From bf2afee14e07de16d3cafc67edbfc2a3cc65e4bc Mon Sep 17 00:00:00 2001 From: Ronnie Sahlberg Date: Fri, 8 Sep 2017 10:37:35 +1000 Subject: cifs: check rsp for NULL before dereferencing in SMB2_open In SMB2_open there are several paths where the SendReceive2 call will return an error before it sets rsp_iov.iov_base thus leaving iov_base uninitialized. Thus we need to check rsp before we dereference it in the call to get_rfc1002_length(). A report of this issue was previously reported in http://www.spinics.net/lists/linux-cifs/msg12846.html RH-bugzilla : 1476151 Version 2 : * Lets properly initialize rsp_iov before we use it. Signed-off-by: Ronnie Sahlberg Reviewed-by: Pavel Shilovsky . Signed-off-by: Steve French Reported-by: Xiaoli Feng CC: Stable --- fs/cifs/smb2pdu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 5531e7ee1210..69a751b038ab 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -1634,7 +1634,7 @@ SMB2_open(const unsigned int xid, struct cifs_open_parms *oparms, __le16 *path, struct cifs_tcon *tcon = oparms->tcon; struct cifs_ses *ses = tcon->ses; struct kvec iov[4]; - struct kvec rsp_iov; + struct kvec rsp_iov = {NULL, 0}; int resp_buftype; int uni_path_len; __le16 *copy_path = NULL; @@ -1763,7 +1763,7 @@ SMB2_open(const unsigned int xid, struct cifs_open_parms *oparms, __le16 *path, if (rc != 0) { cifs_stats_fail_inc(tcon, SMB2_CREATE_HE); - if (err_buf) + if (err_buf && rsp) *err_buf = kmemdup(rsp, get_rfc1002_length(rsp) + 4, GFP_KERNEL); goto creat_exit; -- cgit v1.2.3 From fd7d56270b526ca3ed0c224362e3c64a0f86687a Mon Sep 17 00:00:00 2001 From: John Ogness Date: Thu, 14 Sep 2017 11:42:17 +0200 Subject: fs/proc: Report eip/esp in /prod/PID/stat for coredumping Commit 0a1eb2d474ed ("fs/proc: Stop reporting eip and esp in /proc/PID/stat") stopped reporting eip/esp because it is racy and dangerous for executing tasks. The comment adds: As far as I know, there are no use programs that make any material use of these fields, so just get rid of them. However, existing userspace core-dump-handler applications (for example, minicoredumper) are using these fields since they provide an excellent cross-platform interface to these valuable pointers. So that commit introduced a user space visible regression. Partially revert the change and make the readout possible for tasks with the proper permissions and only if the target task has the PF_DUMPCORE flag set. Fixes: 0a1eb2d474ed ("fs/proc: Stop reporting eip and esp in> /proc/PID/stat") Reported-by: Marco Felsch Signed-off-by: John Ogness Reviewed-by: Andy Lutomirski Cc: Tycho Andersen Cc: Kees Cook Cc: Peter Zijlstra Cc: Brian Gerst Cc: stable@vger.kernel.org Cc: Tetsuo Handa Cc: Borislav Petkov Cc: Al Viro Cc: Linux API Cc: Andrew Morton Cc: Linus Torvalds Link: http://lkml.kernel.org/r/87poatfwg6.fsf@linutronix.de Signed-off-by: Thomas Gleixner --- fs/proc/array.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index 88c355574aa0..525157ca25cb 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -62,6 +62,7 @@ #include #include #include +#include #include #include #include @@ -421,7 +422,15 @@ static int do_task_stat(struct seq_file *m, struct pid_namespace *ns, * esp and eip are intentionally zeroed out. There is no * non-racy way to read them without freezing the task. * Programs that need reliable values can use ptrace(2). + * + * The only exception is if the task is core dumping because + * a program is not able to use ptrace(2) in that case. It is + * safe because the task has stopped executing permanently. */ + if (permitted && (task->flags & PF_DUMPCORE)) { + eip = KSTK_EIP(task); + esp = KSTK_ESP(task); + } } get_task_comm(tcomm, task); -- cgit v1.2.3 From 9764c02fcbad40001fd3f63558d918e4d519bb75 Mon Sep 17 00:00:00 2001 From: Steve French Date: Sun, 17 Sep 2017 10:41:35 -0500 Subject: SMB3: Add support for multidialect negotiate (SMB2.1 and later) With the need to discourage use of less secure dialect, SMB1 (CIFS), we temporarily upgraded the dialect to SMB3 in 4.13, but since there are various servers which only support SMB2.1 (2.1 is more secure than CIFS/SMB1) but not optimal for a default dialect - add support for multidialect negotiation. cifs.ko will now request SMB2.1 or later (ie SMB2.1 or SMB3.0, SMB3.02) and the server will pick the latest most secure one it can support. In addition since we are sending multidialect negotiate, add support for secure negotiate to validate that a man in the middle didn't downgrade us. Signed-off-by: Steve French Reviewed-by: Pavel Shilovsky CC: Stable # 4.13+ --- fs/cifs/cifsglob.h | 6 ++++ fs/cifs/connect.c | 24 ++++++++++----- fs/cifs/smb2ops.c | 40 +++++++++++++++++++++++++ fs/cifs/smb2pdu.c | 85 +++++++++++++++++++++++++++++++++++++++++++++++------- fs/cifs/smb2pdu.h | 2 +- 5 files changed, 139 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 808486c29f0d..de5b2e1fcce5 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -188,6 +188,8 @@ enum smb_version { #ifdef CONFIG_CIFS_SMB311 Smb_311, #endif /* SMB311 */ + Smb_3any, + Smb_default, Smb_version_err }; @@ -1701,6 +1703,10 @@ extern struct smb_version_values smb20_values; #define SMB21_VERSION_STRING "2.1" extern struct smb_version_operations smb21_operations; extern struct smb_version_values smb21_values; +#define SMBDEFAULT_VERSION_STRING "default" +extern struct smb_version_values smbdefault_values; +#define SMB3ANY_VERSION_STRING "3" +extern struct smb_version_values smb3any_values; #define SMB30_VERSION_STRING "3.0" extern struct smb_version_operations smb30_operations; extern struct smb_version_values smb30_values; diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 5aa2d278ca84..8d38b22afb2b 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -301,6 +301,8 @@ static const match_table_t cifs_smb_version_tokens = { { Smb_311, SMB311_VERSION_STRING }, { Smb_311, ALT_SMB311_VERSION_STRING }, #endif /* SMB311 */ + { Smb_3any, SMB3ANY_VERSION_STRING }, + { Smb_default, SMBDEFAULT_VERSION_STRING }, { Smb_version_err, NULL } }; @@ -1148,6 +1150,14 @@ cifs_parse_smb_version(char *value, struct smb_vol *vol) vol->vals = &smb311_values; break; #endif /* SMB311 */ + case Smb_3any: + vol->ops = &smb30_operations; /* currently identical with 3.0 */ + vol->vals = &smb3any_values; + break; + case Smb_default: + vol->ops = &smb30_operations; /* currently identical with 3.0 */ + vol->vals = &smbdefault_values; + break; default: cifs_dbg(VFS, "Unknown vers= option specified: %s\n", value); return 1; @@ -1274,9 +1284,9 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, vol->actimeo = CIFS_DEF_ACTIMEO; - /* FIXME: add autonegotiation for SMB3 or later rather than just SMB3 */ - vol->ops = &smb30_operations; /* both secure and accepted widely */ - vol->vals = &smb30_values; + /* offer SMB2.1 and later (SMB3 etc). Secure and widely accepted */ + vol->ops = &smb30_operations; + vol->vals = &smbdefault_values; vol->echo_interval = SMB_ECHO_INTERVAL_DEFAULT; @@ -1988,11 +1998,10 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, if (got_version == false) pr_warn("No dialect specified on mount. Default has changed to " - "a more secure dialect, SMB3 (vers=3.0), from CIFS " + "a more secure dialect, SMB2.1 or later (e.g. SMB3), from CIFS " "(SMB1). To use the less secure SMB1 dialect to access " - "old servers which do not support SMB3 specify vers=1.0" - " on mount. For somewhat newer servers such as Windows " - "7 try vers=2.1.\n"); + "old servers which do not support SMB3 (or SMB2.1) specify vers=1.0" + " on mount.\n"); kfree(mountdata_copy); return 0; @@ -2133,6 +2142,7 @@ static int match_server(struct TCP_Server_Info *server, struct smb_vol *vol) if (vol->nosharesock) return 0; + /* BB update this for smb3any and default case */ if ((server->vals != vol->vals) || (server->ops != vol->ops)) return 0; diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c index fb2934b9b97c..405a9bf13aac 100644 --- a/fs/cifs/smb2ops.c +++ b/fs/cifs/smb2ops.c @@ -3110,6 +3110,46 @@ struct smb_version_values smb21_values = { .create_lease_size = sizeof(struct create_lease), }; +struct smb_version_values smb3any_values = { + .version_string = SMB3ANY_VERSION_STRING, + .protocol_id = SMB302_PROT_ID, /* doesn't matter, send protocol array */ + .req_capabilities = SMB2_GLOBAL_CAP_DFS | SMB2_GLOBAL_CAP_LEASING | SMB2_GLOBAL_CAP_LARGE_MTU | SMB2_GLOBAL_CAP_PERSISTENT_HANDLES | SMB2_GLOBAL_CAP_ENCRYPTION, + .large_lock_type = 0, + .exclusive_lock_type = SMB2_LOCKFLAG_EXCLUSIVE_LOCK, + .shared_lock_type = SMB2_LOCKFLAG_SHARED_LOCK, + .unlock_lock_type = SMB2_LOCKFLAG_UNLOCK, + .header_size = sizeof(struct smb2_hdr), + .max_header_size = MAX_SMB2_HDR_SIZE, + .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .lock_cmd = SMB2_LOCK, + .cap_unix = 0, + .cap_nt_find = SMB2_NT_FIND, + .cap_large_files = SMB2_LARGE_FILES, + .signing_enabled = SMB2_NEGOTIATE_SIGNING_ENABLED | SMB2_NEGOTIATE_SIGNING_REQUIRED, + .signing_required = SMB2_NEGOTIATE_SIGNING_REQUIRED, + .create_lease_size = sizeof(struct create_lease_v2), +}; + +struct smb_version_values smbdefault_values = { + .version_string = SMBDEFAULT_VERSION_STRING, + .protocol_id = SMB302_PROT_ID, /* doesn't matter, send protocol array */ + .req_capabilities = SMB2_GLOBAL_CAP_DFS | SMB2_GLOBAL_CAP_LEASING | SMB2_GLOBAL_CAP_LARGE_MTU | SMB2_GLOBAL_CAP_PERSISTENT_HANDLES | SMB2_GLOBAL_CAP_ENCRYPTION, + .large_lock_type = 0, + .exclusive_lock_type = SMB2_LOCKFLAG_EXCLUSIVE_LOCK, + .shared_lock_type = SMB2_LOCKFLAG_SHARED_LOCK, + .unlock_lock_type = SMB2_LOCKFLAG_UNLOCK, + .header_size = sizeof(struct smb2_hdr), + .max_header_size = MAX_SMB2_HDR_SIZE, + .read_rsp_size = sizeof(struct smb2_read_rsp) - 1, + .lock_cmd = SMB2_LOCK, + .cap_unix = 0, + .cap_nt_find = SMB2_NT_FIND, + .cap_large_files = SMB2_LARGE_FILES, + .signing_enabled = SMB2_NEGOTIATE_SIGNING_ENABLED | SMB2_NEGOTIATE_SIGNING_REQUIRED, + .signing_required = SMB2_NEGOTIATE_SIGNING_REQUIRED, + .create_lease_size = sizeof(struct create_lease_v2), +}; + struct smb_version_values smb30_values = { .version_string = SMB30_VERSION_STRING, .protocol_id = SMB30_PROT_ID, diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 69a751b038ab..5c16591a128e 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -491,10 +491,25 @@ SMB2_negotiate(const unsigned int xid, struct cifs_ses *ses) req->hdr.sync_hdr.SessionId = 0; - req->Dialects[0] = cpu_to_le16(ses->server->vals->protocol_id); - - req->DialectCount = cpu_to_le16(1); /* One vers= at a time for now */ - inc_rfc1001_len(req, 2); + if (strcmp(ses->server->vals->version_string, + SMB3ANY_VERSION_STRING) == 0) { + req->Dialects[0] = cpu_to_le16(SMB30_PROT_ID); + req->Dialects[1] = cpu_to_le16(SMB302_PROT_ID); + req->DialectCount = cpu_to_le16(2); + inc_rfc1001_len(req, 4); + } else if (strcmp(ses->server->vals->version_string, + SMBDEFAULT_VERSION_STRING) == 0) { + req->Dialects[0] = cpu_to_le16(SMB21_PROT_ID); + req->Dialects[1] = cpu_to_le16(SMB30_PROT_ID); + req->Dialects[2] = cpu_to_le16(SMB302_PROT_ID); + req->DialectCount = cpu_to_le16(3); + inc_rfc1001_len(req, 6); + } else { + /* otherwise send specific dialect */ + req->Dialects[0] = cpu_to_le16(ses->server->vals->protocol_id); + req->DialectCount = cpu_to_le16(1); + inc_rfc1001_len(req, 2); + } /* only one of SMB2 signing flags may be set in SMB2 request */ if (ses->sign) @@ -528,16 +543,42 @@ SMB2_negotiate(const unsigned int xid, struct cifs_ses *ses) */ if (rc == -EOPNOTSUPP) { cifs_dbg(VFS, "Dialect not supported by server. Consider " - "specifying vers=1.0 or vers=2.1 on mount for accessing" + "specifying vers=1.0 or vers=2.0 on mount for accessing" " older servers\n"); goto neg_exit; } else if (rc != 0) goto neg_exit; + if (strcmp(ses->server->vals->version_string, + SMB3ANY_VERSION_STRING) == 0) { + if (rsp->DialectRevision == cpu_to_le16(SMB20_PROT_ID)) { + cifs_dbg(VFS, + "SMB2 dialect returned but not requested\n"); + return -EIO; + } else if (rsp->DialectRevision == cpu_to_le16(SMB21_PROT_ID)) { + cifs_dbg(VFS, + "SMB2.1 dialect returned but not requested\n"); + return -EIO; + } + } else if (strcmp(ses->server->vals->version_string, + SMBDEFAULT_VERSION_STRING) == 0) { + if (rsp->DialectRevision == cpu_to_le16(SMB20_PROT_ID)) { + cifs_dbg(VFS, + "SMB2 dialect returned but not requested\n"); + return -EIO; + } else if (rsp->DialectRevision == cpu_to_le16(SMB21_PROT_ID)) { + /* ops set to 3.0 by default for default so update */ + ses->server->ops = &smb21_operations; + } + } else if (rsp->DialectRevision != ses->server->vals->protocol_id) { + /* if requested single dialect ensure returned dialect matched */ + cifs_dbg(VFS, "Illegal 0x%x dialect returned: not requested\n", + cpu_to_le16(rsp->DialectRevision)); + return -EIO; + } + cifs_dbg(FYI, "mode 0x%x\n", rsp->SecurityMode); - /* BB we may eventually want to match the negotiated vs. requested - dialect, even though we are only requesting one at a time */ if (rsp->DialectRevision == cpu_to_le16(SMB20_PROT_ID)) cifs_dbg(FYI, "negotiated smb2.0 dialect\n"); else if (rsp->DialectRevision == cpu_to_le16(SMB21_PROT_ID)) @@ -558,6 +599,8 @@ SMB2_negotiate(const unsigned int xid, struct cifs_ses *ses) } server->dialect = le16_to_cpu(rsp->DialectRevision); + /* BB: add check that dialect was valid given dialect(s) we asked for */ + /* SMB2 only has an extended negflavor */ server->negflavor = CIFS_NEGFLAVOR_EXTENDED; /* set it to the maximum buffer size value we can send with 1 credit */ @@ -606,6 +649,7 @@ int smb3_validate_negotiate(const unsigned int xid, struct cifs_tcon *tcon) struct validate_negotiate_info_req vneg_inbuf; struct validate_negotiate_info_rsp *pneg_rsp; u32 rsplen; + u32 inbuflen; /* max of 4 dialects */ cifs_dbg(FYI, "validate negotiate\n"); @@ -634,9 +678,30 @@ int smb3_validate_negotiate(const unsigned int xid, struct cifs_tcon *tcon) else vneg_inbuf.SecurityMode = 0; - vneg_inbuf.DialectCount = cpu_to_le16(1); - vneg_inbuf.Dialects[0] = - cpu_to_le16(tcon->ses->server->vals->protocol_id); + + if (strcmp(tcon->ses->server->vals->version_string, + SMB3ANY_VERSION_STRING) == 0) { + vneg_inbuf.Dialects[0] = cpu_to_le16(SMB30_PROT_ID); + vneg_inbuf.Dialects[1] = cpu_to_le16(SMB302_PROT_ID); + vneg_inbuf.DialectCount = cpu_to_le16(2); + /* structure is big enough for 3 dialects, sending only 2 */ + inbuflen = sizeof(struct validate_negotiate_info_req) - 2; + } else if (strcmp(tcon->ses->server->vals->version_string, + SMBDEFAULT_VERSION_STRING) == 0) { + vneg_inbuf.Dialects[0] = cpu_to_le16(SMB21_PROT_ID); + vneg_inbuf.Dialects[1] = cpu_to_le16(SMB30_PROT_ID); + vneg_inbuf.Dialects[2] = cpu_to_le16(SMB302_PROT_ID); + vneg_inbuf.DialectCount = cpu_to_le16(3); + /* structure is big enough for 3 dialects */ + inbuflen = sizeof(struct validate_negotiate_info_req); + } else { + /* otherwise specific dialect was requested */ + vneg_inbuf.Dialects[0] = + cpu_to_le16(tcon->ses->server->vals->protocol_id); + vneg_inbuf.DialectCount = cpu_to_le16(1); + /* structure is big enough for 3 dialects, sending only 1 */ + inbuflen = sizeof(struct validate_negotiate_info_req) - 4; + } rc = SMB2_ioctl(xid, tcon, NO_FILE_ID, NO_FILE_ID, FSCTL_VALIDATE_NEGOTIATE_INFO, true /* is_fsctl */, diff --git a/fs/cifs/smb2pdu.h b/fs/cifs/smb2pdu.h index 393ed5f4e1b6..6c9653a130c8 100644 --- a/fs/cifs/smb2pdu.h +++ b/fs/cifs/smb2pdu.h @@ -716,7 +716,7 @@ struct validate_negotiate_info_req { __u8 Guid[SMB2_CLIENT_GUID_SIZE]; __le16 SecurityMode; __le16 DialectCount; - __le16 Dialects[1]; /* dialect (someday maybe list) client asked for */ + __le16 Dialects[3]; /* BB expand this if autonegotiate > 3 dialects */ } __packed; struct validate_negotiate_info_rsp { -- cgit v1.2.3 From 1368f155a972eab037f81b5dea82afd544227552 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 5 Sep 2017 11:24:15 +0200 Subject: cifs: hide unused functions The newly added SMB2+ attribute support causes unused function warnings when CONFIG_CIFS_XATTR is disabled: fs/cifs/smb2ops.c:563:1: error: 'smb2_set_ea' defined but not used [-Werror=unused-function] smb2_set_ea(const unsigned int xid, struct cifs_tcon *tcon, fs/cifs/smb2ops.c:513:1: error: 'smb2_query_eas' defined but not used [-Werror=unused-function] smb2_query_eas(const unsigned int xid, struct cifs_tcon *tcon, This adds another #ifdef around the affected functions. Fixes: 5517554e4313 ("cifs: Add support for writing attributes on SMB2+") Fixes: 95907fea4fd8 ("cifs: Add support for reading attributes on SMB2+") Signed-off-by: Arnd Bergmann Acked-by: Geert Uytterhoeven Signed-off-by: Steve French --- fs/cifs/smb2ops.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c index 405a9bf13aac..0dafdbae1f8c 100644 --- a/fs/cifs/smb2ops.c +++ b/fs/cifs/smb2ops.c @@ -426,6 +426,7 @@ smb2_query_file_info(const unsigned int xid, struct cifs_tcon *tcon, return rc; } +#ifdef CONFIG_CIFS_XATTR static ssize_t move_smb2_ea_to_cifs(char *dst, size_t dst_size, struct smb2_file_full_ea_info *src, size_t src_size, @@ -613,6 +614,7 @@ smb2_set_ea(const unsigned int xid, struct cifs_tcon *tcon, return rc; } +#endif static bool smb2_can_echo(struct TCP_Server_Info *server) -- cgit v1.2.3 From 94a9daeaece415ce37ccb413b49a580e68e3477b Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 12 Sep 2017 18:13:11 -0500 Subject: Update version of cifs module Signed-off-by: Steve French Reviewed-by: Pavel Shilovsky --- fs/cifs/cifsfs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h index 30bf89b1fd9a..5a10e566f0e6 100644 --- a/fs/cifs/cifsfs.h +++ b/fs/cifs/cifsfs.h @@ -149,5 +149,5 @@ extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg); extern const struct export_operations cifs_export_ops; #endif /* CONFIG_CIFS_NFSD_EXPORT */ -#define CIFS_VERSION "2.09" +#define CIFS_VERSION "2.10" #endif /* _CIFSFS_H */ -- cgit v1.2.3 From 0a51fb7174f2b7866b4d7a4a5c23b685b674beb6 Mon Sep 17 00:00:00 2001 From: Konstantin Khlebnikov Date: Sat, 9 Sep 2017 12:19:22 +0300 Subject: quota: add missing lock into __dquot_transfer() Lock dq_dqb_lock around dquot_decr_inodes() Signed-off-by: Konstantin Khlebnikov Fixes: 7b9ca4c61bc2 ("quota: Reduce contention on dq_data_lock") Signed-off-by: Jan Kara --- fs/quota/dquot.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c index 8381db9db6d9..50b0556a124f 100644 --- a/fs/quota/dquot.c +++ b/fs/quota/dquot.c @@ -1980,7 +1980,9 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to) ret = dquot_add_space(transfer_to[cnt], cur_space, rsv_space, 0, &warn_to[cnt]); if (ret) { + spin_lock(&transfer_to[cnt]->dq_dqb_lock); dquot_decr_inodes(transfer_to[cnt], inode_usage); + spin_unlock(&transfer_to[cnt]->dq_dqb_lock); goto over_quota; } } -- cgit v1.2.3 From 0ab0b271bf75073cb254b5ea0593aceae5a42bd3 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Sat, 16 Sep 2017 22:32:12 +0200 Subject: isofs: fix build regression The new isofs_show_options() function fails to build when CONFIG_NLS is disabled: fs/isofs/inode.c: In function 'isofs_show_options': fs/isofs/inode.c:518:44: error: 'CONFIG_NLS_DEFAULT' undeclared (first use in this function) fs/isofs/inode.c:518:44: note: each undeclared identifier is reported only once for each function it appears in This adds a check for CONFIG_JOLIET (which selects NLS), matching the other uses of the iocharset handling in this file. Fixes: 6fecb86a44f5 ("isofs: Implement show_options") Signed-off-by: Arnd Bergmann Signed-off-by: Jan Kara --- fs/isofs/inode.