ip, map_start_fsb, count_fsb, 0, &first_block, nres, imap, &nimaps, - &free_list); + &dfops); if (error) goto trans_cancel; - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto trans_cancel; @@ -822,7 +822,7 @@ xfs_iomap_write_allocate( } trans_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); xfs_trans_cancel(tp); error0: xfs_iunlock(ip, XFS_ILOCK_EXCL); @@ -843,7 +843,7 @@ xfs_iomap_write_unwritten( int nimaps; xfs_trans_t *tp; xfs_bmbt_irec_t imap; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsize_t i_size; uint resblks; int error; @@ -887,11 +887,11 @@ xfs_iomap_write_unwritten( /* * Modify the unwritten extent state of the buffer. */ - xfs_defer_init(&free_list, &firstfsb); + xfs_defer_init(&dfops, &firstfsb); nimaps = 1; error = xfs_bmapi_write(tp, ip, offset_fsb, count_fsb, XFS_BMAPI_CONVERT, &firstfsb, resblks, - &imap, &nimaps, &free_list); + &imap, &nimaps, &dfops); if (error) goto error_on_bmapi_transaction; @@ -910,7 +910,7 @@ xfs_iomap_write_unwritten( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); } - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto error_on_bmapi_transaction; @@ -937,7 +937,7 @@ xfs_iomap_write_unwritten( return 0; error_on_bmapi_transaction: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_ILOCK_EXCL); return error; diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index c761a6a75904..802bcc326d9f 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -770,7 +770,7 @@ xfs_growfs_rt_alloc( xfs_daddr_t d; /* disk block address */ int error; /* error return value */ xfs_fsblock_t firstblock;/* first block allocated in xaction */ - struct xfs_defer_ops flist; /* list of freed blocks */ + struct xfs_defer_ops dfops; /* list of freed blocks */ xfs_fsblock_t fsbno; /* filesystem block for bno */ struct xfs_bmbt_irec map; /* block map output */ int nmap; /* number of block maps */ @@ -795,14 +795,14 @@ xfs_growfs_rt_alloc( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - xfs_defer_init(&flist, &firstblock); + xfs_defer_init(&dfops, &firstblock); /* * Allocate blocks to the bitmap file. */ nmap = 1; error = xfs_bmapi_write(tp, ip, oblocks, nblocks - oblocks, XFS_BMAPI_METADATA, &firstblock, - resblks, &map, &nmap, &flist); + resblks, &map, &nmap, &dfops); if (!error && nmap < 1) error = -ENOSPC; if (error) @@ -810,7 +810,7 @@ xfs_growfs_rt_alloc( /* * Free any blocks freed up in the transaction, then commit. */ - error = xfs_defer_finish(&tp, &flist, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; error = xfs_trans_commit(tp); @@ -863,7 +863,7 @@ xfs_growfs_rt_alloc( return 0; out_bmap_cancel: - xfs_defer_cancel(&flist); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); return error; diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 3b005ec33886..58142aeeeea6 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -173,7 +173,7 @@ xfs_symlink( struct xfs_inode *ip = NULL; int error = 0; int pathlen; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; xfs_fsblock_t first_block; bool unlock_dp_on_error = false; xfs_fileoff_t first_fsb; @@ -270,7 +270,7 @@ xfs_symlink( * Initialize the bmap freelist prior to calling either * bmapi or the directory create code. */ - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); /* * Allocate an inode for the symlink. @@ -314,7 +314,7 @@ xfs_symlink( error = xfs_bmapi_write(tp, ip, first_fsb, fs_blocks, XFS_BMAPI_METADATA, &first_block, resblks, - mval, &nmaps, &free_list); + mval, &nmaps, &dfops); if (error) goto out_bmap_cancel; @@ -362,7 +362,7 @@ xfs_symlink( * Create the directory entry for the symlink. */ error = xfs_dir_createname(tp, dp, link_name, ip->i_ino, - &first_block, &free_list, resblks); + &first_block, &dfops, resblks); if (error) goto out_bmap_cancel; xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); @@ -377,7 +377,7 @@ xfs_symlink( xfs_trans_set_sync(tp); } - error = xfs_defer_finish(&tp, &free_list, NULL); + error = xfs_defer_finish(&tp, &dfops, NULL); if (error) goto out_bmap_cancel; @@ -393,7 +393,7 @@ xfs_symlink( return 0; out_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); out_trans_cancel: xfs_trans_cancel(tp); out_release_inode: @@ -427,7 +427,7 @@ xfs_inactive_symlink_rmt( int done; int error; xfs_fsblock_t first_block; - struct xfs_defer_ops free_list; + struct xfs_defer_ops dfops; int i; xfs_mount_t *mp; xfs_bmbt_irec_t mval[XFS_SYMLINK_MAPS]; @@ -466,7 +466,7 @@ xfs_inactive_symlink_rmt( * Find the block(s) so we can inval and unmap them. */ done = 0; - xfs_defer_init(&free_list, &first_block); + xfs_defer_init(&dfops, &first_block); nmaps = ARRAY_SIZE(mval); error = xfs_bmapi_read(ip, 0, xfs_symlink_blocks(mp, size), mval, &nmaps, 0); @@ -486,17 +486,17 @@ xfs_inactive_symlink_rmt( xfs_trans_binval(tp, bp); } /* - * Unmap the dead block(s) to the free_list. + * Unmap the dead block(s) to the dfops. */ error = xfs_bunmapi(tp, ip, 0, size, 0, nmaps, - &first_block, &free_list, &done); + &first_block, &dfops, &done); if (error) goto error_bmap_cancel; ASSERT(done); /* * Commit the first transaction. This logs the EFI and the inode. */ - error = xfs_defer_finish(&tp, &free_list, ip); + error = xfs_defer_finish(&tp, &dfops, ip); if (error) goto error_bmap_cancel; /* @@ -526,7 +526,7 @@ xfs_inactive_symlink_rmt( return 0; error_bmap_cancel: - xfs_defer_cancel(&free_list); + xfs_defer_cancel(&dfops); error_trans_cancel: xfs_trans_cancel(tp); error_unlock: -- cgit v1.2.3 From dc42375d5f55d2c48d363957621d7aa913bdc5a1 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:23:49 +1000 Subject: xfs: refactor redo intent item processing Refactor the EFI intent item recovery (and cancellation) functions into a general function that scans the AIL and an intent item type specific handler. Move the function that recovers a single EFI item into the extent free item code. We'll want the generalized function when we start wiring up more redo item types. Furthermore, ensure that log recovery only replays the redo items that were in the AIL prior to recovery by checking the item LSN against the largest LSN seen during log scanning. As written this should never happen, but we can be defensive anyway. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_extfree_item.c | 65 +++++++++++++++++ fs/xfs/xfs_extfree_item.h | 3 + fs/xfs/xfs_log_recover.c | 182 +++++++++++++++++++++------------------------- 3 files changed, 151 insertions(+), 99 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index ab779460ecbf..8c49f883934a 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -20,6 +20,7 @@ #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_bit.h" #include "xfs_mount.h" #include "xfs_trans.h" #include "xfs_trans_priv.h" @@ -486,3 +487,67 @@ xfs_efd_init( return efdp; } + +/* + * Process an extent free intent item that was recovered from + * the log. We need to free the extents that it describes. + */ +int +xfs_efi_recover( + struct xfs_mount *mp, + struct xfs_efi_log_item *efip) +{ + struct xfs_efd_log_item *efdp; + struct xfs_trans *tp; + int i; + int error = 0; + xfs_extent_t *extp; + xfs_fsblock_t startblock_fsb; + + ASSERT(!test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)); + + /* + * First check the validity of the extents described by the + * EFI. If any are bad, then assume that all are bad and + * just toss the EFI. + */ + for (i = 0; i < efip->efi_format.efi_nextents; i++) { + extp = &(efip->efi_format.efi_extents[i]); + startblock_fsb = XFS_BB_TO_FSB(mp, + XFS_FSB_TO_DADDR(mp, extp->ext_start)); + if ((startblock_fsb == 0) || + (extp->ext_len == 0) || + (startblock_fsb >= mp->m_sb.sb_dblocks) || + (extp->ext_len >= mp->m_sb.sb_agblocks)) { + /* + * This will pull the EFI from the AIL and + * free the memory associated with it. + */ + set_bit(XFS_EFI_RECOVERED, &efip->efi_flags); + xfs_efi_release(efip); + return -EIO; + } + } + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); + if (error) + return error; + efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); + + for (i = 0; i < efip->efi_format.efi_nextents; i++) { + extp = &(efip->efi_format.efi_extents[i]); + error = xfs_trans_free_extent(tp, efdp, extp->ext_start, + extp->ext_len); + if (error) + goto abort_error; + + } + + set_bit(XFS_EFI_RECOVERED, &efip->efi_flags); + error = xfs_trans_commit(tp); + return error; + +abort_error: + xfs_trans_cancel(tp); + return error; +} diff --git a/fs/xfs/xfs_extfree_item.h b/fs/xfs/xfs_extfree_item.h index 8fa8651705e1..a32c794a86b7 100644 --- a/fs/xfs/xfs_extfree_item.h +++ b/fs/xfs/xfs_extfree_item.h @@ -98,4 +98,7 @@ int xfs_efi_copy_format(xfs_log_iovec_t *buf, void xfs_efi_item_free(xfs_efi_log_item_t *); void xfs_efi_release(struct xfs_efi_log_item *); +int xfs_efi_recover(struct xfs_mount *mp, + struct xfs_efi_log_item *efip); + #endif /* __XFS_EXTFREE_ITEM_H__ */ diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 080b54b13154..af608aade6df 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -24,7 +24,6 @@ #include "xfs_bit.h" #include "xfs_sb.h" #include "xfs_mount.h" -#include "xfs_defer.h" #include "xfs_da_format.h" #include "xfs_da_btree.h" #include "xfs_inode.h" @@ -4165,126 +4164,112 @@ xlog_recover_process_data( return 0; } -/* - * Process an extent free intent item that was recovered from - * the log. We need to free the extents that it describes. - */ +/* Recover the EFI if necessary. */ STATIC int xlog_recover_process_efi( - xfs_mount_t *mp, - xfs_efi_log_item_t *efip) + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) { - xfs_efd_log_item_t *efdp; - xfs_trans_t *tp; - int i; - int error = 0; - xfs_extent_t *extp; - xfs_fsblock_t startblock_fsb; - - ASSERT(!test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)); + struct xfs_efi_log_item *efip; + int error; /* - * First check the validity of the extents described by the - * EFI. If any are bad, then assume that all are bad and - * just toss the EFI. + * Skip EFIs that we've already processed. */ - for (i = 0; i < efip->efi_format.efi_nextents; i++) { - extp = &(efip->efi_format.efi_extents[i]); - startblock_fsb = XFS_BB_TO_FSB(mp, - XFS_FSB_TO_DADDR(mp, extp->ext_start)); - if ((startblock_fsb == 0) || - (extp->ext_len == 0) || - (startblock_fsb >= mp->m_sb.sb_dblocks) || - (extp->ext_len >= mp->m_sb.sb_agblocks)) { - /* - * This will pull the EFI from the AIL and - * free the memory associated with it. - */ - set_bit(XFS_EFI_RECOVERED, &efip->efi_flags); - xfs_efi_release(efip); - return -EIO; - } - } + efip = container_of(lip, struct xfs_efi_log_item, efi_item); + if (test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)) + return 0; - error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); - if (error) - return error; - efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); + spin_unlock(&ailp->xa_lock); + error = xfs_efi_recover(mp, efip); + spin_lock(&ailp->xa_lock); - for (i = 0; i < efip->efi_format.efi_nextents; i++) { - extp = &(efip->efi_format.efi_extents[i]); - error = xfs_trans_free_extent(tp, efdp, extp->ext_start, - extp->ext_len); - if (error) - goto abort_error; + return error; +} - } +/* Release the EFI since we're cancelling everything. */ +STATIC void +xlog_recover_cancel_efi( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_efi_log_item *efip; - set_bit(XFS_EFI_RECOVERED, &efip->efi_flags); - error = xfs_trans_commit(tp); - return error; + efip = container_of(lip, struct xfs_efi_log_item, efi_item); -abort_error: - xfs_trans_cancel(tp); - return error; + spin_unlock(&ailp->xa_lock); + xfs_efi_release(efip); + spin_lock(&ailp->xa_lock); +} + +/* Is this log item a deferred action intent? */ +static inline bool xlog_item_is_intent(struct xfs_log_item *lip) +{ + switch (lip->li_type) { + case XFS_LI_EFI: + return true; + default: + return false; + } } /* - * When this is called, all of the EFIs which did not have - * corresponding EFDs should be in the AIL. What we do now - * is free the extents associated with each one. + * When this is called, all of the log intent items which did not have + * corresponding log done items should be in the AIL. What we do now + * is update the data structures associated with each one. * - * Since we process the EFIs in normal transactions, they - * will be removed at some point after the commit. This prevents - * us from just walking down the list processing each one. - * We'll use a flag in the EFI to skip those that we've already - * processed and use the AIL iteration mechanism's generation - * count to try to speed this up at least a bit. + * Since we process the log intent items in normal transactions, they + * will be removed at some point after the commit. This prevents us + * from just walking down the list processing each one. We'll use a + * flag in the intent item to skip those that we've already processed + * and use the AIL iteration mechanism's generation count to try to + * speed this up at least a bit. * - * When we start, we know that the EFIs are the only things in - * the AIL. As we process them, however, other items are added - * to the AIL. Since everything added to the AIL must come after - * everything already in the AIL, we stop processing as soon as - * we see something other than an EFI in the AIL. + * When we start, we know that the intents are the only things in the + * AIL. As we process them, however, other items are added to the + * AIL. */ STATIC int -xlog_recover_process_efis( +xlog_recover_process_intents( struct xlog *log) { struct xfs_log_item *lip; - struct xfs_efi_log_item *efip; int error = 0; struct xfs_ail_cursor cur; struct xfs_ail *ailp; + xfs_lsn_t last_lsn; ailp = log->l_ailp; spin_lock(&ailp->xa_lock); lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); + last_lsn = xlog_assign_lsn(log->l_curr_cycle, log->l_curr_block); while (lip != NULL) { /* - * We're done when we see something other than an EFI. - * There should be no EFIs left in the AIL now. + * We're done when we see something other than an intent. + * There should be no intents left in the AIL now. */ - if (lip->li_type != XFS_LI_EFI) { + if (!xlog_item_is_intent(lip)) { #ifdef DEBUG for (; lip; lip = xfs_trans_ail_cursor_next(ailp, &cur)) - ASSERT(lip->li_type != XFS_LI_EFI); + ASSERT(!xlog_item_is_intent(lip)); #endif break; } /* - * Skip EFIs that we've already processed. + * We should never see a redo item with a LSN higher than + * the last transaction we found in the log at the start + * of recovery. */ - efip = container_of(lip, struct xfs_efi_log_item, efi_item); - if (test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)) { - lip = xfs_trans_ail_cursor_next(ailp, &cur); - continue; - } + ASSERT(XFS_LSN_CMP(last_lsn, lip->li_lsn) >= 0); - spin_unlock(&ailp->xa_lock); - error = xlog_recover_process_efi(log->l_mp, efip); - spin_lock(&ailp->xa_lock); + switch (lip->li_type) { + case XFS_LI_EFI: + error = xlog_recover_process_efi(log->l_mp, ailp, lip); + break; + } if (error) goto out; lip = xfs_trans_ail_cursor_next(ailp, &cur); @@ -4296,15 +4281,14 @@ out: } /* - * A cancel occurs when the mount has failed and we're bailing out. Release all - * pending EFIs so they don't pin the AIL. + * A cancel occurs when the mount has failed and we're bailing out. + * Release all pending log intent items so they don't pin the AIL. */ STATIC int -xlog_recover_cancel_efis( +xlog_recover_cancel_intents( struct xlog *log) { struct xfs_log_item *lip; - struct xfs_efi_log_item *efip; int error = 0; struct xfs_ail_cursor cur; struct xfs_ail *ailp; @@ -4314,22 +4298,22 @@ xlog_recover_cancel_efis( lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); while (lip != NULL) { /* - * We're done when we see something other than an EFI. - * There should be no EFIs left in the AIL now. + * We're done when we see something other than an intent. + * There should be no intents left in the AIL now. */ - if (lip->li_type != XFS_LI_EFI) { + if (!xlog_item_is_intent(lip)) { #ifdef DEBUG for (; lip; lip = xfs_trans_ail_cursor_next(ailp, &cur)) - ASSERT(lip->li_type != XFS_LI_EFI); + ASSERT(!xlog_item_is_intent(lip)); #endif break; } - efip = container_of(lip, struct xfs_efi_log_item, efi_item); - - spin_unlock(&ailp->xa_lock); - xfs_efi_release(efip); - spin_lock(&ailp->xa_lock); + switch (lip->li_type) { + case XFS_LI_EFI: + xlog_recover_cancel_efi(log->l_mp, ailp, lip); + break; + } lip = xfs_trans_ail_cursor_next(ailp, &cur); } @@ -5140,16 +5124,16 @@ xlog_recover_finish( */ if (log->l_flags & XLOG_RECOVERY_NEEDED) { int error; - error = xlog_recover_process_efis(log); + error = xlog_recover_process_intents(log); if (error) { - xfs_alert(log->l_mp, "Failed to recover EFIs"); + xfs_alert(log->l_mp, "Failed to recover intents"); return error; } /* - * Sync the log to get all the EFIs out of the AIL. + * Sync the log to get all the intents out of the AIL. * This isn't absolutely necessary, but it helps in * case the unlink transactions would have problems - * pushing the EFIs out of the way. + * pushing the intents out of the way. */ xfs_log_force(log->l_mp, XFS_LOG_SYNC); @@ -5174,7 +5158,7 @@ xlog_recover_cancel( int error = 0; if (log->l_flags & XLOG_RECOVERY_NEEDED) - error = xlog_recover_cancel_efis(log); + error = xlog_recover_cancel_intents(log); return error; } -- cgit v1.2.3 From ba9e780246a15a35f8ebe5b60f4a11bb58e85bda Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:26:33 +1000 Subject: xfs: add tracepoints and error injection for deferred extent freeing Add a couple of tracepoints for the deferred extent free operation and a site for injecting errors while finishing the operation. This makes it easier to debug deferred ops and test log redo. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 7 +++++++ fs/xfs/libxfs/xfs_bmap.c | 2 ++ fs/xfs/xfs_error.h | 4 +++- fs/xfs/xfs_trace.h | 5 ++++- 4 files changed, 16 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 5993f87eee3e..22ac3f136f21 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -2702,6 +2702,13 @@ xfs_free_extent( ASSERT(len != 0); + trace_xfs_bmap_free_deferred(mp, agno, 0, agbno, len); + + if (XFS_TEST_ERROR(false, mp, + XFS_ERRTAG_FREE_EXTENT, + XFS_RANDOM_FREE_EXTENT)) + return -EIO; + error = xfs_free_extent_fix_freelist(tp, agno, &agbp); if (error) return error; diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 540a6b7928af..8e14ff45c73a 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -596,6 +596,8 @@ xfs_bmap_add_free( new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP); new->xefi_startblock = bno; new->xefi_blockcount = (xfs_extlen_t)len; + trace_xfs_bmap_free_defer(mp, XFS_FSB_TO_AGNO(mp, bno), 0, + XFS_FSB_TO_AGBNO(mp, bno), len); xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list); } diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h index 2e4f67f68856..da6f951435c9 100644 --- a/fs/xfs/xfs_error.h +++ b/fs/xfs/xfs_error.h @@ -90,7 +90,8 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_ERRTAG_STRATCMPL_IOERR 19 #define XFS_ERRTAG_DIOWRITE_IOERR 20 #define XFS_ERRTAG_BMAPIFORMAT 21 -#define XFS_ERRTAG_MAX 22 +#define XFS_ERRTAG_FREE_EXTENT 22 +#define XFS_ERRTAG_MAX 23 /* * Random factors for above tags, 1 means always, 2 means 1/2 time, etc. @@ -117,6 +118,7 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_RANDOM_STRATCMPL_IOERR (XFS_RANDOM_DEFAULT/10) #define XFS_RANDOM_DIOWRITE_IOERR (XFS_RANDOM_DEFAULT/10) #define XFS_RANDOM_BMAPIFORMAT XFS_RANDOM_DEFAULT +#define XFS_RANDOM_FREE_EXTENT 1 #ifdef DEBUG extern int xfs_error_test_active; diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index a45b030d4592..939caf5f5a45 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2417,9 +2417,12 @@ DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_cancel); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); -DEFINE_PHYS_EXTENT_DEFERRED_EVENT(xfs_defer_phys_extent); DEFINE_MAP_EXTENT_DEFERRED_EVENT(xfs_defer_map_extent); +#define DEFINE_BMAP_FREE_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT +DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer); +DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH -- cgit v1.2.3 From df3954ff72590fd20b68261a0c939e40fa3579ea Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:29:42 +1000 Subject: xfs: increase XFS_BTREE_MAXLEVELS to fit the rmapbt By my calculations, a 1,073,741,824 block AG with a 1k block size can attain a maximum height of 9. Assuming a record size of 24 bytes, a key/ptr size of 44 bytes, and half-full btree nodes, we'd need 53,687,092 blocks for the records and ~6 million blocks for the keys. That requires a btree of height 9 based on the following derivation: Block size = 1024b sblock CRC header = 56b == 1024-56 = 968 bytes for tree data rmapbt record = 24b == 40 records per leaf block rmapbt ptr/key = 44b == 22 ptr/keys per block Worst case, each block is half full, so 20 records and 11 ptrs per block. 1073741824 rmap records / 20 records per block == 53687092 leaf blocks 53687092 leaves / 11 ptrs per block == 4880645 level 1 blocks == 443695 level 2 blocks == 40336 level 3 blocks == 3667 level 4 blocks == 334 level 5 blocks == 31 level 6 blocks == 3 level 7 blocks == 1 level 8 block Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 154eefdcb730..ac5cd6ef6859 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -133,7 +133,7 @@ do { \ } \ } while (0) -#define XFS_BTREE_MAXLEVELS 8 /* max of all btrees */ +#define XFS_BTREE_MAXLEVELS 9 /* max of all btrees */ struct xfs_btree_ops { /* size of the key and record structures */ -- cgit v1.2.3 From b87049444ac4a6515ba0427d16a73438b646435b Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:30:32 +1000 Subject: xfs: introduce rmap btree definitions Originally-From: Dave Chinner Add new per-ag rmap btree definitions to the per-ag structures. The rmap btree will sit in the empty slots on disk after the free space btrees, and hence form a part of the array of space management btrees. This requires the definition of the btree to be contiguous with the free space btrees. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 6 ++++++ fs/xfs/libxfs/xfs_btree.c | 4 ++-- fs/xfs/libxfs/xfs_btree.h | 3 +++ fs/xfs/libxfs/xfs_format.h | 22 +++++++++++++++++----- fs/xfs/libxfs/xfs_types.h | 4 ++-- 5 files changed, 30 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 22ac3f136f21..c02d8e8419ce 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -2272,6 +2272,10 @@ xfs_agf_verify( be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) > XFS_BTREE_MAXLEVELS) return false; + if (xfs_sb_version_hasrmapbt(&mp->m_sb) && + be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS) + return false; + /* * during growfs operations, the perag is not fully initialised, * so we can't use it for any useful checking. growfs ensures we can't @@ -2403,6 +2407,8 @@ xfs_alloc_read_agf( be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]); pag->pagf_levels[XFS_BTNUM_CNTi] = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]); + pag->pagf_levels[XFS_BTNUM_RMAPi] = + be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]); spin_lock_init(&pag->pagb_lock); pag->pagb_count = 0; pag->pagb_tree = RB_ROOT; diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 4fc46529753d..4066b01a1807 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -44,9 +44,9 @@ kmem_zone_t *xfs_btree_cur_zone; * Btree magic numbers. */ static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = { - { XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC, + { XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, 0, XFS_BMAP_MAGIC, XFS_IBT_MAGIC, XFS_FIBT_MAGIC }, - { XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, + { XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, XFS_RMAP_CRC_MAGIC, XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC } }; #define xfs_btree_magic(cur) \ diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index ac5cd6ef6859..25f1b69e7515 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -77,6 +77,7 @@ union xfs_btree_rec { #define XFS_BTNUM_BMAP ((xfs_btnum_t)XFS_BTNUM_BMAPi) #define XFS_BTNUM_INO ((xfs_btnum_t)XFS_BTNUM_INOi) #define XFS_BTNUM_FINO ((xfs_btnum_t)XFS_BTNUM_FINOi) +#define XFS_BTNUM_RMAP ((xfs_btnum_t)XFS_BTNUM_RMAPi) /* * For logging record fields. @@ -109,6 +110,7 @@ do { \ case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(__mp, bmbt, stat); break; \ case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \ case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \ + case XFS_BTNUM_RMAP: break; \ case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ } \ } while (0) @@ -129,6 +131,7 @@ do { \ __XFS_BTREE_STATS_ADD(__mp, ibt, stat, val); break; \ case XFS_BTNUM_FINO: \ __XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \ + case XFS_BTNUM_RMAP: break; \ case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ } \ } while (0) diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index adb204d40f22..74f4d3da598c 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -455,6 +455,7 @@ xfs_sb_has_compat_feature( } #define XFS_SB_FEAT_RO_COMPAT_FINOBT (1 << 0) /* free inode btree */ +#define XFS_SB_FEAT_RO_COMPAT_RMAPBT (1 << 1) /* reverse map btree */ #define XFS_SB_FEAT_RO_COMPAT_ALL \ (XFS_SB_FEAT_RO_COMPAT_FINOBT) #define XFS_SB_FEAT_RO_COMPAT_UNKNOWN ~XFS_SB_FEAT_RO_COMPAT_ALL @@ -538,6 +539,12 @@ static inline bool xfs_sb_version_hasmetauuid(struct xfs_sb *sbp) (sbp->sb_features_incompat & XFS_SB_FEAT_INCOMPAT_META_UUID); } +static inline bool xfs_sb_version_hasrmapbt(struct xfs_sb *sbp) +{ + return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) && + (sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_RMAPBT); +} + /* * end of superblock version macros */ @@ -598,10 +605,10 @@ xfs_is_quota_inode(struct xfs_sb *sbp, xfs_ino_t ino) #define XFS_AGI_GOOD_VERSION(v) ((v) == XFS_AGI_VERSION) /* - * Btree number 0 is bno, 1 is cnt. This value gives the size of the + * Btree number 0 is bno, 1 is cnt, 2 is rmap. This value gives the size of the * arrays below. */ -#define XFS_BTNUM_AGF ((int)XFS_BTNUM_CNTi + 1) +#define XFS_BTNUM_AGF ((int)XFS_BTNUM_RMAPi + 1) /* * The second word of agf_levels in the first a.g. overlaps the EFS @@ -618,12 +625,10 @@ typedef struct xfs_agf { __be32 agf_seqno; /* sequence # starting from 0 */ __be32 agf_length; /* size in blocks of a.g. */ /* - * Freespace information + * Freespace and rmap information */ __be32 agf_roots[XFS_BTNUM_AGF]; /* root blocks */ - __be32 agf_spare0; /* spare field */ __be32 agf_levels[XFS_BTNUM_AGF]; /* btree levels */ - __be32 agf_spare1; /* spare field */ __be32 agf_flfirst; /* first freelist block's index */ __be32 agf_fllast; /* last freelist block's index */ @@ -1307,6 +1312,13 @@ typedef __be32 xfs_inobt_ptr_t; #define XFS_IBT_BLOCK(mp) ((xfs_agblock_t)(XFS_CNT_BLOCK(mp) + 1)) #define XFS_FIBT_BLOCK(mp) ((xfs_agblock_t)(XFS_IBT_BLOCK(mp) + 1)) +/* + * Reverse mapping btree format definitions + * + * There is a btree for the reverse map per allocation group + */ +#define XFS_RMAP_CRC_MAGIC 0x524d4233 /* 'RMB3' */ + /* * The first data block of an AG depends on whether the filesystem was formatted * with the finobt feature. If so, account for the finobt reserved root btree diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h index b79dc66b2ecd..3d503647f26b 100644 --- a/fs/xfs/libxfs/xfs_types.h +++ b/fs/xfs/libxfs/xfs_types.h @@ -108,8 +108,8 @@ typedef enum { } xfs_lookup_t; typedef enum { - XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_BMAPi, XFS_BTNUM_INOi, - XFS_BTNUM_FINOi, XFS_BTNUM_MAX + XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi, + XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_MAX } xfs_btnum_t; struct xfs_name { -- cgit v1.2.3 From 00f4e4f9073cb6d455c27dc8e92b421edcdc5011 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:31:11 +1000 Subject: xfs: add rmap btree stats infrastructure Originally-From: Dave Chinner The rmap btree will require the same stats as all the other generic btrees, so add all the code for that now. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.h | 5 +++-- fs/xfs/xfs_stats.c | 1 + fs/xfs/xfs_stats.h | 18 +++++++++++++++++- 3 files changed, 21 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 25f1b69e7515..a480947f094b 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -110,7 +110,7 @@ do { \ case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(__mp, bmbt, stat); break; \ case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \ case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \ - case XFS_BTNUM_RMAP: break; \ + case XFS_BTNUM_RMAP: __XFS_BTREE_STATS_INC(__mp, rmap, stat); break; \ case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ } \ } while (0) @@ -131,7 +131,8 @@ do { \ __XFS_BTREE_STATS_ADD(__mp, ibt, stat, val); break; \ case XFS_BTNUM_FINO: \ __XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \ - case XFS_BTNUM_RMAP: break; \ + case XFS_BTNUM_RMAP: \ + __XFS_BTREE_STATS_ADD(__mp, rmap, stat, val); break; \ case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ } \ } while (0) diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c index 8686df6c7609..f04f547fe5b2 100644 --- a/fs/xfs/xfs_stats.c +++ b/fs/xfs/xfs_stats.c @@ -61,6 +61,7 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf) { "bmbt2", XFSSTAT_END_BMBT_V2 }, { "ibt2", XFSSTAT_END_IBT_V2 }, { "fibt2", XFSSTAT_END_FIBT_V2 }, + { "rmapbt", XFSSTAT_END_RMAP_V2 }, /* we print both series of quota information together */ { "qm", XFSSTAT_END_QM }, }; diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h index 483b0eff1988..657865f51e78 100644 --- a/fs/xfs/xfs_stats.h +++ b/fs/xfs/xfs_stats.h @@ -197,7 +197,23 @@ struct xfsstats { __uint32_t xs_fibt_2_alloc; __uint32_t xs_fibt_2_free; __uint32_t xs_fibt_2_moves; -#define XFSSTAT_END_XQMSTAT (XFSSTAT_END_FIBT_V2+6) +#define XFSSTAT_END_RMAP_V2 (XFSSTAT_END_FIBT_V2+15) + __uint32_t xs_rmap_2_lookup; + __uint32_t xs_rmap_2_compare; + __uint32_t xs_rmap_2_insrec; + __uint32_t xs_rmap_2_delrec; + __uint32_t xs_rmap_2_newroot; + __uint32_t xs_rmap_2_killroot; + __uint32_t xs_rmap_2_increment; + __uint32_t xs_rmap_2_decrement; + __uint32_t xs_rmap_2_lshift; + __uint32_t xs_rmap_2_rshift; + __uint32_t xs_rmap_2_split; + __uint32_t xs_rmap_2_join; + __uint32_t xs_rmap_2_alloc; + __uint32_t xs_rmap_2_free; + __uint32_t xs_rmap_2_moves; +#define XFSSTAT_END_XQMSTAT (XFSSTAT_END_RMAP_V2+6) __uint32_t xs_qm_dqreclaims; __uint32_t xs_qm_dqreclaim_misses; __uint32_t xs_qm_dquot_dups; -- cgit v1.2.3 From 8018026ef29756af6144e2e2e8dffc9c2ed0d6f7 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:31:47 +1000 Subject: xfs: rmap btree add more reserved blocks Originally-From: Dave Chinner XFS reserves a small amount of space in each AG for the minimum number of free blocks needed for operation. Adding the rmap btree increases the number of reserved blocks, but it also increases the complexity of the calculation as the free inode btree is optional (like the rmbt). Rather than calculate the prealloc blocks every time we need to check it, add a function to calculate it at mount time and store it in the struct xfs_mount, and convert the XFS_PREALLOC_BLOCKS macro just to use the xfs-mount variable directly. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 11 +++++++++++ fs/xfs/libxfs/xfs_alloc.h | 2 ++ fs/xfs/libxfs/xfs_format.h | 9 +-------- fs/xfs/xfs_fsops.c | 6 +++--- fs/xfs/xfs_mount.c | 2 ++ fs/xfs/xfs_mount.h | 1 + 6 files changed, 20 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index c02d8e8419ce..43c7e53a6a8e 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -50,6 +50,17 @@ STATIC int xfs_alloc_ag_vextent_size(xfs_alloc_arg_t *); STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *, xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *); +xfs_extlen_t +xfs_prealloc_blocks( + struct xfs_mount *mp) +{ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + return XFS_RMAP_BLOCK(mp) + 1; + if (xfs_sb_version_hasfinobt(&mp->m_sb)) + return XFS_FIBT_BLOCK(mp) + 1; + return XFS_IBT_BLOCK(mp) + 1; +} + /* * Lookup the record equal to [bno, len] in the btree given by cur. */ diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index cf268b2d0b6c..20b54aa87026 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -232,4 +232,6 @@ int xfs_alloc_fix_freelist(struct xfs_alloc_arg *args, int flags); int xfs_free_extent_fix_freelist(struct xfs_trans *tp, xfs_agnumber_t agno, struct xfs_buf **agbp); +xfs_extlen_t xfs_prealloc_blocks(struct xfs_mount *mp); + #endif /* __XFS_ALLOC_H__ */ diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index 74f4d3da598c..3dbed011c29a 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -1319,18 +1319,11 @@ typedef __be32 xfs_inobt_ptr_t; */ #define XFS_RMAP_CRC_MAGIC 0x524d4233 /* 'RMB3' */ -/* - * The first data block of an AG depends on whether the filesystem was formatted - * with the finobt feature. If so, account for the finobt reserved root btree - * block. - */ -#define XFS_PREALLOC_BLOCKS(mp) \ +#define XFS_RMAP_BLOCK(mp) \ (xfs_sb_version_hasfinobt(&((mp)->m_sb)) ? \ XFS_FIBT_BLOCK(mp) + 1 : \ XFS_IBT_BLOCK(mp) + 1) - - /* * BMAP Btree format definitions * diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 4c191d2194e8..0a3b822d9f8c 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -243,7 +243,7 @@ xfs_growfs_data_private( agf->agf_flfirst = cpu_to_be32(1); agf->agf_fllast = 0; agf->agf_flcount = 0; - tmpsize = agsize - XFS_PREALLOC_BLOCKS(mp); + tmpsize = agsize - mp->m_ag_prealloc_blocks; agf->agf_freeblks = cpu_to_be32(tmpsize); agf->agf_longest = cpu_to_be32(tmpsize); if (xfs_sb_version_hascrc(&mp->m_sb)) @@ -340,7 +340,7 @@ xfs_growfs_data_private( agno, 0); arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1); - arec->ar_startblock = cpu_to_be32(XFS_PREALLOC_BLOCKS(mp)); + arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks); arec->ar_blockcount = cpu_to_be32( agsize - be32_to_cpu(arec->ar_startblock)); @@ -369,7 +369,7 @@ xfs_growfs_data_private( agno, 0); arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1); - arec->ar_startblock = cpu_to_be32(XFS_PREALLOC_BLOCKS(mp)); + arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks); arec->ar_blockcount = cpu_to_be32( agsize - be32_to_cpu(arec->ar_startblock)); nfree += be32_to_cpu(arec->ar_blockcount); diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 5f3ba88985d5..24ae96b4078f 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -231,6 +231,8 @@ xfs_initialize_perag( if (maxagi) *maxagi = index; + + mp->m_ag_prealloc_blocks = xfs_prealloc_blocks(mp); return 0; out_unwind: diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index c1b798c72126..0537b1f28f90 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -119,6 +119,7 @@ typedef struct xfs_mount { uint m_ag_maxlevels; /* XFS_AG_MAXLEVELS */ uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */ uint m_in_maxlevels; /* max inobt btree levels. */ + xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */ struct radix_tree_root m_perag_tree; /* per-ag accounting info */ spinlock_t m_perag_lock; /* lock for m_perag_tree */ struct mutex m_growlock; /* growfs mutex */ -- cgit v1.2.3 From 340785cca16246f82ccaf11740d885017a9e9341 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:33:42 +1000 Subject: xfs: add owner field to extent allocation and freeing For the rmap btree to work, we have to feed the extent owner information to the the allocation and freeing functions. This information is what will end up in the rmap btree that tracks allocated extents. While we technically don't need the owner information when freeing extents, passing it allows us to validate that the extent we are removing from the rmap btree actually belonged to the owner we expected it to belong to. We also define a special set of owner values for internal metadata that would otherwise have no owner. This allows us to tell the difference between metadata owned by different per-ag btrees, as well as static fs metadata (e.g. AG headers) and internal journal blocks. There are also a couple of special cases we need to take care of - during EFI recovery, we don't actually know who the original owner was, so we need to pass a wildcard to indicate that we aren't checking the owner for validity. We also need special handling in growfs, as we "free" the space in the last AG when extending it, but because it's new space it has no actual owner... While touching the xfs_bmap_add_free() function, re-order the parameters to put the struct xfs_mount first. Extend the owner field to include both the owner type and some sort of index within the owner. The index field will be used to support reverse mappings when reflink is enabled. When we're freeing extents from an EFI, we don't have the owner information available (rmap updates have their own redo items). xfs_free_extent therefore doesn't need to do an rmap update. Make sure that the log replay code signals this correctly. This is based upon a patch originally from Dave Chinner. It has been extended to add more owner information with the intent of helping recovery operations when things go wrong (e.g. offset of user data block in a file). [dchinner: de-shout the xfs_rmap_*_owner helpers] [darrick: minor style fixes suggested by Christoph Hellwig] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 26 ++++++++++------ fs/xfs/libxfs/xfs_alloc.h | 8 +++-- fs/xfs/libxfs/xfs_bmap.c | 25 +++++++++++---- fs/xfs/libxfs/xfs_bmap.h | 4 ++- fs/xfs/libxfs/xfs_bmap_btree.c | 7 ++++- fs/xfs/libxfs/xfs_format.h | 28 +++++++++++++++++ fs/xfs/libxfs/xfs_ialloc.c | 8 +++-- fs/xfs/libxfs/xfs_ialloc_btree.c | 8 ++++- fs/xfs/libxfs/xfs_rmap.h | 67 ++++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_extfree_item.c | 6 +++- fs/xfs/xfs_fsops.c | 17 +++++++--- fs/xfs/xfs_trans.h | 2 +- fs/xfs/xfs_trans_extfree.c | 8 +++-- 13 files changed, 181 insertions(+), 33 deletions(-) create mode 100644 fs/xfs/libxfs/xfs_rmap.h (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 43c7e53a6a8e..6335b8b85349 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -36,6 +36,7 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" +#include "xfs_rmap.h" struct workqueue_struct *xfs_alloc_wq; @@ -1589,14 +1590,15 @@ error0: /* * Free the extent starting at agno/bno for length. */ -STATIC int /* error */ +STATIC int xfs_free_ag_extent( - xfs_trans_t *tp, /* transaction pointer */ - xfs_buf_t *agbp, /* buffer for a.g. freelist header */ - xfs_agnumber_t agno, /* allocation group number */ - xfs_agblock_t bno, /* starting block number */ - xfs_extlen_t len, /* length of extent */ - int isfl) /* set if is freelist blocks - no sb acctg */ + xfs_trans_t *tp, + xfs_buf_t *agbp, + xfs_agnumber_t agno, + xfs_agblock_t bno, + xfs_extlen_t len, + struct xfs_owner_info *oinfo, + int isfl) { xfs_btree_cur_t *bno_cur; /* cursor for by-block btree */ xfs_btree_cur_t *cnt_cur; /* cursor for by-size btree */ @@ -2005,13 +2007,15 @@ xfs_alloc_fix_freelist( * back on the free list? Maybe we should only do this when space is * getting low or the AGFL is more than half full? */ + xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG); while (pag->pagf_flcount > need) { struct xfs_buf *bp; error = xfs_alloc_get_freelist(tp, agbp, &bno, 0); if (error) goto out_agbp_relse; - error = xfs_free_ag_extent(tp, agbp, args->agno, bno, 1, 1); + error = xfs_free_ag_extent(tp, agbp, args->agno, bno, 1, + &targs.oinfo, 1); if (error) goto out_agbp_relse; bp = xfs_btree_get_bufs(mp, tp, args->agno, bno, 0); @@ -2021,6 +2025,7 @@ xfs_alloc_fix_freelist( memset(&targs, 0, sizeof(targs)); targs.tp = tp; targs.mp = mp; + xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG); targs.agbp = agbp; targs.agno = args->agno; targs.alignment = targs.minlen = targs.prod = targs.isfl = 1; @@ -2709,7 +2714,8 @@ int /* error */ xfs_free_extent( struct xfs_trans *tp, /* transaction pointer */ xfs_fsblock_t bno, /* starting block number of extent */ - xfs_extlen_t len) /* length of extent */ + xfs_extlen_t len, /* length of extent */ + struct xfs_owner_info *oinfo) /* extent owner */ { struct xfs_mount *mp = tp->t_mountp; struct xfs_buf *agbp; @@ -2737,7 +2743,7 @@ xfs_free_extent( agbno + len <= be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_length), err); - error = xfs_free_ag_extent(tp, agbp, agno, agbno, len, 0); + error = xfs_free_ag_extent(tp, agbp, agno, agbno, len, oinfo, 0); if (error) goto err; diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index 20b54aa87026..88053422fda0 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -123,6 +123,7 @@ typedef struct xfs_alloc_arg { char isfl; /* set if is freelist blocks - !acctg */ char userdata; /* mask defining userdata treatment */ xfs_fsblock_t firstblock; /* io first block allocated */ + struct xfs_owner_info oinfo; /* owner of blocks being allocated */ } xfs_alloc_arg_t; /* @@ -208,9 +209,10 @@ xfs_alloc_vextent( */ int /* error */ xfs_free_extent( - struct xfs_trans *tp, /* transaction pointer */ - xfs_fsblock_t bno, /* starting block number of extent */ - xfs_extlen_t len); /* length of extent */ + struct xfs_trans *tp, /* transaction pointer */ + xfs_fsblock_t bno, /* starting block number of extent */ + xfs_extlen_t len, /* length of extent */ + struct xfs_owner_info *oinfo);/* extent owner */ int /* error */ xfs_alloc_lookup_ge( diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 8e14ff45c73a..919069fc7b23 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -46,6 +46,7 @@ #include "xfs_symlink.h" #include "xfs_attr_leaf.h" #include "xfs_filestream.h" +#include "xfs_rmap.h" kmem_zone_t *xfs_bmap_free_item_zone; @@ -571,10 +572,11 @@ xfs_bmap_validate_ret( */ void xfs_bmap_add_free( - struct xfs_mount *mp, /* mount point structure */ - struct xfs_defer_ops *dfops, /* list of extents */ - xfs_fsblock_t bno, /* fs block number of extent */ - xfs_filblks_t len) /* length of extent */ + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + xfs_fsblock_t bno, + xfs_filblks_t len, + struct xfs_owner_info *oinfo) { struct xfs_extent_free_item *new; /* new element */ #ifdef DEBUG @@ -593,9 +595,14 @@ xfs_bmap_add_free( ASSERT(agbno + len <= mp->m_sb.sb_agblocks); #endif ASSERT(xfs_bmap_free_item_zone != NULL); + new = kmem_zone_alloc(xfs_bmap_free_item_zone, KM_SLEEP); new->xefi_startblock = bno; new->xefi_blockcount = (xfs_extlen_t)len; + if (oinfo) + new->xefi_oinfo = *oinfo; + else + xfs_rmap_skip_owner_update(&new->xefi_oinfo); trace_xfs_bmap_free_defer(mp, XFS_FSB_TO_AGNO(mp, bno), 0, XFS_FSB_TO_AGBNO(mp, bno), len); xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_FREE, &new->xefi_list); @@ -628,6 +635,7 @@ xfs_bmap_btree_to_extents( xfs_mount_t *mp; /* mount point structure */ __be64 *pp; /* ptr to block address */ struct xfs_btree_block *rblock;/* root btree block */ + struct xfs_owner_info oinfo; mp = ip->i_mount; ifp = XFS_IFORK_PTR(ip, whichfork); @@ -651,7 +659,8 @@ xfs_bmap_btree_to_extents( cblock = XFS_BUF_TO_BLOCK(cbp); if ((error = xfs_btree_check_block(cur, cblock, 0, cbp))) return error; - xfs_bmap_add_free(mp, cur->bc_private.b.dfops, cbno, 1); + xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, whichfork); + xfs_bmap_add_free(mp, cur->bc_private.b.dfops, cbno, 1, &oinfo); ip->i_d.di_nblocks--; xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L); xfs_trans_binval(tp, cbp); @@ -732,6 +741,7 @@ xfs_bmap_extents_to_btree( memset(&args, 0, sizeof(args)); args.tp = tp; args.mp = mp; + xfs_rmap_ino_bmbt_owner(&args.oinfo, ip->i_ino, whichfork); args.firstblock = *firstblock; if (*firstblock == NULLFSBLOCK) { args.type = XFS_ALLOCTYPE_START_BNO; @@ -878,6 +888,7 @@ xfs_bmap_local_to_extents( memset(&args, 0, sizeof(args)); args.tp = tp; args.mp = ip->i_mount; + xfs_rmap_ino_owner(&args.oinfo, ip->i_ino, whichfork, 0); args.firstblock = *firstblock; /* * Allocate a block. We know we need only one, since the @@ -3660,6 +3671,7 @@ xfs_bmap_btalloc( args.tp = ap->tp; args.mp = mp; args.fsbno = ap->blkno; + xfs_rmap_skip_owner_update(&args.oinfo); /* Trim the allocation back to the maximum an AG can fit. */ args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp)); @@ -4839,6 +4851,7 @@ xfs_bmap_del_extent( nblks = 0; do_fx = 0; } + /* * Set flag value to use in switch statement. * Left-contig is 2, right-contig is 1. @@ -5026,7 +5039,7 @@ xfs_bmap_del_extent( */ if (do_fx) xfs_bmap_add_free(mp, dfops, del->br_startblock, - del->br_blockcount); + del->br_blockcount, NULL); /* * Adjust inode # blocks in the file. */ diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index dc01bb85a6e8..254034f96941 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -67,6 +67,7 @@ struct xfs_extent_free_item xfs_fsblock_t xefi_startblock;/* starting fs block number */ xfs_extlen_t xefi_blockcount;/* number of blocks in extent */ struct list_head xefi_list; + struct xfs_owner_info xefi_oinfo; /* extent owner */ }; #define XFS_BMAP_MAX_NMAP 4 @@ -165,7 +166,8 @@ void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); void xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork); void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *dfops, - xfs_fsblock_t bno, xfs_filblks_t len); + xfs_fsblock_t bno, xfs_filblks_t len, + struct xfs_owner_info *oinfo); void xfs_bmap_compute_maxlevels(struct xfs_mount *mp, int whichfork); int xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip, xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork); diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 0df5318cefff..9e34ca413e53 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -35,6 +35,7 @@ #include "xfs_quota.h" #include "xfs_trace.h" #include "xfs_cksum.h" +#include "xfs_rmap.h" /* * Determine the extent state. @@ -447,6 +448,8 @@ xfs_bmbt_alloc_block( args.mp = cur->bc_mp; args.fsbno = cur->bc_private.b.firstblock; args.firstblock = args.fsbno; + xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_private.b.ip->i_ino, + cur->bc_private.b.whichfork); if (args.fsbno == NULLFSBLOCK) { args.fsbno = be64_to_cpu(start->l); @@ -526,8 +529,10 @@ xfs_bmbt_free_block( struct xfs_inode *ip = cur->bc_private.b.ip; struct xfs_trans *tp = cur->bc_tp; xfs_fsblock_t fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp)); + struct xfs_owner_info oinfo; - xfs_bmap_add_free(mp, cur->bc_private.b.dfops, fsbno, 1); + xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_private.b.whichfork); + xfs_bmap_add_free(mp, cur->bc_private.b.dfops, fsbno, 1, &oinfo); ip->i_d.di_nblocks--; xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index 3dbed011c29a..eecfedf19be5 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -1319,6 +1319,34 @@ typedef __be32 xfs_inobt_ptr_t; */ #define XFS_RMAP_CRC_MAGIC 0x524d4233 /* 'RMB3' */ +/* + * Ownership info for an extent. This is used to create reverse-mapping + * entries. + */ +#define XFS_OWNER_INFO_ATTR_FORK (1 << 0) +#define XFS_OWNER_INFO_BMBT_BLOCK (1 << 1) +struct xfs_owner_info { + uint64_t oi_owner; + xfs_fileoff_t oi_offset; + unsigned int oi_flags; +}; + +/* + * Special owner types. + * + * Seeing as we only support up to 8EB, we have the upper bit of the owner field + * to tell us we have a special owner value. We use these for static metadata + * allocated at mkfs/growfs time, as well as for freespace management metadata. + */ +#define XFS_RMAP_OWN_NULL (-1ULL) /* No owner, for growfs */ +#define XFS_RMAP_OWN_UNKNOWN (-2ULL) /* Unknown owner, for EFI recovery */ +#define XFS_RMAP_OWN_FS (-3ULL) /* static fs metadata */ +#define XFS_RMAP_OWN_LOG (-4ULL) /* static fs metadata */ +#define XFS_RMAP_OWN_AG (-5ULL) /* AG freespace btree blocks */ +#define XFS_RMAP_OWN_INOBT (-6ULL) /* Inode btree blocks */ +#define XFS_RMAP_OWN_INODES (-7ULL) /* Inode chunk */ +#define XFS_RMAP_OWN_MIN (-8ULL) /* guard */ + #define XFS_RMAP_BLOCK(mp) \ (xfs_sb_version_hasfinobt(&((mp)->m_sb)) ? \ XFS_FIBT_BLOCK(mp) + 1 : \ diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index f0e1f713b0be..51b4e0de1fdc 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -40,6 +40,7 @@ #include "xfs_icache.h" #include "xfs_trace.h" #include "xfs_log.h" +#include "xfs_rmap.h" /* @@ -615,6 +616,7 @@ xfs_ialloc_ag_alloc( args.tp = tp; args.mp = tp->t_mountp; args.fsbno = NULLFSBLOCK; + xfs_rmap_ag_owner(&args.oinfo, XFS_RMAP_OWN_INODES); #ifdef DEBUG /* randomly do sparse inode allocations */ @@ -1825,12 +1827,14 @@ xfs_difree_inode_chunk( int nextbit; xfs_agblock_t agbno; int contigblk; + struct xfs_owner_info oinfo; DECLARE_BITMAP(holemask, XFS_INOBT_HOLEMASK_BITS); + xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INODES); if (!xfs_inobt_issparse(rec->ir_holemask)) { /* not sparse, calculate extent info directly */ xfs_bmap_add_free(mp, dfops, XFS_AGB_TO_FSB(mp, agno, sagbno), - mp->m_ialloc_blks); + mp->m_ialloc_blks, &oinfo); return; } @@ -1874,7 +1878,7 @@ xfs_difree_inode_chunk( ASSERT(agbno % mp->m_sb.sb_spino_align == 0); ASSERT(contigblk % mp->m_sb.sb_spino_align == 0); xfs_bmap_add_free(mp, dfops, XFS_AGB_TO_FSB(mp, agno, agbno), - contigblk); + contigblk, &oinfo); /* reset range to current bit and carry on... */ startidx = endidx = nextbit; diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index a48f4482004c..c83691edf6b2 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -32,6 +32,7 @@ #include "xfs_trace.h" #include "xfs_cksum.h" #include "xfs_trans.h" +#include "xfs_rmap.h" STATIC int @@ -96,6 +97,7 @@ xfs_inobt_alloc_block( memset(&args, 0, sizeof(args)); args.tp = cur->bc_tp; args.mp = cur->bc_mp; + xfs_rmap_ag_owner(&args.oinfo, XFS_RMAP_OWN_INOBT); args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, sbno); args.minlen = 1; args.maxlen = 1; @@ -125,8 +127,12 @@ xfs_inobt_free_block( struct xfs_btree_cur *cur, struct xfs_buf *bp) { + struct xfs_owner_info oinfo; + + xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INOBT); return xfs_free_extent(cur->bc_tp, - XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp)), 1); + XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp)), 1, + &oinfo); } STATIC int diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h new file mode 100644 index 000000000000..b30e7ad6c7cc --- /dev/null +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -0,0 +1,67 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_RMAP_H__ +#define __XFS_RMAP_H__ + +static inline void +xfs_rmap_ag_owner( + struct xfs_owner_info *oi, + uint64_t owner) +{ + oi->oi_owner = owner; + oi->oi_offset = 0; + oi->oi_flags = 0; +} + +static inline void +xfs_rmap_ino_bmbt_owner( + struct xfs_owner_info *oi, + xfs_ino_t ino, + int whichfork) +{ + oi->oi_owner = ino; + oi->oi_offset = 0; + oi->oi_flags = XFS_OWNER_INFO_BMBT_BLOCK; + if (whichfork == XFS_ATTR_FORK) + oi->oi_flags |= XFS_OWNER_INFO_ATTR_FORK; +} + +static inline void +xfs_rmap_ino_owner( + struct xfs_owner_info *oi, + xfs_ino_t ino, + int whichfork, + xfs_fileoff_t offset) +{ + oi->oi_owner = ino; + oi->oi_offset = offset; + oi->oi_flags = 0; + if (whichfork == XFS_ATTR_FORK) + oi->oi_flags |= XFS_OWNER_INFO_ATTR_FORK; +} + +static inline void +xfs_rmap_skip_owner_update( + struct xfs_owner_info *oi) +{ + oi->oi_owner = XFS_RMAP_OWN_UNKNOWN; +} + +#endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index 8c49f883934a..a8c461735ecb 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -27,6 +27,8 @@ #include "xfs_buf_item.h" #include "xfs_extfree_item.h" #include "xfs_log.h" +#include "xfs_btree.h" +#include "xfs_rmap.h" kmem_zone_t *xfs_efi_zone; @@ -503,6 +505,7 @@ xfs_efi_recover( int error = 0; xfs_extent_t *extp; xfs_fsblock_t startblock_fsb; + struct xfs_owner_info oinfo; ASSERT(!test_bit(XFS_EFI_RECOVERED, &efip->efi_flags)); @@ -534,10 +537,11 @@ xfs_efi_recover( return error; efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); + xfs_rmap_skip_owner_update(&oinfo); for (i = 0; i < efip->efi_format.efi_nextents; i++) { extp = &(efip->efi_format.efi_extents[i]); error = xfs_trans_free_extent(tp, efdp, extp->ext_start, - extp->ext_len); + extp->ext_len, &oinfo); if (error) goto abort_error; diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 0a3b822d9f8c..b625b61d1244 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -41,6 +41,7 @@ #include "xfs_trace.h" #include "xfs_log.h" #include "xfs_filestream.h" +#include "xfs_rmap.h" /* * File system operations @@ -436,6 +437,8 @@ xfs_growfs_data_private( * There are new blocks in the old last a.g. */ if (new) { + struct xfs_owner_info oinfo; + /* * Change the agi length. */ @@ -463,14 +466,20 @@ xfs_growfs_data_private( be32_to_cpu(agi->agi_length)); xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH); + /* * Free the new space. + * + * XFS_RMAP_OWN_NULL is used here to tell the rmap btree that + * this doesn't actually exist in the rmap btree. */ - error = xfs_free_extent(tp, XFS_AGB_TO_FSB(mp, agno, - be32_to_cpu(agf->agf_length) - new), new); - if (error) { + xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_NULL); + error = xfs_free_extent(tp, + XFS_AGB_TO_FSB(mp, agno, + be32_to_cpu(agf->agf_length) - new), + new, &oinfo); + if (error) goto error0; - } } /* diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index a5fe1874f02f..9bc368bb5eb4 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -217,7 +217,7 @@ struct xfs_efd_log_item *xfs_trans_get_efd(struct xfs_trans *, uint); int xfs_trans_free_extent(struct xfs_trans *, struct xfs_efd_log_item *, xfs_fsblock_t, - xfs_extlen_t); + xfs_extlen_t, struct xfs_owner_info *); int xfs_trans_commit(struct xfs_trans *); int __xfs_trans_roll(struct xfs_trans **, struct xfs_inode *, int *); int xfs_trans_roll(struct xfs_trans **, struct xfs_inode *); diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index b48490290b27..3aca3a7578c2 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -120,13 +120,14 @@ xfs_trans_free_extent( struct xfs_trans *tp, struct xfs_efd_log_item *efdp, xfs_fsblock_t start_block, - xfs_extlen_t ext_len) + xfs_extlen_t ext_len, + struct xfs_owner_info *oinfo) { uint next_extent; struct xfs_extent *extp; int error; - error = xfs_free_extent(tp, start_block, ext_len); + error = xfs_free_extent(tp, start_block, ext_len, oinfo); /* * Mark the transaction dirty, even on error. This ensures the @@ -213,7 +214,8 @@ xfs_extent_free_finish_item( free = container_of(item, struct xfs_extent_free_item, xefi_list); error = xfs_trans_free_extent(tp, done_item, free->xefi_startblock, - free->xefi_blockcount); + free->xefi_blockcount, + &free->xefi_oinfo); kmem_free(free); return error; } -- cgit v1.2.3 From 673930c34a4500c616cf9b2bbe1ae131ead2e155 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:33:43 +1000 Subject: xfs: introduce rmap extent operation stubs Originally-From: Dave Chinner Add the stubs into the extent allocation and freeing paths that the rmap btree implementation will hook into. While doing this, add the trace points that will be used to track rmap btree extent manipulations. [darrick.wong@oracle.com: Extend the stubs to take full owner info.] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/Makefile | 1 + fs/xfs/libxfs/xfs_alloc.c | 19 ++++++++-- fs/xfs/libxfs/xfs_rmap.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_rmap.h | 11 ++++++ fs/xfs/xfs_trace.h | 77 ++++++++++++++++++++++++++++++++++++++++ 5 files changed, 195 insertions(+), 2 deletions(-) create mode 100644 fs/xfs/libxfs/xfs_rmap.c (limited to 'fs') diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 01857b096cae..0498291829c6 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -52,6 +52,7 @@ xfs-y += $(addprefix libxfs/, \ xfs_inode_fork.o \ xfs_inode_buf.o \ xfs_log_rlimit.o \ + xfs_rmap.o \ xfs_sb.o \ xfs_symlink_remote.o \ xfs_trans_resv.o \ diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 6335b8b85349..dba7ce4b4be8 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -27,6 +27,7 @@ #include "xfs_defer.h" #include "xfs_inode.h" #include "xfs_btree.h" +#include "xfs_rmap.h" #include "xfs_alloc_btree.h" #include "xfs_alloc.h" #include "xfs_extent_busy.h" @@ -36,7 +37,6 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" -#include "xfs_rmap.h" struct workqueue_struct *xfs_alloc_wq; @@ -649,6 +649,14 @@ xfs_alloc_ag_vextent( ASSERT(!args->wasfromfl || !args->isfl); ASSERT(args->agbno % args->alignment == 0); + /* if not file data, insert new block into the reverse map btree */ + if (args->oinfo.oi_owner != XFS_RMAP_OWN_UNKNOWN) { + error = xfs_rmap_alloc(args->tp, args->agbp, args->agno, + args->agbno, args->len, &args->oinfo); + if (error) + return error; + } + if (!args->wasfromfl) { error = xfs_alloc_update_counters(args->tp, args->pag, args->agbp, @@ -1615,12 +1623,19 @@ xfs_free_ag_extent( xfs_extlen_t nlen; /* new length of freespace */ xfs_perag_t *pag; /* per allocation group data */ + bno_cur = cnt_cur = NULL; mp = tp->t_mountp; + + if (oinfo->oi_owner != XFS_RMAP_OWN_UNKNOWN) { + error = xfs_rmap_free(tp, agbp, agno, bno, len, oinfo); + if (error) + goto error0; + } + /* * Allocate and initialize a cursor for the by-block btree. */ bno_cur = xfs_allocbt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_BNO); - cnt_cur = NULL; /* * Look for a neighboring block on the left (lower block numbers) * that is contiguous with this space. diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c new file mode 100644 index 000000000000..b522bfc436a2 --- /dev/null +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -0,0 +1,89 @@ +/* + * Copyright (c) 2014 Red Hat, Inc. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_bit.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_da_format.h" +#include "xfs_da_btree.h" +#include "xfs_btree.h" +#include "xfs_trans.h" +#include "xfs_alloc.h" +#include "xfs_rmap.h" +#include "xfs_trans_space.h" +#include "xfs_trace.h" +#include "xfs_error.h" +#include "xfs_extent_busy.h" + +int +xfs_rmap_free( + struct xfs_trans *tp, + struct xfs_buf *agbp, + xfs_agnumber_t agno, + xfs_agblock_t bno, + xfs_extlen_t len, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = tp->t_mountp; + int error = 0; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return 0; + + trace_xfs_rmap_unmap(mp, agno, bno, len, false, oinfo); + if (1) + goto out_error; + trace_xfs_rmap_unmap_done(mp, agno, bno, len, false, oinfo); + return 0; + +out_error: + trace_xfs_rmap_unmap_error(mp, agno, error, _RET_IP_); + return error; +} + +int +xfs_rmap_alloc( + struct xfs_trans *tp, + struct xfs_buf *agbp, + xfs_agnumber_t agno, + xfs_agblock_t bno, + xfs_extlen_t len, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = tp->t_mountp; + int error = 0; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return 0; + + trace_xfs_rmap_map(mp, agno, bno, len, false, oinfo); + if (1) + goto out_error; + trace_xfs_rmap_map_done(mp, agno, bno, len, false, oinfo); + return 0; + +out_error: + trace_xfs_rmap_map_error(mp, agno, error, _RET_IP_); + return error; +} diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index b30e7ad6c7cc..dbf0301e9ae1 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -64,4 +64,15 @@ xfs_rmap_skip_owner_update( oi->oi_owner = XFS_RMAP_OWN_UNKNOWN; } +/* Reverse mapping functions. */ + +struct xfs_buf; + +int xfs_rmap_alloc(struct xfs_trans *tp, struct xfs_buf *agbp, + xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, + struct xfs_owner_info *oinfo); +int xfs_rmap_free(struct xfs_trans *tp, struct xfs_buf *agbp, + xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, + struct xfs_owner_info *oinfo); + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 939caf5f5a45..79e3812499ba 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2423,6 +2423,83 @@ DEFINE_MAP_EXTENT_DEFERRED_EVENT(xfs_defer_map_extent); DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer); DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred); +/* rmap tracepoints */ +DECLARE_EVENT_CLASS(xfs_rmap_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten, + struct xfs_owner_info *oinfo), + TP_ARGS(mp, agno, agbno, len, unwritten, oinfo), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + __field(uint64_t, owner) + __field(uint64_t, offset) + __field(unsigned long, flags) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->agbno = agbno; + __entry->len = len; + __entry->owner = oinfo->oi_owner; + __entry->offset = oinfo->oi_offset; + __entry->flags = oinfo->oi_flags; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%lx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->agbno, + __entry->len, + __entry->owner, + __entry->offset, + __entry->flags) +); +#define DEFINE_RMAP_EVENT(name) \ +DEFINE_EVENT(xfs_rmap_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + xfs_agblock_t agbno, xfs_extlen_t len, bool unwritten, \ + struct xfs_owner_info *oinfo), \ + TP_ARGS(mp, agno, agbno, len, unwritten, oinfo)) + +/* simple AG-based error/%ip tracepoint class */ +DECLARE_EVENT_CLASS(xfs_ag_error_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int error, + unsigned long caller_ip), + TP_ARGS(mp, agno, error, caller_ip), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(int, error) + __field(unsigned long, caller_ip) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->error = error; + __entry->caller_ip = caller_ip; + ), + TP_printk("dev %d:%d agno %u error %d caller %ps", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->error, + (char *)__entry->caller_ip) +); + +#define DEFINE_AG_ERROR_EVENT(name) \ +DEFINE_EVENT(xfs_ag_error_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int error, \ + unsigned long caller_ip), \ + TP_ARGS(mp, agno, error, caller_ip)) + +DEFINE_RMAP_EVENT(xfs_rmap_unmap); +DEFINE_RMAP_EVENT(xfs_rmap_unmap_done); +DEFINE_AG_ERROR_EVENT(xfs_rmap_unmap_error); +DEFINE_RMAP_EVENT(xfs_rmap_map); +DEFINE_RMAP_EVENT(xfs_rmap_map_done); +DEFINE_AG_ERROR_EVENT(xfs_rmap_map_error); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH -- cgit v1.2.3 From 035e00acb5c719bd003639b90716a7e94e023b73 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:36:07 +1000 Subject: xfs: define the on-disk rmap btree format Originally-From: Dave Chinner Now we have all the surrounding call infrastructure in place, we can start filling out the rmap btree implementation. Start with the on-disk btree format; add everything needed to read, write and manipulate rmap btree blocks. This prepares the way for adding the btree operations implementation. [darrick: record owner and offset info in rmap btree] [darrick: fork, bmbt and unwritten state in rmap btree] [darrick: flags are a separate field in xfs_rmap_irec] [darrick: calculate maxlevels separately] [darrick: move the 'unwritten' bit into unused parts of rm_offset] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/Makefile | 1 + fs/xfs/libxfs/xfs_btree.c | 3 + fs/xfs/libxfs/xfs_btree.h | 22 +++-- fs/xfs/libxfs/xfs_format.h | 73 ++++++++++++++++ fs/xfs/libxfs/xfs_rmap.h | 67 +++++++++++++++ fs/xfs/libxfs/xfs_rmap_btree.c | 184 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_rmap_btree.h | 55 ++++++++++++ fs/xfs/libxfs/xfs_sb.c | 6 ++ fs/xfs/libxfs/xfs_shared.h | 2 + fs/xfs/xfs_mount.c | 2 + fs/xfs/xfs_mount.h | 3 + fs/xfs/xfs_ondisk.h | 3 + fs/xfs/xfs_trace.h | 2 + 13 files changed, 415 insertions(+), 8 deletions(-) create mode 100644 fs/xfs/libxfs/xfs_rmap_btree.c create mode 100644 fs/xfs/libxfs/xfs_rmap_btree.h (limited to 'fs') diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 0498291829c6..b76e937f0729 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -53,6 +53,7 @@ xfs-y += $(addprefix libxfs/, \ xfs_inode_buf.o \ xfs_log_rlimit.o \ xfs_rmap.o \ + xfs_rmap_btree.o \ xfs_sb.o \ xfs_symlink_remote.o \ xfs_trans_resv.o \ diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 4066b01a1807..fe5f27e4cdc5 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -1213,6 +1213,9 @@ xfs_btree_set_refs( case XFS_BTNUM_BMAP: xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF); break; + case XFS_BTNUM_RMAP: + xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF); + break; default: ASSERT(0); } diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index a480947f094b..573d3a0c5797 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -38,10 +38,11 @@ union xfs_btree_ptr { }; union xfs_btree_key { - xfs_bmbt_key_t bmbt; - xfs_bmdr_key_t bmbr; /* bmbt root block */ - xfs_alloc_key_t alloc; - xfs_inobt_key_t inobt; + struct xfs_bmbt_key bmbt; + xfs_bmdr_key_t bmbr; /* bmbt root block */ + xfs_alloc_key_t alloc; + struct xfs_inobt_key inobt; + struct xfs_rmap_key rmap; }; /* @@ -56,13 +57,18 @@ union xfs_btree_bigkey { xfs_bmdr_key_t bmbr; /* bmbt root block */ xfs_alloc_key_t alloc; struct xfs_inobt_key inobt; + struct { + struct xfs_rmap_key rmap; + struct xfs_rmap_key rmap_hi; + }; }; union xfs_btree_rec { - xfs_bmbt_rec_t bmbt; - xfs_bmdr_rec_t bmbr; /* bmbt root block */ - xfs_alloc_rec_t alloc; - xfs_inobt_rec_t inobt; + struct xfs_bmbt_rec bmbt; + xfs_bmdr_rec_t bmbr; /* bmbt root block */ + struct xfs_alloc_rec alloc; + struct xfs_inobt_rec inobt; + struct xfs_rmap_rec rmap; }; /* diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index eecfedf19be5..f6a0641e64c8 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -1347,6 +1347,79 @@ struct xfs_owner_info { #define XFS_RMAP_OWN_INODES (-7ULL) /* Inode chunk */ #define XFS_RMAP_OWN_MIN (-8ULL) /* guard */ +#define XFS_RMAP_NON_INODE_OWNER(owner) (!!((owner) & (1ULL << 63))) + +/* + * Data record structure + */ +struct xfs_rmap_rec { + __be32 rm_startblock; /* extent start block */ + __be32 rm_blockcount; /* extent length */ + __be64 rm_owner; /* extent owner */ + __be64 rm_offset; /* offset within the owner */ +}; + +/* + * rmap btree record + * rm_offset:63 is the attribute fork flag + * rm_offset:62 is the bmbt block flag + * rm_offset:61 is the unwritten extent flag (same as l0:63 in bmbt) + * rm_offset:54-60 aren't used and should be zero + * rm_offset:0-53 is the block offset within the inode + */ +#define XFS_RMAP_OFF_ATTR_FORK ((__uint64_t)1ULL << 63) +#define XFS_RMAP_OFF_BMBT_BLOCK ((__uint64_t)1ULL << 62) +#define XFS_RMAP_OFF_UNWRITTEN ((__uint64_t)1ULL << 61) + +#define XFS_RMAP_LEN_MAX ((__uint32_t)~0U) +#define XFS_RMAP_OFF_FLAGS (XFS_RMAP_OFF_ATTR_FORK | \ + XFS_RMAP_OFF_BMBT_BLOCK | \ + XFS_RMAP_OFF_UNWRITTEN) +#define XFS_RMAP_OFF_MASK ((__uint64_t)0x3FFFFFFFFFFFFFULL) + +#define XFS_RMAP_OFF(off) ((off) & XFS_RMAP_OFF_MASK) + +#define XFS_RMAP_IS_BMBT_BLOCK(off) (!!((off) & XFS_RMAP_OFF_BMBT_BLOCK)) +#define XFS_RMAP_IS_ATTR_FORK(off) (!!((off) & XFS_RMAP_OFF_ATTR_FORK)) +#define XFS_RMAP_IS_UNWRITTEN(len) (!!((off) & XFS_RMAP_OFF_UNWRITTEN)) + +#define RMAPBT_STARTBLOCK_BITLEN 32 +#define RMAPBT_BLOCKCOUNT_BITLEN 32 +#define RMAPBT_OWNER_BITLEN 64 +#define RMAPBT_ATTRFLAG_BITLEN 1 +#define RMAPBT_BMBTFLAG_BITLEN 1 +#define RMAPBT_EXNTFLAG_BITLEN 1 +#define RMAPBT_UNUSED_OFFSET_BITLEN 7 +#define RMAPBT_OFFSET_BITLEN 54 + +#define XFS_RMAP_ATTR_FORK (1 << 0) +#define XFS_RMAP_BMBT_BLOCK (1 << 1) +#define XFS_RMAP_UNWRITTEN (1 << 2) +#define XFS_RMAP_KEY_FLAGS (XFS_RMAP_ATTR_FORK | \ + XFS_RMAP_BMBT_BLOCK) +#define XFS_RMAP_REC_FLAGS (XFS_RMAP_UNWRITTEN) +struct xfs_rmap_irec { + xfs_agblock_t rm_startblock; /* extent start block */ + xfs_extlen_t rm_blockcount; /* extent length */ + __uint64_t rm_owner; /* extent owner */ + __uint64_t rm_offset; /* offset within the owner */ + unsigned int rm_flags; /* state flags */ +}; + +/* + * Key structure + * + * We don't use the length for lookups + */ +struct xfs_rmap_key { + __be32 rm_startblock; /* extent start block */ + __be64 rm_owner; /* extent owner */ + __be64 rm_offset; /* offset within the owner */ +} __attribute__((packed)); + +/* btree pointer type */ +typedef __be32 xfs_rmap_ptr_t; + #define XFS_RMAP_BLOCK(mp) \ (xfs_sb_version_hasfinobt(&((mp)->m_sb)) ? \ XFS_FIBT_BLOCK(mp) + 1 : \ diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index dbf0301e9ae1..e7a670459fea 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -68,6 +68,73 @@ xfs_rmap_skip_owner_update( struct xfs_buf; +static inline __u64 +xfs_rmap_irec_offset_pack( + const struct xfs_rmap_irec *irec) +{ + __u64 x; + + x = XFS_RMAP_OFF(irec->rm_offset); + if (irec->rm_flags & XFS_RMAP_ATTR_FORK) + x |= XFS_RMAP_OFF_ATTR_FORK; + if (irec->rm_flags & XFS_RMAP_BMBT_BLOCK) + x |= XFS_RMAP_OFF_BMBT_BLOCK; + if (irec->rm_flags & XFS_RMAP_UNWRITTEN) + x |= XFS_RMAP_OFF_UNWRITTEN; + return x; +} + +static inline int +xfs_rmap_irec_offset_unpack( + __u64 offset, + struct xfs_rmap_irec *irec) +{ + if (offset & ~(XFS_RMAP_OFF_MASK | XFS_RMAP_OFF_FLAGS)) + return -EFSCORRUPTED; + irec->rm_offset = XFS_RMAP_OFF(offset); + if (offset & XFS_RMAP_OFF_ATTR_FORK) + irec->rm_flags |= XFS_RMAP_ATTR_FORK; + if (offset & XFS_RMAP_OFF_BMBT_BLOCK) + irec->rm_flags |= XFS_RMAP_BMBT_BLOCK; + if (offset & XFS_RMAP_OFF_UNWRITTEN) + irec->rm_flags |= XFS_RMAP_UNWRITTEN; + return 0; +} + +static inline void +xfs_owner_info_unpack( + struct xfs_owner_info *oinfo, + uint64_t *owner, + uint64_t *offset, + unsigned int *flags) +{ + unsigned int r = 0; + + *owner = oinfo->oi_owner; + *offset = oinfo->oi_offset; + if (oinfo->oi_flags & XFS_OWNER_INFO_ATTR_FORK) + r |= XFS_RMAP_ATTR_FORK; + if (oinfo->oi_flags & XFS_OWNER_INFO_BMBT_BLOCK) + r |= XFS_RMAP_BMBT_BLOCK; + *flags = r; +} + +static inline void +xfs_owner_info_pack( + struct xfs_owner_info *oinfo, + uint64_t owner, + uint64_t offset, + unsigned int flags) +{ + oinfo->oi_owner = owner; + oinfo->oi_offset = XFS_RMAP_OFF(offset); + oinfo->oi_flags = 0; + if (flags & XFS_RMAP_ATTR_FORK) + oinfo->oi_flags |= XFS_OWNER_INFO_ATTR_FORK; + if (flags & XFS_RMAP_BMBT_BLOCK) + oinfo->oi_flags |= XFS_OWNER_INFO_BMBT_BLOCK; +} + int xfs_rmap_alloc(struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, struct xfs_owner_info *oinfo); diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c new file mode 100644 index 000000000000..a9ddc191aa2d --- /dev/null +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -0,0 +1,184 @@ +/* + * Copyright (c) 2014 Red Hat, Inc. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_bit.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_inode.h" +#include "xfs_trans.h" +#include "xfs_alloc.h" +#include "xfs_btree.h" +#include "xfs_rmap_btree.h" +#include "xfs_trace.h" +#include "xfs_cksum.h" +#include "xfs_error.h" +#include "xfs_extent_busy.h" + +static struct xfs_btree_cur * +xfs_rmapbt_dup_cursor( + struct xfs_btree_cur *cur) +{ + return xfs_rmapbt_init_cursor(cur->bc_mp, cur->bc_tp, + cur->bc_private.a.agbp, cur->bc_private.a.agno); +} + +static bool +xfs_rmapbt_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); + struct xfs_perag *pag = bp->b_pag; + unsigned int level; + + /* + * magic number and level verification + * + * During growfs operations, we can't verify the exact level or owner as + * the perag is not fully initialised and hence not attached to the + * buffer. In this case, check against the maximum tree depth. + * + * Similarly, during log recovery we will have a perag structure + * attached, but the agf information will not yet have been initialised + * from the on disk AGF. Again, we can only check against maximum limits + * in this case. + */ + if (block->bb_magic != cpu_to_be32(XFS_RMAP_CRC_MAGIC)) + return false; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return false; + if (!xfs_btree_sblock_v5hdr_verify(bp)) + return false; + + level = be16_to_cpu(block->bb_level); + if (pag && pag->pagf_init) { + if (level >= pag->pagf_levels[XFS_BTNUM_RMAPi]) + return false; + } else if (level >= mp->m_rmap_maxlevels) + return false; + + return xfs_btree_sblock_verify(bp, mp->m_rmap_mxr[level != 0]); +} + +static void +xfs_rmapbt_read_verify( + struct xfs_buf *bp) +{ + if (!xfs_btree_sblock_verify_crc(bp)) + xfs_buf_ioerror(bp, -EFSBADCRC); + else if (!xfs_rmapbt_verify(bp)) + xfs_buf_ioerror(bp, -EFSCORRUPTED); + + if (bp->b_error) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_verifier_error(bp); + } +} + +static void +xfs_rmapbt_write_verify( + struct xfs_buf *bp) +{ + if (!xfs_rmapbt_verify(bp)) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_buf_ioerror(bp, -EFSCORRUPTED); + xfs_verifier_error(bp); + return; + } + xfs_btree_sblock_calc_crc(bp); + +} + +const struct xfs_buf_ops xfs_rmapbt_buf_ops = { + .name = "xfs_rmapbt", + .verify_read = xfs_rmapbt_read_verify, + .verify_write = xfs_rmapbt_write_verify, +}; + +static const struct xfs_btree_ops xfs_rmapbt_ops = { + .rec_len = sizeof(struct xfs_rmap_rec), + .key_len = 2 * sizeof(struct xfs_rmap_key), + + .dup_cursor = xfs_rmapbt_dup_cursor, + .buf_ops = &xfs_rmapbt_buf_ops, + + .get_leaf_keys = xfs_btree_get_leaf_keys_overlapped, + .get_node_keys = xfs_btree_get_node_keys_overlapped, + .update_keys = xfs_btree_update_keys_overlapped, +}; + +/* + * Allocate a new allocation btree cursor. + */ +struct xfs_btree_cur * +xfs_rmapbt_init_cursor( + struct xfs_mount *mp, + struct xfs_trans *tp, + struct xfs_buf *agbp, + xfs_agnumber_t agno) +{ + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + struct xfs_btree_cur *cur; + + cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); + cur->bc_tp = tp; + cur->bc_mp = mp; + cur->bc_btnum = XFS_BTNUM_RMAP; + cur->bc_flags = XFS_BTREE_CRC_BLOCKS; + cur->bc_blocklog = mp->m_sb.sb_blocklog; + cur->bc_ops = &xfs_rmapbt_ops; + cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]); + + cur->bc_private.a.agbp = agbp; + cur->bc_private.a.agno = agno; + + return cur; +} + +/* + * Calculate number of records in an rmap btree block. + */ +int +xfs_rmapbt_maxrecs( + struct xfs_mount *mp, + int blocklen, + int leaf) +{ + blocklen -= XFS_RMAP_BLOCK_LEN; + + if (leaf) + return blocklen / sizeof(struct xfs_rmap_rec); + return blocklen / + (sizeof(struct xfs_rmap_key) + sizeof(xfs_rmap_ptr_t)); +} + +/* Compute the maximum height of an rmap btree. */ +void +xfs_rmapbt_compute_maxlevels( + struct xfs_mount *mp) +{ + mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(mp, + mp->m_rmap_mnr, mp->m_sb.sb_agblocks); +} diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h new file mode 100644 index 000000000000..a3a6b7d476c7 --- /dev/null +++ b/fs/xfs/libxfs/xfs_rmap_btree.h @@ -0,0 +1,55 @@ +/* + * Copyright (c) 2014 Red Hat, Inc. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ +#ifndef __XFS_RMAP_BTREE_H__ +#define __XFS_RMAP_BTREE_H__ + +struct xfs_buf; +struct xfs_btree_cur; +struct xfs_mount; + +/* rmaps only exist on crc enabled filesystems */ +#define XFS_RMAP_BLOCK_LEN XFS_BTREE_SBLOCK_CRC_LEN + +/* + * Record, key, and pointer address macros for btree blocks. + * + * (note that some of these may appear unused, but they are used in userspace) + */ +#define XFS_RMAP_REC_ADDR(block, index) \ + ((struct xfs_rmap_rec *) \ + ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ + (((index) - 1) * sizeof(struct xfs_rmap_rec)))) + +#define XFS_RMAP_KEY_ADDR(block, index) \ + ((struct xfs_rmap_key *) \ + ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ + ((index) - 1) * sizeof(struct xfs_rmap_key))) + +#define XFS_RMAP_PTR_ADDR(block, index, maxrecs) \ + ((xfs_rmap_ptr_t *) \ + ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ + (maxrecs) * sizeof(struct xfs_rmap_key) + \ + ((index) - 1) * sizeof(xfs_rmap_ptr_t))) + +struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp, + struct xfs_trans *tp, struct xfs_buf *bp, + xfs_agnumber_t agno); +int xfs_rmapbt_maxrecs(struct xfs_mount *mp, int blocklen, int leaf); +extern void xfs_rmapbt_compute_maxlevels(struct xfs_mount *mp); + +#endif /* __XFS_RMAP_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 2194e5d4eea8..fb33e70e5c9a 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -37,6 +37,7 @@ #include "xfs_alloc_btree.h" #include "xfs_ialloc_btree.h" #include "xfs_log.h" +#include "xfs_rmap_btree.h" /* * Physical superblock buffer manipulations. Shared with libxfs in userspace. @@ -730,6 +731,11 @@ xfs_sb_mount_common( mp->m_bmap_dmnr[0] = mp->m_bmap_dmxr[0] / 2; mp->m_bmap_dmnr[1] = mp->m_bmap_dmxr[1] / 2; + mp->m_rmap_mxr[0] = xfs_rmapbt_maxrecs(mp, sbp->sb_blocksize, 1); + mp->m_rmap_mxr[1] = xfs_rmapbt_maxrecs(mp, sbp->sb_blocksize, 0); + mp->m_rmap_mnr[0] = mp->m_rmap_mxr[0] / 2; + mp->m_rmap_mnr[1] = mp->m_rmap_mxr[1] / 2; + mp->m_bsize = XFS_FSB_TO_BB(mp, 1); mp->m_ialloc_inos = (int)MAX((__uint16_t)XFS_INODES_PER_CHUNK, sbp->sb_inopblock); diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h index 16002b5ec4eb..0c5b30bd884c 100644 --- a/fs/xfs/libxfs/xfs_shared.h +++ b/fs/xfs/libxfs/xfs_shared.h @@ -38,6 +38,7 @@ extern const struct xfs_buf_ops xfs_agi_buf_ops; extern const struct xfs_buf_ops xfs_agf_buf_ops; extern const struct xfs_buf_ops xfs_agfl_buf_ops; extern const struct xfs_buf_ops xfs_allocbt_buf_ops; +extern const struct xfs_buf_ops xfs_rmapbt_buf_ops; extern const struct xfs_buf_ops xfs_attr3_leaf_buf_ops; extern const struct xfs_buf_ops xfs_attr3_rmt_buf_ops; extern const struct xfs_buf_ops xfs_bmbt_buf_ops; @@ -116,6 +117,7 @@ int xfs_log_calc_minimum_size(struct xfs_mount *); #define XFS_INO_BTREE_REF 3 #define XFS_ALLOC_BTREE_REF 2 #define XFS_BMAP_BTREE_REF 2 +#define XFS_RMAP_BTREE_REF 2 #define XFS_DIR_BTREE_REF 2 #define XFS_INO_REF 2 #define XFS_ATTR_BTREE_REF 1 diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 24ae96b4078f..0041866d51ca 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -42,6 +42,7 @@ #include "xfs_trace.h" #include "xfs_icache.h" #include "xfs_sysfs.h" +#include "xfs_rmap_btree.h" static DEFINE_MUTEX(xfs_uuid_table_mutex); @@ -682,6 +683,7 @@ xfs_mountfs( xfs_bmap_compute_maxlevels(mp, XFS_DATA_FORK); xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK); xfs_ialloc_compute_maxlevels(mp); + xfs_rmapbt_compute_maxlevels(mp); xfs_set_maxicount(mp); diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 0537b1f28f90..0ed0f29ddaf5 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -116,9 +116,12 @@ typedef struct xfs_mount { uint m_bmap_dmnr[2]; /* min bmap btree records */ uint m_inobt_mxr[2]; /* max inobt btree records */ uint m_inobt_mnr[2]; /* min inobt btree records */ + uint m_rmap_mxr[2]; /* max rmap btree records */ + uint m_rmap_mnr[2]; /* min rmap btree records */ uint m_ag_maxlevels; /* XFS_AG_MAXLEVELS */ uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */ uint m_in_maxlevels; /* max inobt btree levels. */ + uint m_rmap_maxlevels; /* max rmap btree levels */ xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */ struct radix_tree_root m_perag_tree; /* per-ag accounting info */ spinlock_t m_perag_lock; /* lock for m_perag_tree */ diff --git a/fs/xfs/xfs_ondisk.h b/fs/xfs/xfs_ondisk.h index 0cc8d8f74356..69e2986a3776 100644 --- a/fs/xfs/xfs_ondisk.h +++ b/fs/xfs/xfs_ondisk.h @@ -49,11 +49,14 @@ xfs_check_ondisk_structs(void) XFS_CHECK_STRUCT_SIZE(struct xfs_dsymlink_hdr, 56); XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_key, 4); XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_rec, 16); + XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_key, 20); + XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_rec, 24); XFS_CHECK_STRUCT_SIZE(struct xfs_timestamp, 8); XFS_CHECK_STRUCT_SIZE(xfs_alloc_key_t, 8); XFS_CHECK_STRUCT_SIZE(xfs_alloc_ptr_t, 4); XFS_CHECK_STRUCT_SIZE(xfs_alloc_rec_t, 8); XFS_CHECK_STRUCT_SIZE(xfs_inobt_ptr_t, 4); + XFS_CHECK_STRUCT_SIZE(xfs_rmap_ptr_t, 4); /* dir/attr trees */ XFS_CHECK_STRUCT_SIZE(struct xfs_attr3_leaf_hdr, 80); diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 79e3812499ba..4c3418b1a244 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2446,6 +2446,8 @@ DECLARE_EVENT_CLASS(xfs_rmap_class, __entry->owner = oinfo->oi_owner; __entry->offset = oinfo->oi_offset; __entry->flags = oinfo->oi_flags; + if (unwritten) + __entry->flags |= XFS_RMAP_UNWRITTEN; ), TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%lx", MAJOR(__entry->dev), MINOR(__entry->dev), -- cgit v1.2.3 From e70d829f8d288cd062d265085420f001828b0683 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:36:08 +1000 Subject: xfs: add rmap btree growfs support Originally-From: Dave Chinner Now we can read and write rmap btree blocks, we can add support to the growfs code to initialise new rmap btree blocks. [darrick.wong@oracle.com: fill out the rmap offset fields] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_fsops.c | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 73 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index b625b61d1244..27e6e8cdd431 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -33,6 +33,7 @@ #include "xfs_btree.h" #include "xfs_alloc_btree.h" #include "xfs_alloc.h" +#include "xfs_rmap_btree.h" #include "xfs_ialloc.h" #include "xfs_fsops.h" #include "xfs_itable.h" @@ -241,6 +242,12 @@ xfs_growfs_data_private( agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp)); agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1); agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1); + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { + agf->agf_roots[XFS_BTNUM_RMAPi] = + cpu_to_be32(XFS_RMAP_BLOCK(mp)); + agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1); + } + agf->agf_flfirst = cpu_to_be32(1); agf->agf_fllast = 0; agf->agf_flcount = 0; @@ -380,6 +387,72 @@ xfs_growfs_data_private( if (error) goto error0; + /* RMAP btree root block */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { + struct xfs_rmap_rec *rrec; + struct xfs_btree_block *block; + + bp = xfs_growfs_get_hdr_buf(mp, + XFS_AGB_TO_DADDR(mp, agno, XFS_RMAP_BLOCK(mp)), + BTOBB(mp->m_sb.sb_blocksize), 0, + &xfs_rmapbt_buf_ops); + if (!bp) { + error = -ENOMEM; + goto error0; + } + + xfs_btree_init_block(mp, bp, XFS_RMAP_CRC_MAGIC, 0, 0, + agno, XFS_BTREE_CRC_BLOCKS); + block = XFS_BUF_TO_BLOCK(bp); + + + /* + * mark the AG header regions as static metadata The BNO + * btree block is the first block after the headers, so + * it's location defines the size of region the static + * metadata consumes. + * + * Note: unlike mkfs, we never have to account for log + * space when growing the data regions + */ + rrec = XFS_RMAP_REC_ADDR(block, 1); + rrec->rm_startblock = 0; + rrec->rm_blockcount = cpu_to_be32(XFS_BNO_BLOCK(mp)); + rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_FS); + rrec->rm_offset = 0; + be16_add_cpu(&block->bb_numrecs, 1); + + /* account freespace btree root blocks */ + rrec = XFS_RMAP_REC_ADDR(block, 2); + rrec->rm_startblock = cpu_to_be32(XFS_BNO_BLOCK(mp)); + rrec->rm_blockcount = cpu_to_be32(2); + rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG); + rrec->rm_offset = 0; + be16_add_cpu(&block->bb_numrecs, 1); + + /* account inode btree root blocks */ + rrec = XFS_RMAP_REC_ADDR(block, 3); + rrec->rm_startblock = cpu_to_be32(XFS_IBT_BLOCK(mp)); + rrec->rm_blockcount = cpu_to_be32(XFS_RMAP_BLOCK(mp) - + XFS_IBT_BLOCK(mp)); + rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_INOBT); + rrec->rm_offset = 0; + be16_add_cpu(&block->bb_numrecs, 1); + + /* account for rmap btree root */ + rrec = XFS_RMAP_REC_ADDR(block, 4); + rrec->rm_startblock = cpu_to_be32(XFS_RMAP_BLOCK(mp)); + rrec->rm_blockcount = cpu_to_be32(1); + rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG); + rrec->rm_offset = 0; + be16_add_cpu(&block->bb_numrecs, 1); + + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) + goto error0; + } + /* * INO btree root block */ -- cgit v1.2.3 From fa30f03cda26783b1294af6e7da9f1142da0f52e Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:37:10 +1000 Subject: xfs: rmap btree transaction reservations The rmap btrees will use the AGFL as the block allocation source, so we need to ensure that the transaction reservations reflect the fact this tree is modified by allocation and freeing. Hence we need to extend all the extent allocation/free reservations used in transactions to handle this. Note that this also gets rid of the unused XFS_ALLOCFREE_LOG_RES macro, as we now do buffer reservations based on the number of buffers logged via xfs_calc_buf_res(). Hence we only need the buffer count calculation now. [darrick: use rmap_maxlevels when calculating log block resv] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_trans_resv.c | 58 +++++++++++++++++++++++++++++------------- fs/xfs/libxfs/xfs_trans_resv.h | 10 -------- 2 files changed, 41 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c index 4c7eb9d7c0c7..301ef2f4dbd6 100644 --- a/fs/xfs/libxfs/xfs_trans_resv.c +++ b/fs/xfs/libxfs/xfs_trans_resv.c @@ -63,6 +63,30 @@ xfs_calc_buf_res( return nbufs * (size + xfs_buf_log_overhead()); } +/* + * Per-extent log reservation for the btree changes involved in freeing or + * allocating an extent. In classic XFS there were two trees that will be + * modified (bnobt + cntbt). With rmap enabled, there are three trees + * (rmapbt). The number of blocks reserved is based on the formula: + * + * num trees * ((2 blocks/level * max depth) - 1) + * + * Keep in mind that max depth is calculated separately for each type of tree. + */ +static uint +xfs_allocfree_log_count( + struct xfs_mount *mp, + uint num_ops) +{ + uint blocks; + + blocks = num_ops * 2 * (2 * mp->m_ag_maxlevels - 1); + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + blocks += num_ops * (2 * mp->m_rmap_maxlevels - 1); + + return blocks; +} + /* * Logging inodes is really tricksy. They are logged in memory format, * which means that what we write into the log doesn't directly translate into @@ -126,7 +150,7 @@ xfs_calc_inode_res( */ STATIC uint xfs_calc_finobt_res( - struct xfs_mount *mp, + struct xfs_mount *mp, int alloc, int modify) { @@ -137,7 +161,7 @@ xfs_calc_finobt_res( res = xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)); if (alloc) - res += xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + res += xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)); if (modify) res += (uint)XFS_FSB_TO_B(mp, 1); @@ -188,10 +212,10 @@ xfs_calc_write_reservation( xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), XFS_FSB_TO_B(mp, 1)) + xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 2), XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 2), XFS_FSB_TO_B(mp, 1)))); } @@ -217,10 +241,10 @@ xfs_calc_itruncate_reservation( xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) + 1, XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 4), XFS_FSB_TO_B(mp, 1)) + xfs_calc_buf_res(5, 0) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)) + xfs_calc_buf_res(2 + mp->m_ialloc_blks + mp->m_in_maxlevels, 0))); @@ -247,7 +271,7 @@ xfs_calc_rename_reservation( xfs_calc_buf_res(2 * XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(7, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 3), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 3), XFS_FSB_TO_B(mp, 1)))); } @@ -286,7 +310,7 @@ xfs_calc_link_reservation( xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)))); } @@ -324,7 +348,7 @@ xfs_calc_remove_reservation( xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(4, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 2), XFS_FSB_TO_B(mp, 1)))); } @@ -371,7 +395,7 @@ xfs_calc_create_resv_alloc( mp->m_sb.sb_sectsize + xfs_calc_buf_res(mp->m_ialloc_blks, XFS_FSB_TO_B(mp, 1)) + xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)); } @@ -399,7 +423,7 @@ xfs_calc_icreate_resv_alloc( return xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) + mp->m_sb.sb_sectsize + xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)) + xfs_calc_finobt_res(mp, 0, 0); } @@ -483,7 +507,7 @@ xfs_calc_ifree_reservation( xfs_calc_buf_res(1, 0) + xfs_calc_buf_res(2 + mp->m_ialloc_blks + mp->m_in_maxlevels, 0) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)) + xfs_calc_finobt_res(mp, 0, 1); } @@ -513,7 +537,7 @@ xfs_calc_growdata_reservation( struct xfs_mount *mp) { return xfs_calc_buf_res(3, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)); } @@ -535,7 +559,7 @@ xfs_calc_growrtalloc_reservation( xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), XFS_FSB_TO_B(mp, 1)) + xfs_calc_inode_res(mp, 1) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)); } @@ -611,7 +635,7 @@ xfs_calc_addafork_reservation( xfs_calc_buf_res(1, mp->m_dir_geo->blksize) + xfs_calc_buf_res(XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK) + 1, XFS_FSB_TO_B(mp, 1)) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 1), XFS_FSB_TO_B(mp, 1)); } @@ -634,7 +658,7 @@ xfs_calc_attrinval_reservation( xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK), XFS_FSB_TO_B(mp, 1))), (xfs_calc_buf_res(9, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 4), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 4), XFS_FSB_TO_B(mp, 1)))); } @@ -701,7 +725,7 @@ xfs_calc_attrrm_reservation( XFS_BM_MAXLEVELS(mp, XFS_ATTR_FORK)) + xfs_calc_buf_res(XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK), 0)), (xfs_calc_buf_res(5, mp->m_sb.sb_sectsize) + - xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 2), + xfs_calc_buf_res(xfs_allocfree_log_count(mp, 2), XFS_FSB_TO_B(mp, 1)))); } diff --git a/fs/xfs/libxfs/xfs_trans_resv.h b/fs/xfs/libxfs/xfs_trans_resv.h index 797815012c0e..0eb46ed6d404 100644 --- a/fs/xfs/libxfs/xfs_trans_resv.h +++ b/fs/xfs/libxfs/xfs_trans_resv.h @@ -67,16 +67,6 @@ struct xfs_trans_resv { /* shorthand way of accessing reservation structure */ #define M_RES(mp) (&(mp)->m_resv) -/* - * Per-extent log reservation for the allocation btree changes - * involved in freeing or allocating an extent. - * 2 trees * (2 blocks/level * max depth - 1) * block size - */ -#define XFS_ALLOCFREE_LOG_RES(mp,nx) \ - ((nx) * (2 * XFS_FSB_TO_B((mp), 2 * (mp)->m_ag_maxlevels - 1))) -#define XFS_ALLOCFREE_LOG_COUNT(mp,nx) \ - ((nx) * (2 * (2 * (mp)->m_ag_maxlevels - 1))) - /* * Per-directory log reservation for any directory change. * dir blocks: (1 btree block per level + data block + free block) * dblock size -- cgit v1.2.3 From 525488520ac69a3612dbceefa573b255a83005e9 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:38:24 +1000 Subject: xfs: rmap btree requires more reserved free space Originally-From: Dave Chinner The rmap btree is allocated from the AGFL, which means we have to ensure ENOSPC is reported to userspace before we run out of free space in each AG. The last allocation in an AG can cause a full height rmap btree split, and that means we have to reserve at least this many blocks *in each AG* to be placed on the AGFL at ENOSPC. Update the various space calculation functions to handle this. Also, because the macros are now executing conditional code and are called quite frequently, convert them to functions that initialise variables in the struct xfs_mount, use the new variables everywhere and document the calculations better. [darrick.wong@oracle.com: don't reserve blocks if !rmap] [dchinner@redhat.com: update m_ag_max_usable after growfs] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 69 +++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_alloc.h | 41 ++++------------------------ fs/xfs/libxfs/xfs_bmap.c | 2 +- fs/xfs/libxfs/xfs_sb.c | 2 ++ fs/xfs/xfs_discard.c | 2 +- fs/xfs/xfs_fsops.c | 5 ++-- fs/xfs/xfs_log_recover.c | 1 + fs/xfs/xfs_mount.c | 2 +- fs/xfs/xfs_mount.h | 2 ++ fs/xfs/xfs_super.c | 2 +- 10 files changed, 86 insertions(+), 42 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index dba7ce4b4be8..73ab1ea1270d 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -62,6 +62,70 @@ xfs_prealloc_blocks( return XFS_IBT_BLOCK(mp) + 1; } +/* + * In order to avoid ENOSPC-related deadlock caused by out-of-order locking of + * AGF buffer (PV 947395), we place constraints on the relationship among + * actual allocations for data blocks, freelist blocks, and potential file data + * bmap btree blocks. However, these restrictions may result in no actual space + * allocated for a delayed extent, for example, a data block in a certain AG is + * allocated but there is no additional block for the additional bmap btree + * block due to a split of the bmap btree of the file. The result of this may + * lead to an infinite loop when the file gets flushed to disk and all delayed + * extents need to be actually allocated. To get around this, we explicitly set + * aside a few blocks which will not be reserved in delayed allocation. + * + * When rmap is disabled, we need to reserve 4 fsbs _per AG_ for the freelist + * and 4 more to handle a potential split of the file's bmap btree. + * + * When rmap is enabled, we must also be able to handle two rmap btree inserts + * to record both the file data extent and a new bmbt block. The bmbt block + * might not be in the same AG as the file data extent. In the worst case + * the bmap btree splits multiple levels and all the new blocks come from + * different AGs, so set aside enough to handle rmap btree splits in all AGs. + */ +unsigned int +xfs_alloc_set_aside( + struct xfs_mount *mp) +{ + unsigned int blocks; + + blocks = 4 + (mp->m_sb.sb_agcount * XFS_ALLOC_AGFL_RESERVE); + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + blocks += mp->m_sb.sb_agcount * mp->m_rmap_maxlevels; + return blocks; +} + +/* + * When deciding how much space to allocate out of an AG, we limit the + * allocation maximum size to the size the AG. However, we cannot use all the + * blocks in the AG - some are permanently used by metadata. These + * blocks are generally: + * - the AG superblock, AGF, AGI and AGFL + * - the AGF (bno and cnt) and AGI btree root blocks, and optionally + * the AGI free inode and rmap btree root blocks. + * - blocks on the AGFL according to xfs_alloc_set_aside() limits + * - the rmapbt root block + * + * The AG headers are sector sized, so the amount of space they take up is + * dependent on filesystem geometry. The others are all single blocks. + */ +unsigned int +xfs_alloc_ag_max_usable( + struct xfs_mount *mp) +{ + unsigned int blocks; + + blocks = XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)); /* ag headers */ + blocks += XFS_ALLOC_AGFL_RESERVE; + blocks += 3; /* AGF, AGI btree root blocks */ + if (xfs_sb_version_hasfinobt(&mp->m_sb)) + blocks++; /* finobt root block */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + blocks++; /* rmap root block */ + + return mp->m_sb.sb_agblocks - blocks; +} + /* * Lookup the record equal to [bno, len] in the btree given by cur. */ @@ -1904,6 +1968,11 @@ xfs_alloc_min_freelist( /* space needed by-size freespace btree */ min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_CNTi] + 1, mp->m_ag_maxlevels); + /* space needed reverse mapping used space btree */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + min_free += min_t(unsigned int, + pag->pagf_levels[XFS_BTNUM_RMAPi] + 1, + mp->m_rmap_maxlevels); return min_free; } diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index 88053422fda0..360f9e425536 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -55,42 +55,6 @@ typedef unsigned int xfs_alloctype_t; #define XFS_ALLOC_FLAG_TRYLOCK 0x00000001 /* use trylock for buffer locking */ #define XFS_ALLOC_FLAG_FREEING 0x00000002 /* indicate caller is freeing extents*/ -/* - * In order to avoid ENOSPC-related deadlock caused by - * out-of-order locking of AGF buffer (PV 947395), we place - * constraints on the relationship among actual allocations for - * data blocks, freelist blocks, and potential file data bmap - * btree blocks. However, these restrictions may result in no - * actual space allocated for a delayed extent, for example, a data - * block in a certain AG is allocated but there is no additional - * block for the additional bmap btree block due to a split of the - * bmap btree of the file. The result of this may lead to an - * infinite loop in xfssyncd when the file gets flushed to disk and - * all delayed extents need to be actually allocated. To get around - * this, we explicitly set aside a few blocks which will not be - * reserved in delayed allocation. Considering the minimum number of - * needed freelist blocks is 4 fsbs _per AG_, a potential split of file's bmap - * btree requires 1 fsb, so we set the number of set-aside blocks - * to 4 + 4*agcount. - */ -#define XFS_ALLOC_SET_ASIDE(mp) (4 + ((mp)->m_sb.sb_agcount * 4)) - -/* - * When deciding how much space to allocate out of an AG, we limit the - * allocation maximum size to the size the AG. However, we cannot use all the - * blocks in the AG - some are permanently used by metadata. These - * blocks are generally: - * - the AG superblock, AGF, AGI and AGFL - * - the AGF (bno and cnt) and AGI btree root blocks - * - 4 blocks on the AGFL according to XFS_ALLOC_SET_ASIDE() limits - * - * The AG headers are sector sized, so the amount of space they take up is - * dependent on filesystem geometry. The others are all single blocks. - */ -#define XFS_ALLOC_AG_MAX_USABLE(mp) \ - ((mp)->m_sb.sb_agblocks - XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)) - 7) - - /* * Argument structure for xfs_alloc routines. * This is turned into a structure to avoid having 20 arguments passed @@ -133,6 +97,11 @@ typedef struct xfs_alloc_arg { #define XFS_ALLOC_INITIAL_USER_DATA (1 << 1)/* special case start of file */ #define XFS_ALLOC_USERDATA_ZERO (1 << 2)/* zero extent on allocation */ +/* freespace limit calculations */ +#define XFS_ALLOC_AGFL_RESERVE 4 +unsigned int xfs_alloc_set_aside(struct xfs_mount *mp); +unsigned int xfs_alloc_ag_max_usable(struct xfs_mount *mp); + xfs_extlen_t xfs_alloc_longest_free_extent(struct xfs_mount *mp, struct xfs_perag *pag, xfs_extlen_t need); unsigned int xfs_alloc_min_freelist(struct xfs_mount *mp, diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 919069fc7b23..c4b89218c7eb 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -3674,7 +3674,7 @@ xfs_bmap_btalloc( xfs_rmap_skip_owner_update(&args.oinfo); /* Trim the allocation back to the maximum an AG can fit. */ - args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp)); + args.maxlen = MIN(ap->length, mp->m_ag_max_usable); args.firstblock = *ap->firstblock; blen = 0; if (nullfb) { diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index fb33e70e5c9a..0e3d4f5ec33c 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -745,6 +745,8 @@ xfs_sb_mount_common( mp->m_ialloc_min_blks = sbp->sb_spino_align; else mp->m_ialloc_min_blks = mp->m_ialloc_blks; + mp->m_alloc_set_aside = xfs_alloc_set_aside(mp); + mp->m_ag_max_usable = xfs_alloc_ag_max_usable(mp); } /* diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c index 272c3f8b6f7d..4ff499aa7338 100644 --- a/fs/xfs/xfs_discard.c +++ b/fs/xfs/xfs_discard.c @@ -179,7 +179,7 @@ xfs_ioc_trim( * matter as trimming blocks is an advisory interface. */ if (range.start >= XFS_FSB_TO_B(mp, mp->m_sb.sb_dblocks) || - range.minlen > XFS_FSB_TO_B(mp, XFS_ALLOC_AG_MAX_USABLE(mp)) || + range.minlen > XFS_FSB_TO_B(mp, mp->m_ag_max_usable) || range.len < mp->m_sb.sb_blocksize) return -EINVAL; diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 27e6e8cdd431..83304d54e6d4 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -584,6 +584,7 @@ xfs_growfs_data_private( } else mp->m_maxicount = 0; xfs_set_low_space_thresholds(mp); + mp->m_alloc_set_aside = xfs_alloc_set_aside(mp); /* update secondary superblocks. */ for (agno = 1; agno < nagcount; agno++) { @@ -721,7 +722,7 @@ xfs_fs_counts( cnt->allocino = percpu_counter_read_positive(&mp->m_icount); cnt->freeino = percpu_counter_read_positive(&mp->m_ifree); cnt->freedata = percpu_counter_read_positive(&mp->m_fdblocks) - - XFS_ALLOC_SET_ASIDE(mp); + mp->m_alloc_set_aside; spin_lock(&mp->m_sb_lock); cnt->freertx = mp->m_sb.sb_frextents; @@ -809,7 +810,7 @@ xfs_reserve_blocks( error = -ENOSPC; do { free = percpu_counter_sum(&mp->m_fdblocks) - - XFS_ALLOC_SET_ASIDE(mp); + mp->m_alloc_set_aside; if (!free) break; diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index af608aade6df..f64d4d06cd95 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -5008,6 +5008,7 @@ xlog_do_recover( xfs_warn(mp, "Failed post-recovery per-ag init: %d", error); return error; } + mp->m_alloc_set_aside = xfs_alloc_set_aside(mp); xlog_recover_check_summary(log); diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 0041866d51ca..faeead671f9f 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1221,7 +1221,7 @@ xfs_mod_fdblocks( batch = XFS_FDBLOCKS_BATCH; __percpu_counter_add(&mp->m_fdblocks, delta, batch); - if (__percpu_counter_compare(&mp->m_fdblocks, XFS_ALLOC_SET_ASIDE(mp), + if (__percpu_counter_compare(&mp->m_fdblocks, mp->m_alloc_set_aside, XFS_FDBLOCKS_BATCH) >= 0) { /* we had space! */ return 0; diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 0ed0f29ddaf5..b36676cde103 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -123,6 +123,8 @@ typedef struct xfs_mount { uint m_in_maxlevels; /* max inobt btree levels. */ uint m_rmap_maxlevels; /* max rmap btree levels */ xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */ + uint m_alloc_set_aside; /* space we can't use */ + uint m_ag_max_usable; /* max space per AG */ struct radix_tree_root m_perag_tree; /* per-ag accounting info */ spinlock_t m_perag_lock; /* lock for m_perag_tree */ struct mutex m_growlock; /* growfs mutex */ diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 95a11beb48c6..449cadf149f9 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1075,7 +1075,7 @@ xfs_fs_statfs( statp->f_blocks = sbp->sb_dblocks - lsize; spin_unlock(&mp->m_sb_lock); - statp->f_bfree = fdblocks - XFS_ALLOC_SET_ASIDE(mp); + statp->f_bfree = fdblocks - mp->m_alloc_set_aside; statp->f_bavail = statp->f_bfree; fakeinos = statp->f_bfree << sbp->sb_inopblog; -- cgit v1.2.3 From 4b8ed67794fe57b23801c65f4ea5b0f0b1f0dbab Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:39:05 +1000 Subject: xfs: add rmap btree operations Originally-From: Dave Chinner Implement the generic btree operations needed to manipulate rmap btree blocks. This is very similar to the per-ag freespace btree implementation, and uses the AGFL for allocation and freeing of blocks. Adapt the rmap btree to store owner offsets within each rmap record, and to handle the primary key being redefined as the tuple [agblk, owner, offset]. The expansion of the primary key is crucial to allowing multiple owners per extent. [darrick: adapt the btree ops to deal with offsets] [darrick: remove init_rec_from_key] [darrick: move unwritten bit to rm_offset] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.h | 1 + fs/xfs/libxfs/xfs_rmap.c | 96 +++++++++++++++ fs/xfs/libxfs/xfs_rmap.h | 9 ++ fs/xfs/libxfs/xfs_rmap_btree.c | 267 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_trace.h | 3 + 5 files changed, 376 insertions(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 573d3a0c5797..9b6d6286466e 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -241,6 +241,7 @@ union xfs_btree_irec { struct xfs_alloc_rec_incore a; struct xfs_bmbt_irec b; struct xfs_inobt_rec_incore i; + struct xfs_rmap_irec r; }; /* diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index b522bfc436a2..ce29d6b7cd58 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -36,6 +36,102 @@ #include "xfs_error.h" #include "xfs_extent_busy.h" +/* + * Lookup the first record less than or equal to [bno, len, owner, offset] + * in the btree given by cur. + */ +int +xfs_rmap_lookup_le( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + uint64_t owner, + uint64_t offset, + unsigned int flags, + int *stat) +{ + cur->bc_rec.r.rm_startblock = bno; + cur->bc_rec.r.rm_blockcount = len; + cur->bc_rec.r.rm_owner = owner; + cur->bc_rec.r.rm_offset = offset; + cur->bc_rec.r.rm_flags = flags; + return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat); +} + +/* + * Lookup the record exactly matching [bno, len, owner, offset] + * in the btree given by cur. + */ +int +xfs_rmap_lookup_eq( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + uint64_t owner, + uint64_t offset, + unsigned int flags, + int *stat) +{ + cur->bc_rec.r.rm_startblock = bno; + cur->bc_rec.r.rm_blockcount = len; + cur->bc_rec.r.rm_owner = owner; + cur->bc_rec.r.rm_offset = offset; + cur->bc_rec.r.rm_flags = flags; + return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat); +} + +/* + * Update the record referred to by cur to the value given + * by [bno, len, owner, offset]. + * This either works (return 0) or gets an EFSCORRUPTED error. + */ +STATIC int +xfs_rmap_update( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *irec) +{ + union xfs_btree_rec rec; + + rec.rmap.rm_startblock = cpu_to_be32(irec->rm_startblock); + rec.rmap.rm_blockcount = cpu_to_be32(irec->rm_blockcount); + rec.rmap.rm_owner = cpu_to_be64(irec->rm_owner); + rec.rmap.rm_offset = cpu_to_be64( + xfs_rmap_irec_offset_pack(irec)); + return xfs_btree_update(cur, &rec); +} + +static int +xfs_rmap_btrec_to_irec( + union xfs_btree_rec *rec, + struct xfs_rmap_irec *irec) +{ + irec->rm_flags = 0; + irec->rm_startblock = be32_to_cpu(rec->rmap.rm_startblock); + irec->rm_blockcount = be32_to_cpu(rec->rmap.rm_blockcount); + irec->rm_owner = be64_to_cpu(rec->rmap.rm_owner); + return xfs_rmap_irec_offset_unpack(be64_to_cpu(rec->rmap.rm_offset), + irec); +} + +/* + * Get the data from the pointed-to record. + */ +int +xfs_rmap_get_rec( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *irec, + int *stat) +{ + union xfs_btree_rec *rec; + int error; + + error = xfs_btree_get_rec(cur, &rec, stat); + if (error || !