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index db692f554158..447a24d77b89 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -514,9 +514,11 @@ static int isofs_show_options(struct seq_file *m, struct dentry *root) if (sbi->s_fmode != ISOFS_INVALID_MODE) seq_printf(m, ",fmode=%o", sbi->s_fmode); +#ifdef CONFIG_JOLIET if (sbi->s_nls_iocharset && strcmp(sbi->s_nls_iocharset->charset, CONFIG_NLS_DEFAULT) != 0) seq_printf(m, ",iocharset=%s", sbi->s_nls_iocharset->charset); +#endif return 0; } -- cgit v1.2.3 From 54640d238760a1a54dfebe039b49682522100186 Mon Sep 17 00:00:00 2001 From: "Eric W. Biederman" Date: Mon, 18 Sep 2017 22:51:14 -0500 Subject: fcntl: Don't set si_code to SI_SIGIO when sig == SIGPOLL When fixing things to avoid ambiguous cases I had a thinko and included SIGPOLL/SIGIO in with all of the other signals that have signal specific si_codes. Which is completely wrong. Fix that. Reported-by: Vince Weaver Signed-off-by: "Eric W. Biederman" --- fs/fcntl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fcntl.c b/fs/fcntl.c index 0491da3b28c3..448a1119f0be 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -749,7 +749,7 @@ static void send_sigio_to_task(struct task_struct *p, * specific si_codes. In that case use SI_SIGIO instead * to remove the ambiguity. */ - if (sig_specific_sicodes(signum)) + if ((signum != SIGPOLL) && sig_specific_sicodes(signum)) si.si_code = SI_SIGIO; /* Make sure we are called with one of the POLL_* -- cgit v1.2.3 From 717e6f2893eb35ce6728c3cacdc297b78d371b31 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Mon, 11 Sep 2017 12:10:08 +0800 Subject: ceph: avoid panic in create_session_open_msg() if utsname() returns NULL utsname() can return NULL while process is exiting. Kernel releases file locks during process exits. We send request to mds when releasing file lock. So it's possible that we open mds session while process is exiting. utsname() is called in create_session_open_msg(). Link: http://tracker.ceph.com/issues/21275 Signed-off-by: "Yan, Zheng" Reviewed-by: Jeff Layton [idryomov@gmail.com: drop utsname.h include from mds_client.c] Signed-off-by: Ilya Dryomov --- fs/ceph/mds_client.c | 7 ++++--- fs/ceph/mds_client.h | 3 +++ 2 files changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 9dd6b836ac9e..84edfc60d87a 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -7,7 +7,6 @@ #include #include #include -#include #include #include "super.h" @@ -884,8 +883,8 @@ static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u6 void *p; const char* metadata[][2] = { - {"hostname", utsname()->nodename}, - {"kernel_version", utsname()->release}, + {"hostname", mdsc->nodename}, + {"kernel_version", init_utsname()->release}, {"entity_id", opt->name ? : ""}, {"root", fsopt->server_path ? : "/"}, {NULL, NULL} @@ -3539,6 +3538,8 @@ int ceph_mdsc_init(struct ceph_fs_client *fsc) init_rwsem(&mdsc->pool_perm_rwsem); mdsc->pool_perm_tree = RB_ROOT; + strncpy(mdsc->nodename, utsname()->nodename, + sizeof(mdsc->nodename) - 1); return 0; } diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h index db57ae98ed34..636d6b2ec49c 100644 --- a/fs/ceph/mds_client.h +++ b/fs/ceph/mds_client.h @@ -8,6 +8,7 @@ #include #include #include +#include #include #include @@ -368,6 +369,8 @@ struct ceph_mds_client { struct rw_semaphore pool_perm_rwsem; struct rb_root pool_perm_tree; + + char nodename[__NEW_UTS_LEN + 1]; }; extern const char *ceph_mds_op_name(int op); -- cgit v1.2.3 From 23586b66d84ba3184b8820277f3fc42761640f87 Mon Sep 17 00:00:00 2001 From: Steve French Date: Mon, 18 Sep 2017 18:18:45 -0500 Subject: Fix SMB3.1.1 guest authentication to Samba Samba rejects SMB3.1.1 dialect (vers=3.1.1) negotiate requests from the kernel client due to the two byte pad at the end of the negotiate contexts. CC: Stable Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg --- fs/cifs/smb2pdu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 5c16591a128e..b0eaebe627e9 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -439,7 +439,7 @@ assemble_neg_contexts(struct smb2_negotiate_req *req) build_encrypt_ctxt((struct smb2_encryption_neg_context *)pneg_ctxt); req->NegotiateContextOffset = cpu_to_le32(OFFSET_OF_NEG_CONTEXT); req->NegotiateContextCount = cpu_to_le16(2); - inc_rfc1001_len(req, 4 + sizeof(struct smb2_preauth_neg_context) + 2 + inc_rfc1001_len(req, 4 + sizeof(struct smb2_preauth_neg_context) + sizeof(struct smb2_encryption_neg_context)); /* calculate hash */ } #else -- cgit v1.2.3 From 590d08d3da45e9fed423b08ab38d71886c07abc8 Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 19 Sep 2017 11:43:47 -0500 Subject: SMB3: Fix endian warning Multi-dialect negotiate patch had a minor endian error. Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg CC: Stable # 4.13+ --- fs/cifs/smb2pdu.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index b0eaebe627e9..b4c58a1db1ae 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -570,10 +570,11 @@ SMB2_negotiate(const unsigned int xid, struct cifs_ses *ses) /* ops set to 3.0 by default for default so update */ ses->server->ops = &smb21_operations; } - } else if (rsp->DialectRevision != ses->server->vals->protocol_id) { + } else if (le16_to_cpu(rsp->DialectRevision) != + ses->server->vals->protocol_id) { /* if requested single dialect ensure returned dialect matched */ cifs_dbg(VFS, "Illegal 0x%x dialect returned: not requested\n", - cpu_to_le16(rsp->DialectRevision)); + le16_to_cpu(rsp->DialectRevision)); return -EIO; } -- cgit v1.2.3 From c721c38957fb19982416f6be71aae7b30630d83b Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 19 Sep 2017 18:40:03 -0500 Subject: SMB3: Warn user if trying to sign connection that authenticated as guest It can be confusing if user ends up authenticated as guest but they requested signing (server will return error validating signed packets) so add log message for this. Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg CC: Stable --- fs/cifs/smb2pdu.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index b4c58a1db1ae..d499ce265c3b 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -1176,6 +1176,8 @@ SMB2_sess_setup(const unsigned int xid, struct cifs_ses *ses, while (sess_data->func) sess_data->func(sess_data); + if ((ses->session_flags & SMB2_SESSION_FLAG_IS_GUEST) && (ses->sign)) + cifs_dbg(VFS, "signing requested but authenticated as guest\n"); rc = sess_data->result; out: kfree(sess_data); -- cgit v1.2.3 From 6e82e929d98552ed5156919e0f532eeba2da704b Mon Sep 17 00:00:00 2001 From: Ronnie Sahlberg Date: Wed, 20 Sep 2017 13:09:23 +1000 Subject: cifs: show 'soft' in the mount options for hard mounts Signed-off-by: Ronnie Sahlberg Signed-off-by: Steve French --- fs/cifs/cifsfs.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 180b3356ff86..a864e6575309 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -461,6 +461,8 @@ cifs_show_options(struct seq_file *s, struct dentry *root) seq_puts(s, ",nocase"); if (tcon->retry) seq_puts(s, ",hard"); + else + seq_puts(s, ",soft"); if (tcon->use_persistent) seq_puts(s, ",persistenthandles"); else if (tcon->use_resilient) -- cgit v1.2.3 From 1fa089ec6d68849bfe60abd141375dc9b21b0ee8 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 20 Sep 2017 16:46:49 -0500 Subject: [SMB3] Update session and share information displayed for debugging SMB2/SMB3 We were not displaying some key fields (session status and capabilities and whether guest authenticated) for SMB2/SMB3 session in /proc/fs/cifs/DebugData. This is needed for real world triage of problems with the (now much more common) SMB3 mounts. Signed-off-by: Steve French --- fs/cifs/cifs_debug.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index 9727e1dcacd5..cbb9534b89b4 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -160,8 +160,13 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) if ((ses->serverDomain == NULL) || (ses->serverOS == NULL) || (ses->serverNOS == NULL)) { - seq_printf(m, "\n%d) entry for %s not fully " - "displayed\n\t", i, ses->serverName); + seq_printf(m, "\n%d) Name: %s Uses: %d Capability: 0x%x\tSession Status: %d\t", + i, ses->serverName, ses->ses_count, + ses->capabilities, ses->status); + if (ses->session_flags & SMB2_SESSION_FLAG_IS_GUEST) + seq_printf(m, "Guest\t"); + else if (ses->session_flags & SMB2_SESSION_FLAG_IS_NULL) + seq_printf(m, "Anonymous\t"); } else { seq_printf(m, "\n%d) Name: %s Domain: %s Uses: %d OS:" -- cgit v1.2.3 From 4d61eda812041ef9c820d7f147884133fd3307bc Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Tue, 19 Sep 2017 16:27:39 +0100 Subject: CIFS: make arrays static const, reduces object code size Don't populate the read-only arrays types[] on the stack, instead make them both static const. Makes the object code smaller by over 200 bytes: Before: text data bss dec hex filename 111503 37696 448 149647 2488f fs/cifs/file.o After: text data bss dec hex filename 111140 37856 448 149444 247c4 fs/cifs/file.o Signed-off-by: Colin Ian King Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg --- fs/cifs/file.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/cifs/file.c b/fs/cifs/file.c index 0786f19d288f..822311919163 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -1102,8 +1102,10 @@ cifs_push_mandatory_locks(struct cifsFileInfo *cfile) struct cifs_tcon *tcon; unsigned int num, max_num, max_buf; LOCKING_ANDX_RANGE *buf, *cur; - int types[] = {LOCKING_ANDX_LARGE_FILES, - LOCKING_ANDX_SHARED_LOCK | LOCKING_ANDX_LARGE_FILES}; + static const int types[] = { + LOCKING_ANDX_LARGE_FILES, + LOCKING_ANDX_SHARED_LOCK | LOCKING_ANDX_LARGE_FILES + }; int i; xid = get_xid(); @@ -1434,8 +1436,10 @@ cifs_unlock_range(struct cifsFileInfo *cfile, struct file_lock *flock, unsigned int xid) { int rc = 0, stored_rc; - int types[] = {LOCKING_ANDX_LARGE_FILES, - LOCKING_ANDX_SHARED_LOCK | LOCKING_ANDX_LARGE_FILES}; + static const int types[] = { + LOCKING_ANDX_LARGE_FILES, + LOCKING_ANDX_SHARED_LOCK | LOCKING_ANDX_LARGE_FILES + }; unsigned int i; unsigned int max_num, num, max_buf; LOCKING_ANDX_RANGE *buf, *cur; -- cgit v1.2.3 From 94183331e815617246b1baa97e0916f358c794bb Mon Sep 17 00:00:00 2001 From: Shu Wang Date: Thu, 7 Sep 2017 16:03:27 +0800 Subject: cifs: release cifs root_cred after exit_cifs memory leak was found by kmemleak. exit_cifs_spnego should be called before cifs module removed, or cifs root_cred will not be released. kmemleak report: unreferenced object 0xffff880070a3ce40 (size 192): backtrace: kmemleak_alloc+0x4a/0xa0 kmem_cache_alloc+0xc7/0x1d0 prepare_kernel_cred+0x20/0x120 init_cifs_spnego+0x2d/0x170 [cifs] 0xffffffffc07801f3 do_one_initcall+0x51/0x1b0 do_init_module+0x60/0x1fd load_module+0x161e/0x1b60 SYSC_finit_module+0xa9/0x100 SyS_finit_module+0xe/0x10 Signed-off-by: Shu Wang Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg CC: Stable --- fs/cifs/cifsfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index a864e6575309..8c8b75d33f31 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -1449,7 +1449,7 @@ exit_cifs(void) exit_cifs_idmap(); #endif #ifdef CONFIG_CIFS_UPCALL - unregister_key_type(&cifs_spnego_key_type); + exit_cifs_spnego(); #endif cifs_destroy_request_bufs(); cifs_destroy_mids(); -- cgit v1.2.3 From f5c4ba816315d3b813af16f5571f86c8d4e897bd Mon Sep 17 00:00:00 2001 From: Shu Wang Date: Fri, 8 Sep 2017 18:48:33 +0800 Subject: cifs: release auth_key.response for reconnect. There is a race that cause cifs reconnect in cifs_mount, - cifs_mount - cifs_get_tcp_session - [ start thread cifs_demultiplex_thread - cifs_read_from_socket: -ECONNABORTED - DELAY_WORK smb2_reconnect_server ] - cifs_setup_session - [ smb2_reconnect_server ] auth_key.response was allocated in cifs_setup_session, and will release when the session destoried. So when session re- connect, auth_key.response should be check and released. Tested with my system: CIFS VFS: Free previous auth_key.response = ffff8800320bbf80 A simple auth_key.response allocation call trace: - cifs_setup_session - SMB2_sess_setup - SMB2_sess_auth_rawntlmssp_authenticate - build_ntlmssp_auth_blob - setup_ntlmv2_rsp Signed-off-by: Shu Wang Signed-off-by: Steve French CC: Stable Reviewed-by: Ronnie Sahlberg --- fs/cifs/connect.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 8d38b22afb2b..0bfc2280436d 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -4154,6 +4154,14 @@ cifs_setup_session(const unsigned int xid, struct cifs_ses *ses, cifs_dbg(FYI, "Security Mode: 0x%x Capabilities: 0x%x TimeAdjust: %d\n", server->sec_mode, server->capabilities, server->timeAdj); + if (ses->auth_key.response) { + cifs_dbg(VFS, "Free previous auth_key.response = %p\n", + ses->auth_key.response); + kfree(ses->auth_key.response); + ses->auth_key.response = NULL; + ses->auth_key.len = 0; + } + if (server->ops->sess_setup) rc = server->ops->sess_setup(xid, ses, nls_info); -- cgit v1.2.3 From 0603c96f3af50e2f9299fa410c224ab1d465e0f9 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 20 Sep 2017 19:57:18 -0500 Subject: SMB: Validate negotiate (to protect against downgrade) even if signing off As long as signing is supported (ie not a guest user connection) and connection is SMB3 or SMB3.02, then validate negotiate (protect against man in the middle downgrade attacks). We had been doing this only when signing was required, not when signing was just enabled, but this more closely matches recommended SMB3 behavior and is better security. Suggested by Metze. Signed-off-by: Steve French Reviewed-by: Jeremy Allison Acked-by: Stefan Metzmacher Reviewed-by: Ronnie Sahlberg CC: Stable --- fs/cifs/smb2pdu.