*stat) + return error; + + return xfs_rmap_btrec_to_irec(rec, irec); +} + int xfs_rmap_free( struct xfs_trans *tp, diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index e7a670459fea..aa39a2a825e7 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -142,4 +142,13 @@ int xfs_rmap_free(struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, struct xfs_owner_info *oinfo); +int xfs_rmap_lookup_le(struct xfs_btree_cur *cur, xfs_agblock_t bno, + xfs_extlen_t len, uint64_t owner, uint64_t offset, + unsigned int flags, int *stat); +int xfs_rmap_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno, + xfs_extlen_t len, uint64_t owner, uint64_t offset, + unsigned int flags, int *stat); +int xfs_rmap_get_rec(struct xfs_btree_cur *cur, struct xfs_rmap_irec *irec, + int *stat); + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c index a9ddc191aa2d..95cb964606d1 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.c +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -29,12 +29,38 @@ #include "xfs_trans.h" #include "xfs_alloc.h" #include "xfs_btree.h" +#include "xfs_rmap.h" #include "xfs_rmap_btree.h" #include "xfs_trace.h" #include "xfs_cksum.h" #include "xfs_error.h" #include "xfs_extent_busy.h" +/* + * Reverse map btree. + * + * This is a per-ag tree used to track the owner(s) of a given extent. With + * reflink it is possible for there to be multiple owners, which is a departure + * from classic XFS. Owner records for data extents are inserted when the + * extent is mapped and removed when an extent is unmapped. Owner records for + * all other block types (i.e. metadata) are inserted when an extent is + * allocated and removed when an extent is freed. There can only be one owner + * of a metadata extent, usually an inode or some other metadata structure like + * an AG btree. + * + * The rmap btree is part of the free space management, so blocks for the tree + * are sourced from the agfl. Hence we need transaction reservation support for + * this tree so that the freelist is always large enough. This also impacts on + * the minimum space we need to leave free in the AG. + * + * The tree is ordered by [ag block, owner, offset]. This is a large key size, + * but it is the only way to enforce unique keys when a block can be owned by + * multiple files at any offset. There's no need to order/search by extent + * size for online updating/management of the tree. It is intended that most + * reverse lookups will be to find the owner(s) of a particular block, or to + * try to recover tree and file data from corrupt primary metadata. + */ + static struct xfs_btree_cur * xfs_rmapbt_dup_cursor( struct xfs_btree_cur *cur) @@ -43,6 +69,172 @@ xfs_rmapbt_dup_cursor( cur->bc_private.a.agbp, cur->bc_private.a.agno); } +STATIC void +xfs_rmapbt_set_root( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr, + int inc) +{ + struct xfs_buf *agbp = cur->bc_private.a.agbp; + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + xfs_agnumber_t seqno = be32_to_cpu(agf->agf_seqno); + int btnum = cur->bc_btnum; + struct xfs_perag *pag = xfs_perag_get(cur->bc_mp, seqno); + + ASSERT(ptr->s != 0); + + agf->agf_roots[btnum] = ptr->s; + be32_add_cpu(&agf->agf_levels[btnum], inc); + pag->pagf_levels[btnum] += inc; + xfs_perag_put(pag); + + xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); +} + +STATIC int +xfs_rmapbt_alloc_block( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *start, + union xfs_btree_ptr *new, + int *stat) +{ + int error; + xfs_agblock_t bno; + + XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); + + /* Allocate the new block from the freelist. If we can't, give up. */ + error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp, + &bno, 1); + if (error) { + XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR); + return error; + } + + trace_xfs_rmapbt_alloc_block(cur->bc_mp, cur->bc_private.a.agno, + bno, 1); + if (bno == NULLAGBLOCK) { + XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); + *stat = 0; + return 0; + } + + xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1, + false); + + xfs_trans_agbtree_delta(cur->bc_tp, 1); + new->s = cpu_to_be32(bno); + + XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); + *stat = 1; + return 0; +} + +STATIC int +xfs_rmapbt_free_block( + struct xfs_btree_cur *cur, + struct xfs_buf *bp) +{ + struct xfs_buf *agbp = cur->bc_private.a.agbp; + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + xfs_agblock_t bno; + int error; + + bno = xfs_daddr_to_agbno(cur->bc_mp, XFS_BUF_ADDR(bp)); + trace_xfs_rmapbt_free_block(cur->bc_mp, cur->bc_private.a.agno, + bno, 1); + error = xfs_alloc_put_freelist(cur->bc_tp, agbp, NULL, bno, 1); + if (error) + return error; + + xfs_extent_busy_insert(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1, + XFS_EXTENT_BUSY_SKIP_DISCARD); + xfs_trans_agbtree_delta(cur->bc_tp, -1); + + return 0; +} + +STATIC int +xfs_rmapbt_get_minrecs( + struct xfs_btree_cur *cur, + int level) +{ + return cur->bc_mp->m_rmap_mnr[level != 0]; +} + +STATIC int +xfs_rmapbt_get_maxrecs( + struct xfs_btree_cur *cur, + int level) +{ + return cur->bc_mp->m_rmap_mxr[level != 0]; +} + +STATIC void +xfs_rmapbt_init_key_from_rec( + union xfs_btree_key *key, + union xfs_btree_rec *rec) +{ + key->rmap.rm_startblock = rec->rmap.rm_startblock; + key->rmap.rm_owner = rec->rmap.rm_owner; + key->rmap.rm_offset = rec->rmap.rm_offset; +} + +STATIC void +xfs_rmapbt_init_rec_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_rec *rec) +{ + rec->rmap.rm_startblock = cpu_to_be32(cur->bc_rec.r.rm_startblock); + rec->rmap.rm_blockcount = cpu_to_be32(cur->bc_rec.r.rm_blockcount); + rec->rmap.rm_owner = cpu_to_be64(cur->bc_rec.r.rm_owner); + rec->rmap.rm_offset = cpu_to_be64( + xfs_rmap_irec_offset_pack(&cur->bc_rec.r)); +} + +STATIC void +xfs_rmapbt_init_ptr_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr) +{ + struct xfs_agf *agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp); + + ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno)); + ASSERT(agf->agf_roots[cur->bc_btnum] != 0); + + ptr->s = agf->agf_roots[cur->bc_btnum]; +} + +STATIC __int64_t +xfs_rmapbt_key_diff( + struct xfs_btree_cur *cur, + union xfs_btree_key *key) +{ + struct xfs_rmap_irec *rec = &cur->bc_rec.r; + struct xfs_rmap_key *kp = &key->rmap; + __u64 x, y; + __int64_t d; + + d = (__int64_t)be32_to_cpu(kp->rm_startblock) - rec->rm_startblock; + if (d) + return d; + + x = be64_to_cpu(kp->rm_owner); + y = rec->rm_owner; + if (x > y) + return 1; + else if (y > x) + return -1; + + x = XFS_RMAP_OFF(be64_to_cpu(kp->rm_offset)); + y = rec->rm_offset; + if (x > y) + return 1; + else if (y > x) + return -1; + return 0; +} + static bool xfs_rmapbt_verify( struct xfs_buf *bp) @@ -117,12 +309,87 @@ const struct xfs_buf_ops xfs_rmapbt_buf_ops = { .verify_write = xfs_rmapbt_write_verify, }; +#if defined(DEBUG) || defined(XFS_WARN) +STATIC int +xfs_rmapbt_keys_inorder( + struct xfs_btree_cur *cur, + union xfs_btree_key *k1, + union xfs_btree_key *k2) +{ + __uint32_t x; + __uint32_t y; + __uint64_t a; + __uint64_t b; + + x = be32_to_cpu(k1->rmap.rm_startblock); + y = be32_to_cpu(k2->rmap.rm_startblock); + if (x < y) + return 1; + else if (x > y) + return 0; + a = be64_to_cpu(k1->rmap.rm_owner); + b = be64_to_cpu(k2->rmap.rm_owner); + if (a < b) + return 1; + else if (a > b) + return 0; + a = XFS_RMAP_OFF(be64_to_cpu(k1->rmap.rm_offset)); + b = XFS_RMAP_OFF(be64_to_cpu(k2->rmap.rm_offset)); + if (a <= b) + return 1; + return 0; +} + +STATIC int +xfs_rmapbt_recs_inorder( + struct xfs_btree_cur *cur, + union xfs_btree_rec *r1, + union xfs_btree_rec *r2) +{ + __uint32_t x; + __uint32_t y; + __uint64_t a; + __uint64_t b; + + x = be32_to_cpu(r1->rmap.rm_startblock); + y = be32_to_cpu(r2->rmap.rm_startblock); + if (x < y) + return 1; + else if (x > y) + return 0; + a = be64_to_cpu(r1->rmap.rm_owner); + b = be64_to_cpu(r2->rmap.rm_owner); + if (a < b) + return 1; + else if (a > b) + return 0; + a = XFS_RMAP_OFF(be64_to_cpu(r1->rmap.rm_offset)); + b = XFS_RMAP_OFF(be64_to_cpu(r2->rmap.rm_offset)); + if (a <= b) + return 1; + return 0; +} +#endif /* DEBUG */ + static const struct xfs_btree_ops xfs_rmapbt_ops = { .rec_len = sizeof(struct xfs_rmap_rec), .key_len = 2 * sizeof(struct xfs_rmap_key), .dup_cursor = xfs_rmapbt_dup_cursor, + .set_root = xfs_rmapbt_set_root, + .alloc_block = xfs_rmapbt_alloc_block, + .free_block = xfs_rmapbt_free_block, + .get_minrecs = xfs_rmapbt_get_minrecs, + .get_maxrecs = xfs_rmapbt_get_maxrecs, + .init_key_from_rec = xfs_rmapbt_init_key_from_rec, + .init_rec_from_cur = xfs_rmapbt_init_rec_from_cur, + .init_ptr_from_cur = xfs_rmapbt_init_ptr_from_cur, + .key_diff = xfs_rmapbt_key_diff, .buf_ops = &xfs_rmapbt_buf_ops, +#if defined(DEBUG) || defined(XFS_WARN) + .keys_inorder = xfs_rmapbt_keys_inorder, + .recs_inorder = xfs_rmapbt_recs_inorder, +#endif .get_leaf_keys = xfs_btree_get_leaf_keys_overlapped, .get_node_keys = xfs_btree_get_node_keys_overlapped, diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 4c3418b1a244..e69912acb23f 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2502,6 +2502,9 @@ DEFINE_RMAP_EVENT(xfs_rmap_map); DEFINE_RMAP_EVENT(xfs_rmap_map_done); DEFINE_AG_ERROR_EVENT(xfs_rmap_map_error); +DEFINE_BUSY_EVENT(xfs_rmapbt_alloc_block); +DEFINE_BUSY_EVENT(xfs_rmapbt_free_block); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH -- cgit v1.2.3 From cfed56ae5f410cd6c1601712a9ed4645b71b170c Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:40:56 +1000 Subject: xfs: support overlapping intervals in the rmap btree Now that the generic btree code supports overlapping intervals, plug in the rmap btree to this functionality. We will need it to find potential left neighbors in xfs_rmap_{alloc,free} later in the patch set. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap_btree.c | 68 ++++++++++++++++++++++++++++++++++++++++-- fs/xfs/libxfs/xfs_rmap_btree.h | 10 +++++-- 2 files changed, 74 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c index 95cb964606d1..232450ce61a8 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.c +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -180,6 +180,35 @@ xfs_rmapbt_init_key_from_rec( key->rmap.rm_offset = rec->rmap.rm_offset; } +/* + * The high key for a reverse mapping record can be computed by shifting + * the startblock and offset to the highest value that would still map + * to that record. In practice this means that we add blockcount-1 to + * the startblock for all records, and if the record is for a data/attr + * fork mapping, we add blockcount-1 to the offset too. + */ +STATIC void +xfs_rmapbt_init_high_key_from_rec( + union xfs_btree_key *key, + union xfs_btree_rec *rec) +{ + __uint64_t off; + int adj; + + adj = be32_to_cpu(rec->rmap.rm_blockcount) - 1; + + key->rmap.rm_startblock = rec->rmap.rm_startblock; + be32_add_cpu(&key->rmap.rm_startblock, adj); + key->rmap.rm_owner = rec->rmap.rm_owner; + key->rmap.rm_offset = rec->rmap.rm_offset; + if (XFS_RMAP_NON_INODE_OWNER(be64_to_cpu(rec->rmap.rm_owner)) || + XFS_RMAP_IS_BMBT_BLOCK(be64_to_cpu(rec->rmap.rm_offset))) + return; + off = be64_to_cpu(key->rmap.rm_offset); + off = (XFS_RMAP_OFF(off) + adj) | (off & ~XFS_RMAP_OFF_MASK); + key->rmap.rm_offset = cpu_to_be64(off); +} + STATIC void xfs_rmapbt_init_rec_from_cur( struct xfs_btree_cur *cur, @@ -235,6 +264,38 @@ xfs_rmapbt_key_diff( return 0; } +STATIC __int64_t +xfs_rmapbt_diff_two_keys( + struct xfs_btree_cur *cur, + union xfs_btree_key *k1, + union xfs_btree_key *k2) +{ + struct xfs_rmap_key *kp1 = &k1->rmap; + struct xfs_rmap_key *kp2 = &k2->rmap; + __int64_t d; + __u64 x, y; + + d = (__int64_t)be32_to_cpu(kp1->rm_startblock) - + be32_to_cpu(kp2->rm_startblock); + if (d) + return d; + + x = be64_to_cpu(kp1->rm_owner); + y = be64_to_cpu(kp2->rm_owner); + if (x > y) + return 1; + else if (y > x) + return -1; + + x = XFS_RMAP_OFF(be64_to_cpu(kp1->rm_offset)); + y = XFS_RMAP_OFF(be64_to_cpu(kp2->rm_offset)); + if (x > y) + return 1; + else if (y > x) + return -1; + return 0; +} + static bool xfs_rmapbt_verify( struct xfs_buf *bp) @@ -382,10 +443,12 @@ static const struct xfs_btree_ops xfs_rmapbt_ops = { .get_minrecs = xfs_rmapbt_get_minrecs, .get_maxrecs = xfs_rmapbt_get_maxrecs, .init_key_from_rec = xfs_rmapbt_init_key_from_rec, + .init_high_key_from_rec = xfs_rmapbt_init_high_key_from_rec, .init_rec_from_cur = xfs_rmapbt_init_rec_from_cur, .init_ptr_from_cur = xfs_rmapbt_init_ptr_from_cur, .key_diff = xfs_rmapbt_key_diff, .buf_ops = &xfs_rmapbt_buf_ops, + .diff_two_keys = xfs_rmapbt_diff_two_keys, #if defined(DEBUG) || defined(XFS_WARN) .keys_inorder = xfs_rmapbt_keys_inorder, .recs_inorder = xfs_rmapbt_recs_inorder, @@ -412,8 +475,9 @@ xfs_rmapbt_init_cursor( cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); cur->bc_tp = tp; cur->bc_mp = mp; + /* Overlapping btree; 2 keys per pointer. */ cur->bc_btnum = XFS_BTNUM_RMAP; - cur->bc_flags = XFS_BTREE_CRC_BLOCKS; + cur->bc_flags = XFS_BTREE_CRC_BLOCKS | XFS_BTREE_OVERLAPPING; cur->bc_blocklog = mp->m_sb.sb_blocklog; cur->bc_ops = &xfs_rmapbt_ops; cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]); @@ -438,7 +502,7 @@ xfs_rmapbt_maxrecs( if (leaf) return blocklen / sizeof(struct xfs_rmap_rec); return blocklen / - (sizeof(struct xfs_rmap_key) + sizeof(xfs_rmap_ptr_t)); + (2 * sizeof(struct xfs_rmap_key) + sizeof(xfs_rmap_ptr_t)); } /* Compute the maximum height of an rmap btree. */ diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h index a3a6b7d476c7..e73a55357dab 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.h +++ b/fs/xfs/libxfs/xfs_rmap_btree.h @@ -38,12 +38,18 @@ struct xfs_mount; #define XFS_RMAP_KEY_ADDR(block, index) \ ((struct xfs_rmap_key *) \ ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ - ((index) - 1) * sizeof(struct xfs_rmap_key))) + ((index) - 1) * 2 * sizeof(struct xfs_rmap_key))) + +#define XFS_RMAP_HIGH_KEY_ADDR(block, index) \ + ((struct xfs_rmap_key *) \ + ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ + sizeof(struct xfs_rmap_key) + \ + ((index) - 1) * 2 * sizeof(struct xfs_rmap_key))) #define XFS_RMAP_PTR_ADDR(block, index, maxrecs) \ ((xfs_rmap_ptr_t *) \ ((char *)(block) + XFS_RMAP_BLOCK_LEN + \ - (maxrecs) * sizeof(struct xfs_rmap_key) + \ + (maxrecs) * 2 * sizeof(struct xfs_rmap_key) + \ ((index) - 1) * sizeof(xfs_rmap_ptr_t))) struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp, -- cgit v1.2.3 From c543838a1e00a5f8791e59ae570b1030d70906f2 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:42:39 +1000 Subject: xfs: teach rmapbt to support interval queries Now that the generic btree code supports querying all records within a range of keys, use that functionality to allow us to ask for all the extents mapped to a range of physical blocks. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap.c | 43 +++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_rmap.h | 9 +++++++++ 2 files changed, 52 insertions(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index ce29d6b7cd58..6638bc725b23 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -183,3 +183,46 @@ out_error: trace_xfs_rmap_map_error(mp, agno, error, _RET_IP_); return error; } + +struct xfs_rmap_query_range_info { + xfs_rmap_query_range_fn fn; + void *priv; +}; + +/* Format btree record and pass to our callback. */ +STATIC int +xfs_rmap_query_range_helper( + struct xfs_btree_cur *cur, + union xfs_btree_rec *rec, + void *priv) +{ + struct xfs_rmap_query_range_info *query = priv; + struct xfs_rmap_irec irec; + int error; + + error = xfs_rmap_btrec_to_irec(rec, &irec); + if (error) + return error; + return query->fn(cur, &irec, query->priv); +} + +/* Find all rmaps between two keys. */ +int +xfs_rmap_query_range( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *low_rec, + struct xfs_rmap_irec *high_rec, + xfs_rmap_query_range_fn fn, + void *priv) +{ + union xfs_btree_irec low_brec; + union xfs_btree_irec high_brec; + struct xfs_rmap_query_range_info query; + + low_brec.r = *low_rec; + high_brec.r = *high_rec; + query.priv = priv; + query.fn = fn; + return xfs_btree_query_range(cur, &low_brec, &high_brec, + xfs_rmap_query_range_helper, &query); +} diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index aa39a2a825e7..34c811a2dacc 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -151,4 +151,13 @@ int xfs_rmap_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno, int xfs_rmap_get_rec(struct xfs_btree_cur *cur, struct xfs_rmap_irec *irec, int *stat); +typedef int (*xfs_rmap_query_range_fn)( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *rec, + void *priv); + +int xfs_rmap_query_range(struct xfs_btree_cur *cur, + struct xfs_rmap_irec *low_rec, struct xfs_rmap_irec *high_rec, + xfs_rmap_query_range_fn fn, void *priv); + #endif /* __XFS_RMAP_H__ */ -- cgit v1.2.3 From aa966d84aa487c4cb7f6fade1a35ca961652a395 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:43:24 +1000 Subject: xfs: add tracepoints for the rmap functions Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_trace.h | 51 +++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 49 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index e69912acb23f..7498c7780d99 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2417,8 +2417,6 @@ DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_cancel); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); -DEFINE_MAP_EXTENT_DEFERRED_EVENT(xfs_defer_map_extent); - #define DEFINE_BMAP_FREE_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer); DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred); @@ -2502,8 +2500,57 @@ DEFINE_RMAP_EVENT(xfs_rmap_map); DEFINE_RMAP_EVENT(xfs_rmap_map_done); DEFINE_AG_ERROR_EVENT(xfs_rmap_map_error); +DECLARE_EVENT_CLASS(xfs_rmapbt_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + xfs_agblock_t agbno, xfs_extlen_t len, + uint64_t owner, uint64_t offset, unsigned int flags), + TP_ARGS(mp, agno, agbno, len, owner, offset, flags), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + __field(uint64_t, owner) + __field(uint64_t, offset) + __field(unsigned int, flags) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->agbno = agbno; + __entry->len = len; + __entry->owner = owner; + __entry->offset = offset; + __entry->flags = flags; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u owner %lld offset %llu flags 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->agbno, + __entry->len, + __entry->owner, + __entry->offset, + __entry->flags) +); +#define DEFINE_RMAPBT_EVENT(name) \ +DEFINE_EVENT(xfs_rmapbt_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + xfs_agblock_t agbno, xfs_extlen_t len, \ + uint64_t owner, uint64_t offset, unsigned int flags), \ + TP_ARGS(mp, agno, agbno, len, owner, offset, flags)) + +#define DEFINE_RMAP_DEFERRED_EVENT DEFINE_MAP_EXTENT_DEFERRED_EVENT +DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_defer); +DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_deferred); + DEFINE_BUSY_EVENT(xfs_rmapbt_alloc_block); DEFINE_BUSY_EVENT(xfs_rmapbt_free_block); +DEFINE_RMAPBT_EVENT(xfs_rmap_update); +DEFINE_RMAPBT_EVENT(xfs_rmap_insert); +DEFINE_RMAPBT_EVENT(xfs_rmap_delete); +DEFINE_AG_ERROR_EVENT(xfs_rmap_insert_error); +DEFINE_AG_ERROR_EVENT(xfs_rmap_delete_error); +DEFINE_AG_ERROR_EVENT(xfs_rmap_update_error); #endif /* _TRACE_XFS_H */ -- cgit v1.2.3 From 0a1b0b3855cf74bb11243076b00178a0f1a0320e Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:44:21 +1000 Subject: xfs: add an extent to the rmap btree Originally-From: Dave Chinner Now all the btree, free space and transaction infrastructure is in place, we can finally add the code to insert reverse mappings to the rmap btree. Freeing will be done in a separate patch, so just the addition operation can be focussed on here. [darrick: handle owner offsets when adding rmaps] [dchinner: remove remaining debug printk statements] [darrick: move unwritten bit to rm_offset] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap.c | 226 +++++++++++++++++++++++++++++++++++++++++++++-- fs/xfs/xfs_trace.h | 2 + 2 files changed, 223 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 6638bc725b23..1c0e2a8943a4 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -31,6 +31,7 @@ #include "xfs_trans.h" #include "xfs_alloc.h" #include "xfs_rmap.h" +#include "xfs_rmap_btree.h" #include "xfs_trans_space.h" #include "xfs_trace.h" #include "xfs_error.h" @@ -158,6 +159,218 @@ out_error: return error; } +/* + * A mergeable rmap must have the same owner and the same values for + * the unwritten, attr_fork, and bmbt flags. The startblock and + * offset are checked separately. + */ +static bool +xfs_rmap_is_mergeable( + struct xfs_rmap_irec *irec, + uint64_t owner, + unsigned int flags) +{ + if (irec->rm_owner == XFS_RMAP_OWN_NULL) + return false; + if (irec->rm_owner != owner) + return false; + if ((flags & XFS_RMAP_UNWRITTEN) ^ + (irec->rm_flags & XFS_RMAP_UNWRITTEN)) + return false; + if ((flags & XFS_RMAP_ATTR_FORK) ^ + (irec->rm_flags & XFS_RMAP_ATTR_FORK)) + return false; + if ((flags & XFS_RMAP_BMBT_BLOCK) ^ + (irec->rm_flags & XFS_RMAP_BMBT_BLOCK)) + return false; + return true; +} + +/* + * When we allocate a new block, the first thing we do is add a reference to + * the extent in the rmap btree. This takes the form of a [agbno, length, + * owner, offset] record. Flags are encoded in the high bits of the offset + * field. + */ +STATIC int +xfs_rmap_map( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec ltrec; + struct xfs_rmap_irec gtrec; + int have_gt; + int have_lt; + int error = 0; + int i; + uint64_t owner; + uint64_t offset; + unsigned int flags = 0; + bool ignore_off; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + ASSERT(owner != 0); + ignore_off = XFS_RMAP_NON_INODE_OWNER(owner) || + (flags & XFS_RMAP_BMBT_BLOCK); + if (unwritten) + flags |= XFS_RMAP_UNWRITTEN; + trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* + * For the initial lookup, look for an exact match or the left-adjacent + * record for our insertion point. This will also give us the record for + * start block contiguity tests. + */ + error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, flags, + &have_lt); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, have_lt == 1, out_error); + + error = xfs_rmap_get_rec(cur, <rec, &have_lt); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, have_lt == 1, out_error); + trace_xfs_rmap_lookup_le_range_result(cur->bc_mp, + cur->bc_private.a.agno, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags); + + if (!xfs_rmap_is_mergeable(<rec, owner, flags)) + have_lt = 0; + + XFS_WANT_CORRUPTED_GOTO(mp, + have_lt == 0 || + ltrec.rm_startblock + ltrec.rm_blockcount <= bno, out_error); + + /* + * Increment the cursor to see if we have a right-adjacent record to our + * insertion point. This will give us the record for end block + * contiguity tests. + */ + error = xfs_btree_increment(cur, 0, &have_gt); + if (error) + goto out_error; + if (have_gt) { + error = xfs_rmap_get_rec(cur, >rec, &have_gt); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, have_gt == 1, out_error); + XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= gtrec.rm_startblock, + out_error); + trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, gtrec.rm_startblock, + gtrec.rm_blockcount, gtrec.rm_owner, + gtrec.rm_offset, gtrec.rm_flags); + if (!xfs_rmap_is_mergeable(>rec, owner, flags)) + have_gt = 0; + } + + /* + * Note: cursor currently points one record to the right of ltrec, even + * if there is no record in the tree to the right. + */ + if (have_lt && + ltrec.rm_startblock + ltrec.rm_blockcount == bno && + (ignore_off || ltrec.rm_offset + ltrec.rm_blockcount == offset)) { + /* + * left edge contiguous, merge into left record. + * + * ltbno ltlen + * orig: |ooooooooo| + * adding: |aaaaaaaaa| + * result: |rrrrrrrrrrrrrrrrrrr| + * bno len + */ + ltrec.rm_blockcount += len; + if (have_gt && + bno + len == gtrec.rm_startblock && + (ignore_off || offset + len == gtrec.rm_offset) && + (unsigned long)ltrec.rm_blockcount + len + + gtrec.rm_blockcount <= XFS_RMAP_LEN_MAX) { + /* + * right edge also contiguous, delete right record + * and merge into left record. + * + * ltbno ltlen gtbno gtlen + * orig: |ooooooooo| |ooooooooo| + * adding: |aaaaaaaaa| + * result: |rrrrrrrrrrrrrrrrrrrrrrrrrrrrr| + */ + ltrec.rm_blockcount += gtrec.rm_blockcount; + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + gtrec.rm_startblock, + gtrec.rm_blockcount, + gtrec.rm_owner, + gtrec.rm_offset, + gtrec.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + } + + /* point the cursor back to the left record and update */ + error = xfs_btree_decrement(cur, 0, &have_gt); + if (error) + goto out_error; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + } else if (have_gt && + bno + len == gtrec.rm_startblock && + (ignore_off || offset + len == gtrec.rm_offset)) { + /* + * right edge contiguous, merge into right record. + * + * gtbno gtlen + * Orig: |ooooooooo| + * adding: |aaaaaaaaa| + * Result: |rrrrrrrrrrrrrrrrrrr| + * bno len + */ + gtrec.rm_startblock = bno; + gtrec.rm_blockcount += len; + if (!ignore_off) + gtrec.rm_offset = offset; + error = xfs_rmap_update(cur, >rec); + if (error) + goto out_error; + } else { + /* + * no contiguous edge with identical owner, insert + * new record at current cursor position. + */ + cur->bc_rec.r.rm_startblock = bno; + cur->bc_rec.r.rm_blockcount = len; + cur->bc_rec.r.rm_owner = owner; + cur->bc_rec.r.rm_offset = offset; + cur->bc_rec.r.rm_flags = flags; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len, + owner, offset, flags); + error = xfs_btree_insert(cur, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + } + + trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +out_error: + if (error) + trace_xfs_rmap_map_error(mp, cur->bc_private.a.agno, + error, _RET_IP_); + return error; +} + +/* + * Add a reference to an extent in the rmap btree. + */ int xfs_rmap_alloc( struct xfs_trans *tp, @@ -168,19 +381,22 @@ xfs_rmap_alloc( struct xfs_owner_info *oinfo) { struct xfs_mount *mp = tp->t_mountp; - int error = 0; + struct xfs_btree_cur *cur; + int error; if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) return 0; - trace_xfs_rmap_map(mp, agno, bno, len, false, oinfo); - if (1) + cur = xfs_rmapbt_init_cursor(mp, tp, agbp, agno); + error = xfs_rmap_map(cur, bno, len, false, oinfo); + if (error) goto out_error; - trace_xfs_rmap_map_done(mp, agno, bno, len, false, oinfo); + + xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); return 0; out_error: - trace_xfs_rmap_map_error(mp, agno, error, _RET_IP_); + xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); return error; } diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 7498c7780d99..e099eda75b68 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2551,6 +2551,8 @@ DEFINE_RMAPBT_EVENT(xfs_rmap_delete); DEFINE_AG_ERROR_EVENT(xfs_rmap_insert_error); DEFINE_AG_ERROR_EVENT(xfs_rmap_delete_error); DEFINE_AG_ERROR_EVENT(xfs_rmap_update_error); +DEFINE_RMAPBT_EVENT(xfs_rmap_lookup_le_range_result); +DEFINE_RMAPBT_EVENT(xfs_rmap_find_right_neighbor_result); #endif /* _TRACE_XFS_H */ -- cgit v1.2.3 From f922cd90b82c5e78a860f194728d4dadc8575106 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 11:45:12 +1000 Subject: xfs: remove an extent from the rmap btree Originally-From: Dave Chinner Now that we have records in the rmap btree, we need to remove them when extents are freed. This needs to find the relevant record in the btree and remove/trim/split it accordingly. [darrick.wong@oracle.com: make rmap routines handle the enlarged keyspace] [dchinner: remove remaining unused debug printks] [darrick: fix a bug when growfs in an AG with an rmap ending at EOFS] Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap.c | 220 +++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 215 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 1c0e2a8943a4..1e48a0b62ba6 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -133,6 +133,212 @@ xfs_rmap_get_rec( return xfs_rmap_btrec_to_irec(rec, irec); } +/* + * Find the extent in the rmap btree and remove it. + * + * The record we find should always be an exact match for the extent that we're + * looking for, since we insert them into the btree without modification. + * + * Special Case #1: when growing the filesystem, we "free" an extent when + * growing the last AG. This extent is new space and so it is not tracked as + * used space in the btree. The growfs code will pass in an owner of + * XFS_RMAP_OWN_NULL to indicate that it expected that there is no owner of this + * extent. We verify that - the extent lookup result in a record that does not + * overlap. + * + * Special Case #2: EFIs do not record the owner of the extent, so when + * recovering EFIs from the log we pass in XFS_RMAP_OWN_UNKNOWN to tell the rmap + * btree to ignore the owner (i.e. wildcard match) so we don't trigger + * corruption checks during log recovery. + */ +STATIC int +xfs_rmap_unmap( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec ltrec; + uint64_t ltoff; + int error = 0; + int i; + uint64_t owner; + uint64_t offset; + unsigned int flags; + bool ignore_off; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + ignore_off = XFS_RMAP_NON_INODE_OWNER(owner) || + (flags & XFS_RMAP_BMBT_BLOCK); + if (unwritten) + flags |= XFS_RMAP_UNWRITTEN; + trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* + * We should always have a left record because there's a static record + * for the AG headers at rm_startblock == 0 created by mkfs/growfs that + * will not ever be removed from the tree. + */ + error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, flags, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + + error = xfs_rmap_get_rec(cur, <rec, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + trace_xfs_rmap_lookup_le_range_result(cur->bc_mp, + cur->bc_private.a.agno, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags); + ltoff = ltrec.rm_offset; + + /* + * For growfs, the incoming extent must be beyond the left record we + * just found as it is new space and won't be used by anyone. This is + * just a corruption check as we don't actually do anything with this + * extent. Note that we need to use >= instead of > because it might + * be the case that the "left" extent goes all the way to EOFS. + */ + if (owner == XFS_RMAP_OWN_NULL) { + XFS_WANT_CORRUPTED_GOTO(mp, bno >= ltrec.rm_startblock + + ltrec.rm_blockcount, out_error); + goto out_done; + } + + /* Make sure the unwritten flag matches. */ + XFS_WANT_CORRUPTED_GOTO(mp, (flags & XFS_RMAP_UNWRITTEN) == + (ltrec.rm_flags & XFS_RMAP_UNWRITTEN), out_error); + + /* Make sure the extent we found covers the entire freeing range. */ + XFS_WANT_CORRUPTED_GOTO(mp, ltrec.rm_startblock <= bno && + ltrec.rm_startblock + ltrec.rm_blockcount >= + bno + len, out_error); + + /* Make sure the owner matches what we expect to find in the tree. */ + XFS_WANT_CORRUPTED_GOTO(mp, owner == ltrec.rm_owner || + XFS_RMAP_NON_INODE_OWNER(owner), out_error); + + /* Check the offset, if necessary. */ + if (!XFS_RMAP_NON_INODE_OWNER(owner)) { + if (flags & XFS_RMAP_BMBT_BLOCK) { + XFS_WANT_CORRUPTED_GOTO(mp, + ltrec.rm_flags & XFS_RMAP_BMBT_BLOCK, + out_error); + } else { + XFS_WANT_CORRUPTED_GOTO(mp, + ltrec.rm_offset <= offset, out_error); + XFS_WANT_CORRUPTED_GOTO(mp, + ltoff + ltrec.rm_blockcount >= offset + len, + out_error); + } + } + + if (ltrec.rm_startblock == bno && ltrec.rm_blockcount == len) { + /* exact match, simply remove the record from rmap tree */ + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + ltrec.rm_startblock, ltrec.rm_blockcount, + ltrec.rm_owner, ltrec.rm_offset, + ltrec.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + } else if (ltrec.rm_startblock == bno) { + /* + * overlap left hand side of extent: move the start, trim the + * length and update the current record. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrrrrrrr| + * bno len + */ + ltrec.rm_startblock += len; + ltrec.rm_blockcount -= len; + if (!ignore_off) + ltrec.rm_offset += len; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + } else if (ltrec.rm_startblock + ltrec.rm_blockcount == bno + len) { + /* + * overlap right hand side of extent: trim the length and update + * the current record. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrrrrrrr| + * bno len + */ + ltrec.rm_blockcount -= len; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + } else { + + /* + * overlap middle of extent: trim the length of the existing + * record to the length of the new left-extent size, increment + * the insertion position so we can insert a new record + * containing the remaining right-extent space. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrr| |rrrr| + * bno len + */ + xfs_extlen_t orig_len = ltrec.rm_blockcount; + + ltrec.rm_blockcount = bno - ltrec.rm_startblock; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto out_error; + + cur->bc_rec.r.rm_startblock = bno + len; + cur->bc_rec.r.rm_blockcount = orig_len - len - + ltrec.rm_blockcount; + cur->bc_rec.r.rm_owner = ltrec.rm_owner; + if (ignore_off) + cur->bc_rec.r.rm_offset = 0; + else + cur->bc_rec.r.rm_offset = offset + len; + cur->bc_rec.r.rm_flags = flags; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, + cur->bc_rec.r.rm_startblock, + cur->bc_rec.r.rm_blockcount, + cur->bc_rec.r.rm_owner, + cur->bc_rec.r.rm_offset, + cur->bc_rec.r.rm_flags); + error = xfs_btree_insert(cur, &i); + if (error) + goto out_error; + } + +out_done: + trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +out_error: + if (error) + trace_xfs_rmap_unmap_error(mp, cur->bc_private.a.agno, + error, _RET_IP_); + return error; +} + +/* + * Remove a reference to an extent in the rmap btree. + */ int xfs_rmap_free( struct xfs_trans *tp, @@ -143,19 +349,23 @@ xfs_rmap_free( struct xfs_owner_info *oinfo) { struct xfs_mount *mp = tp->t_mountp; - int error = 0; + struct xfs_btree_cur *cur; + int error; if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) return 0; - trace_xfs_rmap_unmap(mp, agno, bno, len, false, oinfo); - if (1) + cur = xfs_rmapbt_init_cursor(mp, tp, agbp, agno); + + error = xfs_rmap_unmap(cur, bno, len, false, oinfo); + if (error) goto out_error; - trace_xfs_rmap_unmap_done(mp, agno, bno, len, false, oinfo); + + xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); return 0; out_error: - trace_xfs_rmap_unmap_error(mp, agno, error, _RET_IP_); + xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); return error; } -- cgit v1.2.3 From fb7d9267692a5cdc01648bf4c8fdca51054bc0f2 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:03:19 +1000 Subject: xfs: convert unwritten status of reverse mappings Provide a function to convert an unwritten rmap extent to a real one and vice versa. [ dchinner: Note that this algorithm and code was derived from the existing bmapbt unwritten extent conversion code in xfs_bmap_add_extent_unwritten_real(). ] Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap.c | 441 +++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_trace.h | 5 + 2 files changed, 446 insertions(+) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 1e48a0b62ba6..743927238bdc 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -610,6 +610,447 @@ out_error: return error; } +#define RMAP_LEFT_CONTIG (1 << 0) +#define RMAP_RIGHT_CONTIG (1 << 1) +#define RMAP_LEFT_FILLING (1 << 2) +#define RMAP_RIGHT_FILLING (1 << 3) +#define RMAP_LEFT_VALID (1 << 6) +#define RMAP_RIGHT_VALID (1 << 7) + +#define LEFT r[0] +#define RIGHT r[1] +#define PREV r[2] +#define NEW r[3] + +/* + * Convert an unwritten extent to a real extent or vice versa. + * Does not handle overlapping extents. + */ +STATIC int +xfs_rmap_convert( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec r[4]; /* neighbor extent entries */ + /* left is 0, right is 1, prev is 2 */ + /* new is 3 */ + uint64_t owner; + uint64_t offset; + uint64_t new_endoff; + unsigned int oldext; + unsigned int newext; + unsigned int flags = 0; + int i; + int state = 0; + int error; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + ASSERT(!(XFS_RMAP_NON_INODE_OWNER(owner) || + (flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK)))); + oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0; + new_endoff = offset + len; + trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* + * For the initial lookup, look for an exact match or the left-adjacent + * record for our insertion point. This will also give us the record for + * start block contiguity tests. + */ + error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, oldext, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + + error = xfs_rmap_get_rec(cur, &PREV, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + trace_xfs_rmap_lookup_le_range_result(cur->bc_mp, + cur->bc_private.a.agno, PREV.rm_startblock, + PREV.rm_blockcount, PREV.rm_owner, + PREV.rm_offset, PREV.rm_flags); + + ASSERT(PREV.rm_offset <= offset); + ASSERT(PREV.rm_offset + PREV.rm_blockcount >= new_endoff); + ASSERT((PREV.rm_flags & XFS_RMAP_UNWRITTEN) == oldext); + newext = ~oldext & XFS_RMAP_UNWRITTEN; + + /* + * Set flags determining what part of the previous oldext allocation + * extent is being replaced by a newext allocation. + */ + if (PREV.rm_offset == offset) + state |= RMAP_LEFT_FILLING; + if (PREV.rm_offset + PREV.rm_blockcount == new_endoff) + state |= RMAP_RIGHT_FILLING; + + /* + * Decrement the cursor to see if we have a left-adjacent record to our + * insertion point. This will give us the record for end block + * contiguity tests. + */ + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + if (i) { + state |= RMAP_LEFT_VALID; + error = xfs_rmap_get_rec(cur, &LEFT, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + XFS_WANT_CORRUPTED_GOTO(mp, + LEFT.rm_startblock + LEFT.rm_blockcount <= bno, + done); + trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, LEFT.rm_startblock, + LEFT.rm_blockcount, LEFT.rm_owner, + LEFT.rm_offset, LEFT.rm_flags); + if (LEFT.rm_startblock + LEFT.rm_blockcount == bno && + LEFT.rm_offset + LEFT.rm_blockcount == offset && + xfs_rmap_is_mergeable(&LEFT, owner, newext)) + state |= RMAP_LEFT_CONTIG; + } + + /* + * Increment the cursor to see if we have a right-adjacent record to our + * insertion point. This will give us the record for end block + * contiguity tests. + */ + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto done; + if (i) { + state |= RMAP_RIGHT_VALID; + error = xfs_rmap_get_rec(cur, &RIGHT, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= RIGHT.rm_startblock, + done); + trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, RIGHT.rm_startblock, + RIGHT.rm_blockcount, RIGHT.rm_owner, + RIGHT.rm_offset, RIGHT.rm_flags); + if (bno + len == RIGHT.rm_startblock && + offset + len == RIGHT.rm_offset && + xfs_rmap_is_mergeable(&RIGHT, owner, newext)) + state |= RMAP_RIGHT_CONTIG; + } + + /* check that left + prev + right is not too long */ + if ((state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) == + (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG) && + (unsigned long)LEFT.rm_blockcount + len + + RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX) + state &= ~RMAP_RIGHT_CONTIG; + + trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state, + _RET_IP_); + + /* reset the cursor back to PREV */ + error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, oldext, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + + /* + * Switch out based on the FILLING and CONTIG state bits. + */ + switch (state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) { + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The left and right neighbors are both contiguous with new. + */ + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + RIGHT.rm_startblock, RIGHT.rm_blockcount, + RIGHT.rm_owner, RIGHT.rm_offset, + RIGHT.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + PREV.rm_startblock, PREV.rm_blockcount, + PREV.rm_owner, PREV.rm_offset, + PREV.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW = LEFT; + NEW.rm_blockcount += PREV.rm_blockcount + RIGHT.rm_blockcount; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The left neighbor is contiguous, the right is not. + */ + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + PREV.rm_startblock, PREV.rm_blockcount, + PREV.rm_owner, PREV.rm_offset, + PREV.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW = LEFT; + NEW.rm_blockcount += PREV.rm_blockcount; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The right neighbor is contiguous, the left is not. + */ + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + trace_xfs_rmap_delete(mp, cur->bc_private.a.agno, + RIGHT.rm_startblock, RIGHT.rm_blockcount, + RIGHT.rm_owner, RIGHT.rm_offset, + RIGHT.rm_flags); + error = xfs_btree_delete(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW = PREV; + NEW.rm_blockcount = len + RIGHT.rm_blockcount; + NEW.rm_flags = newext; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING: + /* + * Setting all of a previous oldext extent to newext. + * Neither the left nor right neighbors are contiguous with + * the new one. + */ + NEW = PREV; + NEW.rm_flags = newext; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG: + /* + * Setting the first part of a previous oldext extent to newext. + * The left neighbor is contiguous. + */ + NEW = PREV; + NEW.rm_offset += len; + NEW.rm_startblock += len; + NEW.rm_blockcount -= len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + error = xfs_btree_decrement(cur, 0, &i); + if (error) + goto done; + NEW = LEFT; + NEW.rm_blockcount += len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING: + /* + * Setting the first part of a previous oldext extent to newext. + * The left neighbor is not contiguous. + */ + NEW = PREV; + NEW.rm_startblock += len; + NEW.rm_offset += len; + NEW.rm_blockcount -= len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + NEW.rm_startblock = bno; + NEW.rm_owner = owner; + NEW.rm_offset = offset; + NEW.rm_blockcount = len; + NEW.rm_flags = newext; + cur->bc_rec.r = NEW; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, + len, owner, offset, newext); + error = xfs_btree_insert(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + break; + + case RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting the last part of a previous oldext extent to newext. + * The right neighbor is contiguous with the new allocation. + */ + NEW = PREV; + NEW.rm_blockcount -= len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + error = xfs_btree_increment(cur, 0, &i); + if (error) + goto done; + NEW = RIGHT; + NEW.rm_offset = offset; + NEW.rm_startblock = bno; + NEW.rm_blockcount += len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_RIGHT_FILLING: + /* + * Setting the last part of a previous oldext extent to newext. + * The right neighbor is not contiguous. + */ + NEW = PREV; + NEW.rm_blockcount -= len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + error = xfs_rmap_lookup_eq(cur, bno, len, owner, offset, + oldext, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done); + NEW.rm_startblock = bno; + NEW.rm_owner = owner; + NEW.rm_offset = offset; + NEW.rm_blockcount = len; + NEW.rm_flags = newext; + cur->bc_rec.r = NEW; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, + len, owner, offset, newext); + error = xfs_btree_insert(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + break; + + case 0: + /* + * Setting the middle part of a previous oldext extent to + * newext. Contiguity is impossible here. + * One extent becomes three extents. + */ + /* new right extent - oldext */ + NEW.rm_startblock = bno + len; + NEW.rm_owner = owner; + NEW.rm_offset = new_endoff; + NEW.rm_blockcount = PREV.rm_offset + PREV.rm_blockcount - + new_endoff; + NEW.rm_flags = PREV.rm_flags; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + /* new left extent - oldext */ + NEW = PREV; + NEW.rm_blockcount = offset - PREV.rm_offset; + cur->bc_rec.r = NEW; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, + NEW.rm_startblock, NEW.rm_blockcount, + NEW.rm_owner, NEW.rm_offset, + NEW.rm_flags); + error = xfs_btree_insert(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + /* + * Reset the cursor to the position of the new extent + * we are about to insert as we can't trust it after + * the previous insert. + */ + error = xfs_rmap_lookup_eq(cur, bno, len, owner, offset, + oldext, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done); + /* new middle extent - newext */ + cur->bc_rec.r.rm_flags &= ~XFS_RMAP_UNWRITTEN; + cur->bc_rec.r.rm_flags |= newext; + trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len, + owner, offset, newext); + error = xfs_btree_insert(cur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + break; + + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_LEFT_FILLING | RMAP_RIGHT_CONTIG: + case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG: + case RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_LEFT_CONTIG: + case RMAP_RIGHT_CONTIG: + /* + * These cases are all impossible. + */ + ASSERT(0); + } + + trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +done: + if (error) + trace_xfs_rmap_convert_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +#undef NEW +#undef LEFT +#undef RIGHT +#undef PREV + struct xfs_rmap_query_range_info { xfs_rmap_query_range_fn fn; void *priv; diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index e099eda75b68..551b7e26980c 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -2499,6 +2499,10 @@ DEFINE_AG_ERROR_EVENT(xfs_rmap_unmap_error); DEFINE_RMAP_EVENT(xfs_rmap_map); DEFINE_RMAP_EVENT(xfs_rmap_map_done); DEFINE_AG_ERROR_EVENT(xfs_rmap_map_error); +DEFINE_RMAP_EVENT(xfs_rmap_convert); +DEFINE_RMAP_EVENT(xfs_rmap_convert_done); +DEFINE_AG_ERROR_EVENT(xfs_rmap_convert_error); +DEFINE_AG_ERROR_EVENT(xfs_rmap_convert_state); DECLARE_EVENT_CLASS(xfs_rmapbt_class, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, @@ -2553,6 +2557,7 @@ DEFINE_AG_ERROR_EVENT(xfs_rmap_delete_error); DEFINE_AG_ERROR_EVENT(xfs_rmap_update_error); DEFINE_RMAPBT_EVENT(xfs_rmap_lookup_le_range_result); DEFINE_RMAPBT_EVENT(xfs_rmap_find_right_neighbor_result); +DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_result); #endif /* _TRACE_XFS_H */ -- cgit v1.2.3 From abf09233817b5ea1241db0c187136d3b4738d218 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:03:58 +1000 Subject: xfs: add rmap btree insert and delete helpers Add a couple of helper functions to encapsulate rmap btree insert and delete operations. Add tracepoints to the update function. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_rmap.c | 47 ++++++++++++++++++++++++++++++++++++++++++++++- fs/xfs/libxfs/xfs_rmap.h | 3 +++ 2 files changed, 49 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 743927238bdc..e8ce97fd0608 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -92,13 +92,58 @@ xfs_rmap_update( struct xfs_rmap_irec *irec) { union xfs_btree_rec rec; + int error; + + trace_xfs_rmap_update(cur->bc_mp, cur->bc_private.a.agno, + irec->rm_startblock, irec->rm_blockcount, + irec->rm_owner, irec->rm_offset, irec->rm_flags); rec.rmap.rm_startblock = cpu_to_be32(irec->rm_startblock); rec.rmap.rm_blockcount = cpu_to_be32(irec->rm_blockcount); rec.rmap.rm_owner = cpu_to_be64(irec->rm_owner); rec.rmap.rm_offset = cpu_to_be64( xfs_rmap_irec_offset_pack(irec)); - return xfs_btree_update(cur, &rec); + error = xfs_btree_update(cur, &rec); + if (error) + trace_xfs_rmap_update_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +int +xfs_rmap_insert( + struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, + xfs_extlen_t len, + uint64_t owner, + uint64_t offset, + unsigned int flags) +{ + int i; + int error; + + trace_xfs_rmap_insert(rcur->bc_mp, rcur->bc_private.a.agno, agbno, + len, owner, offset, flags); + + error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(rcur->bc_mp, i == 0, done); + + rcur->bc_rec.r.rm_startblock = agbno; + rcur->bc_rec.r.rm_blockcount = len; + rcur->bc_rec.r.rm_owner = owner; + rcur->bc_rec.r.rm_offset = offset; + rcur->bc_rec.r.rm_flags = flags; + error = xfs_btree_insert(rcur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(rcur->bc_mp, i == 1, done); +done: + if (error) + trace_xfs_rmap_insert_error(rcur->bc_mp, + rcur->bc_private.a.agno, error, _RET_IP_); + return error; } static int diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index 34c811a2dacc..92ac067da7dd 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -148,6 +148,9 @@ int xfs_rmap_lookup_le(struct xfs_btree_cur *cur, xfs_agblock_t bno, int xfs_rmap_lookup_eq(struct xfs_btree_cur *cur, xfs_agblock_t bno, xfs_extlen_t len, uint64_t owner, uint64_t offset, unsigned int flags, int *stat); +int xfs_rmap_insert(struct xfs_btree_cur *rcur, xfs_agblock_t agbno, + xfs_extlen_t len, uint64_t owner, uint64_t offset, + unsigned int flags); int xfs_rmap_get_rec(struct xfs_btree_cur *cur, struct xfs_rmap_irec *irec, int *stat); -- cgit v1.2.3 From 5880f2d78ff17c6ee7c7f6d4071bfd13090c264c Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:04:45 +1000 Subject: xfs: create rmap update intent log items Create rmap update intent/done log items to record redo information in the log. Because we need to roll transactions between updating the bmbt mapping and updating the reverse mapping, we also have to track the status of the metadata updates that will be recorded in the post-roll transactions, just in case we crash before committing the final transaction. This mechanism enables log recovery to finish what was already started. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/Makefile | 1 + fs/xfs/libxfs/xfs_log_format.h | 64 +++++- fs/xfs/libxfs/xfs_rmap.h | 19 ++ fs/xfs/xfs_rmap_item.c | 459 +++++++++++++++++++++++++++++++++++++++++ fs/xfs/xfs_rmap_item.h | 100 +++++++++ fs/xfs/xfs_super.c | 21 ++ 6 files changed, 662 insertions(+), 2 deletions(-) create mode 100644 fs/xfs/xfs_rmap_item.c create mode 100644 fs/xfs/xfs_rmap_item.h (limited to 'fs') diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index b76e937f0729..6c9039384d34 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -103,6 +103,7 @@ xfs-y += xfs_log.o \ xfs_extfree_item.o \ xfs_icreate_item.o \ xfs_inode_item.o \ + xfs_rmap_item.o \ xfs_log_recover.o \ xfs_trans_ail.o \ xfs_trans_buf.o \ diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h index e8f49c029ff0..a8d794d1ae9a 100644 --- a/fs/xfs/libxfs/xfs_log_format.h +++ b/fs/xfs/libxfs/xfs_log_format.h @@ -110,7 +110,9 @@ static inline uint xlog_get_cycle(char *ptr) #define XLOG_REG_TYPE_COMMIT 18 #define XLOG_REG_TYPE_TRANSHDR 19 #define XLOG_REG_TYPE_ICREATE 20 -#define XLOG_REG_TYPE_MAX 20 +#define XLOG_REG_TYPE_RUI_FORMAT 21 +#define XLOG_REG_TYPE_RUD_FORMAT 22 +#define XLOG_REG_TYPE_MAX 22 /* * Flags to log operation header @@ -227,6 +229,8 @@ typedef struct xfs_trans_header { #define XFS_LI_DQUOT 0x123d #define XFS_LI_QUOTAOFF 0x123e #define XFS_LI_ICREATE 0x123f +#define XFS_LI_RUI 0x1240 /* rmap update intent */ +#define XFS_LI_RUD 0x1241 #define XFS_LI_TYPE_DESC \ { XFS_LI_EFI, "XFS_LI_EFI" }, \ @@ -236,7 +240,9 @@ typedef struct xfs_trans_header { { XFS_LI_BUF, "XFS_LI_BUF" }, \ { XFS_LI_DQUOT, "XFS_LI_DQUOT" }, \ { XFS_LI_QUOTAOFF, "XFS_LI_QUOTAOFF" }, \ - { XFS_LI_ICREATE, "XFS_LI_ICREATE" } + { XFS_LI_ICREATE, "XFS_LI_ICREATE" }, \ + { XFS_LI_RUI, "XFS_LI_RUI" }, \ + { XFS_LI_RUD, "XFS_LI_RUD" } /* * Inode Log Item Format definitions. @@ -603,6 +609,60 @@ typedef struct xfs_efd_log_format_64 { xfs_extent_64_t efd_extents[1]; /* array of extents freed */ } xfs_efd_log_format_64_t; +/* + * RUI/RUD (reverse mapping) log format definitions + */ +struct xfs_map_extent { + __uint64_t me_owner; + __uint64_t me_startblock; + __uint64_t me_startoff; + __uint32_t me_len; + __uint32_t me_flags; +}; + +/* rmap me_flags: upper bits are flags, lower byte is type code */ +#define XFS_RMAP_EXTENT_MAP 1 +#define XFS_RMAP_EXTENT_UNMAP 3 +#define XFS_RMAP_EXTENT_CONVERT 5 +#define XFS_RMAP_EXTENT_ALLOC 7 +#define XFS_RMAP_EXTENT_FREE 8 +#define XFS_RMAP_EXTENT_TYPE_MASK 0xFF + +#define XFS_RMAP_EXTENT_ATTR_FORK (1U << 31) +#define XFS_RMAP_EXTENT_BMBT_BLOCK (1U << 30) +#define XFS_RMAP_EXTENT_UNWRITTEN (1U << 29) + +#define XFS_RMAP_EXTENT_FLAGS (XFS_RMAP_EXTENT_TYPE_MASK | \ + XFS_RMAP_EXTENT_ATTR_FORK | \ + XFS_RMAP_EXTENT_BMBT_BLOCK | \ + XFS_RMAP_EXTENT_UNWRITTEN) + +/* + * This is the structure used to lay out an rui log item in the + * log. The rui_extents field is a variable size array whose + * size is given by rui_nextents. + */ +struct xfs_rui_log_format { + __uint16_t rui_type; /* rui log item type */ + __uint16_t rui_size; /* size of this item */ + __uint32_t rui_nextents; /* # extents to free */ + __uint64_t rui_id; /* rui identifier */ + struct xfs_map_extent rui_extents[1]; /* array of extents to rmap */ +}; + +/* + * This is the structure used to lay out an rud log item in the + * log. The rud_extents array is a variable size array whose + * size is given by rud_nextents; + */ +struct xfs_rud_log_format { + __uint16_t rud_type; /* rud log item type */ + __uint16_t rud_size; /* size of this item */ + __uint32_t rud_nextents; /* # of extents freed */ + __uint64_t rud_rui_id; /* id of corresponding rui */ + struct xfs_map_extent rud_extents[1]; /* array of extents rmapped */ +}; + /* * Dquot Log format definitions. * diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index 92ac067da7dd..c4b1133e2ff9 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -163,4 +163,23 @@ int xfs_rmap_query_range(struct xfs_btree_cur *cur, struct xfs_rmap_irec *low_rec, struct xfs_rmap_irec *high_rec, xfs_rmap_query_range_fn fn, void *priv); +enum xfs_rmap_intent_type { + XFS_RMAP_MAP, + XFS_RMAP_MAP_SHARED, + XFS_RMAP_UNMAP, + XFS_RMAP_UNMAP_SHARED, + XFS_RMAP_CONVERT, + XFS_RMAP_CONVERT_SHARED, + XFS_RMAP_ALLOC, + XFS_RMAP_FREE, +}; + +struct xfs_rmap_intent { + struct list_head ri_list; + enum xfs_rmap_intent_type ri_type; + __uint64_t ri_owner; + int ri_whichfork; + struct xfs_bmbt_irec ri_bmap; +}; + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c new file mode 100644 index 000000000000..5398b8478f02 --- /dev/null +++ b/fs/xfs/xfs_rmap_item.c @@ -0,0 +1,459 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_buf_item.h" +#include "xfs_rmap_item.h" +#include "xfs_log.h" + + +kmem_zone_t *xfs_rui_zone; +kmem_zone_t *xfs_rud_zone; + +static inline struct xfs_rui_log_item *RUI_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_rui_log_item, rui_item); +} + +void +xfs_rui_item_free( + struct xfs_rui_log_item *ruip) +{ + if (ruip->rui_format.rui_nextents > XFS_RUI_MAX_FAST_EXTENTS) + kmem_free(ruip); + else + kmem_zone_free(xfs_rui_zone, ruip); +} + +/* + * This returns the number of iovecs needed to log the given rui item. + * We only need 1 iovec for an rui item. It just logs the rui_log_format + * structure. + */ +static inline int +xfs_rui_item_sizeof( + struct xfs_rui_log_item *ruip) +{ + return sizeof(struct xfs_rui_log_format) + + (ruip->rui_format.rui_nextents - 1) * + sizeof(struct xfs_map_extent); +} + +STATIC void +xfs_rui_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + *nvecs += 1; + *nbytes += xfs_rui_item_sizeof(RUI_ITEM(lip)); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given rui log item. We use only 1 iovec, and we point that + * at the rui_log_format structure embedded in the rui item. + * It is at this point that we assert that all of the extent + * slots in the rui item have been filled. + */ +STATIC void +xfs_rui_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_rui_log_item *ruip = RUI_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + ASSERT(atomic_read(&ruip->rui_next_extent) == + ruip->rui_format.rui_nextents); + + ruip->rui_format.rui_type = XFS_LI_RUI; + ruip->rui_format.rui_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUI_FORMAT, &ruip->rui_format, + xfs_rui_item_sizeof(ruip)); +} + +/* + * Pinning has no meaning for an rui item, so just return. + */ +STATIC void +xfs_rui_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * The unpin operation is the last place an RUI is manipulated in the log. It is + * either inserted in the AIL or aborted in the event of a log I/O error. In + * either case, the RUI transaction has been successfully committed to make it + * this far. Therefore, we expect whoever committed the RUI to either construct + * and commit the RUD or drop the RUD's reference in the event of error. Simply + * drop the log's RUI reference now that the log is done with it. + */ +STATIC void +xfs_rui_item_unpin( + struct xfs_log_item *lip, + int remove) +{ + struct xfs_rui_log_item *ruip = RUI_ITEM(lip); + + xfs_rui_release(ruip); +} + +/* + * RUI items have no locking or pushing. However, since RUIs are pulled from + * the AIL when their corresponding RUDs are committed to disk, their situation + * is very similar to being pinned. Return XFS_ITEM_PINNED so that the caller + * will eventually flush the log. This should help in getting the RUI out of + * the AIL. + */ +STATIC uint +xfs_rui_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The RUI has been either committed or aborted if the transaction has been + * cancelled. If the transaction was cancelled, an RUD isn't going to be + * constructed and thus we free the RUI here directly. + */ +STATIC void +xfs_rui_item_unlock( + struct xfs_log_item *lip) +{ + if (lip->li_flags & XFS_LI_ABORTED) + xfs_rui_item_free(RUI_ITEM(lip)); +} + +/* + * The RUI is logged only once and cannot be moved in the log, so simply return + * the lsn at which it's been logged. + */ +STATIC xfs_lsn_t +xfs_rui_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + return lsn; +} + +/* + * The RUI dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_rui_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all rui log items. + */ +static const struct xfs_item_ops xfs_rui_item_ops = { + .iop_size = xfs_rui_item_size, + .iop_format = xfs_rui_item_format, + .iop_pin = xfs_rui_item_pin, + .iop_unpin = xfs_rui_item_unpin, + .iop_unlock = xfs_rui_item_unlock, + .iop_committed = xfs_rui_item_committed, + .iop_push = xfs_rui_item_push, + .iop_committing = xfs_rui_item_committing, +}; + +/* + * Allocate and initialize an rui item with the given number of extents. + */ +struct xfs_rui_log_item * +xfs_rui_init( + struct xfs_mount *mp, + uint nextents) + +{ + struct xfs_rui_log_item *ruip; + uint size; + + ASSERT(nextents > 0); + if (nextents > XFS_RUI_MAX_FAST_EXTENTS) { + size = (uint)(sizeof(struct xfs_rui_log_item) + + ((nextents - 1) * sizeof(struct xfs_map_extent))); + ruip = kmem_zalloc(size, KM_SLEEP); + } else { + ruip = kmem_zone_zalloc(xfs_rui_zone, KM_SLEEP); + } + + xfs_log_item_init(mp, &ruip->rui_item, XFS_LI_RUI, &xfs_rui_item_ops); + ruip->rui_format.rui_nextents = nextents; + ruip->rui_format.rui_id = (uintptr_t)(void *)ruip; + atomic_set(&ruip->rui_next_extent, 0); + atomic_set(&ruip->rui_refcount, 2); + + return ruip; +} + +/* + * Copy an RUI format buffer from the given buf, and into the destination + * RUI format structure. The RUI/RUD items were designed not to need any + * special alignment handling. + */ +int +xfs_rui_copy_format( + struct xfs_log_iovec *buf, + struct xfs_rui_log_format *dst_rui_fmt) +{ + struct xfs_rui_log_format *src_rui_fmt; + uint len; + + src_rui_fmt = buf->i_addr; + len = sizeof(struct xfs_rui_log_format) + + (src_rui_fmt->rui_nextents - 1) * + sizeof(struct xfs_map_extent); + + if (buf->i_len != len) + return -EFSCORRUPTED; + + memcpy((char *)dst_rui_fmt, (char *)src_rui_fmt, len); + return 0; +} + +/* + * Freeing the RUI requires that we remove it from the AIL if it has already + * been placed there. However, the RUI may not yet have been placed in the AIL + * when called by xfs_rui_release() from RUD processing due to the ordering of + * committed vs unpin operations in bulk insert operations. Hence the reference + * count to ensure only the last caller frees the RUI. + */ +void +xfs_rui_release( + struct xfs_rui_log_item *ruip) +{ + if (atomic_dec_and_test(&ruip->rui_refcount)) { + xfs_trans_ail_remove(&ruip->rui_item, SHUTDOWN_LOG_IO_ERROR); + xfs_rui_item_free(ruip); + } +} + +static inline struct xfs_rud_log_item *RUD_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_rud_log_item, rud_item); +} + +STATIC void +xfs_rud_item_free(struct xfs_rud_log_item *rudp) +{ + if (rudp->rud_format.rud_nextents > XFS_RUD_MAX_FAST_EXTENTS) + kmem_free(rudp); + else + kmem_zone_free(xfs_rud_zone, rudp); +} + +/* + * This returns the number of iovecs needed to log the given rud item. + * We only need 1 iovec for an rud item. It just logs the rud_log_format + * structure. + */ +static inline int +xfs_rud_item_sizeof( + struct xfs_rud_log_item *rudp) +{ + return sizeof(struct xfs_rud_log_format) + + (rudp->rud_format.rud_nextents - 1) * + sizeof(struct xfs_map_extent); +} + +STATIC void +xfs_rud_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + *nvecs += 1; + *nbytes += xfs_rud_item_sizeof(RUD_ITEM(lip)); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given rud log item. We use only 1 iovec, and we point that + * at the rud_log_format structure embedded in the rud item. + * It is at this point that we assert that all of the extent + * slots in the rud item have been filled. + */ +STATIC void +xfs_rud_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_rud_log_item *rudp = RUD_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + ASSERT(rudp->rud_next_extent == rudp->rud_format.rud_nextents); + + rudp->rud_format.rud_type = XFS_LI_RUD; + rudp->rud_format.rud_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUD_FORMAT, &rudp->rud_format, + xfs_rud_item_sizeof(rudp)); +} + +/* + * Pinning has no meaning for an rud item, so just return. + */ +STATIC void +xfs_rud_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * Since pinning has no meaning for an rud item, unpinning does + * not either. + */ +STATIC void +xfs_rud_item_unpin( + struct xfs_log_item *lip, + int remove) +{ +} + +/* + * There isn't much you can do to push on an rud item. It is simply stuck + * waiting for the log to be flushed to disk. + */ +STATIC uint +xfs_rud_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The RUD is either committed or aborted if the transaction is cancelled. If + * the transaction is cancelled, drop our reference to the RUI and free the + * RUD. + */ +STATIC void +xfs_rud_item_unlock( + struct xfs_log_item *lip) +{ + struct xfs_rud_log_item *rudp = RUD_ITEM(lip); + + if (lip->li_flags & XFS_LI_ABORTED) { + xfs_rui_release(rudp->rud_ruip); + xfs_rud_item_free(rudp); + } +} + +/* + * When the rud item is committed to disk, all we need to do is delete our + * reference to our partner rui item and then free ourselves. Since we're + * freeing ourselves we must return -1 to keep the transaction code from + * further referencing this item. + */ +STATIC xfs_lsn_t +xfs_rud_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + struct xfs_rud_log_item *rudp = RUD_ITEM(lip); + + /* + * Drop the RUI reference regardless of whether the RUD has been + * aborted. Once the RUD transaction is constructed, it is the sole + * responsibility of the RUD to release the RUI (even if the RUI is + * aborted due to log I/O error). + */ + xfs_rui_release(rudp->rud_ruip); + xfs_rud_item_free(rudp); + + return (xfs_lsn_t)-1; +} + +/* + * The RUD dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_rud_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all rud log items. + */ +static const struct xfs_item_ops xfs_rud_item_ops = { + .iop_size = xfs_rud_item_size, + .iop_format = xfs_rud_item_format, + .iop_pin = xfs_rud_item_pin, + .iop_unpin = xfs_rud_item_unpin, + .iop_unlock = xfs_rud_item_unlock, + .iop_committed = xfs_rud_item_committed, + .iop_push = xfs_rud_item_push, + .iop_committing = xfs_rud_item_committing, +}; + +/* + * Allocate and initialize an rud item with the given number of extents. + */ +struct xfs_rud_log_item * +xfs_rud_init( + struct xfs_mount *mp, + struct xfs_rui_log_item *ruip, + uint nextents) + +{ + struct xfs_rud_log_item *rudp; + uint size; + + ASSERT(nextents > 0); + if (nextents > XFS_RUD_MAX_FAST_EXTENTS) { + size = (uint)(sizeof(struct xfs_rud_log_item) + + ((nextents - 1) * sizeof(struct xfs_map_extent))); + rudp = kmem_zalloc(size, KM_SLEEP); + } else { + rudp = kmem_zone_zalloc(xfs_rud_zone, KM_SLEEP); + } + + xfs_log_item_init(mp, &rudp->rud_item, XFS_LI_RUD, &xfs_rud_item_ops); + rudp->rud_ruip = ruip; + rudp->rud_format.rud_nextents = nextents; + rudp->rud_format.rud_rui_id = ruip->rui_format.rui_id; + + return rudp; +} diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h new file mode 100644 index 000000000000..bd36ab50c0fe --- /dev/null +++ b/fs/xfs/xfs_rmap_item.h @@ -0,0 +1,100 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_RMAP_ITEM_H__ +#define __XFS_RMAP_ITEM_H__ + +/* + * There are (currently) three pairs of rmap btree redo item types: map, unmap, + * and convert. The common abbreviations for these are RUI (rmap update + * intent) and RUD (rmap update done). The redo item type is encoded in the + * flags field of each xfs_map_extent. + * + * *I items should be recorded in the *first* of a series of rolled + * transactions, and the *D items should be recorded in the same transaction + * that records the associated rmapbt updates. Typically, the first + * transaction will record a bmbt update, followed by some number of + * transactions containing rmapbt updates, and finally transactions with any + * bnobt/cntbt updates. + * + * Should the system crash after the commit of the first transaction but + * before the commit of the final transaction in a series, log recovery will + * use the redo information recorded by the intent items to replay the + * (rmapbt/bnobt/cntbt) metadata updates in the non-first transaction. + */ + +/* kernel only RUI/RUD definitions */ + +struct xfs_mount; +struct kmem_zone; + +/* + * Max number of extents in fast allocation path. + */ +#define XFS_RUI_MAX_FAST_EXTENTS 16 + +/* + * Define RUI flag bits. Manipulated by set/clear/test_bit operators. + */ +#define XFS_RUI_RECOVERED 1 + +/* + * This is the "rmap update intent" log item. It is used to log the fact that + * some reverse mappings need to change. It is used in conjunction with the + * "rmap update done" log item described below. + * + * These log items follow the same rules as struct xfs_efi_log_item; see the + * comments about that structure (in xfs_extfree_item.h) for more details. + */ +struct xfs_rui_log_item { + struct xfs_log_item rui_item; + atomic_t rui_refcount; + atomic_t rui_next_extent; + unsigned long rui_flags; /* misc flags */ + struct xfs_rui_log_format rui_format; +}; + +/* + * This is the "rmap update done" log item. It is used to log the fact that + * some rmapbt updates mentioned in an earlier rui item have been performed. + */ +struct xfs_rud_log_item { + struct xfs_log_item rud_item; + struct xfs_rui_log_item *rud_ruip; + uint rud_next_extent; + struct xfs_rud_log_format rud_format; +}; + +/* + * Max number of extents in fast allocation path. + */ +#define XFS_RUD_MAX_FAST_EXTENTS 16 + +extern struct kmem_zone *xfs_rui_zone; +extern struct kmem_zone *xfs_rud_zone; + +struct xfs_rui_log_item *xfs_rui_init(struct xfs_mount *, uint); +struct xfs_rud_log_item *xfs_rud_init(struct xfs_mount *, + struct xfs_rui_log_item *, uint); +int xfs_rui_copy_format(struct xfs_log_iovec *buf, + struct xfs_rui_log_format *dst_rui_fmt); +void xfs_rui_item_free(struct xfs_rui_log_item *); +void xfs_rui_release(struct xfs_rui_log_item *); + +#endif /* __XFS_RMAP_ITEM_H__ */ diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 449cadf149f9..654a0924b3d3 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -46,6 +46,7 @@ #include "xfs_quota.h" #include "xfs_sysfs.h" #include "xfs_ondisk.h" +#include "xfs_rmap_item.h" #include #include @@ -1765,8 +1766,26 @@ xfs_init_zones(void) if (!xfs_icreate_zone) goto out_destroy_ili_zone; + xfs_rud_zone = kmem_zone_init((sizeof(struct xfs_rud_log_item) + + ((XFS_RUD_MAX_FAST_EXTENTS - 1) * + sizeof(struct xfs_map_extent))), + "xfs_rud_item"); + if (!xfs_rud_zone) + goto out_destroy_icreate_zone; + + xfs_rui_zone = kmem_zone_init((sizeof(struct xfs_rui_log_item) + + ((XFS_RUI_MAX_FAST_EXTENTS - 1) * + sizeof(struct xfs_map_extent))), + "xfs_rui_item"); + if (!xfs_rui_zone) + goto out_destroy_rud_zone; + return 0; + out_destroy_rud_zone: + kmem_zone_destroy(xfs_rud_zone); + out_destroy_icreate_zone: + kmem_zone_destroy(xfs_icreate_zone); out_destroy_ili_zone: kmem_zone_destroy(xfs_ili_zone); out_destroy_inode_zone: @@ -1805,6 +1824,8 @@ xfs_destroy_zones(void) * destroy caches. */ rcu_barrier(); + kmem_zone_destroy(xfs_rui_zone); + kmem_zone_destroy(xfs_rud_zone); kmem_zone_destroy(xfs_icreate_zone); kmem_zone_destroy(xfs_ili_zone); kmem_zone_destroy(xfs_inode_zone); -- cgit v1.2.3 From 9e88b5d86703bb836d33dcd0c70eb59e049c80c7 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:09:48 +1000 Subject: xfs: log rmap intent items Provide a mechanism for higher levels to create RUI/RUD items, submit them to the log, and a stub function to deal with recovered RUI items. These parts will be connected to the rmapbt in a later patch. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/Makefile | 1 + fs/xfs/xfs_log_recover.c | 154 +++++++++++++++++++++++++++++++++++ fs/xfs/xfs_rmap_item.c | 60 ++++++++++++++ fs/xfs/xfs_rmap_item.h | 1 + fs/xfs/xfs_trans.h | 17 ++++ fs/xfs/xfs_trans_rmap.c | 205 +++++++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 438 insertions(+) create mode 100644 fs/xfs/xfs_trans_rmap.c (limited to 'fs') diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index 6c9039384d34..3124210d502b 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -109,6 +109,7 @@ xfs-y += xfs_log.o \ xfs_trans_buf.o \ xfs_trans_extfree.o \ xfs_trans_inode.o \ + xfs_trans_rmap.o \ # optional features xfs-$(CONFIG_XFS_QUOTA) += xfs_dquot.o \ diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index f64d4d06cd95..365aef655edb 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -43,6 +43,7 @@ #include "xfs_bmap_btree.h" #include "xfs_error.h" #include "xfs_dir2.h" +#include "xfs_rmap_item.h" #define BLK_AVG(blk1, blk2) ((blk1+blk2) >> 1) @@ -1911,6 +1912,8 @@ xlog_recover_reorder_trans( case XFS_LI_QUOTAOFF: case XFS_LI_EFD: case XFS_LI_EFI: + case XFS_LI_RUI: + case XFS_LI_RUD: trace_xfs_log_recover_item_reorder_tail(log, trans, item, pass); list_move_tail(&item->ri_list, &inode_list); @@ -3414,6 +3417,101 @@ xlog_recover_efd_pass2( return 0; } +/* + * This routine is called to create an in-core extent rmap update + * item from the rui format structure which was logged on disk. + * It allocates an in-core rui, copies the extents from the format + * structure into it, and adds the rui to the AIL with the given + * LSN. + */ +STATIC int +xlog_recover_rui_pass2( + struct xlog *log, + struct xlog_recover_item *item, + xfs_lsn_t lsn) +{ + int error; + struct xfs_mount *mp = log->l_mp; + struct xfs_rui_log_item *ruip; + struct xfs_rui_log_format *rui_formatp; + + rui_formatp = item->ri_buf[0].i_addr; + + ruip = xfs_rui_init(mp, rui_formatp->rui_nextents); + error = xfs_rui_copy_format(&item->ri_buf[0], &ruip->rui_format); + if (error) { + xfs_rui_item_free(ruip); + return error; + } + atomic_set(&ruip->rui_next_extent, rui_formatp->rui_nextents); + + spin_lock(&log->l_ailp->xa_lock); + /* + * The RUI has two references. One for the RUD and one for RUI to ensure + * it makes it into the AIL. Insert the RUI into the AIL directly and + * drop the RUI reference. Note that xfs_trans_ail_update() drops the + * AIL lock. + */ + xfs_trans_ail_update(log->l_ailp, &ruip->rui_item, lsn); + xfs_rui_release(ruip); + return 0; +} + + +/* + * This routine is called when an RUD format structure is found in a committed + * transaction in the log. Its purpose is to cancel the corresponding RUI if it + * was still in the log. To do this it searches the AIL for the RUI with an id + * equal to that in the RUD format structure. If we find it we drop the RUD + * reference, which removes the RUI from the AIL and frees it. + */ +STATIC int +xlog_recover_rud_pass2( + struct xlog *log, + struct xlog_recover_item *item) +{ + struct xfs_rud_log_format *rud_formatp; + struct xfs_rui_log_item *ruip = NULL; + struct xfs_log_item *lip; + __uint64_t rui_id; + struct xfs_ail_cursor cur; + struct xfs_ail *ailp = log->l_ailp; + + rud_formatp = item->ri_buf[0].i_addr; + ASSERT(item->ri_buf[0].i_len == (sizeof(struct xfs_rud_log_format) + + ((rud_formatp->rud_nextents - 1) * + sizeof(struct xfs_map_extent)))); + rui_id = rud_formatp->rud_rui_id; + + /* + * Search for the RUI with the id in the RUD format structure in the + * AIL. + */ + spin_lock(&ailp->xa_lock); + lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); + while (lip != NULL) { + if (lip->li_type == XFS_LI_RUI) { + ruip = (struct xfs_rui_log_item *)lip; + if (ruip->rui_format.rui_id == rui_id) { + /* + * Drop the RUD reference to the RUI. This + * removes the RUI from the AIL and frees it. + */ + spin_unlock(&ailp->xa_lock); + xfs_rui_release(ruip); + spin_lock(&ailp->xa_lock); + break; + } + } + lip = xfs_trans_ail_cursor_next(ailp, &cur); + } + + xfs_trans_ail_cursor_done(&cur); + spin_unlock(&ailp->xa_lock); + + return 0; +} + /* * This routine is called when an inode create format structure is found in a * committed transaction in the log. It's purpose is to initialise the inodes @@ -3639,6 +3737,8 @@ xlog_recover_ra_pass2( case XFS_LI_EFI: case XFS_LI_EFD: case XFS_LI_QUOTAOFF: + case XFS_LI_RUI: + case XFS_LI_RUD: default: break; } @@ -3662,6 +3762,8 @@ xlog_recover_commit_pass1( case XFS_LI_EFD: case XFS_LI_DQUOT: case XFS_LI_ICREATE: + case XFS_LI_RUI: + case XFS_LI_RUD: /* nothing to do in pass 1 */ return 0; default: @@ -3692,6 +3794,10 @@ xlog_recover_commit_pass2( return xlog_recover_efi_pass2(log, item, trans->r_lsn); case XFS_LI_EFD: return xlog_recover_efd_pass2(log, item); + case XFS_LI_RUI: + return xlog_recover_rui_pass2(log, item, trans->r_lsn); + case XFS_LI_RUD: + return xlog_recover_rud_pass2(log, item); case XFS_LI_DQUOT: return xlog_recover_dquot_pass2(log, buffer_list, item, trans->r_lsn); @@ -4204,11 +4310,52 @@ xlog_recover_cancel_efi( spin_lock(&ailp->xa_lock); } +/* Recover the RUI if necessary. */ +STATIC int +xlog_recover_process_rui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_rui_log_item *ruip; + int error; + + /* + * Skip RUIs that we've already processed. + */ + ruip = container_of(lip, struct xfs_rui_log_item, rui_item); + if (test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags)) + return 0; + + spin_unlock(&ailp->xa_lock); + error = xfs_rui_recover(mp, ruip); + spin_lock(&ailp->xa_lock); + + return error; +} + +/* Release the RUI since we're cancelling everything. */ +STATIC void +xlog_recover_cancel_rui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_rui_log_item *ruip; + + ruip = container_of(lip, struct xfs_rui_log_item, rui_item); + + spin_unlock(&ailp->xa_lock); + xfs_rui_release(ruip); + spin_lock(&ailp->xa_lock); +} + /* Is this log item a deferred action intent? */ static inline bool xlog_item_is_intent(struct xfs_log_item *lip) { switch (lip->li_type) { case XFS_LI_EFI: + case XFS_LI_RUI: return true; default: return false; @@ -4269,6 +4416,9 @@ xlog_recover_process_intents( case XFS_LI_EFI: error = xlog_recover_process_efi(log->l_mp, ailp, lip); break; + case XFS_LI_RUI: + error = xlog_recover_process_rui(log->l_mp, ailp, lip); + break; } if (error) goto out; @@ -4313,6 +4463,9 @@ xlog_recover_cancel_intents( case XFS_LI_EFI: xlog_recover_cancel_efi(log->l_mp, ailp, lip); break; + case XFS_LI_RUI: + xlog_recover_cancel_rui(log->l_mp, ailp, lip); + break; } lip = xfs_trans_ail_cursor_next(ailp, &cur); @@ -5130,6 +5283,7 @@ xlog_recover_finish( xfs_alert(log->l_mp, "Failed to recover intents"); return error; } + /* * Sync the log to get all the intents out of the AIL. * This isn't absolutely necessary, but it helps in diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 5398b8478f02..fecd1e4d688d 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -22,6 +22,7 @@ #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_bit.h" #include "xfs_mount.h" #include "xfs_trans.h" #include "xfs_trans_priv.h" @@ -457,3 +458,62 @@ xfs_rud_init( return rudp; } + +/* + * Process an rmap update intent item that was recovered from the log. + * We need to update the rmapbt. + */ +int +xfs_rui_recover( + struct xfs_mount *mp, + struct xfs_rui_log_item *ruip) +{ + int i; + int error = 0; + struct xfs_map_extent *rmap; + xfs_fsblock_t startblock_fsb; + bool op_ok; + + ASSERT(!test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags)); + + /* + * First check the validity of the extents described by the + * RUI. If any are bad, then assume that all are bad and + * just toss the RUI. + */ + for (i = 0; i < ruip->rui_format.rui_nextents; i++) { + rmap = &(ruip->rui_format.rui_extents[i]); + startblock_fsb = XFS_BB_TO_FSB(mp, + XFS_FSB_TO_DADDR(mp, rmap->me_startblock)); + switch (rmap->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) { + case XFS_RMAP_EXTENT_MAP: + case XFS_RMAP_EXTENT_UNMAP: + case XFS_RMAP_EXTENT_CONVERT: + case XFS_RMAP_EXTENT_ALLOC: + case XFS_RMAP_EXTENT_FREE: + op_ok = true; + break; + default: + op_ok = false; + break; + } + if (!op_ok || (startblock_fsb == 0) || + (rmap->me_len == 0) || + (startblock_fsb >= mp->m_sb.sb_dblocks) || + (rmap->me_len >= mp->m_sb.sb_agblocks) || + (rmap->me_flags & ~XFS_RMAP_EXTENT_FLAGS)) { + /* + * This will pull the RUI from the AIL and + * free the memory associated with it. + */ + set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags); + xfs_rui_release(ruip); + return -EIO; + } + } + + /* XXX: do nothing for now */ + set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags); + xfs_rui_release(ruip); + return error; +} diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h index bd36ab50c0fe..59ef3ecaa493 100644 --- a/fs/xfs/xfs_rmap_item.h +++ b/fs/xfs/xfs_rmap_item.h @@ -96,5 +96,6 @@ int xfs_rui_copy_format(struct xfs_log_iovec *buf, struct xfs_rui_log_format *dst_rui_fmt); void xfs_rui_item_free(struct xfs_rui_log_item *); void xfs_rui_release(struct xfs_rui_log_item *); +int xfs_rui_recover(struct xfs_mount *mp, struct xfs_rui_log_item *ruip); #endif /* __XFS_RMAP_ITEM_H__ */ diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 9bc368bb5eb4..02265817f016 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -233,4 +233,21 @@ void xfs_trans_buf_copy_type(struct xfs_buf *dst_bp, extern kmem_zone_t *xfs_trans_zone; extern kmem_zone_t *xfs_log_item_desc_zone; +enum xfs_rmap_intent_type; + +struct xfs_rui_log_item *xfs_trans_get_rui(struct xfs_trans *tp, uint nextents); +void xfs_trans_log_start_rmap_update(struct xfs_trans *tp, + struct xfs_rui_log_item *ruip, enum xfs_rmap_intent_type type, + __uint64_t owner, int whichfork, xfs_fileoff_t startoff, + xfs_fsblock_t startblock, xfs_filblks_t blockcount, + xfs_exntst_t state); + +struct xfs_rud_log_item *xfs_trans_get_rud(struct xfs_trans *tp, + struct xfs_rui_log_item *ruip, uint nextents); +int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp, + struct xfs_rud_log_item *rudp, enum xfs_rmap_intent_type type, + __uint64_t owner, int whichfork, xfs_fileoff_t startoff, + xfs_fsblock_t startblock, xfs_filblks_t blockcount, + xfs_exntst_t state); + #endif /* __XFS_TRANS_H__ */ diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c new file mode 100644 index 000000000000..e3a5172ad0c0 --- /dev/null +++ b/fs/xfs/xfs_trans_rmap.c @@ -0,0 +1,205 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_rmap_item.h" +#include "xfs_alloc.h" +#include "xfs_rmap.h" + +/* + * This routine is called to allocate an "rmap update intent" + * log item that will hold nextents worth of extents. The + * caller must use all nextents extents, because we are not + * flexible about this at all. + */ +struct xfs_rui_log_item * +xfs_trans_get_rui( + struct xfs_trans *tp, + uint nextents) +{ + struct xfs_rui_log_item *ruip; + + ASSERT(tp != NULL); + ASSERT(nextents > 0); + + ruip = xfs_rui_init(tp->t_mountp, nextents); + ASSERT(ruip != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &ruip->rui_item); + return ruip; +} + +/* Set the map extent flags for this reverse mapping. */ +static void +xfs_trans_set_rmap_flags( + struct xfs_map_extent *rmap, + enum xfs_rmap_intent_type type, + int whichfork, + xfs_exntst_t state) +{ + rmap->me_flags = 0; + if (state == XFS_EXT_UNWRITTEN) + rmap->me_flags |= XFS_RMAP_EXTENT_UNWRITTEN; + if (whichfork == XFS_ATTR_FORK) + rmap->me_flags |= XFS_RMAP_EXTENT_ATTR_FORK; + switch (type) { + case XFS_RMAP_MAP: + rmap->me_flags |= XFS_RMAP_EXTENT_MAP; + break; + case XFS_RMAP_UNMAP: + rmap->me_flags |= XFS_RMAP_EXTENT_UNMAP; + break; + case XFS_RMAP_CONVERT: + rmap->me_flags |= XFS_RMAP_EXTENT_CONVERT; + break; + case XFS_RMAP_ALLOC: + rmap->me_flags |= XFS_RMAP_EXTENT_ALLOC; + break; + case XFS_RMAP_FREE: + rmap->me_flags |= XFS_RMAP_EXTENT_FREE; + break; + default: + ASSERT(0); + } +} + +/* + * This routine is called to indicate that the described reverse + * mapping is to be logged as needing to be updated. It should be + * called once for each mapping. + */ +void +xfs_trans_log_start_rmap_update( + struct xfs_trans *tp, + struct xfs_rui_log_item *ruip, + enum xfs_rmap_intent_type type, + __uint64_t owner, + int whichfork, + xfs_fileoff_t startoff, + xfs_fsblock_t startblock, + xfs_filblks_t blockcount, + xfs_exntst_t state) +{ + uint next_extent; + struct xfs_map_extent *rmap; + + tp->t_flags |= XFS_TRANS_DIRTY; + ruip->rui_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + /* + * atomic_inc_return gives us the value after the increment; + * we want to use it as an array index so we need to subtract 1 from + * it. + */ + next_extent = atomic_inc_return(&ruip->rui_next_extent) - 1; + ASSERT(next_extent < ruip->rui_format.rui_nextents); + rmap = &(ruip->rui_format.rui_extents[next_extent]); + rmap->me_owner = owner; + rmap->me_startblock = startblock; + rmap->me_startoff = startoff; + rmap->me_len = blockcount; + xfs_trans_set_rmap_flags(rmap, type, whichfork, state); +} + + +/* + * This routine is called to allocate an "rmap update done" + * log item that will hold nextents worth of extents. The + * caller must use all nextents extents, because we are not + * flexible about this at all. + */ +struct xfs_rud_log_item * +xfs_trans_get_rud( + struct xfs_trans *tp, + struct xfs_rui_log_item *ruip, + uint nextents) +{ + struct xfs_rud_log_item *rudp; + + ASSERT(tp != NULL); + ASSERT(nextents > 0); + + rudp = xfs_rud_init(tp->t_mountp, ruip, nextents); + ASSERT(rudp != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &rudp->rud_item); + return rudp; +} + +/* + * Finish an rmap update and log it to the RUD. Note that the transaction is + * marked dirty regardless of whether the rmap update succeeds or fails to + * support the RUI/RUD lifecycle rules. + */ +int +xfs_trans_log_finish_rmap_update( + struct xfs_trans *tp, + struct xfs_rud_log_item *rudp, + enum xfs_rmap_intent_type type, + __uint64_t owner, + int whichfork, + xfs_fileoff_t startoff, + xfs_fsblock_t startblock, + xfs_filblks_t blockcount, + xfs_exntst_t state) +{ + uint next_extent; + struct xfs_map_extent *rmap; + int error; + + /* XXX: actually finish the rmap update here */ + error = -EFSCORRUPTED; + + /* + * Mark the transaction dirty, even on error. This ensures the + * transaction is aborted, which: + * + * 1.) releases the RUI and frees the RUD + * 2.) shuts down the filesystem + */ + tp->t_flags |= XFS_TRANS_DIRTY; + rudp->rud_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + next_extent = rudp->rud_next_extent; + ASSERT(next_extent < rudp->rud_format.rud_nextents); + rmap = &(rudp->rud_format.rud_extents[next_extent]); + rmap->me_owner = owner; + rmap->me_startblock = startblock; + rmap->me_startoff = startoff; + rmap->me_len = blockcount; + xfs_trans_set_rmap_flags(rmap, type, whichfork, state); + rudp->rud_next_extent++; + + return error; +} -- cgit v1.2.3 From f8dbebef98f0b960a0e91d6b8d45c288c377797b Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:11:01 +1000 Subject: xfs: enable the xfs_defer mechanism to process rmaps to update Connect the xfs_defer mechanism with the pieces that we'll need to handle deferred rmap updates. We'll wire up the existing code to our new deferred mechanism later. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_defer.h | 1 + fs/xfs/xfs_super.c | 1 + fs/xfs/xfs_trans.h | 11 ++-- fs/xfs/xfs_trans_rmap.c | 131 ++++++++++++++++++++++++++++++++++++++++++++-- 4 files changed, 134 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h index b9b5a92ffa6c..cc3981c48296 100644 --- a/fs/xfs/libxfs/xfs_defer.h +++ b/fs/xfs/libxfs/xfs_defer.h @@ -51,6 +51,7 @@ struct xfs_defer_pending { * find all the space it needs. */ enum xfs_defer_ops_type { + XFS_DEFER_OPS_TYPE_RMAP, XFS_DEFER_OPS_TYPE_FREE, XFS_DEFER_OPS_TYPE_MAX, }; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 654a0924b3d3..9f8090951f09 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1876,6 +1876,7 @@ init_xfs_fs(void) XFS_BUILD_OPTIONS " enabled\n"); xfs_extent_free_init_defer_op(); + xfs_rmap_update_init_defer_op(); xfs_dir_startup(); diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 02265817f016..07f4550825f6 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -33,6 +33,8 @@ struct xfs_trans; struct xfs_trans_res; struct xfs_dquot_acct; struct xfs_busy_extent; +struct xfs_rud_log_item; +struct xfs_rui_log_item; typedef struct xfs_log_item { struct list_head li_ail; /* AIL pointers */ @@ -233,15 +235,10 @@ void xfs_trans_buf_copy_type(struct xfs_buf *dst_bp, extern kmem_zone_t *xfs_trans_zone; extern kmem_zone_t *xfs_log_item_desc_zone; +/* rmap updates */ enum xfs_rmap_intent_type; -struct xfs_rui_log_item *xfs_trans_get_rui(struct xfs_trans *tp, uint nextents); -void xfs_trans_log_start_rmap_update(struct xfs_trans *tp, - struct xfs_rui_log_item *ruip, enum xfs_rmap_intent_type type, - __uint64_t owner, int whichfork, xfs_fileoff_t startoff, - xfs_fsblock_t startblock, xfs_filblks_t blockcount, - xfs_exntst_t state); - +void xfs_rmap_update_init_defer_op(void); struct xfs_rud_log_item *xfs_trans_get_rud(struct xfs_trans *tp, struct xfs_rui_log_item *ruip, uint nextents); int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp, diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c index e3a5172ad0c0..baab99077f77 100644 --- a/fs/xfs/xfs_trans_rmap.c +++ b/fs/xfs/xfs_trans_rmap.c @@ -37,7 +37,7 @@ * caller must use all nextents extents, because we are not * flexible about this at all. */ -struct xfs_rui_log_item * +STATIC struct xfs_rui_log_item * xfs_trans_get_rui( struct xfs_trans *tp, uint nextents) @@ -96,7 +96,7 @@ xfs_trans_set_rmap_flags( * mapping is to be logged as needing to be updated. It should be * called once for each mapping. */ -void +STATIC void xfs_trans_log_start_rmap_update( struct xfs_trans *tp, struct xfs_rui_log_item *ruip, @@ -129,7 +129,6 @@ xfs_trans_log_start_rmap_update( xfs_trans_set_rmap_flags(rmap, type, whichfork, state); } - /* * This routine is called to allocate an "rmap update done" * log item that will hold nextents worth of extents. The @@ -203,3 +202,129 @@ xfs_trans_log_finish_rmap_update( return error; } + +/* Sort rmap intents by AG. */ +static int +xfs_rmap_update_diff_items( + void *priv, + struct list_head *a, + struct list_head *b) +{ + struct xfs_mount *mp = priv; + struct xfs_rmap_intent *ra; + struct xfs_rmap_intent *rb; + + ra = container_of(a, struct xfs_rmap_intent, ri_list); + rb = container_of(b, struct xfs_rmap_intent, ri_list); + return XFS_FSB_TO_AGNO(mp, ra->ri_bmap.br_startblock) - + XFS_FSB_TO_AGNO(mp, rb->ri_bmap.br_startblock); +} + +/* Get an RUI. */ +STATIC void * +xfs_rmap_update_create_intent( + struct xfs_trans *tp, + unsigned int count) +{ + return xfs_trans_get_rui(tp, count); +} + +/* Log rmap updates in the intent item. */ +STATIC void +xfs_rmap_update_log_item( + struct xfs_trans *tp, + void *intent, + struct list_head *item) +{ + struct xfs_rmap_intent *rmap; + + rmap = container_of(item, struct xfs_rmap_intent, ri_list); + xfs_trans_log_start_rmap_update(tp, intent, rmap->ri_type, + rmap->ri_owner, rmap->ri_whichfork, + rmap->ri_bmap.br_startoff, + rmap->ri_bmap.br_startblock, + rmap->ri_bmap.br_blockcount, + rmap->ri_bmap.br_state); +} + +/* Get an RUD so we can process all the deferred rmap updates. */ +STATIC void * +xfs_rmap_update_create_done( + struct xfs_trans *tp, + void *intent, + unsigned int count) +{ + return xfs_trans_get_rud(tp, intent, count); +} + +/* Process a deferred rmap update. */ +STATIC int +xfs_rmap_update_finish_item( + struct xfs_trans *tp, + struct xfs_defer_ops *dop, + struct list_head *item, + void *done_item, + void **state) +{ + struct xfs_rmap_intent *rmap; + int error; + + rmap = container_of(item, struct xfs_rmap_intent, ri_list); + error = xfs_trans_log_finish_rmap_update(tp, done_item, + rmap->ri_type, + rmap->ri_owner, rmap->ri_whichfork, + rmap->ri_bmap.br_startoff, + rmap->ri_bmap.br_startblock, + rmap->ri_bmap.br_blockcount, + rmap->ri_bmap.br_state); + kmem_free(rmap); + return error; +} + +/* Clean up after processing deferred rmaps. */ +STATIC void +xfs_rmap_update_finish_cleanup( + struct xfs_trans *tp, + void *state, + int error) +{ +} + +/* Abort all pending RUIs. */ +STATIC void +xfs_rmap_update_abort_intent( + void *intent) +{ + xfs_rui_release(intent); +} + +/* Cancel a deferred rmap update. */ +STATIC void +xfs_rmap_update_cancel_item( + struct list_head *item) +{ + struct xfs_rmap_intent *rmap; + + rmap = container_of(item, struct xfs_rmap_intent, ri_list); + kmem_free(rmap); +} + +static const struct xfs_defer_op_type xfs_rmap_update_defer_type = { + .type = XFS_DEFER_OPS_TYPE_RMAP, + .max_items = XFS_RUI_MAX_FAST_EXTENTS, + .diff_items = xfs_rmap_update_diff_items, + .create_intent = xfs_rmap_update_create_intent, + .abort_intent = xfs_rmap_update_abort_intent, + .log_item = xfs_rmap_update_log_item, + .create_done = xfs_rmap_update_create_done, + .finish_item = xfs_rmap_update_finish_item, + .finish_cleanup = xfs_rmap_update_finish_cleanup, + .cancel_item = xfs_rmap_update_cancel_item, +}; + +/* Register the deferred op type. */ +void +xfs_rmap_update_init_defer_op(void) +{ + xfs_defer_init_op_type(&xfs_rmap_update_defer_type); +} -- cgit v1.2.3 From 9c19464469556a0cd342fc40a24926ab46d7d243 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:16:05 +1000 Subject: xfs: propagate bmap updates to rmapbt When we map, unmap, or convert an extent in a file's data or attr fork, schedule a respective update in the rmapbt. Previous versions of this patch required a 1:1 correspondence between bmap and rmap, but this is no longer true as we now have ability to make interval queries against the rmapbt. We use the deferred operations code to handle redo operations atomically and deadlock free. This plumbs in all five rmap actions (map, unmap, convert extent, alloc, free); we'll use the first three now for file data, and reflink will want the last two. We also add an error injection site to test log recovery. Finally, we need to fix the bmap shift extent code to adjust the rmaps correctly. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_bmap.c | 55 ++++++++-- fs/xfs/libxfs/xfs_rmap.c | 259 +++++++++++++++++++++++++++++++++++++++++++++++ fs/xfs/libxfs/xfs_rmap.h | 24 +++++ fs/xfs/xfs_bmap_util.c | 1 + fs/xfs/xfs_error.h | 4 +- fs/xfs/xfs_rmap_item.c | 57 ++++++++++- fs/xfs/xfs_trans.h | 3 +- fs/xfs/xfs_trans_rmap.c | 13 ++- 8 files changed, 400 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index c4b89218c7eb..b060bca93402 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -2179,6 +2179,11 @@ xfs_bmap_add_extent_delay_real( ASSERT(0); } + /* add reverse mapping */ + error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip, whichfork, new); + if (error) + goto done; + /* convert to a btree if necessary */ if (xfs_bmap_needs_btree(bma->ip, whichfork)) { int tmp_logflags; /* partial log flag return val */ @@ -2715,6 +2720,11 @@ xfs_bmap_add_extent_unwritten_real( ASSERT(0); } + /* update reverse mappings */ + error = xfs_rmap_convert_extent(mp, dfops, ip, XFS_DATA_FORK, new); + if (error) + goto done; + /* convert to a btree if necessary */ if (xfs_bmap_needs_btree(ip, XFS_DATA_FORK)) { int tmp_logflags; /* partial log flag return val */ @@ -3107,6 +3117,11 @@ xfs_bmap_add_extent_hole_real( break; } + /* add reverse mapping */ + error = xfs_rmap_map_extent(mp, bma->dfops, bma->ip, whichfork, new); + if (error) + goto done; + /* convert to a btree if necessary */ if (xfs_bmap_needs_btree(bma->ip, whichfork)) { int tmp_logflags; /* partial log flag return val */ @@ -5034,6 +5049,14 @@ xfs_bmap_del_extent( ++*idx; break; } + + /* remove reverse mapping */ + if (!delay) { + error = xfs_rmap_unmap_extent(mp, dfops, ip, whichfork, del); + if (error) + goto done; + } + /* * If we need to, add to list of extents to delete. */ @@ -5573,7 +5596,8 @@ xfs_bmse_shift_one( struct xfs_bmbt_rec_host *gotp, struct xfs_btree_cur *cur, int *logflags, - enum shift_direction direction) + enum shift_direction direction, + struct xfs_defer_ops *dfops) { struct xfs_ifork *ifp; struct xfs_mount *mp; @@ -5621,9 +5645,13 @@ xfs_bmse_shift_one( /* check whether to merge the extent or shift it down */ if (xfs_bmse_can_merge(&adj_irec, &got, offset_shift_fsb)) { - return xfs_bmse_merge(ip, whichfork, offset_shift_fsb, - *current_ext, gotp, adj_irecp, - cur, logflags); + error = xfs_bmse_merge(ip, whichfork, offset_shift_fsb, + *current_ext, gotp, adj_irecp, + cur, logflags); + if (error) + return error; + adj_irec = got; + goto update_rmap; } } else { startoff = got.br_startoff + offset_shift_fsb; @@ -5660,9 +5688,10 @@ update_current_ext: (*current_ext)--; xfs_bmbt_set_startoff(gotp, startoff); *logflags |= XFS_ILOG_CORE; + adj_irec = got; if (!cur) { *logflags |= XFS_ILOG_DEXT; - return 0; + goto update_rmap; } error = xfs_bmbt_lookup_eq(cur, got.br_startoff, got.br_startblock, @@ -5672,8 +5701,18 @@ update_current_ext: XFS_WANT_CORRUPTED_RETURN(mp, i == 1); got.br_startoff = startoff; - return xfs_bmbt_update(cur, got.br_startoff, got.br_startblock, - got.br_blockcount, got.br_state); + error = xfs_bmbt_update(cur, got.br_startoff, got.br_startblock, + got.br_blockcount, got.br_state); + if (error) + return error; + +update_rmap: + /* update reverse mapping */ + error = xfs_rmap_unmap_extent(mp, dfops, ip, whichfork, &adj_irec); + if (error) + return error; + adj_irec.br_startoff = startoff; + return xfs_rmap_map_extent(mp, dfops, ip, whichfork, &adj_irec); } /* @@ -5801,7 +5840,7 @@ xfs_bmap_shift_extents( while (nexts++ < num_exts) { error = xfs_bmse_shift_one(ip, whichfork, offset_shift_fsb, ¤t_ext, gotp, cur, &logflags, - direction); + direction, dfops); if (error) goto del_cursor; /* diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index e8ce97fd0608..73d05407d663 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -36,6 +36,8 @@ #include "xfs_trace.h" #include "xfs_error.h" #include "xfs_extent_busy.h" +#include "xfs_bmap.h" +#include "xfs_inode.h" /* * Lookup the first record less than or equal to [bno, len, owner, offset] @@ -1138,3 +1140,260 @@ xfs_rmap_query_range( return xfs_btree_query_range(cur, &low_brec, &high_brec, xfs_rmap_query_range_helper, &query); } + +/* Clean up after calling xfs_rmap_finish_one. */ +void +xfs_rmap_finish_one_cleanup( + struct xfs_trans *tp, + struct xfs_btree_cur *rcur, + int error) +{ + struct xfs_buf *agbp; + + if (rcur == NULL) + return; + agbp = rcur->bc_private.a.agbp; + xfs_btree_del_cursor(rcur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); + if (error) + xfs_trans_brelse(tp, agbp); +} + +/* + * Process one of the deferred rmap operations. We pass back the + * btree cursor to maintain our lock on the rmapbt between calls. + * This saves time and eliminates a buffer deadlock between the + * superblock and the AGF because we'll always grab them in the same + * order. + */ +int +xfs_rmap_finish_one( + struct xfs_trans *tp, + enum xfs_rmap_intent_type type, + __uint64_t owner, + int whichfork, + xfs_fileoff_t startoff, + xfs_fsblock_t startblock, + xfs_filblks_t blockcount, + xfs_exntst_t state, + struct xfs_btree_cur **pcur) +{ + struct xfs_mount *mp = tp->t_mountp; + struct xfs_btree_cur *rcur; + struct xfs_buf *agbp = NULL; + int error = 0; + xfs_agnumber_t agno; + struct xfs_owner_info oinfo; + xfs_agblock_t bno; + bool unwritten; + + agno = XFS_FSB_TO_AGNO(mp, startblock); + ASSERT(agno != NULLAGNUMBER); + bno = XFS_FSB_TO_AGBNO(mp, startblock); + + trace_xfs_rmap_deferred(mp, agno, type, bno, owner, whichfork, + startoff, blockcount, state); + + if (XFS_TEST_ERROR(false, mp, + XFS_ERRTAG_RMAP_FINISH_ONE, + XFS_RANDOM_RMAP_FINISH_ONE)) + return -EIO; + + /* + * If we haven't gotten a cursor or the cursor AG doesn't match + * the startblock, get one now. + */ + rcur = *pcur; + if (rcur != NULL && rcur->bc_private.a.agno != agno) { + xfs_rmap_finish_one_cleanup(tp, rcur, 0); + rcur = NULL; + *pcur = NULL; + } + if (rcur == NULL) { + /* + * Refresh the freelist before we start changing the + * rmapbt, because a shape change could cause us to + * allocate blocks. + */ + error = xfs_free_extent_fix_freelist(tp, agno, &agbp); + if (error) + return error; + if (!agbp) + return -EFSCORRUPTED; + + rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, agno); + if (!rcur) { + error = -ENOMEM; + goto out_cur; + } + } + *pcur = rcur; + + xfs_rmap_ino_owner(&oinfo, owner, whichfork, startoff); + unwritten = state == XFS_EXT_UNWRITTEN; + bno = XFS_FSB_TO_AGBNO(rcur->bc_mp, startblock); + + switch (type) { + case XFS_RMAP_ALLOC: + case XFS_RMAP_MAP: + error = xfs_rmap_map(rcur, bno, blockcount, unwritten, &oinfo); + break; + case XFS_RMAP_FREE: + case XFS_RMAP_UNMAP: + error = xfs_rmap_unmap(rcur, bno, blockcount, unwritten, + &oinfo); + break; + case XFS_RMAP_CONVERT: + error = xfs_rmap_convert(rcur, bno, blockcount, !unwritten, + &oinfo); + break; + default: + ASSERT(0); + error = -EFSCORRUPTED; + } + return error; + +out_cur: + xfs_trans_brelse(tp, agbp); + + return error; +} + +/* + * Don't defer an rmap if we aren't an rmap filesystem. + */ +static bool +xfs_rmap_update_is_needed( + struct xfs_mount *mp) +{ + return xfs_sb_version_hasrmapbt(&mp->m_sb); +} + +/* + * Record a rmap intent; the list is kept sorted first by AG and then by + * increasing age. + */ +static int +__xfs_rmap_add( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + enum xfs_rmap_intent_type type, + __uint64_t owner, + int whichfork, + struct xfs_bmbt_irec *bmap) +{ + struct xfs_rmap_intent *ri; + + trace_xfs_rmap_defer(mp, XFS_FSB_TO_AGNO(mp, bmap->br_startblock), + type, + XFS_FSB_TO_AGBNO(mp, bmap->br_startblock), + owner, whichfork, + bmap->br_startoff, + bmap->br_blockcount, + bmap->br_state); + + ri = kmem_alloc(sizeof(struct xfs_rmap_intent), KM_SLEEP | KM_NOFS); + INIT_LIST_HEAD(&ri->ri_list); + ri->ri_type = type; + ri->ri_owner = owner; + ri->ri_whichfork = whichfork; + ri->ri_bmap = *bmap; + + xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_RMAP, &ri->ri_list); + return 0; +} + +/* Map an extent into a file. */ +int +xfs_rmap_map_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_rmap_update_is_needed(mp)) + return 0; + + return __xfs_rmap_add(mp, dfops, XFS_RMAP_MAP, ip->i_ino, + whichfork, PREV); +} + +/* Unmap an extent out of a file. */ +int +xfs_rmap_unmap_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_rmap_update_is_needed(mp)) + return 0; + + return __xfs_rmap_add(mp, dfops, XFS_RMAP_UNMAP, ip->i_ino, + whichfork, PREV); +} + +/* Convert a data fork extent from unwritten to real or vice versa. */ +int +xfs_rmap_convert_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_rmap_update_is_needed(mp)) + return 0; + + return __xfs_rmap_add(mp, dfops, XFS_RMAP_CONVERT, ip->i_ino, + whichfork, PREV); +} + +/* Schedule the creation of an rmap for non-file data. */ +int +xfs_rmap_alloc_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + xfs_agnumber_t agno, + xfs_agblock_t bno, + xfs_extlen_t len, + __uint64_t owner) +{ + struct xfs_bmbt_irec bmap; + + if (!xfs_rmap_update_is_needed(mp)) + return 0; + + bmap.br_startblock = XFS_AGB_TO_FSB(mp, agno, bno); + bmap.br_blockcount = len; + bmap.br_startoff = 0; + bmap.br_state = XFS_EXT_NORM; + + return __xfs_rmap_add(mp, dfops, XFS_RMAP_ALLOC, owner, + XFS_DATA_FORK, &bmap); +} + +/* Schedule the deletion of an rmap for non-file data. */ +int +xfs_rmap_free_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + xfs_agnumber_t agno, + xfs_agblock_t bno, + xfs_extlen_t len, + __uint64_t owner) +{ + struct xfs_bmbt_irec bmap; + + if (!xfs_rmap_update_is_needed(mp)) + return 0; + + bmap.br_startblock = XFS_AGB_TO_FSB(mp, agno, bno); + bmap.br_blockcount = len; + bmap.br_startoff = 0; + bmap.br_state = XFS_EXT_NORM; + + return __xfs_rmap_add(mp, dfops, XFS_RMAP_FREE, owner, + XFS_DATA_FORK, &bmap); +} diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index c4b1133e2ff9..71cf99a4acba 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -182,4 +182,28 @@ struct xfs_rmap_intent { struct xfs_bmbt_irec ri_bmap; }; +/* functions for updating the rmapbt based on bmbt map/unmap operations */ +int xfs_rmap_map_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, int whichfork, + struct xfs_bmbt_irec *imap); +int xfs_rmap_unmap_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, int whichfork, + struct xfs_bmbt_irec *imap); +int xfs_rmap_convert_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, int whichfork, + struct xfs_bmbt_irec *imap); +int xfs_rmap_alloc_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, + __uint64_t owner); +int xfs_rmap_free_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + xfs_agnumber_t agno, xfs_agblock_t bno, xfs_extlen_t len, + __uint64_t owner); + +void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp, + struct xfs_btree_cur *rcur, int error); +int xfs_rmap_finish_one(struct xfs_trans *tp, enum xfs_rmap_intent_type type, + __uint64_t owner, int whichfork, xfs_fileoff_t startoff, + xfs_fsblock_t startblock, xfs_filblks_t blockcount, + xfs_exntst_t state, struct xfs_btree_cur **pcur); + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 3bbe46bdbc6e..a819d7baa346 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -41,6 +41,7 @@ #include "xfs_trace.h" #include "xfs_icache.h" #include "xfs_log.h" +#include "xfs_rmap_btree.h" /* Kernel only BMAP related definitions and functions */ diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h index da6f951435c9..3d224702fbc0 100644 --- a/fs/xfs/xfs_error.h +++ b/fs/xfs/xfs_error.h @@ -91,7 +91,8 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_ERRTAG_DIOWRITE_IOERR 20 #define XFS_ERRTAG_BMAPIFORMAT 21 #define XFS_ERRTAG_FREE_EXTENT 22 -#define XFS_ERRTAG_MAX 23 +#define XFS_ERRTAG_RMAP_FINISH_ONE 23 +#define XFS_ERRTAG_MAX 24 /* * Random factors for above tags, 1 means always, 2 means 1/2 time, etc. @@ -119,6 +120,7 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_RANDOM_DIOWRITE_IOERR (XFS_RANDOM_DEFAULT/10) #define XFS_RANDOM_BMAPIFORMAT XFS_RANDOM_DEFAULT #define XFS_RANDOM_FREE_EXTENT 1 +#define XFS_RANDOM_RMAP_FINISH_ONE 1 #ifdef DEBUG extern int xfs_error_test_active; diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index fecd1e4d688d..6d6cc3b8d44f 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -24,11 +24,13 @@ #include "xfs_trans_resv.h" #include "xfs_bit.h" #include "xfs_mount.h" +#include "xfs_defer.h" #include "xfs_trans.h" #include "xfs_trans_priv.h" #include "xfs_buf_item.h" #include "xfs_rmap_item.h" #include "xfs_log.h" +#include "xfs_rmap.h" kmem_zone_t *xfs_rui_zone; @@ -473,6 +475,12 @@ xfs_rui_recover( struct xfs_map_extent *rmap; xfs_fsblock_t startblock_fsb; bool op_ok; + struct xfs_rud_log_item *rudp; + enum xfs_rmap_intent_type type; + int whichfork; + xfs_exntst_t state; + struct xfs_trans *tp; + struct xfs_btree_cur *rcur = NULL; ASSERT(!test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags)); @@ -512,8 +520,53 @@ xfs_rui_recover( } } - /* XXX: do nothing for now */ + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); + if (error) + return error; + rudp = xfs_trans_get_rud(tp, ruip, ruip->rui_format.rui_nextents); + + for (i = 0; i < ruip->rui_format.rui_nextents; i++) { + rmap = &(ruip->rui_format.rui_extents[i]); + state = (rmap->me_flags & XFS_RMAP_EXTENT_UNWRITTEN) ? + XFS_EXT_UNWRITTEN : XFS_EXT_NORM; + whichfork = (rmap->me_flags & XFS_RMAP_EXTENT_ATTR_FORK) ? + XFS_ATTR_FORK : XFS_DATA_FORK; + switch (rmap->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) { + case XFS_RMAP_EXTENT_MAP: + type = XFS_RMAP_MAP; + break; + case XFS_RMAP_EXTENT_UNMAP: + type = XFS_RMAP_UNMAP; + break; + case XFS_RMAP_EXTENT_CONVERT: + type = XFS_RMAP_CONVERT; + break; + case XFS_RMAP_EXTENT_ALLOC: + type = XFS_RMAP_ALLOC; + break; + case XFS_RMAP_EXTENT_FREE: + type = XFS_RMAP_FREE; + break; + default: + error = -EFSCORRUPTED; + goto abort_error; + } + error = xfs_trans_log_finish_rmap_update(tp, rudp, type, + rmap->me_owner, whichfork, + rmap->me_startoff, rmap->me_startblock, + rmap->me_len, state, &rcur); + if (error) + goto abort_error; + + } + + xfs_rmap_finish_one_cleanup(tp, rcur, error); set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags); - xfs_rui_release(ruip); + error = xfs_trans_commit(tp); + return error; + +abort_error: + xfs_rmap_finish_one_cleanup(tp, rcur, error); + xfs_trans_cancel(tp); return error; } diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 07f4550825f6..bb4b84f9347e 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -35,6 +35,7 @@ struct xfs_dquot_acct; struct xfs_busy_extent; struct xfs_rud_log_item; struct xfs_rui_log_item; +struct xfs_btree_cur; typedef struct xfs_log_item { struct list_head li_ail; /* AIL pointers */ @@ -245,6 +246,6 @@ int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp, struct xfs_rud_log_item *rudp, enum xfs_rmap_intent_type type, __uint64_t owner, int whichfork, xfs_fileoff_t startoff, xfs_fsblock_t startblock, xfs_filblks_t blockcount, - xfs_exntst_t state); + xfs_exntst_t state, struct xfs_btree_cur **pcur); #endif /* __XFS_TRANS_H__ */ diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c index baab99077f77..83414764e1ab 100644 --- a/fs/xfs/xfs_trans_rmap.c +++ b/fs/xfs/xfs_trans_rmap.c @@ -171,14 +171,15 @@ xfs_trans_log_finish_rmap_update( xfs_fileoff_t startoff, xfs_fsblock_t startblock, xfs_filblks_t blockcount, - xfs_exntst_t state) + xfs_exntst_t state, + struct xfs_btree_cur **pcur) { uint next_extent; struct xfs_map_extent *rmap; int error; - /* XXX: actually finish the rmap update here */ - error = -EFSCORRUPTED; + error = xfs_rmap_finish_one(tp, type, owner, whichfork, startoff, + startblock, blockcount, state, pcur); /* * Mark the transaction dirty, even on error. This ensures the @@ -276,7 +277,8 @@ xfs_rmap_update_finish_item( rmap->ri_bmap.br_startoff, rmap->ri_bmap.br_startblock, rmap->ri_bmap.br_blockcount, - rmap->ri_bmap.br_state); + rmap->ri_bmap.br_state, + (struct xfs_btree_cur **)state); kmem_free(rmap); return error; } @@ -288,6 +290,9 @@ xfs_rmap_update_finish_cleanup( void *state, int error) { + struct xfs_btree_cur *rcur = state; + + xfs_rmap_finish_one_cleanup(tp, rcur, error); } /* Abort all pending RUIs. */ -- cgit v1.2.3 From 5d650e90a101557a7a652989c6d5eb657ae2476b Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:16:44 +1000 Subject: xfs: add rmap btree geometry feature flag Originally-From: Dave Chinner So xfs_info and other userspace utilities know the filesystem is using this feature. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_fs.h | 1 + fs/xfs/xfs_fsops.c | 4 +++- 2 files changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h index f5ec9c5ccae6..79455058b752 100644 --- a/fs/xfs/libxfs/xfs_fs.h +++ b/fs/xfs/libxfs/xfs_fs.h @@ -206,6 +206,7 @@ typedef struct xfs_fsop_resblks { #define XFS_FSOP_GEOM_FLAGS_FTYPE 0x10000 /* inode directory types */ #define XFS_FSOP_GEOM_FLAGS_FINOBT 0x20000 /* free inode btree */ #define XFS_FSOP_GEOM_FLAGS_SPINODES 0x40000 /* sparse inode chunks */ +#define XFS_FSOP_GEOM_FLAGS_RMAPBT 0x80000 /* Reverse mapping btree */ /* * Minimum and maximum sizes need for growth checks. diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 83304d54e6d4..0f96847b90e1 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -106,7 +106,9 @@ xfs_fs_geometry( (xfs_sb_version_hasfinobt(&mp->m_sb) ? XFS_FSOP_GEOM_FLAGS_FINOBT : 0) | (xfs_sb_version_hassparseinodes(&mp->m_sb) ? - XFS_FSOP_GEOM_FLAGS_SPINODES : 0); + XFS_FSOP_GEOM_FLAGS_SPINODES : 0) | + (xfs_sb_version_hasrmapbt(&mp->m_sb) ? + XFS_FSOP_GEOM_FLAGS_RMAPBT : 0); geo->logsectsize = xfs_sb_version_hassector(&mp->m_sb) ? mp->m_sb.sb_logsectsize : BBSIZE; geo->rtsectsize = mp->m_sb.sb_blocksize; -- cgit v1.2.3 From a650e8f98e5d9df8a5866d73aa9a796601f1abfc Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:17:11 +1000 Subject: xfs: add rmap btree block detection to log recovery Originally-From: Dave Chinner So such blocks can be correctly identified and have their operations structures attached to validate recovery has not resulted in a correct block. Signed-off-by: Dave Chinner Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_log_recover.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 365aef655edb..e51fd2bdacb6 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2231,6 +2231,7 @@ xlog_recover_get_buf_lsn( case XFS_ABTC_CRC_MAGIC: case XFS_ABTB_MAGIC: case XFS_ABTC_MAGIC: + case XFS_RMAP_CRC_MAGIC: case XFS_IBT_CRC_MAGIC: case XFS_IBT_MAGIC: { struct xfs_btree_block *btb = blk; @@ -2399,6 +2400,9 @@ xlog_recover_validate_buf_type( case XFS_BMAP_MAGIC: bp->b_ops = &xfs_bmbt_buf_ops; break; + case XFS_RMAP_CRC_MAGIC: + bp->b_ops = &xfs_rmapbt_buf_ops; + break; default: xfs_warn(mp, "Bad btree block magic!"); ASSERT(0); -- cgit v1.2.3 From 2b0eeb5e74ced62a94b01c65ddec9be7e5edcea1 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:18:07 +1000 Subject: xfs: disable XFS_IOC_SWAPEXT when rmap btree is enabled Swapping extents between two inodes requires the owner to be updated in the rmap tree for all the extents that are swapped. This code does not yet exist, so switch off the XFS_IOC_SWAPEXT ioctl until support has been implemented. This will need to be done before the rmap btree code can have the experimental tag removed. This functionality will be provided in a (much) later patch, using some of the reflink deferred block remapping functionality to accomlish extent swapping with rmap updates. Signed-off-by: Dave Chinner Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/xfs_bmap_util.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index a819d7baa346..4ece4f2ffc72 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -1535,6 +1535,10 @@ xfs_swap_extents( __uint64_t tmp; int lock_flags; + /* XXX: we can't do this with rmap, will fix later */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + return -EOPNOTSUPP; + tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL); if (!tempifp) { error = -ENOMEM; -- cgit v1.2.3 From 04f130605ff6fb01a93a0885607921df9c463eed Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:19:53 +1000 Subject: xfs: don't update rmapbt when fixing agfl Allow a caller of xfs_alloc_fix_freelist to disable rmapbt updates when fixing the AG freelist. xfs_repair needs this during phase 5 to be able to adjust the freelist while it's reconstructing the rmap btree; the missing entries will be added back at the very end of phase 5 once the AGFL contents settle down. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 18 ++++++++++++++---- fs/xfs/libxfs/xfs_alloc.h | 3 +++ 2 files changed, 17 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 73ab1ea1270d..a0dc2999778d 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -2090,9 +2090,21 @@ xfs_alloc_fix_freelist( * anything other than extra overhead when we need to put more blocks * back on the free list? Maybe we should only do this when space is * getting low or the AGFL is more than half full? + * + * The NOSHRINK flag prevents the AGFL from being shrunk if it's too + * big; the NORMAP flag prevents AGFL expand/shrink operations from + * updating the rmapbt. Both flags are used in xfs_repair while we're + * rebuilding the rmapbt, and neither are used by the kernel. They're + * both required to ensure that rmaps are correctly recorded for the + * regenerated AGFL, bnobt, and cntbt. See repair/phase5.c and + * repair/rmap.c in xfsprogs for details. */ - xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG); - while (pag->pagf_flcount > need) { + memset(&targs, 0, sizeof(targs)); + if (flags & XFS_ALLOC_FLAG_NORMAP) + xfs_rmap_skip_owner_update(&targs.oinfo); + else + xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG); + while (!(flags & XFS_ALLOC_FLAG_NOSHRINK) && pag->pagf_flcount > need) { struct xfs_buf *bp; error = xfs_alloc_get_freelist(tp, agbp, &bno, 0); @@ -2106,10 +2118,8 @@ xfs_alloc_fix_freelist( xfs_trans_binval(tp, bp); } - memset(&targs, 0, sizeof(targs)); targs.tp = tp; targs.mp = mp; - xfs_rmap_ag_owner(&targs.oinfo, XFS_RMAP_OWN_AG); targs.agbp = agbp; targs.agno = args->agno; targs.alignment = targs.minlen = targs.prod = targs.isfl = 1; diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index 360f9e425536..6fe2d6b7cfe9 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -54,6 +54,9 @@ typedef unsigned int xfs_alloctype_t; */ #define XFS_ALLOC_FLAG_TRYLOCK 0x00000001 /* use trylock for buffer locking */ #define XFS_ALLOC_FLAG_FREEING 0x00000002 /* indicate caller is freeing extents*/ +#define XFS_ALLOC_FLAG_NORMAP 0x00000004 /* don't modify the rmapbt */ +#define XFS_ALLOC_FLAG_NOSHRINK 0x00000008 /* don't shrink the freelist */ + /* * Argument structure for xfs_alloc routines. -- cgit v1.2.3 From 1c0607ace9bd639d22ad1bd453ffeb7d55913f88 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:20:57 +1000 Subject: xfs: enable the rmap btree functionality Originally-From: Dave Chinner Add the feature flag to the supported matrix so that the kernel can mount and use rmap btree enabled filesystems Signed-off-by: Dave Chinner [darrick.wong@oracle.com: move the experimental tag] Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_format.h | 3 ++- fs/xfs/xfs_super.c | 4 ++++ 2 files changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index f6a0641e64c8..f814d42c73b2 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -457,7 +457,8 @@ xfs_sb_has_compat_feature( #define XFS_SB_FEAT_RO_COMPAT_FINOBT (1 << 0) /* free inode btree */ #define XFS_SB_FEAT_RO_COMPAT_RMAPBT (1 << 1) /* reverse map btree */ #define XFS_SB_FEAT_RO_COMPAT_ALL \ - (XFS_SB_FEAT_RO_COMPAT_FINOBT) + (XFS_SB_FEAT_RO_COMPAT_FINOBT | \ + XFS_SB_FEAT_RO_COMPAT_RMAPBT) #define XFS_SB_FEAT_RO_COMPAT_UNKNOWN ~XFS_SB_FEAT_RO_COMPAT_ALL static inline bool xfs_sb_has_ro_compat_feature( diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 9f8090951f09..45773df1dd0b 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1574,6 +1574,10 @@ xfs_fs_fill_super( } } + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + xfs_alert(mp, + "EXPERIMENTAL reverse mapping btree feature enabled. Use at your own risk!"); + error = xfs_mountfs(mp); if (error) goto out_filestream_unmount; -- cgit v1.2.3 From 973b83194bf12f7e315aace57ae2096ff7b82360 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:22:12 +1000 Subject: xfs: remove the get*keys and update_keys btree ops pointers These are internal btree functions; we don't need them to be dispatched via function pointers. Make them static again and just check the overlapped flag to figure out what we need to do. The strategy behind this patch was suggested by Christoph. Signed-off-by: Darrick J. Wong Suggested-by: Christoph Hellwig Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc_btree.c | 4 - fs/xfs/libxfs/xfs_bmap_btree.c | 4 - fs/xfs/libxfs/xfs_btree.c | 153 ++++++++++++++++----------------------- fs/xfs/libxfs/xfs_btree.h | 24 ------ fs/xfs/libxfs/xfs_ialloc_btree.c | 8 -- fs/xfs/libxfs/xfs_rmap_btree.c | 4 - 6 files changed, 61 insertions(+), 136 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c index c60eeb826b9d..5ba2dac5e67c 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.c +++ b/fs/xfs/libxfs/xfs_alloc_btree.c @@ -403,10 +403,6 @@ static const struct xfs_btree_ops xfs_allocbt_ops = { .keys_inorder = xfs_allocbt_keys_inorder, .recs_inorder = xfs_allocbt_recs_inorder, #endif - - .get_leaf_keys = xfs_btree_get_leaf_keys, - .get_node_keys = xfs_btree_get_node_keys, - .update_keys = xfs_btree_update_keys, }; /* diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 9e34ca413e53..cd85274e810c 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -763,10 +763,6 @@ static const struct xfs_btree_ops xfs_bmbt_ops = { .keys_inorder = xfs_bmbt_keys_inorder, .recs_inorder = xfs_bmbt_recs_inorder, #endif - - .get_leaf_keys = xfs_btree_get_leaf_keys, - .get_node_keys = xfs_btree_get_node_keys, - .update_keys = xfs_btree_update_keys, }; /* diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index fe5f27e4cdc5..2a13d418ed4b 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -1951,29 +1951,6 @@ error0: return error; } -/* Determine the low key of a leaf block (simple) */ -void -xfs_btree_get_leaf_keys( - struct xfs_btree_cur *cur, - struct xfs_btree_block *block, - union xfs_btree_key *key) -{ - union xfs_btree_rec *rec; - - rec = xfs_btree_rec_addr(cur, 1, block); - cur->bc_ops->init_key_from_rec(key, rec); -} - -/* Determine the low key of a node block (simple) */ -void -xfs_btree_get_node_keys( - struct xfs_btree_cur *cur, - struct xfs_btree_block *block, - union xfs_btree_key *key) -{ - memcpy(key, xfs_btree_key_addr(cur, 1, block), cur->bc_ops->key_len); -} - /* Find the high key storage area from a regular key. */ STATIC union xfs_btree_key * xfs_btree_high_key_from_key( @@ -1985,60 +1962,67 @@ xfs_btree_high_key_from_key( (cur->bc_ops->key_len / 2)); } -/* Determine the low and high keys of a leaf block (overlapped) */ -void -xfs_btree_get_leaf_keys_overlapped( +/* Determine the low (and high if overlapped) keys of a leaf block */ +STATIC void +xfs_btree_get_leaf_keys( struct xfs_btree_cur *cur, struct xfs_btree_block *block, union xfs_btree_key *key) { - int n; - union xfs_btree_rec *rec; union xfs_btree_key max_hkey; union xfs_btree_key hkey; + union xfs_btree_rec *rec; union xfs_btree_key *high; + int n; - ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); rec = xfs_btree_rec_addr(cur, 1, block); cur->bc_ops->init_key_from_rec(key, rec); - cur->bc_ops->init_high_key_from_rec(&max_hkey, rec); - for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { - rec = xfs_btree_rec_addr(cur, n, block); - cur->bc_ops->init_high_key_from_rec(&hkey, rec); - if (cur->bc_ops->diff_two_keys(cur, &hkey, &max_hkey) > 0) - max_hkey = hkey; - } + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + + cur->bc_ops->init_high_key_from_rec(&max_hkey, rec); + for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { + rec = xfs_btree_rec_addr(cur, n, block); + cur->bc_ops->init_high_key_from_rec(&hkey, rec); + if (cur->bc_ops->diff_two_keys(cur, &hkey, &max_hkey) + > 0) + max_hkey = hkey; + } - high = xfs_btree_high_key_from_key(cur, key); - memcpy(high, &max_hkey, cur->bc_ops->key_len / 2); + high = xfs_btree_high_key_from_key(cur, key); + memcpy(high, &max_hkey, cur->bc_ops->key_len / 2); + } } -/* Determine the low and high keys of a node block (overlapped) */ -void -xfs_btree_get_node_keys_overlapped( +/* Determine the low (and high if overlapped) keys of a node block */ +STATIC void +xfs_btree_get_node_keys( struct xfs_btree_cur *cur, struct xfs_btree_block *block, union xfs_btree_key *key) { - int n; union xfs_btree_key *hkey; union xfs_btree_key *max_hkey; union xfs_btree_key *high; + int n; - ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); - memcpy(key, xfs_btree_key_addr(cur, 1, block), - cur->bc_ops->key_len / 2); - - max_hkey = xfs_btree_high_key_addr(cur, 1, block); - for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { - hkey = xfs_btree_high_key_addr(cur, n, block); - if (cur->bc_ops->diff_two_keys(cur, hkey, max_hkey) > 0) - max_hkey = hkey; - } + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + memcpy(key, xfs_btree_key_addr(cur, 1, block), + cur->bc_ops->key_len / 2); + + max_hkey = xfs_btree_high_key_addr(cur, 1, block); + for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { + hkey = xfs_btree_high_key_addr(cur, n, block); + if (cur->bc_ops->diff_two_keys(cur, hkey, max_hkey) > 0) + max_hkey = hkey; + } - high = xfs_btree_high_key_from_key(cur, key); - memcpy(high, max_hkey, cur->bc_ops->key_len / 2); + high = xfs_btree_high_key_from_key(cur, key); + memcpy(high, max_hkey, cur->bc_ops->key_len / 2); + } else { + memcpy(key, xfs_btree_key_addr(cur, 1, block), + cur->bc_ops->key_len); + } } /* Derive the keys for any btree block. */ @@ -2049,9 +2033,9 @@ xfs_btree_get_keys( union xfs_btree_key *key) { if (be16_to_cpu(block->bb_level) == 0) - cur->bc_ops->get_leaf_keys(cur, block, key); + xfs_btree_get_leaf_keys(cur, block, key); else - cur->bc_ops->get_node_keys(cur, block, key); + xfs_btree_get_node_keys(cur, block, key); } /* @@ -2125,28 +2109,12 @@ __xfs_btree_updkeys( xfs_btree_log_keys(cur, bp, ptr, ptr); if (level + 1 >= cur->bc_nlevels) break; - cur->bc_ops->get_node_keys(cur, block, lkey); + xfs_btree_get_node_keys(cur, block, lkey); } return 0; } -/* - * Update all the keys from some level in cursor back to the root, stopping - * when we find a key pair that don't need updating. - */ -int -xfs_btree_update_keys_overlapped( - struct xfs_btree_cur *cur, - int level) -{ - struct xfs_buf *bp; - struct xfs_btree_block *block; - - block = xfs_btree_get_block(cur, level, &bp); - return __xfs_btree_updkeys(cur, level, block, bp, false); -} - /* Update all the keys from some level in cursor back to the root. */ STATIC int xfs_btree_updkeys_force( @@ -2163,7 +2131,7 @@ xfs_btree_updkeys_force( /* * Update the parent keys of the given level, progressing towards the root. */ -int +STATIC int xfs_btree_update_keys( struct xfs_btree_cur *cur, int level) @@ -2174,20 +2142,21 @@ xfs_btree_update_keys( union xfs_btree_key key; int ptr; - ASSERT(!(cur->bc_flags & XFS_BTREE_OVERLAPPING)); + ASSERT(level >= 0); + + block = xfs_btree_get_block(cur, level, &bp); + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) + return __xfs_btree_updkeys(cur, level, block, bp, false); XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); XFS_BTREE_TRACE_ARGIK(cur, level, keyp); - ASSERT(level >= 0); - /* * Go up the tree from this level toward the root. * At each level, update the key value to the value input. * Stop when we reach a level where the cursor isn't pointing * at the first entry in the block. */ - block = xfs_btree_get_block(cur, level, &bp); xfs_btree_get_keys(cur, block, &key); for (level++, ptr = 1; ptr == 1 && level < cur->bc_nlevels; level++) { #ifdef DEBUG @@ -2257,7 +2226,7 @@ xfs_btree_update( /* Pass new key value up to our parent. */ if (xfs_btree_needs_key_update(cur, ptr)) { - error = cur->bc_ops->update_keys(cur, 0); + error = xfs_btree_update_keys(cur, 0); if (error) goto error0; } @@ -2447,13 +2416,13 @@ xfs_btree_lshift( goto error1; /* Update the parent keys of the right block. */ - error = cur->bc_ops->update_keys(cur, level); + error = xfs_btree_update_keys(cur, level); if (error) goto error1; /* Update the parent high keys of the left block, if needed. */ if (tcur->bc_flags & XFS_BTREE_OVERLAPPING) { - error = tcur->bc_ops->update_keys(tcur, level); + error = xfs_btree_update_keys(tcur, level); if (error) goto error1; } @@ -2633,13 +2602,13 @@ xfs_btree_rshift( /* Update the parent high keys of the left block, if needed. */ if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { - error = cur->bc_ops->update_keys(cur, level); + error = xfs_btree_update_keys(cur, level); if (error) goto error1; } /* Update the parent keys of the right block. */ - error = cur->bc_ops->update_keys(tcur, level); + error = xfs_btree_update_keys(tcur, level); if (error) goto error1; @@ -2778,7 +2747,7 @@ __xfs_btree_split( xfs_btree_log_ptrs(cur, rbp, 1, rrecs); /* Stash the keys of the new block for later insertion. */ - cur->bc_ops->get_node_keys(cur, right, key); + xfs_btree_get_node_keys(cur, right, key); } else { /* It's a leaf. Move records. */ union xfs_btree_rec *lrp; /* left record pointer */ @@ -2792,7 +2761,7 @@ __xfs_btree_split( xfs_btree_log_recs(cur, rbp, 1, rrecs); /* Stash the keys of the new block for later insertion. */ - cur->bc_ops->get_leaf_keys(cur, right, key); + xfs_btree_get_leaf_keys(cur, right, key); } /* @@ -2822,7 +2791,7 @@ __xfs_btree_split( /* Update the parent high keys of the left block, if needed. */ if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { - error = cur->bc_ops->update_keys(cur, level); + error = xfs_btree_update_keys(cur, level); if (error) goto error0; } @@ -3143,9 +3112,9 @@ xfs_btree_new_root( * Get the keys for the left block's keys and put them directly * in the parent block. Do the same for the right block. */ - cur->bc_ops->get_node_keys(cur, left, + xfs_btree_get_node_keys(cur, left, xfs_btree_key_addr(cur, 1, new)); - cur->bc_ops->get_node_keys(cur, right, + xfs_btree_get_node_keys(cur, right, xfs_btree_key_addr(cur, 2, new)); } else { /* @@ -3153,9 +3122,9 @@ xfs_btree_new_root( * directly in the parent block. Do the same for the right * block. */ - cur->bc_ops->get_leaf_keys(cur, left, + xfs_btree_get_leaf_keys(cur, left, xfs_btree_key_addr(cur, 1, new)); - cur->bc_ops->get_leaf_keys(cur, right, + xfs_btree_get_leaf_keys(cur, right, xfs_btree_key_addr(cur, 2, new)); } xfs_btree_log_keys(cur, nbp, 1, 2); @@ -3434,7 +3403,7 @@ xfs_btree_insrec( if (bp && bp->b_bn != old_bn) { xfs_btree_get_keys(cur, block, lkey); } else if (xfs_btree_needs_key_update(cur, optr)) { - error = cur->bc_ops->update_keys(cur, level); + error = xfs_btree_update_keys(cur, level); if (error) goto error0; } @@ -3880,7 +3849,7 @@ xfs_btree_delrec( * key values above us in the tree. */ if (xfs_btree_needs_key_update(cur, ptr)) { - error = cur->bc_ops->update_keys(cur, level); + error = xfs_btree_update_keys(cur, level); if (error) goto error0; } diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 9b6d6286466e..04d0865e5e6d 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -214,19 +214,6 @@ struct xfs_btree_ops { union xfs_btree_rec *r1, union xfs_btree_rec *r2); #endif - - /* derive the low & high keys from the records in a leaf block */ - void (*get_leaf_keys)(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, - union xfs_btree_key *key); - - /* derive the low & high keys from the keys in a node block */ - void (*get_node_keys)(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, - union xfs_btree_key *key); - - /* update the parent keys of given btree level */ - int (*update_keys)(struct xfs_btree_cur *cur, int level); }; /* @@ -527,17 +514,6 @@ bool xfs_btree_sblock_verify(struct xfs_buf *bp, unsigned int max_recs); uint xfs_btree_compute_maxlevels(struct xfs_mount *mp, uint *limits, unsigned long len); -void xfs_btree_get_leaf_keys(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, union xfs_btree_key *key); -void xfs_btree_get_node_keys(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, union xfs_btree_key *key); -int xfs_btree_update_keys(struct xfs_btree_cur *cur, int level); -void xfs_btree_get_leaf_keys_overlapped(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, union xfs_btree_key *key); -void xfs_btree_get_node_keys_overlapped(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, union xfs_btree_key *key); -int xfs_btree_update_keys_overlapped(struct xfs_btree_cur *cur, int level); - /* return codes */ #define XFS_BTREE_QUERY_RANGE_CONTINUE 0 /* keep iterating */ #define XFS_BTREE_QUERY_RANGE_ABORT 1 /* stop iterating */ diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index c83691edf6b2..31ca2208c03d 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -320,10 +320,6 @@ static const struct xfs_btree_ops xfs_inobt_ops = { .keys_inorder = xfs_inobt_keys_inorder, .recs_inorder = xfs_inobt_recs_inorder, #endif - - .get_leaf_keys = xfs_btree_get_leaf_keys, - .get_node_keys = xfs_btree_get_node_keys, - .update_keys = xfs_btree_update_keys, }; static const struct xfs_btree_ops xfs_finobt_ops = { @@ -345,10 +341,6 @@ static const struct xfs_btree_ops xfs_finobt_ops = { .keys_inorder = xfs_inobt_keys_inorder, .recs_inorder = xfs_inobt_recs_inorder, #endif - - .get_leaf_keys = xfs_btree_get_leaf_keys, - .get_node_keys = xfs_btree_get_node_keys, - .update_keys = xfs_btree_update_keys, }; /* diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c index 232450ce61a8..bc1faebc84ec 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.c +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -453,10 +453,6 @@ static const struct xfs_btree_ops xfs_rmapbt_ops = { .keys_inorder = xfs_rmapbt_keys_inorder, .recs_inorder = xfs_rmapbt_recs_inorder, #endif - - .get_leaf_keys = xfs_btree_get_leaf_keys_overlapped, - .get_node_keys = xfs_btree_get_node_keys_overlapped, - .update_keys = xfs_btree_update_keys_overlapped, }; /* -- cgit v1.2.3 From 1f704b2b47822435765aee16f120ae06cc40e78c Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:22:45 +1000 Subject: xfs: remove unnecesary lshift/rshift key initialization In the lshift/rshift functions we don't use the key variable for anything now, so remove the variable and its initializer. The update_keys functions figure out the key for a block on their own. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 14 -------------- 1 file changed, 14 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 2a13d418ed4b..40c19f82190c 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2249,7 +2249,6 @@ xfs_btree_lshift( int level, int *stat) /* success/failure */ { - union xfs_btree_key key; /* btree key */ struct xfs_buf *lbp; /* left buffer pointer */ struct xfs_btree_block *left; /* left btree block */ int lrecs; /* left record count */ @@ -2392,13 +2391,6 @@ xfs_btree_lshift( xfs_btree_rec_addr(cur, 2, right), -1, rrecs); xfs_btree_log_recs(cur, rbp, 1, rrecs); - - /* - * If it's the first record in the block, we'll need a key - * structure to pass up to the next level (updkey). - */ - cur->bc_ops->init_key_from_rec(&key, - xfs_btree_rec_addr(cur, 1, right)); } /* @@ -2461,7 +2453,6 @@ xfs_btree_rshift( int level, int *stat) /* success/failure */ { - union xfs_btree_key key; /* btree key */ struct xfs_buf *lbp; /* left buffer pointer */ struct xfs_btree_block *left; /* left btree block */ struct xfs_buf *rbp; /* right buffer pointer */ @@ -2570,11 +2561,6 @@ xfs_btree_rshift( /* Now put the new data in, and log it. */ xfs_btree_copy_recs(cur, rrp, lrp, 1); xfs_btree_log_recs(cur, rbp, 1, rrecs + 1); - - cur->bc_ops->init_key_from_rec(&key, rrp); - - ASSERT(cur->bc_ops->recs_inorder(cur, rrp, - xfs_btree_rec_addr(cur, 2, right))); } /* -- cgit v1.2.3 From c1d22ae89cf6086d6a457b3b9241fcb36ebddd14 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:26:22 +1000 Subject: xfs: in btree_lshift, only allocate temporary cursor when needed We only need the temporary cursor in _btree_lshift if we're shifting in an overlapped btree. Therefore, factor that into a single block of code so we avoid unnecessary cursor duplication. Also fix use of the wrong cursor when checking for corruption in xfs_btree_rshift(). Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_btree.c | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 40c19f82190c..b5c213a051cd 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -2397,29 +2397,29 @@ xfs_btree_lshift( * Using a temporary cursor, update the parent key values of the * block on the left. */ - error = xfs_btree_dup_cursor(cur, &tcur); - if (error) - goto error0; - i = xfs_btree_firstrec(tcur, level); - XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0); - - error = xfs_btree_decrement(tcur, level, &i); - if (error) - goto error1; + if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + error = xfs_btree_dup_cursor(cur, &tcur); + if (error) + goto error0; + i = xfs_btree_firstrec(tcur, level); + XFS_WANT_CORRUPTED_GOTO(tcur->bc_mp, i == 1, error0); - /* Update the parent keys of the right block. */ - error = xfs_btree_update_keys(cur, level); - if (error) - goto error1; + error = xfs_btree_decrement(tcur, level, &i); + if (error) + goto error1; - /* Update the parent high keys of the left block, if needed. */ - if (tcur->bc_flags & XFS_BTREE_OVERLAPPING) { + /* Update the parent high keys of the left block, if needed. */ error = xfs_btree_update_keys(tcur, level); if (error) goto error1; + + xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR); } - xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR); + /* Update the parent keys of the right block. */ + error = xfs_btree_update_keys(cur, level); + if (error) + goto error0; /* Slide the cursor value left one. */ cur->bc_ptrs[level]--; @@ -2580,7 +2580,7 @@ xfs_btree_rshift( if (error) goto error0; i = xfs_btree_lastrec(tcur, level); - XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, error0); + XFS_WANT_CORRUPTED_GOTO(tcur->bc_mp, i == 1, error0); error = xfs_btree_increment(tcur, level, &i); if (error) -- cgit v1.2.3 From 722e251770306ee325151b28e40b5d7e5497d687 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:28:43 +1000 Subject: xfs: remove the extents array from the rmap update done log item Nothing ever uses the extent array in the rmap update done redo item, so remove it before it is fixed in the on-disk log format. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_log_format.h | 3 +-- fs/xfs/xfs_log_recover.c | 4 +--- fs/xfs/xfs_rmap_item.c | 50 ++++++------------------------------------ fs/xfs/xfs_rmap_item.h | 8 +------ fs/xfs/xfs_super.c | 4 +--- fs/xfs/xfs_trans.h | 2 +- fs/xfs/xfs_trans_rmap.c | 33 +++------------------------- 7 files changed, 15 insertions(+), 89 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h index a8d794d1ae9a..a6eed43fa7cd 100644 --- a/fs/xfs/libxfs/xfs_log_format.h +++ b/fs/xfs/libxfs/xfs_log_format.h @@ -658,9 +658,8 @@ struct xfs_rui_log_format { struct xfs_rud_log_format { __uint16_t rud_type; /* rud log item type */ __uint16_t rud_size; /* size of this item */ - __uint32_t rud_nextents; /* # of extents freed */ + __uint32_t __pad; __uint64_t rud_rui_id; /* id of corresponding rui */ - struct xfs_map_extent rud_extents[1]; /* array of extents rmapped */ }; /* diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index e51fd2bdacb6..e8638fd2c0c3 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3482,9 +3482,7 @@ xlog_recover_rud_pass2( struct xfs_ail *ailp = log->l_ailp; rud_formatp = item->ri_buf[0].i_addr; - ASSERT(item->ri_buf[0].i_len == (sizeof(struct xfs_rud_log_format) + - ((rud_formatp->rud_nextents - 1) * - sizeof(struct xfs_map_extent)))); + ASSERT(item->ri_buf[0].i_len == sizeof(struct xfs_rud_log_format)); rui_id = rud_formatp->rud_rui_id; /* diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 6d6cc3b8d44f..7e4743219430 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -272,29 +272,6 @@ static inline struct xfs_rud_log_item *RUD_ITEM(struct xfs_log_item *lip) return container_of(lip, struct xfs_rud_log_item, rud_item); } -STATIC void -xfs_rud_item_free(struct xfs_rud_log_item *rudp) -{ - if (rudp->rud_format.rud_nextents > XFS_RUD_MAX_FAST_EXTENTS) - kmem_free(rudp); - else - kmem_zone_free(xfs_rud_zone, rudp); -} - -/* - * This returns the number of iovecs needed to log the given rud item. - * We only need 1 iovec for an rud item. It just logs the rud_log_format - * structure. - */ -static inline int -xfs_rud_item_sizeof( - struct xfs_rud_log_item *rudp) -{ - return sizeof(struct xfs_rud_log_format) + - (rudp->rud_format.rud_nextents - 1) * - sizeof(struct xfs_map_extent); -} - STATIC void xfs_rud_item_size( struct xfs_log_item *lip, @@ -302,7 +279,7 @@ xfs_rud_item_size( int *nbytes) { *nvecs += 1; - *nbytes += xfs_rud_item_sizeof(RUD_ITEM(lip)); + *nbytes += sizeof(struct xfs_rud_log_format); } /* @@ -320,13 +297,11 @@ xfs_rud_item_format( struct xfs_rud_log_item *rudp = RUD_ITEM(lip); struct xfs_log_iovec *vecp = NULL; - ASSERT(rudp->rud_next_extent == rudp->rud_format.rud_nextents); - rudp->rud_format.rud_type = XFS_LI_RUD; rudp->rud_format.rud_size = 1; xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUD_FORMAT, &rudp->rud_format, - xfs_rud_item_sizeof(rudp)); + sizeof(struct xfs_rud_log_format)); } /* @@ -374,7 +349,7 @@ xfs_rud_item_unlock( if (lip->li_flags & XFS_LI_ABORTED) { xfs_rui_release(rudp->rud_ruip); - xfs_rud_item_free(rudp); + kmem_zone_free(xfs_rud_zone, rudp); } } @@ -398,7 +373,7 @@ xfs_rud_item_committed( * aborted due to log I/O error). */ xfs_rui_release(rudp->rud_ruip); - xfs_rud_item_free(rudp); + kmem_zone_free(xfs_rud_zone, rudp); return (xfs_lsn_t)-1; } @@ -437,25 +412,14 @@ static const struct xfs_item_ops xfs_rud_item_ops = { struct xfs_rud_log_item * xfs_rud_init( struct xfs_mount *mp, - struct xfs_rui_log_item *ruip, - uint nextents) + struct xfs_rui_log_item *ruip) { struct xfs_rud_log_item *rudp; - uint size; - - ASSERT(nextents > 0); - if (nextents > XFS_RUD_MAX_FAST_EXTENTS) { - size = (uint)(sizeof(struct xfs_rud_log_item) + - ((nextents - 1) * sizeof(struct xfs_map_extent))); - rudp = kmem_zalloc(size, KM_SLEEP); - } else { - rudp = kmem_zone_zalloc(xfs_rud_zone, KM_SLEEP); - } + rudp = kmem_zone_zalloc(xfs_rud_zone, KM_SLEEP); xfs_log_item_init(mp, &rudp->rud_item, XFS_LI_RUD, &xfs_rud_item_ops); rudp->rud_ruip = ruip; - rudp->rud_format.rud_nextents = nextents; rudp->rud_format.rud_rui_id = ruip->rui_format.rui_id; return rudp; @@ -523,7 +487,7 @@ xfs_rui_recover( error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); if (error) return error; - rudp = xfs_trans_get_rud(tp, ruip, ruip->rui_format.rui_nextents); + rudp = xfs_trans_get_rud(tp, ruip); for (i = 0; i < ruip->rui_format.rui_nextents; i++) { rmap = &(ruip->rui_format.rui_extents[i]); diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h index 59ef3ecaa493..aefcc3a318a5 100644 --- a/fs/xfs/xfs_rmap_item.h +++ b/fs/xfs/xfs_rmap_item.h @@ -77,21 +77,15 @@ struct xfs_rui_log_item { struct xfs_rud_log_item { struct xfs_log_item rud_item; struct xfs_rui_log_item *rud_ruip; - uint rud_next_extent; struct xfs_rud_log_format rud_format; }; -/* - * Max number of extents in fast allocation path. - */ -#define XFS_RUD_MAX_FAST_EXTENTS 16 - extern struct kmem_zone *xfs_rui_zone; extern struct kmem_zone *xfs_rud_zone; struct xfs_rui_log_item *xfs_rui_init(struct xfs_mount *, uint); struct xfs_rud_log_item *xfs_rud_init(struct xfs_mount *, - struct xfs_rui_log_item *, uint); + struct xfs_rui_log_item *); int xfs_rui_copy_format(struct xfs_log_iovec *buf, struct xfs_rui_log_format *dst_rui_fmt); void xfs_rui_item_free(struct xfs_rui_log_item *); diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 45773df1dd0b..24ef83ef04de 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1770,9 +1770,7 @@ xfs_init_zones(void) if (!xfs_icreate_zone) goto out_destroy_ili_zone; - xfs_rud_zone = kmem_zone_init((sizeof(struct xfs_rud_log_item) + - ((XFS_RUD_MAX_FAST_EXTENTS - 1) * - sizeof(struct xfs_map_extent))), + xfs_rud_zone = kmem_zone_init(sizeof(struct xfs_rud_log_item), "xfs_rud_item"); if (!xfs_rud_zone) goto out_destroy_icreate_zone; diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index bb4b84f9347e..e2bf86aad33d 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -241,7 +241,7 @@ enum xfs_rmap_intent_type; void xfs_rmap_update_init_defer_op(void); struct xfs_rud_log_item *xfs_trans_get_rud(struct xfs_trans *tp, - struct xfs_rui_log_item *ruip, uint nextents); + struct xfs_rui_log_item *ruip); int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp, struct xfs_rud_log_item *rudp, enum xfs_rmap_intent_type type, __uint64_t owner, int whichfork, xfs_fileoff_t startoff, diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c index 83414764e1ab..35650d6b2606 100644 --- a/fs/xfs/xfs_trans_rmap.c +++ b/fs/xfs/xfs_trans_rmap.c @@ -129,29 +129,14 @@ xfs_trans_log_start_rmap_update( xfs_trans_set_rmap_flags(rmap, type, whichfork, state); } -/* - * This routine is called to allocate an "rmap update done" - * log item that will hold nextents worth of extents. The - * caller must use all nextents extents, because we are not - * flexible about this at all. - */ struct xfs_rud_log_item * xfs_trans_get_rud( struct xfs_trans *tp, - struct xfs_rui_log_item *ruip, - uint nextents) + struct xfs_rui_log_item *ruip) { struct xfs_rud_log_item *rudp; - ASSERT(tp != NULL); - ASSERT(nextents > 0); - - rudp = xfs_rud_init(tp->t_mountp, ruip, nextents); - ASSERT(rudp != NULL); - - /* - * Get a log_item_desc to point at the new item. - */ + rudp = xfs_rud_init(tp->t_mountp, ruip); xfs_trans_add_item(tp, &rudp->rud_item); return rudp; } @@ -174,8 +159,6 @@ xfs_trans_log_finish_rmap_update( xfs_exntst_t state, struct xfs_btree_cur **pcur) { - uint next_extent; - struct xfs_map_extent *rmap; int error; error = xfs_rmap_finish_one(tp, type, owner, whichfork, startoff, @@ -191,16 +174,6 @@ xfs_trans_log_finish_rmap_update( tp->t_flags |= XFS_TRANS_DIRTY; rudp->rud_item.li_desc->lid_flags |= XFS_LID_DIRTY; - next_extent = rudp->rud_next_extent; - ASSERT(next_extent < rudp->rud_format.rud_nextents); - rmap = &(rudp->rud_format.rud_extents[next_extent]); - rmap->me_owner = owner; - rmap->me_startblock = startblock; - rmap->me_startoff = startoff; - rmap->me_len = blockcount; - xfs_trans_set_rmap_flags(rmap, type, whichfork, state); - rudp->rud_next_extent++; - return error; } @@ -255,7 +228,7 @@ xfs_rmap_update_create_done( void *intent, unsigned int count) { - return xfs_trans_get_rud(tp, intent, count); + return xfs_trans_get_rud(tp, intent); } /* Process a deferred rmap update. */ -- cgit v1.2.3 From e127fafd1d3c46532903d530dfa69822d3d35d8a Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:29:32 +1000 Subject: xfs: remove unnecessary parentheses from log redo item recovery functions Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_extfree_item.c | 12 ++++++------ fs/xfs/xfs_rmap_item.c | 12 ++++++------ 2 files changed, 12 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index a8c461735ecb..d7bc14906af8 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -515,13 +515,13 @@ xfs_efi_recover( * just toss the EFI. */ for (i = 0; i < efip->efi_format.efi_nextents; i++) { - extp = &(efip->efi_format.efi_extents[i]); + extp = &efip->efi_format.efi_extents[i]; startblock_fsb = XFS_BB_TO_FSB(mp, XFS_FSB_TO_DADDR(mp, extp->ext_start)); - if ((startblock_fsb == 0) || - (extp->ext_len == 0) || - (startblock_fsb >= mp->m_sb.sb_dblocks) || - (extp->ext_len >= mp->m_sb.sb_agblocks)) { + if (startblock_fsb == 0 || + extp->ext_len == 0 || + startblock_fsb >= mp->m_sb.sb_dblocks || + extp->ext_len >= mp->m_sb.sb_agblocks) { /* * This will pull the EFI from the AIL and * free the memory associated with it. @@ -539,7 +539,7 @@ xfs_efi_recover( xfs_rmap_skip_owner_update(&oinfo); for (i = 0; i < efip->efi_format.efi_nextents; i++) { - extp = &(efip->efi_format.efi_extents[i]); + extp = &efip->efi_format.efi_extents[i]; error = xfs_trans_free_extent(tp, efdp, extp->ext_start, extp->ext_len, &oinfo); if (error) diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 7e4743219430..2500f28689d5 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -454,7 +454,7 @@ xfs_rui_recover( * just toss the RUI. */ for (i = 0; i < ruip->rui_format.rui_nextents; i++) { - rmap = &(ruip->rui_format.rui_extents[i]); + rmap = &ruip->rui_format.rui_extents[i]; startblock_fsb = XFS_BB_TO_FSB(mp, XFS_FSB_TO_DADDR(mp, rmap->me_startblock)); switch (rmap->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) { @@ -469,10 +469,10 @@ xfs_rui_recover( op_ok = false; break; } - if (!op_ok || (startblock_fsb == 0) || - (rmap->me_len == 0) || - (startblock_fsb >= mp->m_sb.sb_dblocks) || - (rmap->me_len >= mp->m_sb.sb_agblocks) || + if (!op_ok || startblock_fsb == 0 || + rmap->me_len == 0 || + startblock_fsb >= mp->m_sb.sb_dblocks || + rmap->me_len >= mp->m_sb.sb_agblocks || (rmap->me_flags & ~XFS_RMAP_EXTENT_FLAGS)) { /* * This will pull the RUI from the AIL and @@ -490,7 +490,7 @@ xfs_rui_recover( rudp = xfs_trans_get_rud(tp, ruip); for (i = 0; i < ruip->rui_format.rui_nextents; i++) { - rmap = &(ruip->rui_format.rui_extents[i]); + rmap = &ruip->rui_format.rui_extents[i]; state = (rmap->me_flags & XFS_RMAP_EXTENT_UNWRITTEN) ? XFS_EXT_UNWRITTEN : XFS_EXT_NORM; whichfork = (rmap->me_flags & XFS_RMAP_EXTENT_ATTR_FORK) ? -- cgit v1.2.3 From 51ce9d000c0a80bddf9a93a4232283cc9a4564c0 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:30:31 +1000 Subject: xfs: collapse single use static functions Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/xfs_trans_extfree.c | 88 +++++++++++++------------------------- fs/xfs/xfs_trans_rmap.c | 103 +++++++++++++++------------------------------ 2 files changed, 63 insertions(+), 128 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index 3aca3a7578c2..7d1e84c1b848 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -29,61 +29,6 @@ #include "xfs_alloc.h" #include "xfs_bmap.h" -/* - * This routine is called to allocate an "extent free intention" - * log item that will hold nextents worth of extents. The - * caller must use all nextents extents, because we are not - * flexible about this at all. - */ -STATIC struct xfs_efi_log_item * -xfs_trans_get_efi(struct xfs_trans *tp, - uint nextents) -{ - struct xfs_efi_log_item *efip; - - ASSERT(tp != NULL); - ASSERT(nextents > 0); - - efip = xfs_efi_init(tp->t_mountp, nextents); - ASSERT(efip != NULL); - - /* - * Get a log_item_desc to point at the new item. - */ - xfs_trans_add_item(tp, &efip->efi_item); - return efip; -} - -/* - * This routine is called to indicate that the described - * extent is to be logged as needing to be freed. It should - * be called once for each extent to be freed. - */ -STATIC void -xfs_trans_log_efi_extent(struct xfs_trans *tp, - struct xfs_efi_log_item *efip, - xfs_fsblock_t start_block, - xfs_extlen_t ext_len) -{ - uint next_extent; - struct xfs_extent *extp; - - tp->t_flags |= XFS_TRANS_DIRTY; - efip->efi_item.li_desc->lid_flags |= XFS_LID_DIRTY; - - /* - * atomic_inc_return gives us the value after the increment; - * we want to use it as an array index so we need to subtract 1 from - * it. - */ - next_extent = atomic_inc_return(&efip->efi_next_extent) - 1; - ASSERT(next_extent < efip->efi_format.efi_nextents); - extp = &(efip->efi_format.efi_extents[next_extent]); - extp->ext_start = start_block; - extp->ext_len = ext_len; -} - - /* * This routine is called to allocate an "extent free done" * log item that will hold nextents worth of extents. The @@ -172,7 +117,19 @@ xfs_extent_free_create_intent( struct xfs_trans *tp, unsigned int count) { - return xfs_trans_get_efi(tp, count); + struct xfs_efi_log_item *efip; + + ASSERT(tp != NULL); + ASSERT(count > 0); + + efip = xfs_efi_init(tp->t_mountp, count); + ASSERT(efip != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &efip->efi_item); + return efip; } /* Log a free extent to the intent item. */ @@ -182,11 +139,26 @@ xfs_extent_free_log_item( void *intent, struct list_head *item) { + struct xfs_efi_log_item *efip = intent; struct xfs_extent_free_item *free; + uint next_extent; + struct xfs_extent *extp; free = container_of(item, struct xfs_extent_free_item, xefi_list); - xfs_trans_log_efi_extent(tp, intent, free->xefi_startblock, - free->xefi_blockcount); + + tp->t_flags |= XFS_TRANS_DIRTY; + efip->efi_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + /* + * atomic_inc_return gives us the value after the increment; + * we want to use it as an array index so we need to subtract 1 from + * it. + */ + next_extent = atomic_inc_return(&efip->efi_next_extent) - 1; + ASSERT(next_extent < efip->efi_format.efi_nextents); + extp = &efip->efi_format.efi_extents[next_extent]; + extp->ext_start = free->xefi_startblock; + extp->ext_len = free->xefi_blockcount; } /* Get an EFD so we can process all the free extents. */ diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c index 35650d6b2606..5a50ef881568 100644 --- a/fs/xfs/xfs_trans_rmap.c +++ b/fs/xfs/xfs_trans_rmap.c @@ -31,32 +31,6 @@ #include "xfs_alloc.h" #include "xfs_rmap.h" -/* - * This routine is called to allocate an "rmap update intent" - * log item that will hold nextents worth of extents. The - * caller must use all nextents extents, because we are not - * flexible about this at all. - */ -STATIC struct xfs_rui_log_item * -xfs_trans_get_rui( - struct xfs_trans *tp, - uint nextents) -{ - struct xfs_rui_log_item *ruip; - - ASSERT(tp != NULL); - ASSERT(nextents > 0); - - ruip = xfs_rui_init(tp->t_mountp, nextents); - ASSERT(ruip != NULL); - - /* - * Get a log_item_desc to point at the new item. - */ - xfs_trans_add_item(tp, &ruip->rui_item); - return ruip; -} - /* Set the map extent flags for this reverse mapping. */ static void xfs_trans_set_rmap_flags( @@ -91,44 +65,6 @@ xfs_trans_set_rmap_flags( } } -/* - * This routine is called to indicate that the described reverse - * mapping is to be logged as needing to be updated. It should be - * called once for each mapping. - */ -STATIC void -xfs_trans_log_start_rmap_update( - struct xfs_trans *tp, - struct xfs_rui_log_item *ruip, - enum xfs_rmap_intent_type type, - __uint64_t owner, - int whichfork, - xfs_fileoff_t startoff, - xfs_fsblock_t startblock, - xfs_filblks_t blockcount, - xfs_exntst_t state) -{ - uint next_extent; - struct xfs_map_extent *rmap; - - tp->t_flags |= XFS_TRANS_DIRTY; - ruip->rui_item.li_desc->lid_flags |= XFS_LID_DIRTY; - - /* - * atomic_inc_return gives us the value after the increment; - * we want to use it as an array index so we need to subtract 1 from - * it. - */ - next_extent = atomic_inc_return(&ruip->rui_next_extent) - 1; - ASSERT(next_extent < ruip->rui_format.rui_nextents); - rmap = &(ruip->rui_format.rui_extents[next_extent]); - rmap->me_owner = owner; - rmap->me_startblock = startblock; - rmap->me_startoff = startoff; - rmap->me_len = blockcount; - xfs_trans_set_rmap_flags(rmap, type, whichfork, state); -} - struct xfs_rud_log_item * xfs_trans_get_rud( struct xfs_trans *tp, @@ -200,7 +136,19 @@ xfs_rmap_update_create_intent( struct xfs_trans *tp, unsigned int count) { - return xfs_trans_get_rui(tp, count); + struct xfs_rui_log_item *ruip; + + ASSERT(tp != NULL); + ASSERT(count > 0); + + ruip = xfs_rui_init(tp->t_mountp, count); + ASSERT(ruip != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &ruip->rui_item); + return ruip; } /* Log rmap updates in the intent item. */ @@ -210,14 +158,29 @@ xfs_rmap_update_log_item( void *intent, struct list_head *item) { + struct xfs_rui_log_item *ruip = intent; struct xfs_rmap_intent *rmap; + uint next_extent; + struct xfs_map_extent *map; rmap = container_of(item, struct xfs_rmap_intent, ri_list); - xfs_trans_log_start_rmap_update(tp, intent, rmap->ri_type, - rmap->ri_owner, rmap->ri_whichfork, - rmap->ri_bmap.br_startoff, - rmap->ri_bmap.br_startblock, - rmap->ri_bmap.br_blockcount, + + tp->t_flags |= XFS_TRANS_DIRTY; + ruip->rui_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + /* + * atomic_inc_return gives us the value after the increment; + * we want to use it as an array index so we need to subtract 1 from + * it. + */ + next_extent = atomic_inc_return(&ruip->rui_next_extent) - 1; + ASSERT(next_extent < ruip->rui_format.rui_nextents); + map = &ruip->rui_format.rui_extents[next_extent]; + map->me_owner = rmap->ri_owner; + map->me_startblock = rmap->ri_bmap.br_startblock; + map->me_startoff = rmap->ri_bmap.br_startoff; + map->me_len = rmap->ri_bmap.br_blockcount; + xfs_trans_set_rmap_flags(map, rmap->ri_type, rmap->ri_whichfork, rmap->ri_bmap.br_state); } -- cgit v1.2.3 From 3481b68285238054be519ad0c8cad5cc2425e26c Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Wed, 3 Aug 2016 12:31:07 +1000 Subject: xfs: move (and rename) the deferred bmap-free tracepoints Rename the deferred bmap-free to extent_free and make them only trigger when we're really running deferred ops. Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Signed-off-by: Dave Chinner --- fs/xfs/libxfs/xfs_alloc.c | 2 -- fs/xfs/xfs_trans_extfree.c | 7 +++++++ 2 files changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index a0dc2999778d..776ae2f325d1 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -2819,8 +2819,6 @@ xfs_free_extent( ASSERT(len != 0); - trace_xfs_bmap_free_deferred(mp, agno, 0, agbno, len); - if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_FREE_EXTENT, XFS_RANDOM_FREE_EXTENT)) diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index 7d1e84c1b848..459ddec137a4 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -21,6 +21,7 @@ #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_bit.h" #include "xfs_mount.h" #include "xfs_defer.h" #include "xfs_trans.h" @@ -28,6 +29,7 @@ #include "xfs_extfree_item.h" #include "xfs_alloc.h" #include "xfs_bmap.h" +#include "xfs_trace.h" /* * This routine is called to allocate an "extent free done" @@ -68,10 +70,15 @@ xfs_trans_free_extent( xfs_extlen_t ext_len, struct xfs_owner_info *oinfo) { + struct xfs_mount *mp = tp->t_mountp; uint next_extent; + xfs_agnumber_t agno = XFS_FSB_TO_AGNO(mp, start_block); + xfs_agblock_t agbno = XFS_FSB_TO_AGBNO(mp, start_block); struct xfs_extent *extp; int error; + trace_xfs_bmap_free_deferred(tp->t_mountp, agno, 0, agbno, ext_len); + error = xfs_free_extent(tp, start_block, ext_len, oinfo); /* -- cgit v1.2.3 From c1892c37769cf89c7e7ba57528ae2ccb5d153c9b Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 3 Aug 2016 13:44:27 +0200 Subject: vfs: fix deadlock in file_remove_privs() on overlayfs file_remove_privs() is called with inode lock on file_inode(), which proceeds to calling notify_change() on file->f_path.dentry. Which triggers the WARN_ON_ONCE(!inode_is_locked(inode)) in addition to deadlocking later when ovl_setattr tries to lock the underlying inode again. Fix this mess by not mixing the layers, but doing everything on underlying dentry/inode. Signed-off-by: Miklos Szeredi Fixes: 07a2daab49c5 ("ovl: Copy up underlying inode's ->i_mode to overlay inode") Cc: --- fs/inode.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 4ccbc21b30ce..68db39050446 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1740,8 +1740,8 @@ static int __remove_privs(struct dentry *dentry, int kill) */ int file_remove_privs(struct file *file) { - struct dentry *dentry = file->f_path.dentry; - struct inode *inode = d_inode(dentry); + struct dentry *dentry = file_dentry(file); + struct inode *inode = file_inode(file); int kill; int error = 0; @@ -1749,7 +1749,7 @@ int file_remove_privs(struct file *file) if (IS_NOSEC(inode)) return 0; - kill = file_needs_remove_privs(file); + kill = dentry_needs_remove_privs(dentry); if (kill < 0) return kill; if (kill) -- cgit v1.2.3 From f0fce87c36aec5c4a895c78b76396f5727047b93 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 3 Aug 2016 13:44:27 +0200 Subject: vfs: make dentry_needs_remove_privs() internal Only used by the vfs. Signed-off-by: Miklos Szeredi --- fs/inode.c | 1 - fs/internal.h | 1 + include/linux/fs.h | 1 - 3 files changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 68db39050446..2029357fcf97 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1720,7 +1720,6 @@ int dentry_needs_remove_privs(struct dentry *dentry) mask |= ATTR_KILL_PRIV; return mask; } -EXPORT_SYMBOL(dentry_needs_remove_privs); static int __remove_privs(struct dentry *dentry, int kill) { diff --git a/fs/internal.h b/fs/internal.h index f57ced528cde..30b568f02613 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -114,6 +114,7 @@ extern int vfs_open(const struct path *, struct file *, const struct cred *); */ extern long prune_icache_sb(struct super_block *sb, struct shrink_control *sc); extern void inode_add_lru(struct inode *inode); +extern int dentry_needs_remove_privs(struct dentry *dentry); /* * fs-writeback.c diff --git a/include/linux/fs.h b/include/linux/fs.h index 3888341913ef..448641bcafc5 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2725,7 +2725,6 @@ extern struct inode *new_inode(struct super_block *sb); extern void free_inode_nonrcu(struct inode *inode); extern int should_remove_suid(struct dentry *); extern int file_remove_privs(struct file *); -extern int dentry_needs_remove_privs(struct dentry *dentry); extern void __insert_inode_hash(struct inode *, unsigned long hashval); static inline void insert_inode_hash(struct inode *inode) -- cgit v1.2.3 From 797fc16d8f00bf681b7f9969c0701725b472b11a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 3 Aug 2016 11:02:48 -0400 Subject: 9p: fix braino introduced in "9p: new helper - v9fs_parent_fid()" In v9fs_vfs_rename() we need to clone the parents' fids, not just find them. Spotted-by: Johannes Berg Signed-off-by: Al Viro --- fs/9p/fid.h | 4 ++++ fs/9p/vfs_inode.c | 4 ++-- 2 files changed, 6 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/9p/fid.h b/fs/9p/fid.h index 12700df0bb51..d4c5002d56ad 100644 --- a/fs/9p/fid.h +++ b/fs/9p/fid.h @@ -31,4 +31,8 @@ static inline struct p9_fid *v9fs_parent_fid(struct dentry *dentry) struct p9_fid *v9fs_fid_clone(struct dentry *dentry); void v9fs_fid_add(struct dentry *dentry, struct p9_fid *fid); struct p9_fid *v9fs_writeback_fid(struct dentry *dentry); +static inline struct p9_fid *clone_fid(struct p9_fid *fid) +{ + return IS_ERR(fid) ? fid : p9_client_walk(fid, 0, NULL, 1); +} #endif diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 7da9a8354fad..26189eebcd1c 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -975,13 +975,13 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (IS_ERR(oldfid)) return PTR_ERR(oldfid); - olddirfid = v9fs_parent_fid(old_dentry); + olddirfid = clone_fid(v9fs_parent_fid(old_dentry)); if (IS_ERR(olddirfid)) { retval = PTR_ERR(olddirfid); goto done; } - newdirfid = v9fs_parent_fid(new_dentry); + newdirfid = clone_fid(v9fs_parent_fid(new_dentry)); if (IS_ERR(newdirfid)) { retval = PTR_ERR(newdirfid); goto clunk_olddir; -- cgit v1.2.3 From 7d50a29fe43848b7d3c8309f8617db3305c0db18 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 3 Aug 2016 11:12:12 -0400 Subject: 9p: use clone_fid() in a bunch of places it cleans the things up Signed-off-by: Al Viro --- fs/9p/fid.c | 26 +------------------------- fs/9p/fid.h | 5 ++++- fs/9p/vfs_inode.c | 2 +- fs/9p/vfs_inode_dotl.c | 2 +- fs/9p/xattr.c | 4 +--- 5 files changed, 8 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/9p/fid.c b/fs/9p/fid.c index 47db55aee7f2..60fb47469c86 100644 --- a/fs/9p/fid.c +++ b/fs/9p/fid.c @@ -257,36 +257,12 @@ struct p9_fid *v9fs_fid_lookup(struct dentry *dentry) return v9fs_fid_lookup_with_uid(dentry, uid, any); } -struct p9_fid *v9fs_fid_clone(struct dentry *dentry) -{ - struct p9_fid *fid, *ret; - - fid = v9fs_fid_lookup(dentry); - if (IS_ERR(fid)) - return fid; - - ret = p9_client_walk(fid, 0, NULL, 1); - return ret; -} - -static struct p9_fid *v9fs_fid_clone_with_uid(struct dentry *dentry, kuid_t uid) -{ - struct p9_fid *fid, *ret; - - fid = v9fs_fid_lookup_with_uid(dentry, uid, 0); - if (IS_ERR(fid)) - return fid; - - ret = p9_client_walk(fid, 0, NULL, 1); - return ret; -} - struct p9_fid *v9fs_writeback_fid(struct dentry *dentry) { int err; struct p9_fid *fid; - fid = v9fs_fid_clone_with_uid(dentry, GLOBAL_ROOT_UID); + fid = clone_fid(v9fs_fid_lookup_with_uid(dentry, GLOBAL_ROOT_UID, 0)); if (IS_ERR(fid)) goto error_out; /* diff --git a/fs/9p/fid.h b/fs/9p/fid.h index d4c5002d56ad..4491bcaf42b8 100644 --- a/fs/9p/fid.h +++ b/fs/9p/fid.h @@ -28,11 +28,14 @@ static inline struct p9_fid *v9fs_parent_fid(struct dentry *dentry) { return v9fs_fid_lookup(dentry->d_parent); } -struct p9_fid *v9fs_fid_clone(struct dentry *dentry); void v9fs_fid_add(struct dentry *dentry, struct p9_fid *fid); struct p9_fid *v9fs_writeback_fid(struct dentry *dentry); static inline struct p9_fid *clone_fid(struct p9_fid *fid) { return IS_ERR(fid) ? fid : p9_client_walk(fid, 0, NULL, 1); } +static inline struct p9_fid *v9fs_fid_clone(struct dentry *dentry) +{ + return clone_fid(v9fs_fid_lookup(dentry)); +} #endif diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 26189eebcd1c..8b1999b528e9 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -661,7 +661,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir, } /* clone a fid to use for creation */ - ofid = p9_client_walk(dfid, 0, NULL, 1); + ofid = clone_fid(dfid); if (IS_ERR(ofid)) { err = PTR_ERR(ofid); p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err); diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index 2ed04c2fe7af..eeabcb0bad12 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -281,7 +281,7 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry, } /* clone a fid to use for creation */ - ofid = p9_client_walk(dfid, 0, NULL, 1); + ofid = clone_fid(dfid); if (IS_ERR(ofid)) { err = PTR_ERR(ofid); p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err); diff --git a/fs/9p/xattr.c b/fs/9p/xattr.c index a6bd349bab23..f329eee6dc93 100644 --- a/fs/9p/xattr.c +++ b/fs/9p/xattr.c @@ -97,8 +97,6 @@ int v9fs_xattr_set(struct dentry *dentry, const char *name, const void *value, size_t value_len, int flags) { struct p9_fid *fid = v9fs_fid_lookup(dentry); - if (IS_ERR(fid)) - return PTR_ERR(fid); return v9fs_fid_xattr_set(fid, name, value, value_len, flags); } @@ -115,7 +113,7 @@ int v9fs_fid_xattr_set(struct p9_fid *fid, const char *name, name, value_len, flags); /* Clone it */ - fid = p9_client_walk(fid, 0, NULL, 1); + fid = clone_fid(fid); if (IS_ERR(fid)) return PTR_ERR(fid); -- cgit v1.2.3 From 4b2e0162e4ab1a976eec0a36ca0bd920b02bfff3 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Wed, 3 Aug 2016 17:33:32 +0200 Subject: fs/proc: Add compiler check for -Wno-override-init to support gcc < 4.2 With gcc < 4.2 (e.g. 4.1.2): CC fs/proc/task_mmu.o cc1: error: unrecognized command line option "-Wno-override-init" To fix this, only enable the compiler option when it is actually supported by the compiler. Fixes: ca52953f5f24 ("fs/proc/task_mmu.c: suppress compilation warnings with W=1") Signed-off-by: Geert Uytterhoeven Acked-by: Valdis Kletnieks Cc: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/proc/Makefile b/fs/proc/Makefile index a8c13605b434..12c6922c913c 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -4,7 +4,7 @@ obj-y += proc.o -CFLAGS_task_mmu.o += -Wno-override-init +CFLAGS_task_mmu.o += $(call cc-option,-Wno-override-init,) proc-y := nommu.o task_nommu.o proc-$(CONFIG_MMU) := task_mmu.o -- cgit v1.2.3 From db20a8925bc420eed033c5d91ff6afa74465e521 Mon Sep 17 00:00:00 2001 From: David Howells Date: Wed, 3 Aug 2016 17:57:39 +0100 Subject: cachefiles: Fix race between inactivating and culling a cache object There's a race between cachefiles_mark_object_inactive() and cachefiles_cull(): (1) cachefiles_cull() can't delete a backing file until the cache object is marked inactive, but as soon as that's the case it's fair game. (2) cachefiles_mark_object_inactive() marks the object as being inactive and *only then* reads the i_blocks on the backing inode - but cachefiles_cull() might've managed to delete it by this point. Fix this by making sure cachefiles_mark_object_inactive() gets any data it needs from the backing inode before deactivating the object. Without this, the following oops may occur: BUG: unable to handle kernel NULL pointer dereference at 0000000000000098 IP: [] cachefiles_mark_object_inactive+0x61/0xb0 [cachefiles] ... CPU: 11 PID: 527 Comm: kworker/u64:4 Tainted: G I ------------ 3.10.0-470.el7.x86_64 #1 Hardware name: Hewlett-Packard HP Z600 Workstation/0B54h, BIOS 786G4 v03.19 03/11/2011 Workqueue: fscache_object fscache_object_work_func [fscache] task: ffff880035edaf10 ti: ffff8800b77c0000 task.ti: ffff8800b77c0000 RIP: 0010:[] cachefiles_mark_object_inactive+0x61/0xb0 [cachefiles] RSP: 0018:ffff8800b77c3d70 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff8800bf6cc400 RCX: 0000000000000034 RDX: 0000000000000000 RSI: ffff880090ffc710 RDI: ffff8800bf761ef8 RBP: ffff8800b77c3d88 R08: 2000000000000000 R09: 0090ffc710000000 R10: ff51005d2ff1c400 R11: 0000000000000000 R12: ffff880090ffc600 R13: ffff8800bf6cc520 R14: ffff8800bf6cc400 R15: ffff8800bf6cc498 FS: 0000000000000000(0000) GS:ffff8800bb8c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 0000000000000098 CR3: 00000000019ba000 CR4: 00000000000007e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Stack: ffff880090ffc600 ffff8800bf6cc400 ffff8800867df140 ffff8800b77c3db0 ffffffffa06c48cb ffff880090ffc600 ffff880090ffc180 ffff880090ffc658 ffff8800b77c3df0 ffffffffa085d846 ffff8800a96b8150 ffff880090ffc600 Call Trace: [] cachefiles_drop_object+0x6b/0xf0 [cachefiles] [] fscache_drop_object+0xd6/0x1e0 [fscache] [] fscache_object_work_func+0xa5/0x200 [fscache] [] process_one_work+0x17b/0x470 [] worker_thread+0x126/0x410 [] ? rescuer_thread+0x460/0x460 [] kthread+0xcf/0xe0 [] ? kthread_create_on_node+0x140/0x140 [] ret_from_fork+0x58/0x90 [] ? kthread_create_on_node+0x140/0x140 The oopsing code shows: callq 0xffffffff810af6a0 mov 0xf8(%r12),%rax mov 0x30(%rax),%rax mov 0x98(%rax),%rax <---- oops here lock add %rax,0x130(%rbx) where this is: d_backing_inode(object->dentry)->i_blocks Fixes: a5b3a80b899bda0f456f1246c4c5a1191ea01519 (CacheFiles: Provide read-and-reset release counters for cachefilesd) Reported-by: Jianhong Yin Signed-off-by: David Howells Reviewed-by: Jeff Layton Reviewed-by: Steve Dickson cc: stable@vger.kernel.org Signed-off-by: Al Viro --- fs/cachefiles/namei.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cachefiles/namei.c b/fs/cachefiles/namei.c index 4ae75006e73b..3f7c2cd41f8f 100644 --- a/fs/cachefiles/namei.c +++ b/fs/cachefiles/namei.c @@ -263,6 +263,8 @@ requeue: void cachefiles_mark_object_inactive(struct cachefiles_cache *cache, struct cachefiles_object *object) { + blkcnt_t i_blocks = d_backing_inode(object->dentry)->i_blocks; + write_lock(&cache->active_lock); rb_erase(&object->active_node, &cache->active_nodes); clear_bit(CACHEFILES_OBJECT_ACTIVE, &object->flags); @@ -273,8 +275,7 @@ void cachefiles_mark_object_inactive(struct cachefiles_cache *cache, /* This object can now be culled, so we need to let the daemon know * that there is something it can remove if it needs to. */ - atomic_long_add(d_backing_inode(object->dentry)->i_blocks, - &cache->b_released); + atomic_long_add(i_blocks, &cache->b_released); if (atomic_inc_return(&cache->f_released)) cachefiles_state_changed(cache); } -- cgit v1.2.3 From 42049bf60db4c01e0432fc861463dcd3208c0d93 Mon Sep 17 00:00:00 2001 From: Chris Mason Date: Wed, 3 Aug 2016 14:05:46 -0700 Subject: Btrfs: fix __MAX_CSUM_ITEMS Jeff Mahoney's cleanup commit (14a1e067b4) wasn't correct for csums on machines where the pagesize >= metadata blocksize. This just reverts the relevant hunks to bring the old math back. Signed-off-by: Chris Mason --- fs/btrfs/file-item.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index 2fc803da4680..d0d571c47d33 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -27,8 +27,9 @@ #include "print-tree.h" #include "compression.h" -#define __MAX_CSUM_ITEMS(r, size) \ - ((unsigned long)(((BTRFS_MAX_ITEM_SIZE(r) * 2) / size) - 1)) +#define __MAX_CSUM_ITEMS(r, size) ((unsigned long)(((BTRFS_LEAF_DATA_SIZE(r) - \ + sizeof(struct btrfs_item) * 2) / \ + size) - 1)) #define MAX_CSUM_ITEMS(r, size) (min_t(u32, __MAX_CSUM_ITEMS(r, size), \ PAGE_SIZE)) -- cgit v1.2.3 From 8a545f185145e3c09348cd74326268ecfc6715a3 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Wed, 13 Jul 2016 13:12:34 +0300 Subject: hostfs: Freeing an ERR_PTR in hostfs_fill_sb_common() We can't pass error pointers to kfree() or it causes an oops. Fixes: 52b209f7b848 ('get rid of hostfs_read_inode()') Signed-off-by: Dan Carpenter Signed-off-by: Richard Weinberger --- fs/hostfs/hostfs_kern.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index 5c57654927a6..90e46cd752fe 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -959,10 +959,11 @@ static int hostfs_fill_sb_common(struct super_block *sb, void *d, int silent) if (S_ISLNK(root_inode->i_mode)) { char *name = follow_link(host_root_path); - if (IS_ERR(name)) + if (IS_ERR(name)) { err = PTR_ERR(name); - else - err = read_name(root_inode, name); + goto out_put; + } + err = read_name(root_inode, name); kfree(name); if (err) goto out_put; -- cgit v1.2.3 From 99a01cdf9d572e91c2805a2b89d459797864e3ca Mon Sep 17 00:00:00 2001 From: Ross Zwisler Date: Wed, 3 Aug 2016 13:46:15 -0700 Subject: block: remove BLK_DEV_DAX config option The functionality for block device DAX was already removed with commit acc93d30d7d4 ("Revert "block: enable dax for raw block devices"") However, we still had a config option hanging around that was always disabled because it depended on CONFIG_BROKEN. This config option was introduced in commit 03cdadb04077 ("block: disable block device DAX by default") This change reverts that commit, removing the dead config option. Link: http://lkml.kernel.org/r/20160729182314.6368-1-ross.zwisler@linux.intel.com Signed-off-by: Ross Zwisler Cc: Dave Hansen Acked-by: Dan Williams Cc: Jens Axboe Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- block/Kconfig | 13 ------------- fs/block_dev.c | 6 +----- 2 files changed, 1 insertion(+), 18 deletions(-) (limited to 'fs') diff --git a/block/Kconfig b/block/Kconfig index 0363cd731320..161491d0a879 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -88,19 +88,6 @@ config BLK_DEV_INTEGRITY T10/SCSI Data Integrity Field or the T13/ATA External Path Protection. If in doubt, say N. -config BLK_DEV_DAX - bool "Block device DAX support" - depends on FS_DAX - depends on BROKEN - help - When DAX support is available (CONFIG_FS_DAX) raw block - devices can also support direct userspace access to the - storage capacity via MMAP(2) similar to a file on a - DAX-enabled filesystem. However, the DAX I/O-path disables - some standard I/O-statistics, and the MMAP(2) path has some - operational differences due to bypassing the page - cache. If in doubt, say N. - config BLK_DEV_THROTTLING bool "Block layer bio throttling support" depends on BLK_CGROUP=y diff --git a/fs/block_dev.c b/fs/block_dev.c index ada42cf42d06..2033a3f91d58 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1275,11 +1275,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) bdev->bd_disk = disk; bdev->bd_queue = disk->queue; bdev->bd_contains = bdev; - if (IS_ENABLED(CONFIG_BLK_DEV_DAX) && - blk_queue_dax(disk->queue)) - bdev->bd_inode->i_flags = S_DAX; - else - bdev->bd_inode->i_flags = 0; + bdev->bd_inode->i_flags = 0; if (!partno) { ret = -ENXIO; -- cgit v1.2.3 From dc5ff2b1d66f21c27a4c37236636dff6946437e4 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 26 Jul 2016 11:38:20 +0200 Subject: writeback: Write dirty times for WB_SYNC_ALL writeback Currently we take care to handle I_DIRTY_TIME in vfs_fsync() and queue_io() so that inodes which have only dirty timestamps are properly written on fsync(2) and sync(2). However there are other call sites - most notably going through write_inode_now() - which expect inode to be clean after WB_SYNC_ALL writeback. This is not currently true as we do not clear I_DIRTY_TIME in __writeback_single_inode() even for WB_SYNC_ALL writeback in all the cases. This then resulted in the following oops because bdev_write_inode() did not clean the inode and writeback code later stumbled over a dirty inode with detached wb. general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN Modules linked in: CPU: 3 PID: 32 Comm: kworker/u10:1 Not tainted 4.6.0-rc3+ #349 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Workqueue: writeback wb_workfn (flush-11:0) task: ffff88006ccf1840 ti: ffff88006cda8000 task.ti: ffff88006cda8000 RIP: 0010:[] [] locked_inode_to_wb_and_lock_list+0xa2/0x750 RSP: 0018:ffff88006cdaf7d0 EFLAGS: 00010246 RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffff88006ccf2050 RDX: 0000000000000000 RSI: 000000114c8a8484 RDI: 0000000000000286 RBP: ffff88006cdaf820 R08: ffff88006ccf1840 R09: 0000000000000000 R10: 000229915090805f R11: 0000000000000001 R12: ffff88006a72f5e0 R13: dffffc0000000000 R14: ffffed000d4e5eed R15: ffffffff8830cf40 FS: 0000000000000000(0000) GS:ffff88006d500000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000003301bf8 CR3: 000000006368f000 CR4: 00000000000006e0 DR0: 0000000000001ec9 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000600 Stack: ffff88006a72f680 ffff88006a72f768 ffff8800671230d8 03ff88006cdaf948 ffff88006a72f668 ffff88006a72f5e0 ffff8800671230d8 ffff88006cdaf948 ffff880065b90cc8 ffff880067123100 ffff88006cdaf970 ffffffff8188e12e Call Trace: [< inline >] inode_to_wb_and_lock_list fs/fs-writeback.c:309 [] writeback_sb_inodes+0x4de/0x1250 fs/fs-writeback.c:1554 [] __writeback_inodes_wb+0x104/0x1e0 fs/fs-writeback.c:1600 [] wb_writeback+0x7ce/0xc90 fs/fs-writeback.c:1709 [< inline >] wb_do_writeback fs/fs-writeback.c:1844 [] wb_workfn+0x2f9/0x1000 fs/fs-writeback.c:1884 [] process_one_work+0x78e/0x15c0 kernel/workqueue.c:2094 [] worker_thread+0xdb/0xfc0 kernel/workqueue.c:2228 [] kthread+0x23f/0x2d0 drivers/block/aoe/aoecmd.c:1303 [] ret_from_fork+0x22/0x50 arch/x86/entry/entry_64.S:392 Code: 05 94 4a a8 06 85 c0 0f 85 03 03 00 00 e8 07 15 d0 ff 41 80 3e 00 0f 85 64 06 00 00 49 8b 9c 24 88 01 00 00 48 89 d8 48 c1 e8 03 <42> 80 3c 28 00 0f 85 17 06 00 00 48 8b 03 48 83 c0 50 48 39 c3 RIP [< inline >] wb_get include/linux/backing-dev-defs.h:212 RIP [] locked_inode_to_wb_and_lock_list+0xa2/0x750 fs/fs-writeback.c:281 RSP ---[ end trace 986a4d314dcb2694 ]--- Fix the problem by making sure __writeback_single_inode() writes inode only with dirty times in WB_SYNC_ALL mode. Reported-by: Dmitry Vyukov Tested-by: Laurent Dufour Signed-off-by: Jan Kara Signed-off-by: Jens Axboe --- fs/fs-writeback.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 56c8fda436c0..4d09d4441e3e 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -1327,6 +1327,7 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc) dirty = inode->i_state & I_DIRTY; if (inode->i_state & I_DIRTY_TIME) { if ((dirty & (I_DIRTY_SYNC | I_DIRTY_DATASYNC)) || + wbc->sync_mode == WB_SYNC_ALL || unlikely(inode->i_state & I_DIRTY_TIME_EXPIRED) || unlikely(time_after(jiffies, (inode->dirtied_time_when + -- cgit v1.2.3 From 20bd723ec6a3261df5e02250cd3a1fbb09a343f2 Mon Sep 17 00:00:00 2001 From: Paolo Valente Date: Wed, 27 Jul 2016 07:22:05 +0200 Subject: block: add missing group association in bio-cloning functions When a bio is cloned, the newly created bio must be associated with the same blkcg as the original bio (if BLK_CGROUP is enabled). If this operation is not performed, then the new bio is not associated with any group, and the group of the current task is returned when the group of the bio is requested. Depending on the cloning frequency, this may cause a large percentage of the bios belonging to a given group to be treated as if belonging to other groups (in most cases as if belonging to the root group). The expected group isolation may thereby be broken. This commit adds the missing association in bio-cloning functions. Fixes: da2f0f74cf7d ("Btrfs: add support for blkio controllers") Cc: stable@vger.kernel.org # v4.3+ Signed-off-by: Paolo Valente Reviewed-by: Nikolay Borisov Reviewed-by: Jeff Moyer Acked-by: Tejun Heo Signed-off-by: Jens Axboe --- block/bio.c | 15 +++++++++++++++ fs/btrfs/extent_io.c | 6 ------ include/linux/bio.h | 3 +++ 3 files changed, 18 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/block/bio.c b/block/bio.c index 54ee3846c3a5..3f76a38a5e2d 100644 --- a/block/bio.c +++ b/block/bio.c @@ -583,6 +583,8 @@ void __bio_clone_fast(struct bio *bio, struct bio *bio_src) bio->bi_rw = bio_src->bi_rw; bio->bi_iter = bio_src->bi_iter; bio->bi_io_vec = bio_src->bi_io_vec; + + bio_clone_blkcg_association(bio, bio_src); } EXPORT_SYMBOL(__bio_clone_fast); @@ -687,6 +689,8 @@ integrity_clone: } } + bio_clone_blkcg_association(bio, bio_src); + return bio; } EXPORT_SYMBOL(bio_clone_bioset); @@ -2004,6 +2008,17 @@ void bio_disassociate_task(struct bio *bio) } } +/** + * bio_clone_blkcg_association - clone blkcg association from src to dst bio + * @dst: destination bio + * @src: source bio + */ +void bio_clone_blkcg_association(struct bio *dst, struct bio *src) +{ + if (src->bi_css) + WARN_ON(bio_associate_blkcg(dst, src->bi_css)); +} + #endif /* CONFIG_BLK_CGROUP */ static void __init biovec_init_slabs(void) diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index cee4cb99b8ce..5850d79806f4 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2697,12 +2697,6 @@ struct bio *btrfs_bio_clone(struct bio *bio, gfp_t gfp_mask) btrfs_bio->csum = NULL; btrfs_bio->csum_allocated = NULL; btrfs_bio->end_io = NULL; - -#ifdef CONFIG_BLK_CGROUP - /* FIXME, put this into bio_clone_bioset */ - if (bio->bi_css) - bio_associate_blkcg(new, bio->bi_css); -#endif } return new; } diff --git a/include/linux/bio.h b/include/linux/bio.h index 583c10810e32..e09a8895fc31 100644 --- a/include/linux/bio.h +++ b/include/linux/bio.h @@ -470,11 +470,14 @@ extern unsigned int bvec_nr_vecs(unsigned short idx); int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css); int bio_associate_current(struct bio *bio); void bio_disassociate_task(struct bio *bio); +void bio_clone_blkcg_association(struct bio *dst, struct bio *src); #else /* CONFIG_BLK_CGROUP */ static inline int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css) { return 0; } static inline int bio_associate_current(struct bio *bio) { return -ENOENT; } static inline void bio_disassociate_task(struct bio *bio) { } +static inline void bio_clone_blkcg_association(struct bio *dst, + struct bio *src) { } #endif /* CONFIG_BLK_CGROUP */ #ifdef CONFIG_HIGHMEM -- cgit v1.2.3 From 1aee6b9a7d947d42ed84baa4cf461e9d943b80f0 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 27 Jul 2016 12:52:21 -0600 Subject: f2fs: drop bio->bi_rw manual assignment Merge 4fc29c1aa375 included this extra line, but it's not needed (or useful) since we'll bio_set_op_attrs() right after to properly set the op and flags for the bio. Signed-off-by: Jens Axboe --- fs/f2fs/data.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e2624275d828..d64d2a515cb2 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -245,7 +245,6 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) bio_put(bio); return -EFAULT; } - bio->bi_rw = fio->op_flags; bio_set_op_attrs(bio, fio->op, fio->op_flags); __submit_bio(fio->sbi, bio, fio->type); -- cgit v1.2.3 From b571bc606e714e448b00920987d77b384a6a9570 Mon Sep 17 00:00:00 2001 From: Shaun Tancheff Date: Sat, 30 Jul 2016 16:45:48 -0500 Subject: Fixup direct bi_rw modifiers bi_rw should be using bio_set_op_attrs to set bi_rw. Signed-off-by: Shaun Tancheff Cc: Chris Mason Cc: Josef Bacik Cc: David Sterba Cc: Mike Christie Reviewed-by: Christoph Hellwig Signed-off-by: Jens Axboe --- fs/btrfs/extent_io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 5850d79806f4..881eb4667051 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2049,7 +2049,7 @@ int repair_io_failure(struct inode *inode, u64 start, u64 length, u64 logical, return -EIO; } bio->bi_bdev = dev->bdev; - bio->bi_rw = WRITE_SYNC; + bio_set_op_attrs(bio, REQ_OP_WRITE, WRITE_SYNC); bio_add_page(bio, page, length, pg_offset); if (btrfsic_submit_bio_wait(bio)) { -- cgit v1.2.3 From abf545484d31b68777a85c5c8f5b4bcde08283eb Mon Sep 17 00:00:00 2001 From: Mike Christie Date: Thu, 4 Aug 2016 14:23:34 -0600 Subject: mm/block: convert rw_page users to bio op use The rw_page users were not converted to use bio/req ops. As a result bdev_write_page is not passing down REQ_OP_WRITE and the IOs will be sent down as reads. Signed-off-by: Mike Christie Fixes: 4e1b2d52a80d ("block, fs, drivers: remove REQ_OP compat defs and related code") Modified by me to: 1) Drop op_flags passing into ->rw_page(), as we don't use it. 2) Make op_is_write() and friends safe to use for !CONFIG_BLOCK Signed-off-by: Jens Axboe --- drivers/block/brd.c | 17 +++++++---------- drivers/block/zram/zram_drv.c | 28 +++++++++++++++------------- drivers/nvdimm/btt.c | 18 +++++++++--------- drivers/nvdimm/pmem.c | 12 ++++++------ fs/block_dev.c | 7 ++++--- fs/mpage.c | 2 +- include/linux/blk_types.h | 22 +++++++++++----------- include/linux/blkdev.h | 2 +- include/linux/fs.h | 3 ++- include/linux/pagemap.h | 2 +- mm/filemap.c | 6 +++--- 11 files changed, 60 insertions(+), 59 deletions(-) (limited to 'fs') diff --git a/drivers/block/brd.c b/drivers/block/brd.c index 3022dad24071..3439b28cce8b 100644 --- a/drivers/block/brd.c +++ b/drivers/block/brd.c @@ -300,20 +300,20 @@ static void copy_from_brd(void *dst, struct brd_device *brd, * Process a single bvec of a bio. */ static int brd_do_bvec(struct brd_device *brd, struct page *page, - unsigned int len, unsigned int off, int rw, + unsigned int len, unsigned int off, int op, sector_t sector) { void *mem; int err = 0; - if (rw != READ) { + if (op_is_write(op)) { err = copy_to_brd_setup(brd, sector, len); if (err) goto out; } mem = kmap_atomic(page); - if (rw == READ) { + if (!op_is_write(op)) { copy_from_brd(mem + off, brd, sector, len); flush_dcache_page(page); } else { @@ -330,7 +330,6 @@ static blk_qc_t brd_make_request(struct request_queue *q, struct bio *bio) { struct block_device *bdev = bio->bi_bdev; struct brd_device *brd = bdev->bd_disk->private_data; - int rw; struct bio_vec bvec; sector_t sector; struct bvec_iter iter; @@ -347,14 +346,12 @@ static blk_qc_t brd_make_request(struct request_queue *q, struct bio *bio) goto out; } - rw = bio_data_dir(bio); - bio_for_each_segment(bvec, bio, iter) { unsigned int len = bvec.bv_len; int err; err = brd_do_bvec(brd, bvec.bv_page, len, - bvec.bv_offset, rw, sector); + bvec.bv_offset, bio_op(bio), sector); if (err) goto io_error; sector += len >> SECTOR_SHIFT; @@ -369,11 +366,11 @@ io_error: } static int brd_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int rw) + struct page *page, int op) { struct brd_device *brd = bdev->bd_disk->private_data; - int err = brd_do_bvec(brd, page, PAGE_SIZE, 0, rw, sector); - page_endio(page, rw & WRITE, err); + int err = brd_do_bvec(brd, page, PAGE_SIZE, 0, op, sector); + page_endio(page, op, err); return err; } diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c index 7454cf188c8e..ca29649c4b08 100644 --- a/drivers/block/zram/zram_drv.c +++ b/drivers/block/zram/zram_drv.c @@ -843,15 +843,15 @@ static void zram_bio_discard(struct zram *zram, u32 index, } static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, - int offset, int rw) + int offset, int op) { unsigned long start_time = jiffies; int ret; - generic_start_io_acct(rw, bvec->bv_len >> SECTOR_SHIFT, + generic_start_io_acct(op, bvec->bv_len >> SECTOR_SHIFT, &zram->disk->part0); - if (rw == READ) { + if (!op_is_write(op)) { atomic64_inc(&zram->stats.num_reads); ret = zram_bvec_read(zram, bvec, index, offset); } else { @@ -859,10 +859,10 @@ static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, ret = zram_bvec_write(zram, bvec, index, offset); } - generic_end_io_acct(rw, &zram->disk->part0, start_time); + generic_end_io_acct(op, &zram->disk->part0, start_time); if (unlikely(ret)) { - if (rw == READ) + if (!op_is_write(op)) atomic64_inc(&zram->stats.failed_reads); else atomic64_inc(&zram->stats.failed_writes); @@ -873,7 +873,7 @@ static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, static void __zram_make_request(struct zram *zram, struct bio *bio) { - int offset, rw; + int offset; u32 index; struct bio_vec bvec; struct bvec_iter iter; @@ -888,7 +888,6 @@ static void __zram_make_request(struct zram *zram, struct bio *bio) return; } - rw = bio_data_dir(bio); bio_for_each_segment(bvec, bio, iter) { int max_transfer_size = PAGE_SIZE - offset; @@ -903,15 +902,18 @@ static void __zram_make_request(struct zram *zram, struct bio *bio) bv.bv_len = max_transfer_size; bv.bv_offset = bvec.bv_offset; - if (zram_bvec_rw(zram, &bv, index, offset, rw) < 0) + if (zram_bvec_rw(zram, &bv, index, offset, + bio_op(bio)) < 0) goto out; bv.bv_len = bvec.bv_len - max_transfer_size; bv.bv_offset += max_transfer_size; - if (zram_bvec_rw(zram, &bv, index + 1, 0, rw) < 0) + if (zram_bvec_rw(zram, &bv, index + 1, 0, + bio_op(bio)) < 0) goto out; } else - if (zram_bvec_rw(zram, &bvec, index, offset, rw) < 0) + if (zram_bvec_rw(zram, &bvec, index, offset, + bio_op(bio)) < 0) goto out; update_position(&index, &offset, &bvec); @@ -968,7 +970,7 @@ static void zram_slot_free_notify(struct block_device *bdev, } static int zram_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int rw) + struct page *page, int op) { int offset, err = -EIO; u32 index; @@ -992,7 +994,7 @@ static int zram_rw_page(struct block_device *bdev, sector_t sector, bv.bv_len = PAGE_SIZE; bv.bv_offset = 0; - err = zram_bvec_rw(zram, &bv, index, offset, rw); + err = zram_bvec_rw(zram, &bv, index, offset, op); put_zram: zram_meta_put(zram); out: @@ -1005,7 +1007,7 @@ out: * (e.g., SetPageError, set_page_dirty and extra works). */ if (err == 0) - page_endio(page, rw, 0); + page_endio(page, op, 0); return err; } diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c index 9dce03f420eb..7cf3bdfaf809 100644 --- a/drivers/nvdimm/btt.c +++ b/drivers/nvdimm/btt.c @@ -1133,11 +1133,11 @@ static int btt_write_pg(struct btt *btt, struct bio_integrity_payload *bip, static int btt_do_bvec(struct btt *btt, struct bio_integrity_payload *bip, struct page *page, unsigned int len, unsigned int off, - int rw, sector_t sector) + int op, sector_t sector) { int ret; - if (rw == READ) { + if (!op_is_write(op)) { ret = btt_read_pg(btt, bip, page, off, sector, len); flush_dcache_page(page); } else { @@ -1155,7 +1155,7 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio) struct bvec_iter iter; unsigned long start; struct bio_vec bvec; - int err = 0, rw; + int err = 0; bool do_acct; /* @@ -1170,7 +1170,6 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio) } do_acct = nd_iostat_start(bio, &start); - rw = bio_data_dir(bio); bio_for_each_segment(bvec, bio, iter) { unsigned int len = bvec.bv_len; @@ -1181,11 +1180,12 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio) BUG_ON(len % btt->sector_size); err = btt_do_bvec(btt, bip, bvec.bv_page, len, bvec.bv_offset, - rw, iter.bi_sector); + bio_op(bio), iter.bi_sector); if (err) { dev_info(&btt->nd_btt->dev, "io error in %s sector %lld, len %d,\n", - (rw == READ) ? "READ" : "WRITE", + (op_is_write(bio_op(bio))) ? "WRITE" : + "READ", (unsigned long long) iter.bi_sector, len); bio->bi_error = err; break; @@ -1200,12 +1200,12 @@ out: } static int btt_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int rw) + struct page *page, int op) { struct btt *btt = bdev->bd_disk->private_data; - btt_do_bvec(btt, NULL, page, PAGE_SIZE, 0, rw, sector); - page_endio(page, rw & WRITE, 0); + btt_do_bvec(btt, NULL, page, PAGE_SIZE, 0, op, sector); + page_endio(page, op, 0); return 0; } diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index b511099457db..d64d92481c1d 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -67,7 +67,7 @@ static void pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset, } static int pmem_do_bvec(struct pmem_device *pmem, struct page *page, - unsigned int len, unsigned int off, int rw, + unsigned int len, unsigned int off, int op, sector_t sector) { int rc = 0; @@ -79,7 +79,7 @@ static int pmem_do_bvec(struct pmem_device *pmem, struct page *page, if (unlikely(is_bad_pmem(&pmem->bb, sector, len))) bad_pmem = true; - if (rw == READ) { + if (!op_is_write(op)) { if (unlikely(bad_pmem)) rc = -EIO; else { @@ -134,7 +134,7 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) do_acct = nd_iostat_start(bio, &start); bio_for_each_segment(bvec, bio, iter) { rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len, - bvec.bv_offset, bio_data_dir(bio), + bvec.bv_offset, bio_op(bio), iter.bi_sector); if (rc) { bio->bi_error = rc; @@ -152,12 +152,12 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) } static int pmem_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int rw) + struct page *page, int op) { struct pmem_device *pmem = bdev->bd_queue->queuedata; int rc; - rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, rw, sector); + rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, op, sector); /* * The ->rw_page interface is subtle and tricky. The core @@ -166,7 +166,7 @@ static int pmem_rw_page(struct block_device *bdev, sector_t sector, * caused by double completion. */ if (rc == 0) - page_endio(page, rw & WRITE, 0); + page_endio(page, op, 0); return rc; } diff --git a/fs/block_dev.c b/fs/block_dev.c index 2033a3f91d58..d402899ba135 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -416,7 +416,8 @@ int bdev_read_page(struct block_device *bdev, sector_t sector, result = blk_queue_enter(bdev->bd_queue, false); if (result) return result; - result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, READ); + result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, + REQ_OP_READ); blk_queue_exit(bdev->bd_queue); return result; } @@ -445,7 +446,6 @@ int bdev_write_page(struct block_device *bdev, sector_t sector, struct page *page, struct writeback_control *wbc) { int result; - int rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE; const struct block_device_operations *ops = bdev->bd_disk->fops; if (!ops->rw_page || bdev_get_integrity(bdev)) @@ -455,7 +455,8 @@ int bdev_write_page(struct block_device *bdev, sector_t sector, return result; set_page_writeback(page); - result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, rw); + result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, + REQ_OP_WRITE); if (result) end_page_writeback(page); else diff --git a/fs/mpage.c b/fs/mpage.c index 2ca1f39c8cba..7a09c55b4bd0 100644 --- a/fs/mpage.c +++ b/fs/mpage.c @@ -50,7 +50,7 @@ static void mpage_end_io(struct bio *bio) bio_for_each_segment_all(bv, bio, i) { struct page *page = bv->bv_page; - page_endio(page, bio_data_dir(bio), bio->bi_error); + page_endio(page, bio_op(bio), bio->bi_error); } bio_put(bio); diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h index f254eb264924..14b28ff2caf8 100644 --- a/include/linux/blk_types.h +++ b/include/linux/blk_types.h @@ -18,6 +18,17 @@ struct cgroup_subsys_state; typedef void (bio_end_io_t) (struct bio *); typedef void (bio_destructor_t) (struct bio *); +enum req_op { + REQ_OP_READ, + REQ_OP_WRITE, + REQ_OP_DISCARD, /* request to discard sectors */ + REQ_OP_SECURE_ERASE, /* request to securely erase sectors */ + REQ_OP_WRITE_SAME, /* write same block many times */ + REQ_OP_FLUSH, /* request for cache flush */ +}; + +#define REQ_OP_BITS 3 + #ifdef CONFIG_BLOCK /* * main unit of I/O for the block layer and lower layers (ie drivers and @@ -228,17 +239,6 @@ enum rq_flag_bits { #define REQ_HASHED (1ULL << __REQ_HASHED) #define REQ_MQ_INFLIGHT (1ULL << __REQ_MQ_INFLIGHT) -enum req_op { - REQ_OP_READ, - REQ_OP_WRITE, - REQ_OP_DISCARD, /* request to discard sectors */ - REQ_OP_SECURE_ERASE, /* request to securely erase sectors */ - REQ_OP_WRITE_SAME, /* write same block many times */ - REQ_OP_FLUSH, /* request for cache flush */ -}; - -#define REQ_OP_BITS 3 - typedef unsigned int blk_qc_t; #define BLK_QC_T_NONE -1U #define BLK_QC_T_SHIFT 16 diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index de7935961c27..ccd68c0d01de 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1672,7 +1672,7 @@ struct blk_dax_ctl { struct block_device_operations { int (*open) (struct block_device *, fmode_t); void (*release) (struct gendisk *, fmode_t); - int (*rw_page)(struct block_device *, sector_t, struct page *, int rw); + int (*rw_page)(struct block_device *, sector_t, struct page *, int op); int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); long (*direct_access)(struct block_device *, sector_t, void **, pfn_t *, diff --git a/include/linux/fs.h b/include/linux/fs.h index f3f0b4c8e8ac..498255e6914e 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2480,12 +2480,13 @@ extern void init_special_inode(struct inode *, umode_t, dev_t); extern void make_bad_inode(struct inode *); extern bool is_bad_inode(struct inode *); -#ifdef CONFIG_BLOCK static inline bool op_is_write(unsigned int op) { return op == REQ_OP_READ ? false : true; } +#ifdef CONFIG_BLOCK + /* * return data direction, READ or WRITE */ diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index 81363b834900..45786374abbd 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -510,7 +510,7 @@ static inline void wait_on_page_writeback(struct page *page) extern void end_page_writeback(struct page *page); void wait_for_stable_page(struct page *page); -void page_endio(struct page *page, int rw, int err); +void page_endio(struct page *page, int op, int err); /* * Add an arbitrary waiter to a page's wait queue diff --git a/mm/filemap.c b/mm/filemap.c index 3083ded98b15..daef091d4c50 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -887,9 +887,9 @@ EXPORT_SYMBOL(end_page_writeback); * After completing I/O on a page, call this routine to update the page * flags appropriately */ -void page_endio(struct page *page, int rw, int err) +void page_endio(struct page *page, int op, int err) { - if (rw == READ) { + if (!op_is_write(op)) { if (!err) { SetPageUptodate(page); } else { @@ -897,7 +897,7 @@ void page_endio(struct page *page, int rw, int err) SetPageError(page); } unlock_page(page); - } else { /* rw == WRITE */ + } else { if (err) { SetPageError(page); if (page->mapping) -- cgit v1.2.3 From 7eed34f18d9f0f96f51bfb73d649a51a20857eb0 Mon Sep 17 00:00:00 2001 From: Oleg Drokin Date: Thu, 14 Jul 2016 23:20:22 -0400 Subject: nfsd: Make creates return EEXIST instead of EACCES When doing a create (mkdir/mknod) on a name, it's worth checking the name exists first before returning EACCES in case the directory is not writeable by the user. This makes return values on the client more consistent regardless of whenever the entry there is cached in the local cache or not. Another positive side effect is certain programs only expect EEXIST in that case even despite POSIX allowing any valid error to be returned. Signed-off-by: Oleg Drokin Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4proc.c | 6 +++++- fs/nfsd/vfs.c | 11 ++++++++++- 2 files changed, 15 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index e0c15f879d89..9d7e1edf0cca 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -605,8 +605,12 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, fh_init(&resfh, NFS4_FHSIZE); + /* + * We just check that parent is accessible here, nfsd_* do their + * own access permission checks + */ status = fh_verify(rqstp, &cstate->current_fh, S_IFDIR, - NFSD_MAY_CREATE); + NFSD_MAY_EXEC); if (status) return status; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 6fbd81ecb410..fda4f86161f8 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1161,7 +1161,11 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, if (isdotent(fname, flen)) goto out; - err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_CREATE); + /* + * Even though it is a create, first let's see if we are even allowed + * to peek inside the parent + */ + err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); if (err) goto out; @@ -1211,6 +1215,11 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, goto out; } + /* Now let's see if we actually have permissions to create */ + err = nfsd_permission(rqstp, fhp->fh_export, dentry, NFSD_MAY_CREATE); + if (err) + goto out; + if (!(iap->ia_valid & ATTR_MODE)) iap->ia_mode = 0; iap->ia_mode = (iap->ia_mode & S_IALLUGO) | type; -- cgit v1.2.3 From 12391d07230ec7368269227207b961b2dcb34180 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 19 Jul 2016 17:25:42 -0400 Subject: nfsd: remove redundant zero-length check from create lookup_one_len already has this check. The only effect of this patch is to return access instead of perm in the 0-length-filename case. I actually prefer nfserr_perm (or _inval?), but I doubt anyone cares. The isdotent check seems redundant too, but I worry that some client might actually care about that strange nfserr_exist error. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfsproc.c | 3 --- fs/nfsd/vfs.c | 3 --- 2 files changed, 6 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfsproc.c b/fs/nfsd/nfsproc.c index 4cd78ef4c95c..6722fe52ea15 100644 --- a/fs/nfsd/nfsproc.c +++ b/fs/nfsd/nfsproc.c @@ -251,9 +251,6 @@ nfsd_proc_create(struct svc_rqst *rqstp, struct nfsd_createargs *argp, /* Check for NFSD_MAY_WRITE in nfsd_create if necessary */ - nfserr = nfserr_acces; - if (!argp->len) - goto done; nfserr = nfserr_exist; if (isdotent(argp->name, argp->len)) goto done; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index fda4f86161f8..fba8e7e521e0 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1154,9 +1154,6 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, __be32 err2; int host_err; - err = nfserr_perm; - if (!flen) - goto out; err = nfserr_exist; if (isdotent(fname, flen)) goto out; -- cgit v1.2.3 From e75b23f9e323b1e0759619c39d5a9f7a3a5d9d2c Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Tue, 19 Jul 2016 17:33:04 -0400 Subject: nfsd: check d_can_lookup in fh_verify of directories Create and other nfsd ops generally assume we can call lookup_one_len on inodes with S_IFDIR set. Al says that this assumption isn't true in general, though it should be for the filesystem objects nfsd sees. Add a check just to make sure our assumption isn't violated. Remove a couple checks for i_op->lookup in create code. Cc: Al Viro Signed-off-by: J. Bruce Fields --- fs/nfsd/nfsfh.c | 14 ++++++++++---- fs/nfsd/vfs.c | 9 --------- 2 files changed, 10 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 27250e279c37..cfe7500d5847 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -59,14 +59,20 @@ static int nfsd_acceptable(void *expv, struct dentry *dentry) * the write call). */ static inline __be32 -nfsd_mode_check(struct svc_rqst *rqstp, umode_t mode, umode_t requested) +nfsd_mode_check(struct svc_rqst *rqstp, struct dentry *dentry, + umode_t requested) { - mode &= S_IFMT; + umode_t mode = d_inode(dentry)->i_mode & S_IFMT; if (requested == 0) /* the caller doesn't care */ return nfs_ok; - if (mode == requested) + if (mode == requested) { + if (mode == S_IFDIR && !d_can_lookup(dentry)) { + WARN_ON_ONCE(1); + return nfserr_notdir; + } return nfs_ok; + } /* * v4 has an error more specific than err_notdir which we should * return in preference to err_notdir: @@ -340,7 +346,7 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, int access) if (error) goto out; - error = nfsd_mode_check(rqstp, d_inode(dentry)->i_mode, type); + error = nfsd_mode_check(rqstp, dentry, type); if (error) goto out; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index fba8e7e521e0..cf14e394a1b0 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1169,9 +1169,6 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, dentry = fhp->fh_dentry; dirp = d_inode(dentry); - err = nfserr_notdir; - if (!dirp->i_op->lookup) - goto out; /* * Check whether the response file handle has been verified yet. * If it has, the parent directory should already be locked. @@ -1310,12 +1307,6 @@ do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, dentry = fhp->fh_dentry; dirp = d_inode(dentry); - /* Get all the sanity checks out of the way before - * we lock the parent. */ - err = nfserr_notdir; - if (!dirp->i_op->lookup) - goto out; - host_err = fh_want_write(fhp); if (host_err) goto out_nfserr; -- cgit v1.2.3 From b44061d0b914c11fb8f8a747c91d111973f56d76 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 20 Jul 2016 16:16:06 -0400 Subject: nfsd: reorganize nfsd_create There's some odd logic in nfsd_create() that allows it to be called with the parent directory either locked or unlocked. The only already-locked caller is NFSv2's nfsd_proc_create(). It's less confusing to split out the unlocked case into a separate function which the NFSv2 code can call directly. Also fix some comments while we're here. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfsproc.c | 4 +- fs/nfsd/vfs.c | 109 ++++++++++++++++++++++++++++-------------------------- fs/nfsd/vfs.h | 3 ++ 3 files changed, 61 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfsproc.c b/fs/nfsd/nfsproc.c index 6722fe52ea15..e9214768cde9 100644 --- a/fs/nfsd/nfsproc.c +++ b/fs/nfsd/nfsproc.c @@ -359,8 +359,8 @@ nfsd_proc_create(struct svc_rqst *rqstp, struct nfsd_createargs *argp, nfserr = 0; if (!inode) { /* File doesn't exist. Create it and set attrs */ - nfserr = nfsd_create(rqstp, dirfhp, argp->name, argp->len, - attr, type, rdev, newfhp); + nfserr = nfsd_create_locked(rqstp, dirfhp, argp->name, + argp->len, attr, type, rdev, newfhp); } else if (type == S_IFREG) { dprintk("nfsd: existing %s, valid=%x, size=%ld\n", argp->name, attr->ia_valid, (long) attr->ia_size); diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index cf14e394a1b0..d21d0827cac4 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1135,16 +1135,9 @@ nfsd_check_ignore_resizing(struct iattr *iap) iap->ia_valid &= ~ATTR_SIZE; } -/* - * Create a file (regular, directory, device, fifo); UNIX sockets - * not yet implemented. - * If the response fh has been verified, the parent directory should - * already be locked. Note that the parent directory is left locked. - * - * N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp - */ +/* The parent directory should already be locked: */ __be32 -nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, +nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, char *fname, int flen, struct iattr *iap, int type, dev_t rdev, struct svc_fh *resfhp) { @@ -1154,50 +1147,15 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, __be32 err2; int host_err; - err = nfserr_exist; - if (isdotent(fname, flen)) - goto out; - - /* - * Even though it is a create, first let's see if we are even allowed - * to peek inside the parent - */ - err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); - if (err) - goto out; - dentry = fhp->fh_dentry; dirp = d_inode(dentry); - /* - * Check whether the response file handle has been verified yet. - * If it has, the parent directory should already be locked. - */ - if (!resfhp->fh_dentry) { - host_err = fh_want_write(fhp); - if (host_err) - goto out_nfserr; - - /* called from nfsd_proc_mkdir, or possibly nfsd3_proc_create */ - fh_lock_nested(fhp, I_MUTEX_PARENT); - dchild = lookup_one_len(fname, dentry, flen); - host_err = PTR_ERR(dchild); - if (IS_ERR(dchild)) - goto out_nfserr; - err = fh_compose(resfhp, fhp->fh_export, dchild, fhp); - if (err) - goto out; - } else { - /* called from nfsd_proc_create */ - dchild = dget(resfhp->fh_dentry); - if (!fhp->fh_locked) { - /* not actually possible */ - printk(KERN_ERR - "nfsd_create: parent %pd2 not locked!\n", + dchild = dget(resfhp->fh_dentry); + if (!fhp->fh_locked) { + WARN_ONCE(1, "nfsd_create: parent %pd2 not locked!\n", dentry); - err = nfserr_io; - goto out; - } + err = nfserr_io; + goto out; } /* * Make sure the child dentry is still negative ... @@ -1225,9 +1183,6 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, goto out; } - /* - * Get the dir op function pointer. - */ err = 0; host_err = 0; switch (type) { @@ -1254,7 +1209,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, /* * nfsd_create_setattr already committed the child. Transactional * filesystems had a chance to commit changes for both parent and - * child * simultaneously making the following commit_metadata a + * child simultaneously making the following commit_metadata a * noop. */ err2 = nfserrno(commit_metadata(fhp)); @@ -1275,6 +1230,54 @@ out_nfserr: goto out; } +/* + * Create a filesystem object (regular, directory, special). + * Note that the parent directory is left locked. + * + * N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp + */ +__be32 +nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, + char *fname, int flen, struct iattr *iap, + int type, dev_t rdev, struct svc_fh *resfhp) +{ + struct dentry *dentry, *dchild = NULL; + struct inode *dirp; + __be32 err; + int host_err; + + if (isdotent(fname, flen)) + return nfserr_exist; + + /* + * Even though it is a create, first let's see if we are even allowed + * to peek inside the parent + */ + err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); + if (err) + return err; + + dentry = fhp->fh_dentry; + dirp = d_inode(dentry); + + host_err = fh_want_write(fhp); + if (host_err) + return nfserrno(host_err); + + fh_lock_nested(fhp, I_MUTEX_PARENT); + dchild = lookup_one_len(fname, dentry, flen); + host_err = PTR_ERR(dchild); + if (IS_ERR(dchild)) + return nfserrno(host_err); + err = fh_compose(resfhp, fhp->fh_export, dchild, fhp); + if (err) { + dput(dchild); + return err; + } + return nfsd_create_locked(rqstp, fhp, fname, flen, iap, type, + rdev, resfhp); +} + #ifdef CONFIG_NFSD_V3 /* diff --git a/fs/nfsd/vfs.h b/fs/nfsd/vfs.h index 2d573ec057f8..3cbb1b33777b 100644 --- a/fs/nfsd/vfs.h +++ b/fs/nfsd/vfs.h @@ -59,6 +59,9 @@ __be32 nfsd4_vfs_fallocate(struct svc_rqst *, struct svc_fh *, __be32 nfsd4_clone_file_range(struct file *, u64, struct file *, u64, u64); #endif /* CONFIG_NFSD_V4 */ +__be32 nfsd_create_locked(struct svc_rqst *, struct svc_fh *, + char *name, int len, struct iattr *attrs, + int type, dev_t rdev, struct svc_fh *res); __be32 nfsd_create(struct svc_rqst *, struct svc_fh *, char *name, int len, struct iattr *attrs, int type, dev_t rdev, struct svc_fh *res); -- cgit v1.2.3 From d03d9fe476ea81ad3be25967ef35a671dbd79b55 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 21 Jul 2016 21:55:42 -0400 Subject: nfsd: remove unnecessary positive-dentry check vfs_{create,mkdir,mknod} each begin with a call to may_create(), which returns EEXIST if the object already exists. This check is therefore unnecessary. (In the NFSv2 case, nfsd_proc_create also has such a check. Contrary to RFC 1094, our code seems to believe that a CREATE of an existing file should succeed. I'm leaving that behavior alone.) Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 10 ---------- 1 file changed, 10 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index d21d0827cac4..c95709486ca4 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1157,17 +1157,7 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, err = nfserr_io; goto out; } - /* - * Make sure the child dentry is still negative ... - */ - err = nfserr_exist; - if (d_really_is_positive(dchild)) { - dprintk("nfsd_create: dentry %pd/%pd not negative!\n", - dentry, dchild); - goto out; - } - /* Now let's see if we actually have permissions to create */ err = nfsd_permission(rqstp, fhp->fh_export, dentry, NFSD_MAY_CREATE); if (err) goto out; -- cgit v1.2.3 From 71423274498169911bf9eedf02d5e7ac0a083801 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 22 Jul 2016 12:03:46 -0400 Subject: nfsd: clean up bad-type check in nfsd_create_locked Minor cleanup, no change in behavior. Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index c95709486ca4..1809624997f0 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1166,13 +1166,6 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, iap->ia_mode = 0; iap->ia_mode = (iap->ia_mode & S_IALLUGO) | type; - err = nfserr_inval; - if (!S_ISREG(type) && !S_ISDIR(type) && !special_file(type)) { - printk(KERN_WARNING "nfsd: bad file type %o in nfsd_create\n", - type); - goto out; - } - err = 0; host_err = 0; switch (type) { @@ -1190,6 +1183,10 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, case S_IFSOCK: host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev); break; + default: + printk(KERN_WARNING "nfsd: bad file type %o in nfsd_create\n", + type); + host_err = -EINVAL; } if (host_err < 0) goto out_nfserr; -- cgit v1.2.3 From fa08139d5e02838ccabd9dcde22663d3b6780cad Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 21 Jul 2016 16:00:12 -0400 Subject: nfsd: drop unnecessary MAY_EXEC check from create We need an fh_verify to make sure we at least have a dentry, but actual permission checks happen later. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4proc.c | 7 +------ fs/nfsd/vfs.c | 6 +----- 2 files changed, 2 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 9d7e1edf0cca..1fb222752b2b 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -605,12 +605,7 @@ nfsd4_create(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, fh_init(&resfh, NFS4_FHSIZE); - /* - * We just check that parent is accessible here, nfsd_* do their - * own access permission checks - */ - status = fh_verify(rqstp, &cstate->current_fh, S_IFDIR, - NFSD_MAY_EXEC); + status = fh_verify(rqstp, &cstate->current_fh, S_IFDIR, NFSD_MAY_NOP); if (status) return status; diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 1809624997f0..6666188c1138 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1236,11 +1236,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, if (isdotent(fname, flen)) return nfserr_exist; - /* - * Even though it is a create, first let's see if we are even allowed - * to peek inside the parent - */ - err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); + err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_NOP); if (err) return err; -- cgit v1.2.3 From 2b11885921a48767b6ace27c481beba3b42371ce Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Wed, 3 Aug 2016 22:05:00 +0300 Subject: nfsd: remove some dead code in nfsd_create_locked() We changed this around in f135af1041f ('nfsd: reorganize nfsd_create') so "dchild" can't be an error pointer any more. Also, dchild can't be NULL here (and dput would already handle this even if it was). Signed-off-by: Dan Carpenter Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 6666188c1138..ba944123167b 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1141,7 +1141,7 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, char *fname, int flen, struct iattr *iap, int type, dev_t rdev, struct svc_fh *resfhp) { - struct dentry *dentry, *dchild = NULL; + struct dentry *dentry, *dchild; struct inode *dirp; __be32 err; __be32 err2; @@ -1208,8 +1208,7 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp, if (!err) err = fh_update(resfhp); out: - if (dchild && !IS_ERR(dchild)) - dput(dchild); + dput(dchild); return err; out_nfserr: -- cgit v1.2.3 From 529182e204db083cb7bda832d1c5c6d9278ba1cb Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Fri, 29 Jul 2016 18:11:32 -0700 Subject: ramoops: use DT reserved-memory bindings Instead of a ramoops-specific node, use a child node of /reserved-memory. This requires that of_platform_device_create() be explicitly called for the node, though, since "/reserved-memory" does not have its own "compatible" property. Suggested-by: Rob Herring Signed-off-by: Kees Cook Acked-by: Rob Herring --- Documentation/devicetree/bindings/misc/ramoops.txt | 48 ---------------------- .../bindings/reserved-memory/ramoops.txt | 48 ++++++++++++++++++++++ Documentation/ramoops.txt | 38 ++++++++++++----- drivers/of/platform.c | 20 ++++++++- fs/pstore/ram.c | 23 ++++------- 5 files changed, 100 insertions(+), 77 deletions(-) delete mode 100644 Documentation/devicetree/bindings/misc/ramoops.txt create mode 100644 Documentation/devicetree/bindings/reserved-memory/ramoops.txt (limited to 'fs') diff --git a/Documentation/devicetree/bindings/misc/ramoops.txt b/Documentation/devicetree/bindings/misc/ramoops.txt deleted file mode 100644 index cd02cec67d38..000000000000 --- a/Documentation/devicetree/bindings/misc/ramoops.txt +++ /dev/null @@ -1,48 +0,0 @@ -Ramoops oops/panic logger -========================= - -ramoops provides persistent RAM storage for oops and panics, so they can be -recovered after a reboot. It is a backend to pstore, so this node is named -"ramoops" after the backend, rather than "pstore" which is the subsystem. - -Parts of this storage may be set aside for other persistent log buffers, such -as kernel log messages, or for optional ECC error-correction data. The total -size of these optional buffers must fit in the reserved region. - -Any remaining space will be used for a circular buffer of oops and panic -records. These records have a configurable size, with a size of 0 indicating -that they should be disabled. - -At least one of "record-size", "console-size", "ftrace-size", or "pmsg-size" -must be set non-zero, but are otherwise optional as listed below. - - -Required properties: - -- compatible: must be "ramoops" - -- memory-region: phandle to a region of memory that is preserved between - reboots - - -Optional properties: - -- ecc-size: enables ECC support and specifies ECC buffer size in bytes - (defaults to 0: no ECC) - -- record-size: maximum size in bytes of each dump done on oops/panic - (defaults to 0: disabled) - -- console-size: size in bytes of log buffer reserved for kernel messages - (defaults to 0: disabled) - -- ftrace-size: size in bytes of log buffer reserved for function tracing and - profiling (defaults to 0: disabled) - -- pmsg-size: size in bytes of log buffer reserved for userspace messages - (defaults to 0: disabled) - -- unbuffered: if present, use unbuffered mappings to map the reserved region - (defaults to buffered mappings) - -- no-dump-oops: if present, only dump panics (defaults to panics and oops) diff --git a/Documentation/devicetree/bindings/reserved-memory/ramoops.txt b/Documentation/devicetree/bindings/reserved-memory/ramoops.txt new file mode 100644 index 000000000000..e81f821a2135 --- /dev/null +++ b/Documentation/devicetree/bindings/reserved-memory/ramoops.txt @@ -0,0 +1,48 @@ +Ramoops oops/panic logger +========================= + +ramoops provides persistent RAM storage for oops and panics, so they can be +recovered after a reboot. This is a child-node of "/reserved-memory", and +is named "ramoops" after the backend, rather than "pstore" which is the +subsystem. + +Parts of this storage may be set aside for other persistent log buffers, such +as kernel log messages, or for optional ECC error-correction data. The total +size of these optional buffers must fit in the reserved region. + +Any remaining space will be used for a circular buffer of oops and panic +records. These records have a configurable size, with a size of 0 indicating +that they should be disabled. + +At least one of "record-size", "console-size", "ftrace-size", or "pmsg-size" +must be set non-zero, but are otherwise optional as listed below. + + +Required properties: + +- compatible: must be "ramoops" + +- reg: region of memory that is preserved between reboots + + +Optional properties: + +- ecc-size: enables ECC support and specifies ECC buffer size in bytes + (defaults to 0: no ECC) + +- record-size: maximum size in bytes of each dump done on oops/panic + (defaults to 0: disabled) + +- console-size: size in bytes of log buffer reserved for kernel messages + (defaults to 0: disabled) + +- ftrace-size: size in bytes of log buffer reserved for function tracing and + profiling (defaults to 0: disabled) + +- pmsg-size: size in bytes of log buffer reserved for userspace messages + (defaults to 0: disabled) + +- unbuffered: if present, use unbuffered mappings to map the reserved region + (defaults to buffered mappings) + +- no-dump-oops: if present, only dump panics (defaults to panics and oops) diff --git a/Documentation/ramoops.txt b/Documentation/ramoops.txt index 9264bcab4099..26b9f31cf65a 100644 --- a/Documentation/ramoops.txt +++ b/Documentation/ramoops.txt @@ -45,18 +45,34 @@ corrupt, but usually it is restorable. 2. Setting the parameters -Setting the ramoops parameters can be done in 3 different manners: - 1. Use the module parameters (which have the names of the variables described - as before). - For quick debugging, you can also reserve parts of memory during boot - and then use the reserved memory for ramoops. For example, assuming a machine - with > 128 MB of memory, the following kernel command line will tell the - kernel to use only the first 128 MB of memory, and place ECC-protected ramoops - region at 128 MB boundary: +Setting the ramoops parameters can be done in several different manners: + + A. Use the module parameters (which have the names of the variables described + as before). For quick debugging, you can also reserve parts of memory during + boot and then use the reserved memory for ramoops. For example, assuming a + machine with > 128 MB of memory, the following kernel command line will tell + the kernel to use only the first 128 MB of memory, and place ECC-protected + ramoops region at 128 MB boundary: "mem=128M ramoops.mem_address=0x8000000 ramoops.ecc=1" - 2. Use Device Tree bindings, as described in - Documentation/device-tree/bindings/misc/ramoops.txt. - 3. Use a platform device and set the platform data. The parameters can then + + B. Use Device Tree bindings, as described in + Documentation/device-tree/bindings/reserved-memory/ramoops.txt. + For example: + + reserved-memory { + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ramoops@8f000000 { + compatible = "ramoops"; + reg = <0 0x8f000000 0 0x100000>; + record-size = <0x4000>; + console-size = <0x4000>; + }; + }; + + C. Use a platform device and set the platform data. The parameters can then be set through that platform data. An example of doing that is: #include diff --git a/drivers/of/platform.c b/drivers/of/platform.c index 765390e3ed8d..8aa197691074 100644 --- a/drivers/of/platform.c +++ b/drivers/of/platform.c @@ -499,8 +499,24 @@ EXPORT_SYMBOL_GPL(of_platform_default_populate); static int __init of_platform_default_populate_init(void) { - if (of_have_populated_dt()) - of_platform_default_populate(NULL, NULL, NULL); + struct device_node *node; + + if (!of_have_populated_dt()) + return -ENODEV; + + /* + * Handle ramoops explicitly, since it is inside /reserved-memory, + * which lacks a "compatible" property. + */ + node = of_find_node_by_path("/reserved-memory"); + if (node) { + node = of_find_compatible_node(node, NULL, "ramoops"); + if (node) + of_platform_device_create(node, NULL, NULL); + } + + /* Populate everything else. */ + of_platform_default_populate(NULL, NULL, NULL); return 0; } diff --git a/fs/pstore/ram.c b/fs/pstore/ram.c index 47516a794011..427eee7570bb 100644 --- a/fs/pstore/ram.c +++ b/fs/pstore/ram.c @@ -486,30 +486,21 @@ static int ramoops_parse_dt(struct platform_device *pdev, struct ramoops_platform_data *pdata) { struct device_node *of_node = pdev->dev.of_node; - struct device_node *mem_region; - struct resource res; + struct resource *res; u32 value; int ret; dev_dbg(&pdev->dev, "using Device Tree\n"); - mem_region = of_parse_phandle(of_node, "memory-region", 0); - if (!mem_region) { - dev_err(&pdev->dev, "no memory-region phandle\n"); - return -ENODEV; - } - - ret = of_address_to_resource(mem_region, 0, &res); - of_node_put(mem_region); - if (ret) { + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { dev_err(&pdev->dev, - "failed to translate memory-region to resource: %d\n", - ret); - return ret; + "failed to locate DT /reserved-memory resource\n"); + return -EINVAL; } - pdata->mem_size = resource_size(&res); - pdata->mem_address = res.start; + pdata->mem_size = resource_size(res); + pdata->mem_address = res->start; pdata->mem_type = of_property_read_bool(of_node, "unbuffered"); pdata->dump_oops = !of_property_read_bool(of_node, "no-dump-oops"); -- cgit v1.2.3 From e976e56423dc1cc01686861fc3e0c6c0ec8cd8b7 Mon Sep 17 00:00:00 2001 From: Hiraku Toyooka Date: Mon, 25 Jul 2016 12:56:55 +0900 Subject: ramoops: use persistent_ram_free() instead of kfree() for freeing prz persistent_ram_zone(=prz) structures are allocated by persistent_ram_new(), which includes vmap() or ioremap(). But they are currently freed by kfree(). This uses persistent_ram_free() for correct this asymmetry usage. Signed-off-by: Hiraku Toyooka Signed-off-by: Nobuhiro Iwamatsu Cc: Mark Salyzyn Cc: Seiji Aguchi Signed-off-by: Kees Cook --- fs/pstore/ram.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/pstore/ram.c b/fs/pstore/ram.c index 427eee7570bb..7a034d62cf8c 100644 --- a/fs/pstore/ram.c +++ b/fs/pstore/ram.c @@ -643,11 +643,11 @@ fail_buf: kfree(cxt->pstore.buf); fail_clear: cxt->pstore.bufsize = 0; - kfree(cxt->mprz); + persistent_ram_free(cxt->mprz); fail_init_mprz: - kfree(cxt->fprz); + persistent_ram_free(cxt->fprz); fail_init_fprz: - kfree(cxt->cprz); + persistent_ram_free(cxt->cprz); fail_init_cprz: ramoops_free_przs(cxt); fail_out: -- cgit v1.2.3 From 337684a1746f93ae107e05d90977b070bb7e39d8 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Tue, 2 Aug 2016 19:58:28 +0800 Subject: fs: return EPERM on immutable inode In most cases, EPERM is returned on immutable inode, and there're only a few places returning EACCES. I noticed this when running LTP on overlayfs, setxattr03 failed due to unexpected EACCES on immutable inode. So converting all EACCES to EPERM on immutable inode. Acked-by: Dave Chinner Signed-off-by: Eryu Guan Signed-off-by: Al Viro Signed-off-by: Linus Torvalds --- fs/gfs2/inode.c | 2 +- fs/namei.c | 2 +- fs/utimes.c | 3 ++- fs/xfs/xfs_ioctl.c | 2 +- 4 files changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index e0621cacf134..e4da0ecd3285 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -1800,7 +1800,7 @@ int gfs2_permission(struct inode *inode, int mask) } if ((mask & MAY_WRITE) && IS_IMMUTABLE(inode)) - error = -EACCES; + error = -EPERM; else error = generic_permission(inode, mask); if (gfs2_holder_initialized(&i_gh)) diff --git a/fs/namei.c b/fs/namei.c index c386a329ab20..adb04146df09 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -410,7 +410,7 @@ int __inode_permission(struct inode *inode, int mask) * Nobody gets write access to an immutable file. */ if (IS_IMMUTABLE(inode)) - return -EACCES; + return -EPERM; /* * Updating mtime will likely cause i_uid and i_gid to be diff --git a/fs/utimes.c b/fs/utimes.c index 85c40f4f373d..794f5f5b1fb5 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -92,10 +92,11 @@ static int utimes_common(struct path *path, struct timespec *times) * then we need to check permissions, because * inode_change_ok() won't do it. */ - error = -EACCES; + error = -EPERM; if (IS_IMMUTABLE(inode)) goto mnt_drop_write_and_out; + error = -EACCES; if (!inode_owner_or_capable(inode)) { error = inode_permission(inode, MAY_WRITE); if (error) diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index cf46658392ce..96a70fd1f5d6 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -232,7 +232,7 @@ xfs_open_by_handle( } if ((fmode & FMODE_WRITE) && IS_IMMUTABLE(inode)) { - error = -EACCES; + error = -EPERM; goto out_dput; } -- cgit v1.2.3 From c11f0c0b5bb949673e4fc16c742f0316ae4ced20 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 5 Aug 2016 08:11:04 -0600 Subject: block/mm: make bdev_ops->rw_page() take a bool for read/write Commit abf545484d31 changed it from an 'rw' flags type to the newer ops based interface, but now we're effectively leaking some bdev internals to the rest of the kernel. Since we only care about whether it's a read or a write at that level, just pass in a bool 'is_write' parameter instead. Then we can also move op_is_write() and friends back under CONFIG_BLOCK protection. Reviewed-by: Mike Christie Signed-off-by: Jens Axboe --- drivers/block/brd.c | 16 ++++++++-------- drivers/block/zram/zram_drv.c | 23 ++++++++++++----------- drivers/nvdimm/btt.c | 12 ++++++------ drivers/nvdimm/pmem.c | 12 ++++++------ fs/block_dev.c | 6 ++---- fs/mpage.c | 2 +- include/linux/blk_types.h | 22 +++++++++++----------- include/linux/blkdev.h | 2 +- include/linux/fs.h | 3 +-- include/linux/pagemap.h | 2 +- mm/filemap.c | 4 ++-- 11 files changed, 51 insertions(+), 53 deletions(-) (limited to 'fs') diff --git a/drivers/block/brd.c b/drivers/block/brd.c index 3439b28cce8b..0c76d4016eeb 100644 --- a/drivers/block/brd.c +++ b/drivers/block/brd.c @@ -300,20 +300,20 @@ static void copy_from_brd(void *dst, struct brd_device *brd, * Process a single bvec of a bio. */ static int brd_do_bvec(struct brd_device *brd, struct page *page, - unsigned int len, unsigned int off, int op, + unsigned int len, unsigned int off, bool is_write, sector_t sector) { void *mem; int err = 0; - if (op_is_write(op)) { + if (is_write) { err = copy_to_brd_setup(brd, sector, len); if (err) goto out; } mem = kmap_atomic(page); - if (!op_is_write(op)) { + if (!is_write) { copy_from_brd(mem + off, brd, sector, len); flush_dcache_page(page); } else { @@ -350,8 +350,8 @@ static blk_qc_t brd_make_request(struct request_queue *q, struct bio *bio) unsigned int len = bvec.bv_len; int err; - err = brd_do_bvec(brd, bvec.bv_page, len, - bvec.bv_offset, bio_op(bio), sector); + err = brd_do_bvec(brd, bvec.bv_page, len, bvec.bv_offset, + op_is_write(bio_op(bio)), sector); if (err) goto io_error; sector += len >> SECTOR_SHIFT; @@ -366,11 +366,11 @@ io_error: } static int brd_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int op) + struct page *page, bool is_write) { struct brd_device *brd = bdev->bd_disk->private_data; - int err = brd_do_bvec(brd, page, PAGE_SIZE, 0, op, sector); - page_endio(page, op, err); + int err = brd_do_bvec(brd, page, PAGE_SIZE, 0, is_write, sector); + page_endio(page, is_write, err); return err; } diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c index ca29649c4b08..04365b17ee67 100644 --- a/drivers/block/zram/zram_drv.c +++ b/drivers/block/zram/zram_drv.c @@ -843,15 +843,16 @@ static void zram_bio_discard(struct zram *zram, u32 index, } static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, - int offset, int op) + int offset, bool is_write) { unsigned long start_time = jiffies; + int rw_acct = is_write ? REQ_OP_WRITE : REQ_OP_READ; int ret; - generic_start_io_acct(op, bvec->bv_len >> SECTOR_SHIFT, + generic_start_io_acct(rw_acct, bvec->bv_len >> SECTOR_SHIFT, &zram->disk->part0); - if (!op_is_write(op)) { + if (!is_write) { atomic64_inc(&zram->stats.num_reads); ret = zram_bvec_read(zram, bvec, index, offset); } else { @@ -859,10 +860,10 @@ static int zram_bvec_rw(struct zram *zram, struct bio_vec *bvec, u32 index, ret = zram_bvec_write(zram, bvec, index, offset); } - generic_end_io_acct(op, &zram->disk->part0, start_time); + generic_end_io_acct(rw_acct, &zram->disk->part0, start_time); if (unlikely(ret)) { - if (!op_is_write(op)) + if (!is_write) atomic64_inc(&zram->stats.failed_reads); else atomic64_inc(&zram->stats.failed_writes); @@ -903,17 +904,17 @@ static void __zram_make_request(struct zram *zram, struct bio *bio) bv.bv_offset = bvec.bv_offset; if (zram_bvec_rw(zram, &bv, index, offset, - bio_op(bio)) < 0) + op_is_write(bio_op(bio))) < 0) goto out; bv.bv_len = bvec.bv_len - max_transfer_size; bv.bv_offset += max_transfer_size; if (zram_bvec_rw(zram, &bv, index + 1, 0, - bio_op(bio)) < 0) + op_is_write(bio_op(bio))) < 0) goto out; } else if (zram_bvec_rw(zram, &bvec, index, offset, - bio_op(bio)) < 0) + op_is_write(bio_op(bio))) < 0) goto out; update_position(&index, &offset, &bvec); @@ -970,7 +971,7 @@ static void zram_slot_free_notify(struct block_device *bdev, } static int zram_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int op) + struct page *page, bool is_write) { int offset, err = -EIO; u32 index; @@ -994,7 +995,7 @@ static int zram_rw_page(struct block_device *bdev, sector_t sector, bv.bv_len = PAGE_SIZE; bv.bv_offset = 0; - err = zram_bvec_rw(zram, &bv, index, offset, op); + err = zram_bvec_rw(zram, &bv, index, offset, is_write); put_zram: zram_meta_put(zram); out: @@ -1007,7 +1008,7 @@ out: * (e.g., SetPageError, set_page_dirty and extra works). */ if (err == 0) - page_endio(page, op, 0); + page_endio(page, is_write, 0); return err; } diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c index 7cf3bdfaf809..88e91666f145 100644 --- a/drivers/nvdimm/btt.c +++ b/drivers/nvdimm/btt.c @@ -1133,11 +1133,11 @@ static int btt_write_pg(struct btt *btt, struct bio_integrity_payload *bip, static int btt_do_bvec(struct btt *btt, struct bio_integrity_payload *bip, struct page *page, unsigned int len, unsigned int off, - int op, sector_t sector) + bool is_write, sector_t sector) { int ret; - if (!op_is_write(op)) { + if (!is_write) { ret = btt_read_pg(btt, bip, page, off, sector, len); flush_dcache_page(page); } else { @@ -1180,7 +1180,7 @@ static blk_qc_t btt_make_request(struct request_queue *q, struct bio *bio) BUG_ON(len % btt->sector_size); err = btt_do_bvec(btt, bip, bvec.bv_page, len, bvec.bv_offset, - bio_op(bio), iter.bi_sector); + op_is_write(bio_op(bio)), iter.bi_sector); if (err) { dev_info(&btt->nd_btt->dev, "io error in %s sector %lld, len %d,\n", @@ -1200,12 +1200,12 @@ out: } static int btt_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int op) + struct page *page, bool is_write) { struct btt *btt = bdev->bd_disk->private_data; - btt_do_bvec(btt, NULL, page, PAGE_SIZE, 0, op, sector); - page_endio(page, op, 0); + btt_do_bvec(btt, NULL, page, PAGE_SIZE, 0, is_write, sector); + page_endio(page, is_write, 0); return 0; } diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index d64d92481c1d..20bae50c231d 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -67,7 +67,7 @@ static void pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset, } static int pmem_do_bvec(struct pmem_device *pmem, struct page *page, - unsigned int len, unsigned int off, int op, + unsigned int len, unsigned int off, bool is_write, sector_t sector) { int rc = 0; @@ -79,7 +79,7 @@ static int pmem_do_bvec(struct pmem_device *pmem, struct page *page, if (unlikely(is_bad_pmem(&pmem->bb, sector, len))) bad_pmem = true; - if (!op_is_write(op)) { + if (!is_write) { if (unlikely(bad_pmem)) rc = -EIO; else { @@ -134,7 +134,7 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) do_acct = nd_iostat_start(bio, &start); bio_for_each_segment(bvec, bio, iter) { rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len, - bvec.bv_offset, bio_op(bio), + bvec.bv_offset, op_is_write(bio_op(bio)), iter.bi_sector); if (rc) { bio->bi_error = rc; @@ -152,12 +152,12 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) } static int pmem_rw_page(struct block_device *bdev, sector_t sector, - struct page *page, int op) + struct page *page, bool is_write) { struct pmem_device *pmem = bdev->bd_queue->queuedata; int rc; - rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, op, sector); + rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, is_write, sector); /* * The ->rw_page interface is subtle and tricky. The core @@ -166,7 +166,7 @@ static int pmem_rw_page(struct block_device *bdev, sector_t sector, * caused by double completion. */ if (rc == 0) - page_endio(page, op, 0); + page_endio(page, is_write, 0); return rc; } diff --git a/fs/block_dev.c b/fs/block_dev.c index d402899ba135..c3cdde87cc8c 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -416,8 +416,7 @@ int bdev_read_page(struct block_device *bdev, sector_t sector, result = blk_queue_enter(bdev->bd_queue, false); if (result) return result; - result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, - REQ_OP_READ); + result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, false); blk_queue_exit(bdev->bd_queue); return result; } @@ -455,8 +454,7 @@ int bdev_write_page(struct block_device *bdev, sector_t sector, return result; set_page_writeback(page); - result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, - REQ_OP_WRITE); + result = ops->rw_page(bdev, sector + get_start_sect(bdev), page, true); if (result) end_page_writeback(page); else diff --git a/fs/mpage.c b/fs/mpage.c index 7a09c55b4bd0..d2413af0823a 100644 --- a/fs/mpage.c +++ b/fs/mpage.c @@ -50,7 +50,7 @@ static void mpage_end_io(struct bio *bio) bio_for_each_segment_all(bv, bio, i) { struct page *page = bv->bv_page; - page_endio(page, bio_op(bio), bio->bi_error); + page_endio(page, op_is_write(bio_op(bio)), bio->bi_error); } bio_put(bio); diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h index 14b28ff2caf8..f254eb264924 100644 --- a/include/linux/blk_types.h +++ b/include/linux/blk_types.h @@ -18,17 +18,6 @@ struct cgroup_subsys_state; typedef void (bio_end_io_t) (struct bio *); typedef void (bio_destructor_t) (struct bio *); -enum req_op { - REQ_OP_READ, - REQ_OP_WRITE, - REQ_OP_DISCARD, /* request to discard sectors */ - REQ_OP_SECURE_ERASE, /* request to securely erase sectors */ - REQ_OP_WRITE_SAME, /* write same block many times */ - REQ_OP_FLUSH, /* request for cache flush */ -}; - -#define REQ_OP_BITS 3 - #ifdef CONFIG_BLOCK /* * main unit of I/O for the block layer and lower layers (ie drivers and @@ -239,6 +228,17 @@ enum rq_flag_bits { #define REQ_HASHED (1ULL << __REQ_HASHED) #define REQ_MQ_INFLIGHT (1ULL << __REQ_MQ_INFLIGHT) +enum req_op { + REQ_OP_READ, + REQ_OP_WRITE, + REQ_OP_DISCARD, /* request to discard sectors */ + REQ_OP_SECURE_ERASE, /* request to securely erase sectors */ + REQ_OP_WRITE_SAME, /* write same block many times */ + REQ_OP_FLUSH, /* request for cache flush */ +}; + +#define REQ_OP_BITS 3 + typedef unsigned int blk_qc_t; #define BLK_QC_T_NONE -1U #define BLK_QC_T_SHIFT 16 diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index ccd68c0d01de..2c210b6a7bcf 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1672,7 +1672,7 @@ struct blk_dax_ctl { struct block_device_operations { int (*open) (struct block_device *, fmode_t); void (*release) (struct gendisk *, fmode_t); - int (*rw_page)(struct block_device *, sector_t, struct page *, int op); + int (*rw_page)(struct block_device *, sector_t, struct page *, bool); int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); long (*direct_access)(struct block_device *, sector_t, void **, pfn_t *, diff --git a/include/linux/fs.h b/include/linux/fs.h index 33f0e96db06f..3523bf62f328 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -2480,13 +2480,12 @@ extern void init_special_inode(struct inode *, umode_t, dev_t); extern void make_bad_inode(struct inode *); extern bool is_bad_inode(struct inode *); +#ifdef CONFIG_BLOCK static inline bool op_is_write(unsigned int op) { return op == REQ_OP_READ ? false : true; } -#ifdef CONFIG_BLOCK - /* * return data direction, READ or WRITE */ diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index 45786374abbd..66a1260b33de 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -510,7 +510,7 @@ static inline void wait_on_page_writeback(struct page *page) extern void end_page_writeback(struct page *page); void wait_for_stable_page(struct page *page); -void page_endio(struct page *page, int op, int err); +void page_endio(struct page *page, bool is_write, int err); /* * Add an arbitrary waiter to a page's wait queue diff --git a/mm/filemap.c b/mm/filemap.c index daef091d4c50..8a287dfc5372 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -887,9 +887,9 @@ EXPORT_SYMBOL(end_page_writeback); * After completing I/O on a page, call this routine to update the page * flags appropriately */ -void page_endio(struct page *page, int op, int err) +void page_endio(struct page *page, bool is_write, int err) { - if (!op_is_write(op)) { + if (!is_write) { if (!err) { SetPageUptodate(page); } else { -- cgit v1.2.3 From 1eff9d322a444245c67515edb52bc0eb68374aa8 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 5 Aug 2016 15:35:16 -0600 Subject: block: rename bio bi_rw to bi_opf Since commit 63a4cc24867d, bio->bi_rw contains flags in the lower portion and the op code in the higher portions. This means that old code that relies on manually setting bi_rw is most likely going to be broken. Instead of letting that brokeness linger, rename the member, to force old and out-of-tree code to break at compile time instead of at runtime. No intended functional changes in this commit. Signed-off-by: Jens Axboe --- Documentation/block/biodoc.txt | 4 ++-- Documentation/device-mapper/dm-flakey.txt | 2 +- block/bio-integrity.c | 2 +- block/bio.c | 6 +++--- block/blk-core.c | 26 +++++++++++++------------- block/blk-merge.c | 8 ++++---- block/blk-mq.c | 10 +++++----- block/blk-throttle.c | 8 ++++---- block/cfq-iosched.c | 4 ++-- drivers/block/drbd/drbd_main.c | 8 ++++---- drivers/block/drbd/drbd_receiver.c | 2 +- drivers/block/drbd/drbd_req.c | 6 +++--- drivers/block/drbd/drbd_worker.c | 2 +- drivers/block/pktcdvd.c | 2 +- drivers/block/umem.c | 2 +- drivers/md/bcache/request.c | 12 ++++++------ drivers/md/bcache/super.c | 2 +- drivers/md/bcache/writeback.h | 2 +- drivers/md/dm-cache-target.c | 8 ++++---- drivers/md/dm-crypt.c | 4 ++-- drivers/md/dm-era-target.c | 2 +- drivers/md/dm-flakey.c | 6 +++--- drivers/md/dm-io.c | 6 +++--- drivers/md/dm-log-writes.c | 4 ++-- drivers/md/dm-mpath.c | 2 +- drivers/md/dm-raid1.c | 10 +++++----- drivers/md/dm-region-hash.c | 4 ++-- drivers/md/dm-snap.c | 6 +++--- drivers/md/dm-stripe.c | 4 ++-- drivers/md/dm-thin.c | 8 ++++---- drivers/md/dm-zero.c | 2 +- drivers/md/dm.c | 10 +++++----- drivers/md/linear.c | 2 +- drivers/md/md.c | 4 ++-- drivers/md/multipath.c | 8 ++++---- drivers/md/raid0.c | 2 +- drivers/md/raid1.c | 6 +++--- drivers/md/raid10.c | 8 ++++---- drivers/md/raid5-cache.c | 2 +- drivers/md/raid5.c | 20 ++++++++++---------- drivers/nvdimm/pmem.c | 4 ++-- fs/btrfs/check-integrity.c | 10 +++++----- fs/btrfs/disk-io.c | 2 +- fs/btrfs/inode.c | 6 +++--- fs/btrfs/volumes.c | 6 +++--- include/linux/bio.h | 4 ++-- include/linux/blk-cgroup.h | 4 ++-- include/linux/blk_types.h | 15 ++++++++------- include/trace/events/bcache.h | 8 ++++---- include/trace/events/block.h | 14 +++++++------- kernel/trace/blktrace.c | 6 +++--- 51 files changed, 158 insertions(+), 157 deletions(-) (limited to 'fs') diff --git a/Documentation/block/biodoc.txt b/Documentation/block/biodoc.txt index 026d13362aca..bcdb2b4c1f12 100644 --- a/Documentation/block/biodoc.txt +++ b/Documentation/block/biodoc.txt @@ -269,7 +269,7 @@ Arjan's proposed request priority scheme allows higher levels some broad requests which haven't aged too much on the queue. Potentially this priority could even be exposed to applications in some manner, providing higher level tunability. Time based aging avoids starvation of lower priority - requests. Some bits in the bi_rw flags field in the bio structure are + requests. Some bits in the bi_opf flags field in the bio structure are intended to be used for this priority information. @@ -432,7 +432,7 @@ struct bio { struct bio *bi_next; /* request queue link */ struct block_device *bi_bdev; /* target device */ unsigned long bi_flags; /* status, command, etc */ - unsigned long bi_rw; /* low bits: r/w, high: priority */ + unsigned long bi_opf; /* low bits: r/w, high: priority */ unsigned int bi_vcnt; /* how may bio_vec's */ struct bvec_iter bi_iter; /* current index into bio_vec array */ diff --git a/Documentation/device-mapper/dm-flakey.txt b/Documentation/device-mapper/dm-flakey.txt index 6ff5c2327227..c43030718cef 100644 --- a/Documentation/device-mapper/dm-flakey.txt +++ b/Documentation/device-mapper/dm-flakey.txt @@ -42,7 +42,7 @@ Optional feature parameters: : Either 'r' to corrupt reads or 'w' to corrupt writes. 'w' is incompatible with drop_writes. : The value (from 0-255) to write. - : Perform the replacement only if bio->bi_rw has all the + : Perform the replacement only if bio->bi_opf has all the selected flags set. Examples: diff --git a/block/bio-integrity.c b/block/bio-integrity.c index f70cc3bdfd01..63f72f00c72e 100644 --- a/block/bio-integrity.c +++ b/block/bio-integrity.c @@ -86,7 +86,7 @@ struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio, bip->bip_bio = bio; bio->bi_integrity = bip; - bio->bi_rw |= REQ_INTEGRITY; + bio->bi_opf |= REQ_INTEGRITY; return bip; err: diff --git a/block/bio.c b/block/bio.c index 3f76a38a5e2d..f39477538fef 100644 --- a/block/bio.c +++ b/block/bio.c @@ -580,7 +580,7 @@ void __bio_clone_fast(struct bio *bio, struct bio *bio_src) */ bio->bi_bdev = bio_src->bi_bdev; bio_set_flag(bio, BIO_CLONED); - bio->bi_rw = bio_src->bi_rw; + bio->bi_opf = bio_src->bi_opf; bio->bi_iter = bio_src->bi_iter; bio->bi_io_vec = bio_src->bi_io_vec; @@ -663,7 +663,7 @@ struct bio *bio_clone_bioset(struct bio *bio_src, gfp_t gfp_mask, if (!bio) return NULL; bio->bi_bdev = bio_src->bi_bdev; - bio->bi_rw = bio_src->bi_rw; + bio->bi_opf = bio_src->bi_opf; bio->bi_iter.bi_sector = bio_src->bi_iter.bi_sector; bio->bi_iter.bi_size = bio_src->bi_iter.bi_size; @@ -873,7 +873,7 @@ int submit_bio_wait(struct bio *bio) init_completion(&ret.event); bio->bi_private = &ret; bio->bi_end_io = submit_bio_wait_endio; - bio->bi_rw |= REQ_SYNC; + bio->bi_opf |= REQ_SYNC; submit_bio(bio); wait_for_completion_io(&ret.event); diff --git a/block/blk-core.c b/block/blk-core.c index a687e9cc16c2..999442ec4601 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -1029,7 +1029,7 @@ static bool blk_rq_should_init_elevator(struct bio *bio) * Flush requests do not use the elevator so skip initialization. * This allows a request to share the flush and elevator data. */ - if (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA)) + if (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA)) return false; return true; @@ -1504,7 +1504,7 @@ EXPORT_SYMBOL_GPL(blk_add_request_payload); bool bio_attempt_back_merge(struct request_queue *q, struct request *req, struct bio *bio) { - const int ff = bio->bi_rw & REQ_FAILFAST_MASK; + const int ff = bio->bi_opf & REQ_FAILFAST_MASK; if (!ll_back_merge_fn(q, req, bio)) return false; @@ -1526,7 +1526,7 @@ bool bio_attempt_back_merge(struct request_queue *q, struct request *req, bool bio_attempt_front_merge(struct request_queue *q, struct request *req, struct bio *bio) { - const int ff = bio->bi_rw & REQ_FAILFAST_MASK; + const int ff = bio->bi_opf & REQ_FAILFAST_MASK; if (!ll_front_merge_fn(q, req, bio)) return false; @@ -1648,8 +1648,8 @@ void init_request_from_bio(struct request *req, struct bio *bio) { req->cmd_type = REQ_TYPE_FS; - req->cmd_flags |= bio->bi_rw & REQ_COMMON_MASK; - if (bio->bi_rw & REQ_RAHEAD) + req->cmd_flags |= bio->bi_opf & REQ_COMMON_MASK; + if (bio->bi_opf & REQ_RAHEAD) req->cmd_flags |= REQ_FAILFAST_MASK; req->errors = 0; @@ -1660,7 +1660,7 @@ void init_request_from_bio(struct request *req, struct bio *bio) static blk_qc_t blk_queue_bio(struct request_queue *q, struct bio *bio) { - const bool sync = !!(bio->bi_rw & REQ_SYNC); + const bool sync = !!(bio->bi_opf & REQ_SYNC); struct blk_plug *plug; int el_ret, rw_flags = 0, where = ELEVATOR_INSERT_SORT; struct request *req; @@ -1681,7 +1681,7 @@ static blk_qc_t blk_queue_bio(struct request_queue *q, struct bio *bio) return BLK_QC_T_NONE; } - if (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA)) { + if (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA)) { spin_lock_irq(q->queue_lock); where = ELEVATOR_INSERT_FLUSH; goto get_rq; @@ -1728,7 +1728,7 @@ get_rq: /* * Add in META/PRIO flags, if set, before we get to the IO scheduler */ - rw_flags |= (bio->bi_rw & (REQ_META | REQ_PRIO)); + rw_flags |= (bio->bi_opf & (REQ_META | REQ_PRIO)); /* * Grab a free request. This is might sleep but can not fail. @@ -1805,7 +1805,7 @@ static void handle_bad_sector(struct bio *bio) printk(KERN_INFO "attempt to access beyond end of device\n"); printk(KERN_INFO "%s: rw=%d, want=%Lu, limit=%Lu\n", bdevname(bio->bi_bdev, b), - bio->bi_rw, + bio->bi_opf, (unsigned long long)bio_end_sector(bio), (long long)(i_size_read(bio->bi_bdev->bd_inode) >> 9)); } @@ -1918,9 +1918,9 @@ generic_make_request_checks(struct bio *bio) * drivers without flush support don't have to worry * about them. */ - if ((bio->bi_rw & (REQ_PREFLUSH | REQ_FUA)) && + if ((bio->bi_opf & (REQ_PREFLUSH | REQ_FUA)) && !test_bit(QUEUE_FLAG_WC, &q->queue_flags)) { - bio->bi_rw &= ~(REQ_PREFLUSH | REQ_FUA); + bio->bi_opf &= ~(REQ_PREFLUSH | REQ_FUA); if (!nr_sectors) { err = 0; goto end_io; @@ -2219,7 +2219,7 @@ unsigned int blk_rq_err_bytes(const struct request *rq) * one. */ for (bio = rq->bio; bio; bio = bio->bi_next) { - if ((bio->bi_rw & ff) != ff) + if ((bio->bi_opf & ff) != ff) break; bytes += bio->bi_iter.bi_size; } @@ -2630,7 +2630,7 @@ bool blk_update_request(struct request *req, int error, unsigned int nr_bytes) /* mixed attributes always follow the first bio */ if (req->cmd_flags & REQ_MIXED_MERGE) { req->cmd_flags &= ~REQ_FAILFAST_MASK; - req->cmd_flags |= req->bio->bi_rw & REQ_FAILFAST_MASK; + req->cmd_flags |= req->bio->bi_opf & REQ_FAILFAST_MASK; } /* diff --git a/block/blk-merge.c b/block/blk-merge.c index 41cbd4878958..3eec75a9e91d 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -186,7 +186,7 @@ void blk_queue_split(struct request_queue *q, struct bio **bio, if (split) { /* there isn't chance to merge the splitted bio */ - split->bi_rw |= REQ_NOMERGE; + split->bi_opf |= REQ_NOMERGE; bio_chain(split, *bio); trace_block_split(q, split, (*bio)->bi_iter.bi_sector); @@ -616,9 +616,9 @@ void blk_rq_set_mixed_merge(struct request *rq) * Distributes the attributs to each bio. */ for (bio = rq->bio; bio; bio = bio->bi_next) { - WARN_ON_ONCE((bio->bi_rw & REQ_FAILFAST_MASK) && - (bio->bi_rw & REQ_FAILFAST_MASK) != ff); - bio->bi_rw |= ff; + WARN_ON_ONCE((bio->bi_opf & REQ_FAILFAST_MASK) && + (bio->bi_opf & REQ_FAILFAST_MASK) != ff); + bio->bi_opf |= ff; } rq->cmd_flags |= REQ_MIXED_MERGE; } diff --git a/block/blk-mq.c b/block/blk-mq.c index 6a63da101bc4..e931a0e8e73d 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1234,7 +1234,7 @@ static struct request *blk_mq_map_request(struct request_queue *q, ctx = blk_mq_get_ctx(q); hctx = q->mq_ops->map_queue(q, ctx->cpu); - if (rw_is_sync(bio_op(bio), bio->bi_rw)) + if (rw_is_sync(bio_op(bio), bio->bi_opf)) op_flags |= REQ_SYNC; trace_block_getrq(q, bio, op); @@ -1302,8 +1302,8 @@ static int blk_mq_direct_issue_request(struct request *rq, blk_qc_t *cookie) */ static blk_qc_t blk_mq_make_request(struct request_queue *q, struct bio *bio) { - const int is_sync = rw_is_sync(bio_op(bio), bio->bi_rw); - const int is_flush_fua = bio->bi_rw & (REQ_PREFLUSH | REQ_FUA); + const int is_sync = rw_is_sync(bio_op(bio), bio->bi_opf); + const int is_flush_fua = bio->bi_opf & (REQ_PREFLUSH | REQ_FUA); struct blk_map_ctx data; struct request *rq; unsigned int request_count = 0; @@ -1396,8 +1396,8 @@ done: */ static blk_qc_t blk_sq_make_request(struct request_queue *q, struct bio *bio) { - const int is_sync = rw_is_sync(bio_op(bio), bio->bi_rw); - const int is_flush_fua = bio->bi_rw & (REQ_PREFLUSH | REQ_FUA); + const int is_sync = rw_is_sync(bio_op(bio), bio->bi_opf); + const int is_flush_fua = bio->bi_opf & (REQ_PREFLUSH | REQ_FUA); struct blk_plug *plug; unsigned int request_count = 0; struct blk_map_ctx data; diff --git a/block/blk-throttle.c b/block/blk-throttle.c index c5494e403239..f1aba26f4719 100644 --- a/block/blk-throttle.c +++ b/block/blk-throttle.c @@ -821,8 +821,8 @@ static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio) * second time when it eventually gets issued. Set it when a bio * is being charged to a tg. */ - if (!(bio->bi_rw & REQ_THROTTLED)) - bio->bi_rw |= REQ_THROTTLED; + if (!(bio->bi_opf & REQ_THROTTLED)) + bio->bi_opf |= REQ_THROTTLED; } /** @@ -1399,7 +1399,7 @@ bool blk_throtl_bio(struct request_queue *q, struct blkcg_gq *blkg, WARN_ON_ONCE(!rcu_read_lock_held()); /* see throtl_charge_bio() */ - if ((bio->bi_rw & REQ_THROTTLED) || !tg->has_rules[rw]) + if ((bio->bi_opf & REQ_THROTTLED) || !tg->has_rules[rw]) goto out; spin_lock_irq(q->queue_lock); @@ -1478,7 +1478,7 @@ out: * being issued. */ if (!throttled) - bio->bi_rw &= ~REQ_THROTTLED; + bio->bi_opf &= ~REQ_THROTTLED; return throttled; } diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index acabba198de9..cc2f6dbd4303 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -918,7 +918,7 @@ static inline struct cfq_data *cic_to_cfqd(struct cfq_io_cq *cic) */ static inline bool cfq_bio_sync(struct bio *bio) { - return bio_data_dir(bio) == READ || (bio->bi_rw & REQ_SYNC); + return bio_data_dir(bio) == READ || (bio->bi_opf & REQ_SYNC); } /* @@ -2565,7 +2565,7 @@ static void cfq_merged_request(struct request_queue *q, struct request *req, static void cfq_bio_merged(struct request_queue *q, struct request *req, struct bio *bio) { - cfqg_stats_update_io_merged(RQ_CFQG(req), bio_op(bio), bio->bi_rw); + cfqg_stats_update_io_merged(RQ_CFQG(req), bio_op(bio), bio->bi_opf); } static void diff --git a/drivers/block/drbd/drbd_main.c b/drivers/block/drbd/drbd_main.c index 0501ae0c517b..100be556e613 100644 --- a/drivers/block/drbd/drbd_main.c +++ b/drivers/block/drbd/drbd_main.c @@ -1663,13 +1663,13 @@ static u32 bio_flags_to_wire(struct drbd_connection *connection, struct bio *bio) { if (connection->agreed_pro_version >= 95) - return (bio->bi_rw & REQ_SYNC ? DP_RW_SYNC : 0) | - (bio->bi_rw & REQ_FUA ? DP_FUA : 0) | - (bio->bi_rw & REQ_PREFLUSH ? DP_FLUSH : 0) | + return (bio->bi_opf & REQ_SYNC ? DP_RW_SYNC : 0) | + (bio->bi_opf & REQ_FUA ? DP_FUA : 0) | + (bio->bi_opf & REQ_PREFLUSH ? DP_FLUSH : 0) | (bio_op(bio) == REQ_OP_WRITE_SAME ? DP_WSAME : 0) | (bio_op(bio) == REQ_OP_DISCARD ? DP_DISCARD : 0); else - return bio->bi_rw & REQ_SYNC ? DP_RW_SYNC : 0; + return bio->bi_opf & REQ_SYNC ? DP_RW_SYNC : 0; } /* Used to send write or TRIM aka REQ_DISCARD requests diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c index df45713dfbe8..942384f34e22 100644 --- a/drivers/block/drbd/drbd_receiver.c +++ b/drivers/block/drbd/drbd_receiver.c @@ -1564,7 +1564,7 @@ static void drbd_issue_peer_wsame(struct drbd_device *device, * drbd_submit_peer_request() * @device: DRBD device. * @peer_req: peer request - * @rw: flag field, see bio->bi_rw + * @rw: flag field, see bio->bi_opf * * May spread the pages to multiple bios, * depending on bio_add_page restrictions. diff --git a/drivers/block/drbd/drbd_req.c b/drivers/block/drbd/drbd_req.c index 66b8e4bb74d8..de279fe4e4fd 100644 --- a/drivers/block/drbd/drbd_req.c +++ b/drivers/block/drbd/drbd_req.c @@ -288,7 +288,7 @@ void drbd_req_complete(struct drbd_request *req, struct bio_and_error *m) */ if (!ok && bio_op(req->master_bio) == REQ_OP_READ && - !(req->master_bio->bi_rw & REQ_RAHEAD) && + !(req->master_bio->bi_opf & REQ_RAHEAD) && !list_empty(&req->tl_requests)) req->rq_state |= RQ_POSTPONED; @@ -1137,7 +1137,7 @@ static int drbd_process_write_request(struct drbd_request *req) * replicating, in which case there is no point. */ if (unlikely(req->i.size == 0)) { /* The only size==0 bios we expect are empty flushes. */ - D_ASSERT(device, req->master_bio->bi_rw & REQ_PREFLUSH); + D_ASSERT(device, req->master_bio->bi_opf & REQ_PREFLUSH); if (remote) _req_mod(req, QUEUE_AS_DRBD_BARRIER); return remote; @@ -1176,7 +1176,7 @@ drbd_submit_req_private_bio(struct drbd_request *req) if (bio_op(bio) != REQ_OP_READ) type = DRBD_FAULT_DT_WR; - else if (bio->bi_rw & REQ_RAHEAD) + else if (bio->bi_opf & REQ_RAHEAD) type = DRBD_FAULT_DT_RA; else type = DRBD_FAULT_DT_RD; diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c index 35dbb3dca47e..c6755c9a0aea 100644 --- a/drivers/block/drbd/drbd_worker.c +++ b/drivers/block/drbd/drbd_worker.c @@ -256,7 +256,7 @@ void drbd_request_endio(struct bio *bio) what = DISCARD_COMPLETED_WITH_ERROR; break; case REQ_OP_READ: - if (bio->bi_rw & REQ_RAHEAD) + if (bio->bi_opf & REQ_RAHEAD) what = READ_AHEAD_COMPLETED_WITH_ERROR; else what = READ_COMPLETED_WITH_ERROR; diff --git a/drivers/block/pktcdvd.c b/drivers/block/pktcdvd.c index 9393bc730acf..90fa4ac149db 100644 --- a/drivers/block/pktcdvd.c +++ b/drivers/block/pktcdvd.c @@ -1157,7 +1157,7 @@ static int pkt_start_recovery(struct packet_data *pkt) bio_reset(pkt->bio); pkt->bio->bi_bdev = pd->bdev; - pkt->bio->bi_rw = REQ_WRITE; + bio_set_op_attrs(pkt->bio, REQ_OP_WRITE, 0); pkt->bio->bi_iter.bi_sector = new_sector; pkt->bio->bi_iter.bi_size = pkt->frames * CD_FRAMESIZE; pkt->bio->bi_vcnt = pkt->frames; diff --git a/drivers/block/umem.c b/drivers/block/umem.c index d0a3e6d4515f..be90e15854ed 100644 --- a/drivers/block/umem.c +++ b/drivers/block/umem.c @@ -535,7 +535,7 @@ static blk_qc_t mm_make_request(struct request_queue *q, struct bio *bio) *card->biotail = bio; bio->bi_next = NULL; card->biotail = &bio->bi_next; - if (bio->bi_rw & REQ_SYNC || !mm_check_plugged(card)) + if (bio->bi_opf & REQ_SYNC || !mm_check_plugged(card)) activate(card); spin_unlock_irq(&card->lock); diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c index 69f16f43f8ab..4b177fe11ebb 100644 --- a/drivers/md/bcache/request.c +++ b/drivers/md/bcache/request.c @@ -208,7 +208,7 @@ static void bch_data_insert_start(struct closure *cl) * Journal writes are marked REQ_PREFLUSH; if the original write was a * flush, it'll wait on the journal write. */ - bio->bi_rw &= ~(REQ_PREFLUSH|REQ_FUA); + bio->bi_opf &= ~(REQ_PREFLUSH|REQ_FUA); do { unsigned i; @@ -405,7 +405,7 @@ static bool check_should_bypass(struct cached_dev *dc, struct bio *bio) if (!congested && mode == CACHE_MODE_WRITEBACK && op_is_write(bio_op(bio)) && - (bio->bi_rw & REQ_SYNC)) + (bio->bi_opf & REQ_SYNC)) goto rescale; spin_lock(&dc->io_lock); @@ -668,7 +668,7 @@ static inline struct search *search_alloc(struct bio *bio, s->iop.write_prio = 0; s->iop.error = 0; s->iop.flags = 0; - s->iop.flush_journal = (bio->bi_rw & (REQ_PREFLUSH|REQ_FUA)) != 0; + s->iop.flush_journal = (bio->bi_opf & (REQ_PREFLUSH|REQ_FUA)) != 0; s->iop.wq = bcache_wq; return s; @@ -796,8 +796,8 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s, goto out_submit; } - if (!(bio->bi_rw & REQ_RAHEAD) && - !(bio->bi_rw & REQ_META) && + if (!(bio->bi_opf & REQ_RAHEAD) && + !(bio->bi_opf & REQ_META) && s->iop.c->gc_stats.in_use < CUTOFF_CACHE_READA) reada = min_t(sector_t, dc->readahead >> 9, bdev_sectors(bio->bi_bdev) - bio_end_sector(bio)); @@ -920,7 +920,7 @@ static void cached_dev_write(struct cached_dev *dc, struct search *s) bch_writeback_add(dc); s->iop.bio = bio; - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { /* Also need to send a flush to the backing device */ struct bio *flush = bio_alloc_bioset(GFP_NOIO, 0, dc->disk.bio_split); diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c index 88ef6d14cce3..95a4ca6ce6ff 100644 --- a/drivers/md/bcache/super.c +++ b/drivers/md/bcache/super.c @@ -347,7 +347,7 @@ static void uuid_io(struct cache_set *c, int op, unsigned long op_flags, for (i = 0; i < KEY_PTRS(k); i++) { struct bio *bio = bch_bbio_alloc(c); - bio->bi_rw = REQ_SYNC|REQ_META|op_flags; + bio->bi_opf = REQ_SYNC | REQ_META | op_flags; bio->bi_iter.bi_size = KEY_SIZE(k) << 9; bio->bi_end_io = uuid_endio; diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h index 073a042aed24..301eaf565167 100644 --- a/drivers/md/bcache/writeback.h +++ b/drivers/md/bcache/writeback.h @@ -57,7 +57,7 @@ static inline bool should_writeback(struct cached_dev *dc, struct bio *bio, if (would_skip) return false; - return bio->bi_rw & REQ_SYNC || + return bio->bi_opf & REQ_SYNC || in_use <= CUTOFF_WRITEBACK; } diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c index 718744db62df..59b2c50562e4 100644 --- a/drivers/md/dm-cache-target.c +++ b/drivers/md/dm-cache-target.c @@ -788,7 +788,7 @@ static void check_if_tick_bio_needed(struct cache *cache, struct bio *bio) spin_lock_irqsave(&cache->lock, flags); if (cache->need_tick_bio && - !(bio->bi_rw & (REQ_FUA | REQ_PREFLUSH)) && + !(bio->bi_opf & (REQ_FUA | REQ_PREFLUSH)) && bio_op(bio) != REQ_OP_DISCARD) { pb->tick = true; cache->need_tick_bio = false; @@ -830,7 +830,7 @@ static dm_oblock_t get_bio_block(struct cache *cache, struct bio *bio) static int bio_triggers_commit(struct cache *cache, struct bio *bio) { - return bio->bi_rw & (REQ_PREFLUSH | REQ_FUA); + return bio->bi_opf & (REQ_PREFLUSH | REQ_FUA); } /* @@ -1069,7 +1069,7 @@ static void dec_io_migrations(struct cache *cache) static bool discard_or_flush(struct bio *bio) { return bio_op(bio) == REQ_OP_DISCARD || - bio->bi_rw & (REQ_PREFLUSH | REQ_FUA); + bio->bi_opf & (REQ_PREFLUSH | REQ_FUA); } static void __cell_defer(struct cache *cache, struct dm_bio_prison_cell *cell) @@ -1980,7 +1980,7 @@ static void process_deferred_bios(struct cache *cache) bio = bio_list_pop(&bios); - if (bio->bi_rw & REQ_PREFLUSH) + if (bio->bi_opf & REQ_PREFLUSH) process_flush_bio(cache, bio); else if (bio_op(bio) == REQ_OP_DISCARD) process_discard_bio(cache, &structs, bio); diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c index 8f2e3e2ffd26..4e9784b4e0ac 100644 --- a/drivers/md/dm-crypt.c +++ b/drivers/md/dm-crypt.c @@ -1136,7 +1136,7 @@ static void clone_init(struct dm_crypt_io *io, struct bio *clone) clone->bi_private = io; clone->bi_end_io = crypt_endio; clone->bi_bdev = cc->dev->bdev; - bio_set_op_attrs(clone, bio_op(io->base_bio), io->base_bio->bi_rw); + bio_set_op_attrs(clone, bio_op(io->base_bio), io->base_bio->bi_opf); } static int kcryptd_io_read(struct dm_crypt_io *io, gfp_t gfp) @@ -1915,7 +1915,7 @@ static int crypt_map(struct dm_target *ti, struct bio *bio) * - for REQ_PREFLUSH device-mapper core ensures that no IO is in-flight * - for REQ_OP_DISCARD caller must use flush if IO ordering matters */ - if (unlikely(bio->bi_rw & REQ_PREFLUSH || + if (unlikely(bio->bi_opf & REQ_PREFLUSH || bio_op(bio) == REQ_OP_DISCARD)) { bio->bi_bdev = cc->dev->bdev; if (bio_sectors(bio)) diff --git a/drivers/md/dm-era-target.c b/drivers/md/dm-era-target.c index 2faf49d8f4d7..bf2b2676cb8a 100644 --- a/drivers/md/dm-era-target.c +++ b/drivers/md/dm-era-target.c @@ -1542,7 +1542,7 @@ static int era_map(struct dm_target *ti, struct bio *bio) /* * REQ_PREFLUSH bios carry no data, so we're not interested in them. */ - if (!(bio->bi_rw & REQ_PREFLUSH) && + if (!(bio->bi_opf & REQ_PREFLUSH) && (bio_data_dir(bio) == WRITE) && !metadata_current_marked(era->md, block)) { defer_bio(era, bio); diff --git a/drivers/md/dm-flakey.c b/drivers/md/dm-flakey.c index 19db13e99466..97e446d54a15 100644 --- a/drivers/md/dm-flakey.c +++ b/drivers/md/dm-flakey.c @@ -16,7 +16,7 @@ #define DM_MSG_PREFIX "flakey" #define all_corrupt_bio_flags_match(bio, fc) \ - (((bio)->bi_rw & (fc)->corrupt_bio_flags) == (fc)->corrupt_bio_flags) + (((bio)->bi_opf & (fc)->corrupt_bio_flags) == (fc)->corrupt_bio_flags) /* * Flakey: Used for testing only, simulates intermittent, @@ -266,9 +266,9 @@ static void corrupt_bio_data(struct bio *bio, struct flakey_c *fc) data[fc->corrupt_bio_byte - 1] = fc->corrupt_bio_value; DMDEBUG("Corrupting data bio=%p by writing %u to byte %u " - "(rw=%c bi_rw=%u bi_sector=%llu cur_bytes=%u)\n", + "(rw=%c bi_opf=%u bi_sector=%llu cur_bytes=%u)\n", bio, fc->corrupt_bio_value, fc->corrupt_bio_byte, - (bio_data_dir(bio) == WRITE) ? 'w' : 'r', bio->bi_rw, + (bio_data_dir(bio) == WRITE) ? 'w' : 'r', bio->bi_opf, (unsigned long long)bio->bi_iter.bi_sector, bio_bytes); } } diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c index daa03e41654a..0bf1a12e35fe 100644 --- a/drivers/md/dm-io.c +++ b/drivers/md/dm-io.c @@ -505,9 +505,9 @@ static int dp_init(struct dm_io_request *io_req, struct dpages *dp, * New collapsed (a)synchronous interface. * * If the IO is asynchronous (i.e. it has notify.fn), you must either unplug - * the queue with blk_unplug() some time later or set REQ_SYNC in io_req->bi_rw. - * If you fail to do one of these, the IO will be submitted to the disk after - * q->unplug_delay, which defaults to 3ms in blk-settings.c. + * the queue with blk_unplug() some time later or set REQ_SYNC in + * io_req->bi_opf. If you fail to do one of these, the IO will be submitted to + * the disk after q->unplug_delay, which defaults to 3ms in blk-settings.c. */ int dm_io(struct dm_io_request *io_req, unsigned num_regions, struct dm_io_region *where, unsigned long *sync_error_bits) diff --git a/drivers/md/dm-log-writes.c b/drivers/md/dm-log-writes.c index b5dbf7a0515e..4ab68033f9d1 100644 --- a/drivers/md/dm-log-writes.c +++ b/drivers/md/dm-log-writes.c @@ -555,8 +555,8 @@ static int log_writes_map(struct dm_target *ti, struct bio *bio) struct bio_vec bv; size_t alloc_size; int i = 0; - bool flush_bio = (bio->bi_rw & REQ_PREFLUSH); - bool fua_bio = (bio->bi_rw & REQ_FUA); + bool flush_bio = (bio->bi_opf & REQ_PREFLUSH); + bool fua_bio = (bio->bi_opf & REQ_FUA); bool discard_bio = (bio_op(bio) == REQ_OP_DISCARD); pb->block = NULL; diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c index d7107d23b897..ac734e5bbe48 100644 --- a/drivers/md/dm-mpath.c +++ b/drivers/md/dm-mpath.c @@ -661,7 +661,7 @@ static int __multipath_map_bio(struct multipath *m, struct bio *bio, struct dm_m bio->bi_error = 0; bio->bi_bdev = pgpath->path.dev->bdev; - bio->bi_rw |= REQ_FAILFAST_TRANSPORT; + bio->bi_opf |= REQ_FAILFAST_TRANSPORT; if (pgpath->pg->ps.type->start_io) pgpath->pg->ps.type->start_io(&pgpath->pg->ps, diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c index dac55b254a09..bdf1606f67bc 100644 --- a/drivers/md/dm-raid1.c +++ b/drivers/md/dm-raid1.c @@ -657,7 +657,7 @@ static void do_write(struct mirror_set *ms, struct bio *bio) struct mirror *m; struct dm_io_request io_req = { .bi_op = REQ_OP_WRITE, - .bi_op_flags = bio->bi_rw & WRITE_FLUSH_FUA, + .bi_op_flags = bio->bi_opf & WRITE_FLUSH_FUA, .mem.type = DM_IO_BIO, .mem.ptr.bio = bio, .notify.fn = write_callback, @@ -704,7 +704,7 @@ static void do_writes(struct mirror_set *ms, struct bio_list *writes) bio_list_init(&requeue); while ((bio = bio_list_pop(writes))) { - if ((bio->bi_rw & REQ_PREFLUSH) || + if ((bio->bi_opf & REQ_PREFLUSH) || (bio_op(bio) == REQ_OP_DISCARD)) { bio_list_add(&sync, bio); continue; @@ -1217,7 +1217,7 @@ static int mirror_map(struct dm_target *ti, struct bio *bio) * If region is not in-sync queue the bio. */ if (!r || (r == -EWOULDBLOCK)) { - if (bio->bi_rw & REQ_RAHEAD) + if (bio->bi_opf & REQ_RAHEAD) return -EWOULDBLOCK; queue_bio(ms, bio, rw); @@ -1253,7 +1253,7 @@ static int mirror_end_io(struct dm_target *ti, struct bio *bio, int error) * We need to dec pending if this was a write. */ if (rw == WRITE) { - if (!(bio->bi_rw & REQ_PREFLUSH) && + if (!(bio->bi_opf & REQ_PREFLUSH) && bio_op(bio) != REQ_OP_DISCARD) dm_rh_dec(ms->rh, bio_record->write_region); return error; @@ -1262,7 +1262,7 @@ static int mirror_end_io(struct dm_target *ti, struct bio *bio, int error) if (error == -EOPNOTSUPP) goto out; - if ((error == -EWOULDBLOCK) && (bio->bi_rw & REQ_RAHEAD)) + if ((error == -EWOULDBLOCK) && (bio->bi_opf & REQ_RAHEAD)) goto out; if (unlikely(error)) { diff --git a/drivers/md/dm-region-hash.c b/drivers/md/dm-region-hash.c index b11813431f31..85c32b22a420 100644 --- a/drivers/md/dm-region-hash.c +++ b/drivers/md/dm-region-hash.c @@ -398,7 +398,7 @@ void dm_rh_mark_nosync(struct dm_region_hash *rh, struct bio *bio) region_t region = dm_rh_bio_to_region(rh, bio); int recovering = 0; - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { rh->flush_failure = 1; return; } @@ -526,7 +526,7 @@ void dm_rh_inc_pending(struct dm_region_hash *rh, struct bio_list *bios) struct bio *bio; for (bio = bios->head; bio; bio = bio->bi_next) { - if (bio->bi_rw & REQ_PREFLUSH || bio_op(bio) == REQ_OP_DISCARD) + if (bio->bi_opf & REQ_PREFLUSH || bio_op(bio) == REQ_OP_DISCARD) continue; rh_inc(rh, dm_rh_bio_to_region(rh, bio)); } diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c index ce2a910709f7..c65feeada864 100644 --- a/drivers/md/dm-snap.c +++ b/drivers/md/dm-snap.c @@ -1680,7 +1680,7 @@ static int snapshot_map(struct dm_target *ti, struct bio *bio) init_tracked_chunk(bio); - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { bio->bi_bdev = s->cow->bdev; return DM_MAPIO_REMAPPED; } @@ -1800,7 +1800,7 @@ static int snapshot_merge_map(struct dm_target *ti, struct bio *bio) init_tracked_chunk(bio); - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { if (!dm_bio_get_target_bio_nr(bio)) bio->bi_bdev = s->origin->bdev; else @@ -2286,7 +2286,7 @@ static int origin_map(struct dm_target *ti, struct bio *bio) bio->bi_bdev = o->dev->bdev; - if (unlikely(bio->bi_rw & REQ_PREFLUSH)) + if (unlikely(bio->bi_opf & REQ_PREFLUSH)) return DM_MAPIO_REMAPPED; if (bio_data_dir(bio) != WRITE) diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c index 83f1d4667195..28193a57bf47 100644 --- a/drivers/md/dm-stripe.c +++ b/drivers/md/dm-stripe.c @@ -286,7 +286,7 @@ static int stripe_map(struct dm_target *ti, struct bio *bio) uint32_t stripe; unsigned target_bio_nr; - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { target_bio_nr = dm_bio_get_target_bio_nr(bio); BUG_ON(target_bio_nr >= sc->stripes); bio->bi_bdev = sc->stripe[target_bio_nr].dev->bdev; @@ -383,7 +383,7 @@ static int stripe_end_io(struct dm_target *ti, struct bio *bio, int error) if (!error) return 0; /* I/O complete */ - if ((error == -EWOULDBLOCK) && (bio->bi_rw & REQ_RAHEAD)) + if ((error == -EWOULDBLOCK) && (bio->bi_opf & REQ_RAHEAD)) return error; if (error == -EOPNOTSUPP) diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c index 197ea2003400..d1c05c12a9db 100644 --- a/drivers/md/dm-thin.c +++ b/drivers/md/dm-thin.c @@ -699,7 +699,7 @@ static void remap_to_origin(struct thin_c *tc, struct bio *bio) static int bio_triggers_commit(struct thin_c *tc, struct bio *bio) { - return (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA)) && + return (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA)) && dm_thin_changed_this_transaction(tc->td); } @@ -870,7 +870,7 @@ static void __inc_remap_and_issue_cell(void *context, struct bio *bio; while ((bio = bio_list_pop(&cell->bios))) { - if (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA) || + if (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA) || bio_op(bio) == REQ_OP_DISCARD) bio_list_add(&info->defer_bios, bio); else { @@ -1717,7 +1717,7 @@ static void __remap_and_issue_shared_cell(void *context, while ((bio = bio_list_pop(&cell->bios))) { if ((bio_data_dir(bio) == WRITE) || - (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA) || + (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA) || bio_op(bio) == REQ_OP_DISCARD)) bio_list_add(&info->defer_bios, bio); else { @@ -2635,7 +2635,7 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio) return DM_MAPIO_SUBMITTED; } - if (bio->bi_rw & (REQ_PREFLUSH | REQ_FUA) || + if (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA) || bio_op(bio) == REQ_OP_DISCARD) { thin_defer_bio_with_throttle(tc, bio); return DM_MAPIO_SUBMITTED; diff --git a/drivers/md/dm-zero.c b/drivers/md/dm-zero.c index 618b8752dcf1..b616f11d8473 100644 --- a/drivers/md/dm-zero.c +++ b/drivers/md/dm-zero.c @@ -37,7 +37,7 @@ static int zero_map(struct dm_target *ti, struct bio *bio) { switch (bio_op(bio)) { case REQ_OP_READ: - if (bio->bi_rw & REQ_RAHEAD) { + if (bio->bi_opf & REQ_RAHEAD) { /* readahead of null bytes only wastes buffer cache */ return -EIO; } diff --git a/drivers/md/dm.c b/drivers/md/dm.c index dfa09e14e847..fa9b1cb4438a 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -798,12 +798,12 @@ static void dec_pending(struct dm_io *io, int error) if (io_error == DM_ENDIO_REQUEUE) return; - if ((bio->bi_rw & REQ_PREFLUSH) && bio->bi_iter.bi_size) { + if ((bio->bi_opf & REQ_PREFLUSH) && bio->bi_iter.bi_size) { /* * Preflush done for flush with data, reissue * without REQ_PREFLUSH. */ - bio->bi_rw &= ~REQ_PREFLUSH; + bio->bi_opf &= ~REQ_PREFLUSH; queue_io(md, bio); } else { /* done with normal IO or empty flush */ @@ -964,7 +964,7 @@ void dm_accept_partial_bio(struct bio *bio, unsigned n_sectors) { struct dm_target_io *tio = container_of(bio, struct dm_target_io, clone); unsigned bi_size = bio->bi_iter.bi_size >> SECTOR_SHIFT; - BUG_ON(bio->bi_rw & REQ_PREFLUSH); + BUG_ON(bio->bi_opf & REQ_PREFLUSH); BUG_ON(bi_size > *tio->len_ptr); BUG_ON(n_sectors > bi_size); *tio->len_ptr -= bi_size - n_sectors; @@ -1252,7 +1252,7 @@ static void __split_and_process_bio(struct mapped_device *md, start_io_acct(ci.io); - if (bio->bi_rw & REQ_PREFLUSH) { + if (bio->bi_opf & REQ_PREFLUSH) { ci.bio = &ci.md->flush_bio; ci.sector_count = 0; error = __send_empty_flush(&ci); @@ -1290,7 +1290,7 @@ static blk_qc_t dm_make_request(struct request_queue *q, struct bio *bio) if (unlikely(test_bit(DMF_BLOCK_IO_FOR_SUSPEND, &md->flags))) { dm_put_live_table(md, srcu_idx); - if (!(bio->bi_rw & REQ_RAHEAD)) + if (!(bio->bi_opf & REQ_RAHEAD)) queue_io(md, bio); else bio_io_error(bio); diff --git a/drivers/md/linear.c b/drivers/md/linear.c index 70ff888d25d0..86f5d435901d 100644 --- a/drivers/md/linear.c +++ b/drivers/md/linear.c @@ -221,7 +221,7 @@ static void linear_make_request(struct mddev *mddev, struct bio *bio) struct bio *split; sector_t start_sector, end_sector, data_offset; - if (unlikely(bio->bi_rw & REQ_PREFLUSH)) { + if (unlikely(bio->bi_opf & REQ_PREFLUSH)) { md_flush_request(mddev, bio); return; } diff --git a/drivers/md/md.c b/drivers/md/md.c index 2c3ab6f5e6be..d646f6e444f0 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -285,7 +285,7 @@ static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio) */ sectors = bio_sectors(bio); /* bio could be mergeable after passing to underlayer */ - bio->bi_rw &= ~REQ_NOMERGE; + bio->bi_opf &= ~REQ_NOMERGE; mddev->pers->make_request(mddev, bio); cpu = part_stat_lock(); @@ -414,7 +414,7 @@ static void md_submit_flush_data(struct work_struct *ws) /* an empty barrier - all done */ bio_endio(bio); else { - bio->bi_rw &= ~REQ_PREFLUSH; + bio->bi_opf &= ~REQ_PREFLUSH; mddev->pers->make_request(mddev, bio); } diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index 4974682842ae..673efbd6fc47 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c @@ -91,7 +91,7 @@ static void multipath_end_request(struct bio *bio) if (!bio->bi_error) multipath_end_bh_io(mp_bh, 0); - else if (!(bio->bi_rw & REQ_RAHEAD)) { + else if (!(bio->bi_opf & REQ_RAHEAD)) { /* * oops, IO error: */ @@ -112,7 +112,7 @@ static void multipath_make_request(struct mddev *mddev, struct bio * bio) struct multipath_bh * mp_bh; struct multipath_info *multipath; - if (unlikely(bio->bi_rw & REQ_PREFLUSH)) { + if (unlikely(bio->bi_opf & REQ_PREFLUSH)) { md_flush_request(mddev, bio); return; } @@ -135,7 +135,7 @@ static void multipath_make_request(struct mddev *mddev, struct bio * bio) mp_bh->bio.bi_iter.bi_sector += multipath->rdev->data_offset; mp_bh->bio.bi_bdev = multipath->rdev->bdev; - mp_bh->bio.bi_rw |= REQ_FAILFAST_TRANSPORT; + mp_bh->bio.bi_opf |= REQ_FAILFAST_TRANSPORT; mp_bh->bio.bi_end_io = multipath_end_request; mp_bh->bio.bi_private = mp_bh; generic_make_request(&mp_bh->bio); @@ -360,7 +360,7 @@ static void multipathd(struct md_thread *thread) bio->bi_iter.bi_sector += conf->multipaths[mp_bh->path].rdev->data_offset; bio->bi_bdev = conf->multipaths[mp_bh->path].rdev->bdev; - bio->bi_rw |= REQ_FAILFAST_TRANSPORT; + bio->bi_opf |= REQ_FAILFAST_TRANSPORT; bio->bi_end_io = multipath_end_request; bio->bi_private = mp_bh; generic_make_request(bio); diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index c3d439083212..258986a2699d 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c @@ -458,7 +458,7 @@ static void raid0_make_request(struct mddev *mddev, struct bio *bio) struct md_rdev *tmp_dev; struct bio *split; - if (unlikely(bio->bi_rw & REQ_PREFLUSH)) { + if (unlikely(bio->bi_opf & REQ_PREFLUSH)) { md_flush_request(mddev, bio); return; } diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 46168ef2e279..21dc00eb1989 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -1043,8 +1043,8 @@ static void raid1_make_request(struct mddev *mddev, struct bio * bio) unsigned long flags; const int op = bio_op(bio); const int rw = bio_data_dir(bio); - const unsigned long do_sync = (bio->bi_rw & REQ_SYNC); - const unsigned long do_flush_fua = (bio->bi_rw & + const unsigned long do_sync = (bio->bi_opf & REQ_SYNC); + const unsigned long do_flush_fua = (bio->bi_opf & (REQ_PREFLUSH | REQ_FUA)); struct md_rdev *blocked_rdev; struct blk_plug_cb *cb; @@ -2318,7 +2318,7 @@ read_more: raid_end_bio_io(r1_bio); } else { const unsigned long do_sync - = r1_bio->master_bio->bi_rw & REQ_SYNC; + = r1_bio->master_bio->bi_opf & REQ_SYNC; if (bio) { r1_bio->bios[r1_bio->read_disk] = mddev->ro ? IO_BLOCKED : NULL; diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index ed29fc899f06..0e4efcd10795 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -1054,8 +1054,8 @@ static void __make_request(struct mddev *mddev, struct bio *bio) int i; const int op = bio_op(bio); const int rw = bio_data_dir(bio); - const unsigned long do_sync = (bio->bi_rw & REQ_SYNC); - const unsigned long do_fua = (bio->bi_rw & REQ_FUA); + const unsigned long do_sync = (bio->bi_opf & REQ_SYNC); + const unsigned long do_fua = (bio->bi_opf & REQ_FUA); unsigned long flags; struct md_rdev *blocked_rdev; struct blk_plug_cb *cb; @@ -1440,7 +1440,7 @@ static void raid10_make_request(struct mddev *mddev, struct bio *bio) struct bio *split; - if (unlikely(bio->bi_rw & REQ_PREFLUSH)) { + if (unlikely(bio->bi_opf & REQ_PREFLUSH)) { md_flush_request(mddev, bio); return; } @@ -2533,7 +2533,7 @@ read_more: return; } - do_sync = (r10_bio->master_bio->bi_rw & REQ_SYNC); + do_sync = (r10_bio->master_bio->bi_opf & REQ_SYNC); slot = r10_bio->read_slot; printk_ratelimited( KERN_ERR diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index 5504ce2bac06..51f76ddbe265 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -536,7 +536,7 @@ int r5l_handle_flush_request(struct r5l_log *log, struct bio *bio) bio_endio(bio); return 0; } - bio->bi_rw &= ~REQ_PREFLUSH; + bio->bi_opf &= ~REQ_PREFLUSH; return -EAGAIN; } diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index d189e894b921..8912407a4dd0 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c @@ -806,7 +806,7 @@ static void stripe_add_to_batch_list(struct r5conf *conf, struct stripe_head *sh dd_idx = 0; while (dd_idx == sh->pd_idx || dd_idx == sh->qd_idx) dd_idx++; - if (head->dev[dd_idx].towrite->bi_rw != sh->dev[dd_idx].towrite->bi_rw || + if (head->dev[dd_idx].towrite->bi_opf != sh->dev[dd_idx].towrite->bi_opf || bio_op(head->dev[dd_idx].towrite) != bio_op(sh->dev[dd_idx].towrite)) goto unlock_out; @@ -1003,7 +1003,7 @@ again: pr_debug("%s: for %llu schedule op %d on disc %d\n", __func__, (unsigned long long)sh->sector, - bi->bi_rw, i); + bi->bi_opf, i); atomic_inc(&sh->count); if (sh != head_sh) atomic_inc(&head_sh->count); @@ -1014,7 +1014,7 @@ again: bi->bi_iter.bi_sector = (sh->sector + rdev->data_offset); if (test_bit(R5_ReadNoMerge, &head_sh->dev[i].flags)) - bi->bi_rw |= REQ_NOMERGE; + bi->bi_opf |= REQ_NOMERGE; if (test_bit(R5_SkipCopy, &sh->dev[i].flags)) WARN_ON(test_bit(R5_UPTODATE, &sh->dev[i].flags)); @@ -1055,7 +1055,7 @@ again: pr_debug("%s: for %llu schedule op %d on " "replacement disc %d\n", __func__, (unsigned long long)sh->sector, - rbi->bi_rw, i); + rbi->bi_opf, i); atomic_inc(&sh->count); if (sh != head_sh) atomic_inc(&head_sh->count); @@ -1088,7 +1088,7 @@ again: if (op_is_write(op)) set_bit(STRIPE_DEGRADED, &sh->state); pr_debug("skip op %d on disc %d for sector %llu\n", - bi->bi_rw, i, (unsigned long long)sh->sector); + bi->bi_opf, i, (unsigned long long)sh->sector); clear_bit(R5_LOCKED, &sh->dev[i].flags); set_bit(STRIPE_HANDLE, &sh->state); } @@ -1619,9 +1619,9 @@ again: while (wbi && wbi->bi_iter.bi_sector < dev->sector + STRIPE_SECTORS) { - if (wbi->bi_rw & REQ_FUA) + if (wbi->bi_opf & REQ_FUA) set_bit(R5_WantFUA, &dev->flags); - if (wbi->bi_rw & REQ_SYNC) + if (wbi->bi_opf & REQ_SYNC) set_bit(R5_SyncIO, &dev->flags); if (bio_op(wbi) == REQ_OP_DISCARD) set_bit(R5_Discard, &dev->flags); @@ -5154,7 +5154,7 @@ static void raid5_make_request(struct mddev *mddev, struct bio * bi) DEFINE_WAIT(w); bool do_prepare; - if (unlikely(bi->bi_rw & REQ_PREFLUSH)) { + if (unlikely(bi->bi_opf & REQ_PREFLUSH)) { int ret = r5l_handle_flush_request(conf->log, bi); if (ret == 0) @@ -5237,7 +5237,7 @@ static void raid5_make_request(struct mddev *mddev, struct bio * bi) (unsigned long long)logical_sector); sh = raid5_get_active_stripe(conf, new_sector, previous, - (bi->bi_rw & REQ_RAHEAD), 0); + (bi->bi_opf & REQ_RAHEAD), 0); if (sh) { if (unlikely(previous)) { /* expansion might have moved on while waiting for a @@ -5305,7 +5305,7 @@ static void raid5_make_request(struct mddev *mddev, struct bio * bi) set_bit(STRIPE_HANDLE, &sh->state); clear_bit(STRIPE_DELAYED, &sh->state); if ((!sh->batch_head || sh == sh->batch_head) && - (bi->bi_rw & REQ_SYNC) && + (bi->bi_opf & REQ_SYNC) && !test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) atomic_inc(&conf->preread_active_stripes); release_stripe_plug(mddev, sh); diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c index 20bae50c231d..571a6c7ee2fc 100644 --- a/drivers/nvdimm/pmem.c +++ b/drivers/nvdimm/pmem.c @@ -128,7 +128,7 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) struct pmem_device *pmem = q->queuedata; struct nd_region *nd_region = to_region(pmem); - if (bio->bi_rw & REQ_FLUSH) + if (bio->bi_opf & REQ_FLUSH) nvdimm_flush(nd_region); do_acct = nd_iostat_start(bio, &start); @@ -144,7 +144,7 @@ static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio) if (do_acct) nd_iostat_end(bio, start); - if (bio->bi_rw & REQ_FUA) + if (bio->bi_opf & REQ_FUA) nvdimm_flush(nd_region); bio_endio(bio); diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index 5d5cae05818d..66789471b49d 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -2945,7 +2945,7 @@ static void __btrfsic_submit_bio(struct bio *bio) printk(KERN_INFO "submit_bio(rw=%d,0x%x, bi_vcnt=%u," " bi_sector=%llu (bytenr %llu), bi_bdev=%p)\n", - bio_op(bio), bio->bi_rw, bio->bi_vcnt, + bio_op(bio), bio->bi_opf, bio->bi_vcnt, (unsigned long long)bio->bi_iter.bi_sector, dev_bytenr, bio->bi_bdev); @@ -2976,18 +2976,18 @@ static void __btrfsic_submit_bio(struct bio *bio) btrfsic_process_written_block(dev_state, dev_bytenr, mapped_datav, bio->bi_vcnt, bio, &bio_is_patched, - NULL, bio->bi_rw); + NULL, bio->bi_opf); while (i > 0) { i--; kunmap(bio->bi_io_vec[i].bv_page); } kfree(mapped_datav); - } else if (NULL != dev_state && (bio->bi_rw & REQ_PREFLUSH)) { + } else if (NULL != dev_state && (bio->bi_opf & REQ_PREFLUSH)) { if (dev_state->state->print_mask & BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH) printk(KERN_INFO "submit_bio(rw=%d,0x%x FLUSH, bdev=%p)\n", - bio_op(bio), bio->bi_rw, bio->bi_bdev); + bio_op(bio), bio->bi_opf, bio->bi_bdev); if (!dev_state->dummy_block_for_bio_bh_flush.is_iodone) { if ((dev_state->state->print_mask & (BTRFSIC_PRINT_MASK_SUBMIT_BIO_BH | @@ -3005,7 +3005,7 @@ static void __btrfsic_submit_bio(struct bio *bio) block->never_written = 0; block->iodone_w_error = 0; block->flush_gen = dev_state->last_flush_gen + 1; - block->submit_bio_bh_rw = bio->bi_rw; + block->submit_bio_bh_rw = bio->bi_opf; block->orig_bio_bh_private = bio->bi_private; block->orig_bio_bh_end_io.bio = bio->bi_end_io; block->next_in_same_bio = NULL; diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 87dad552e39a..59febfb8d04a 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -870,7 +870,7 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, atomic_inc(&fs_info->nr_async_submits); - if (bio->bi_rw & REQ_SYNC) + if (bio->bi_opf & REQ_SYNC) btrfs_set_work_high_priority(&async->work); btrfs_queue_work(fs_info->workers, &async->work); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b0f421f332ae..2f5975954ccf 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -8209,7 +8209,7 @@ static void btrfs_end_dio_bio(struct bio *bio) if (err) btrfs_warn(BTRFS_I(dip->inode)->root->fs_info, "direct IO failed ino %llu rw %d,%u sector %#Lx len %u err no %d", - btrfs_ino(dip->inode), bio_op(bio), bio->bi_rw, + btrfs_ino(dip->inode), bio_op(bio), bio->bi_opf, (unsigned long long)bio->bi_iter.bi_sector, bio->bi_iter.bi_size, err); @@ -8373,7 +8373,7 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip, if (!bio) return -ENOMEM; - bio_set_op_attrs(bio, bio_op(orig_bio), orig_bio->bi_rw); + bio_set_op_attrs(bio, bio_op(orig_bio), orig_bio->bi_opf); bio->bi_private = dip; bio->bi_end_io = btrfs_end_dio_bio; btrfs_io_bio(bio)->logical = file_offset; @@ -8411,7 +8411,7 @@ next_block: start_sector, GFP_NOFS); if (!bio) goto out_err; - bio_set_op_attrs(bio, bio_op(orig_bio), orig_bio->bi_rw); + bio_set_op_attrs(bio, bio_op(orig_bio), orig_bio->bi_opf); bio->bi_private = dip; bio->bi_end_io = btrfs_end_dio_bio; btrfs_io_bio(bio)->logical = file_offset; diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index bb0addce7558..51f125508771 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6012,7 +6012,7 @@ static void btrfs_end_bio(struct bio *bio) else btrfs_dev_stat_inc(dev, BTRFS_DEV_STAT_READ_ERRS); - if ((bio->bi_rw & WRITE_FLUSH) == WRITE_FLUSH) + if ((bio->bi_opf & WRITE_FLUSH) == WRITE_FLUSH) btrfs_dev_stat_inc(dev, BTRFS_DEV_STAT_FLUSH_ERRS); btrfs_dev_stat_print_on_error(dev); @@ -6089,7 +6089,7 @@ static noinline void btrfs_schedule_bio(struct btrfs_root *root, bio->bi_next = NULL; spin_lock(&device->io_lock); - if (bio->bi_rw & REQ_SYNC) + if (bio->bi_opf & REQ_SYNC) pending_bios = &device->pending_sync_bios; else pending_bios = &device->pending_bios; @@ -6127,7 +6127,7 @@ static void submit_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio, rcu_read_lock(); name = rcu_dereference(dev->name); pr_debug("btrfs_map_bio: rw %d 0x%x, sector=%llu, dev=%lu " - "(%s id %llu), size=%u\n", bio_op(bio), bio->bi_rw, + "(%s id %llu), size=%u\n", bio_op(bio), bio->bi_opf, (u64)bio->bi_iter.bi_sector, (u_long)dev->bdev->bd_dev, name->str, dev->devid, bio->bi_iter.bi_size); rcu_read_unlock(); diff --git a/include/linux/bio.h b/include/linux/bio.h index e09a8895fc31..59ffaa68b11b 100644 --- a/include/linux/bio.h +++ b/include/linux/bio.h @@ -95,7 +95,7 @@ static inline bool bio_is_rw(struct bio *bio) static inline bool bio_mergeable(struct bio *bio) { - if (bio->bi_rw & REQ_NOMERGE_FLAGS) + if (bio->bi_opf & REQ_NOMERGE_FLAGS) return false; return true; @@ -318,7 +318,7 @@ struct bio_integrity_payload { static inline struct bio_integrity_payload *bio_integrity(struct bio *bio) { - if (bio->bi_rw & REQ_INTEGRITY) + if (bio->bi_opf & REQ_INTEGRITY) return bio->bi_integrity; return NULL; diff --git a/include/linux/blk-cgroup.h b/include/linux/blk-cgroup.h index f77150a4a96a..10648e300c93 100644 --- a/include/linux/blk-cgroup.h +++ b/include/linux/blk-cgroup.h @@ -714,9 +714,9 @@ static inline bool blkcg_bio_issue_check(struct request_queue *q, if (!throtl) { blkg = blkg ?: q->root_blkg; - blkg_rwstat_add(&blkg->stat_bytes, bio_op(bio), bio->bi_rw, + blkg_rwstat_add(&blkg->stat_bytes, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); - blkg_rwstat_add(&blkg->stat_ios, bio_op(bio), bio->bi_rw, 1); + blkg_rwstat_add(&blkg->stat_ios, bio_op(bio), bio->bi_opf, 1); } rcu_read_unlock(); diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h index f254eb264924..436f43f87da9 100644 --- a/include/linux/blk_types.h +++ b/include/linux/blk_types.h @@ -27,8 +27,9 @@ struct bio { struct bio *bi_next; /* request queue link */ struct block_device *bi_bdev; int bi_error; - unsigned int bi_rw; /* bottom bits req flags, - * top bits REQ_OP + unsigned int bi_opf; /* bottom bits req flags, + * top bits REQ_OP. Use + * accessors. */ unsigned short bi_flags; /* status, command, etc */ unsigned short bi_ioprio; @@ -89,13 +90,13 @@ struct bio { }; #define BIO_OP_SHIFT (8 * sizeof(unsigned int) - REQ_OP_BITS) -#define bio_op(bio) ((bio)->bi_rw >> BIO_OP_SHIFT) +#define bio_op(bio) ((bio)->bi_opf >> BIO_OP_SHIFT) #define bio_set_op_attrs(bio, op, op_flags) do { \ WARN_ON(op >= (1 << REQ_OP_BITS)); \ - (bio)->bi_rw &= ((1 << BIO_OP_SHIFT) - 1); \ - (bio)->bi_rw |= ((unsigned int) (op) << BIO_OP_SHIFT); \ - (bio)->bi_rw |= op_flags; \ + (bio)->bi_opf &= ((1 << BIO_OP_SHIFT) - 1); \ + (bio)->bi_opf |= ((unsigned int) (op) << BIO_OP_SHIFT); \ + (bio)->bi_opf |= op_flags; \ } while (0) #define BIO_RESET_BYTES offsetof(struct bio, bi_max_vecs) @@ -138,7 +139,7 @@ struct bio { /* * Request flags. For use in the cmd_flags field of struct request, and in - * bi_rw of struct bio. Note that some flags are only valid in either one. + * bi_opf of struct bio. Note that some flags are only valid in either one. */ enum rq_flag_bits { /* common flags */ diff --git a/include/trace/events/bcache.h b/include/trace/events/bcache.h index 65673d8b81ac..d336b890e31f 100644 --- a/include/trace/events/bcache.h +++ b/include/trace/events/bcache.h @@ -27,7 +27,7 @@ DECLARE_EVENT_CLASS(bcache_request, __entry->sector = bio->bi_iter.bi_sector; __entry->orig_sector = bio->bi_iter.bi_sector - 16; __entry->nr_sector = bio->bi_iter.bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); ), @@ -102,7 +102,7 @@ DECLARE_EVENT_CLASS(bcache_bio, __entry->dev = bio->bi_bdev->bd_dev; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio->bi_iter.bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); ), @@ -138,7 +138,7 @@ TRACE_EVENT(bcache_read, __entry->dev = bio->bi_bdev->bd_dev; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio->bi_iter.bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); __entry->cache_hit = hit; __entry->bypass = bypass; @@ -170,7 +170,7 @@ TRACE_EVENT(bcache_write, __entry->inode = inode; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio->bi_iter.bi_size >> 9; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); __entry->writeback = writeback; __entry->bypass = bypass; diff --git a/include/trace/events/block.h b/include/trace/events/block.h index 5a2a7592068f..8f3a163b8166 100644 --- a/include/trace/events/block.h +++ b/include/trace/events/block.h @@ -274,7 +274,7 @@ TRACE_EVENT(block_bio_bounce, bio->bi_bdev->bd_dev : 0; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio_sectors(bio); - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); memcpy(__entry->comm, current->comm, TASK_COMM_LEN); ), @@ -313,7 +313,7 @@ TRACE_EVENT(block_bio_complete, __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio_sectors(bio); __entry->error = error; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); ), @@ -341,7 +341,7 @@ DECLARE_EVENT_CLASS(block_bio_merge, __entry->dev = bio->bi_bdev->bd_dev; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio_sectors(bio); - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); memcpy(__entry->comm, current->comm, TASK_COMM_LEN); ), @@ -409,7 +409,7 @@ TRACE_EVENT(block_bio_queue, __entry->dev = bio->bi_bdev->bd_dev; __entry->sector = bio->bi_iter.bi_sector; __entry->nr_sector = bio_sectors(bio); - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); memcpy(__entry->comm, current->comm, TASK_COMM_LEN); ), @@ -439,7 +439,7 @@ DECLARE_EVENT_CLASS(block_get_rq, __entry->sector = bio ? bio->bi_iter.bi_sector : 0; __entry->nr_sector = bio ? bio_sectors(bio) : 0; blk_fill_rwbs(__entry->rwbs, bio ? bio_op(bio) : 0, - bio ? bio->bi_rw : 0, __entry->nr_sector); + bio ? bio->bi_opf : 0, __entry->nr_sector); memcpy(__entry->comm, current->comm, TASK_COMM_LEN); ), @@ -573,7 +573,7 @@ TRACE_EVENT(block_split, __entry->dev = bio->bi_bdev->bd_dev; __entry->sector = bio->bi_iter.bi_sector; __entry->new_sector = new_sector; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); memcpy(__entry->comm, current->comm, TASK_COMM_LEN); ), @@ -617,7 +617,7 @@ TRACE_EVENT(block_bio_remap, __entry->nr_sector = bio_sectors(bio); __entry->old_dev = dev; __entry->old_sector = from; - blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_rw, + blk_fill_rwbs(__entry->rwbs, bio_op(bio), bio->bi_opf, bio->bi_iter.bi_size); ), diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index fb345cd11883..7598e6ca817a 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -776,7 +776,7 @@ static void blk_add_trace_bio(struct request_queue *q, struct bio *bio, return; __blk_add_trace(bt, bio->bi_iter.bi_sector, bio->bi_iter.bi_size, - bio_op(bio), bio->bi_rw, what, error, 0, NULL); + bio_op(bio), bio->bi_opf, what, error, 0, NULL); } static void blk_add_trace_bio_bounce(void *ignore, @@ -881,7 +881,7 @@ static void blk_add_trace_split(void *ignore, __be64 rpdu = cpu_to_be64(pdu); __blk_add_trace(bt, bio->bi_iter.bi_sector, - bio->bi_iter.bi_size, bio_op(bio), bio->bi_rw, + bio->bi_iter.bi_size, bio_op(bio), bio->bi_opf, BLK_TA_SPLIT, bio->bi_error, sizeof(rpdu), &rpdu); } @@ -915,7 +915,7 @@ static void blk_add_trace_bio_remap(void *ignore, r.sector_from = cpu_to_be64(from); __blk_add_trace(bt, bio->bi_iter.bi_sector, bio->bi_iter.bi_size, - bio_op(bio), bio->bi_rw, BLK_TA_REMAP, bio->bi_error, + bio_op(bio), bio->bi_opf, BLK_TA_REMAP, bio->bi_error, sizeof(r), &r); } -- cgit v1.2.3 From c4159a75b64c0e67caededf4d7372c1b58a5f42a Mon Sep 17 00:00:00 2001 From: Vladimir Davydov Date: Mon, 8 Aug 2016 23:03:12 +0300 Subject: mm: memcontrol: only mark charged pages with PageKmemcg To distinguish non-slab pages charged to kmemcg we mark them PageKmemcg, which sets page->_mapcount to -512. Currently, we set/clear PageKmemcg in __alloc_pages_nodemask()/free_pages_prepare() for any page allocated with __GFP_ACCOUNT, including those that aren't actually charged to any cgroup, i.e. allocated from the root cgroup context. To avoid overhead in case cgroups are not used, we only do that if memcg_kmem_enabled() is true. The latter is set iff there are kmem-enabled memory cgroups (online or offline). The root cgroup is not considered kmem-enabled. As a result, if a page is allocated with __GFP_ACCOUNT for the root cgroup when there are kmem-enabled memory cgroups and is freed after all kmem-enabled memory cgroups were removed, e.g. # no memory cgroups has been created yet, create one mkdir /sys/fs/cgroup/memory/test # run something allocating pages with __GFP_ACCOUNT, e.g. # a program using pipe dmesg | tail # remove the memory cgroup rmdir /sys/fs/cgroup/memory/test we'll get bad page state bug complaining about page->_mapcount != -1: BUG: Bad page state in process swapper/0 pfn:1fd945c page:ffffea007f651700 count:0 mapcount:-511 mapping: (null) index:0x0 flags: 0x1000000000000000() To avoid that, let's mark with PageKmemcg only those pages that are actually charged to and hence pin a non-root memory cgroup. Fixes: 4949148ad433 ("mm: charge/uncharge kmemcg from generic page allocator paths") Reported-and-tested-by: Eric Dumazet Signed-off-by: Vladimir Davydov Signed-off-by: Linus Torvalds --- fs/pipe.c | 4 +--- mm/memcontrol.c | 14 ++++++++++++-- mm/page_alloc.c | 14 +++++--------- 3 files changed, 18 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/pipe.c b/fs/pipe.c index 4b32928f5426..4ebe6b2e5217 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -144,10 +144,8 @@ static int anon_pipe_buf_steal(struct pipe_inode_info *pipe, struct page *page = buf->page; if (page_count(page) == 1) { - if (memcg_kmem_enabled()) { + if (memcg_kmem_enabled()) memcg_kmem_uncharge(page, 0); - __ClearPageKmemcg(page); - } __SetPageLocked(page); return 0; } diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 66beca1ad92f..e74d7080ec9e 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -2337,8 +2337,11 @@ int memcg_kmem_charge(struct page *page, gfp_t gfp, int order) return 0; memcg = get_mem_cgroup_from_mm(current->mm); - if (!mem_cgroup_is_root(memcg)) + if (!mem_cgroup_is_root(memcg)) { ret = memcg_kmem_charge_memcg(page, gfp, order, memcg); + if (!ret) + __SetPageKmemcg(page); + } css_put(&memcg->css); return ret; } @@ -2365,6 +2368,11 @@ void memcg_kmem_uncharge(struct page *page, int order) page_counter_uncharge(&memcg->memsw, nr_pages); page->mem_cgroup = NULL; + + /* slab pages do not have PageKmemcg flag set */ + if (PageKmemcg(page)) + __ClearPageKmemcg(page); + css_put_many(&memcg->css, nr_pages); } #endif /* !CONFIG_SLOB */ @@ -5537,8 +5545,10 @@ static void uncharge_list(struct list_head *page_list) else nr_file += nr_pages; pgpgout++; - } else + } else { nr_kmem += 1 << compound_order(page); + __ClearPageKmemcg(page); + } page->mem_cgroup = NULL; } while (next != page_list); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index fb975cec3518..ee744fa3b93d 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1008,10 +1008,8 @@ static __always_inline bool free_pages_prepare(struct page *page, } if (PageMappingFlags(page)) page->mapping = NULL; - if (memcg_kmem_enabled() && PageKmemcg(page)) { + if (memcg_kmem_enabled() && PageKmemcg(page)) memcg_kmem_uncharge(page, order); - __ClearPageKmemcg(page); - } if (check_free) bad += free_pages_check(page); if (bad) @@ -3756,12 +3754,10 @@ no_zone: } out: - if (memcg_kmem_enabled() && (gfp_mask & __GFP_ACCOUNT) && page) { - if (unlikely(memcg_kmem_charge(page, gfp_mask, order))) { - __free_pages(page, order); - page = NULL; - } else - __SetPageKmemcg(page); + if (memcg_kmem_enabled() && (gfp_mask & __GFP_ACCOUNT) && page && + unlikely(memcg_kmem_charge(page, gfp_mask, order) != 0)) { + __free_pages(page, order); + page = NULL; } if (kmemcheck_enabled && page) -- cgit v1.2.3