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index d499ce265c3b..6f0e6343c15e 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -656,15 +656,22 @@ int smb3_validate_negotiate(const unsigned int xid, struct cifs_tcon *tcon) /* * validation ioctl must be signed, so no point sending this if we - * can not sign it. We could eventually change this to selectively + * can not sign it (ie are not known user). Even if signing is not + * required (enabled but not negotiated), in those cases we selectively * sign just this, the first and only signed request on a connection. - * This is good enough for now since a user who wants better security - * would also enable signing on the mount. Having validation of - * negotiate info for signed connections helps reduce attack vectors + * Having validation of negotiate info helps reduce attack vectors. */ - if (tcon->ses->server->sign == false) + if (tcon->ses->session_flags & SMB2_SESSION_FLAG_IS_GUEST) return 0; /* validation requires signing */ + if (tcon->ses->user_name == NULL) { + cifs_dbg(FYI, "Can't validate negotiate: null user mount\n"); + return 0; /* validation requires signing */ + } + + if (tcon->ses->session_flags & SMB2_SESSION_FLAG_IS_NULL) + cifs_dbg(VFS, "Unexpected null user (anonymous) auth flag sent by server\n"); + vneg_inbuf.Capabilities = cpu_to_le32(tcon->ses->server->vals->req_capabilities); memcpy(vneg_inbuf.Guid, tcon->ses->server->client_guid, -- cgit v1.2.3 From 6e70e26dc52be62c1f39f81b5f71fa5e643677aa Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 21 Sep 2017 21:32:29 -0500 Subject: SMB3: handle new statx fields We weren't returning the creation time or the two easily supported attributes (ENCRYPTED or COMPRESSED) for the getattr call to allow statx to return these fields. Signed-off-by: Steve French Reviewed-by: Ronnie Sahlberg \ Acked-by: Jeff Layton CC: Stable Reviewed-by: Pavel Shilovsky --- fs/cifs/inode.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'fs') diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c index a8693632235f..7c732cb44164 100644 --- a/fs/cifs/inode.c +++ b/fs/cifs/inode.c @@ -234,6 +234,8 @@ cifs_unix_basic_to_fattr(struct cifs_fattr *fattr, FILE_UNIX_BASIC_INFO *info, fattr->cf_atime = cifs_NTtimeToUnix(info->LastAccessTime); fattr->cf_mtime = cifs_NTtimeToUnix(info->LastModificationTime); fattr->cf_ctime = cifs_NTtimeToUnix(info->LastStatusChange); + /* old POSIX extensions don't get create time */ + fattr->cf_mode = le64_to_cpu(info->Permissions); /* @@ -2024,6 +2026,19 @@ int cifs_getattr(const struct path *path, struct kstat *stat, stat->blksize = CIFS_MAX_MSGSIZE; stat->ino = CIFS_I(inode)->uniqueid; + /* old CIFS Unix Extensions doesn't return create time */ + if (CIFS_I(inode)->createtime) { + stat->result_mask |= STATX_BTIME; + stat->btime = + cifs_NTtimeToUnix(cpu_to_le64(CIFS_I(inode)->createtime)); + } + + stat->attributes_mask |= (STATX_ATTR_COMPRESSED | STATX_ATTR_ENCRYPTED); + if (CIFS_I(inode)->cifsAttrs & FILE_ATTRIBUTE_COMPRESSED) + stat->attributes |= STATX_ATTR_COMPRESSED; + if (CIFS_I(inode)->cifsAttrs & FILE_ATTRIBUTE_ENCRYPTED) + stat->attributes |= STATX_ATTR_ENCRYPTED; + /* * If on a multiuser mount without unix extensions or cifsacl being * enabled, and the admin hasn't overridden them, set the ownership -- cgit v1.2.3 From 1013e760d10e614dc10b5624ce9fc41563ba2e65 Mon Sep 17 00:00:00 2001 From: Steve French Date: Fri, 22 Sep 2017 01:40:27 -0500 Subject: SMB3: Don't ignore O_SYNC/O_DSYNC and O_DIRECT flags Signed-off-by: Steve French CC: Stable Reviewed-by: Ronnie Sahlberg Reviewed-by: Pavel Shilovsky --- fs/cifs/file.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs') diff --git a/fs/cifs/file.c b/fs/cifs/file.c index 822311919163..92fdf9c35de2 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -224,6 +224,13 @@ cifs_nt_open(char *full_path, struct inode *inode, struct cifs_sb_info *cifs_sb, if (backup_cred(cifs_sb)) create_options |= CREATE_OPEN_BACKUP_INTENT; + /* O_SYNC also has bit for O_DSYNC so following check picks up either */ + if (f_flags & O_SYNC) + create_options |= CREATE_WRITE_THROUGH; + + if (f_flags & O_DIRECT) + create_options |= CREATE_NO_BUFFER; + oparms.tcon = tcon; oparms.cifs_sb = cifs_sb; oparms.desired_access = desired_access; -- cgit v1.2.3 From 332391a9935da939319e473b4680e173df75afcf Mon Sep 17 00:00:00 2001 From: Lukas Czerner Date: Thu, 21 Sep 2017 08:16:29 -0600 Subject: fs: Fix page cache inconsistency when mixing buffered and AIO DIO Currently when mixing buffered reads and asynchronous direct writes it is possible to end up with the situation where we have stale data in the page cache while the new data is already written to disk. This is permanent until the affected pages are flushed away. Despite the fact that mixing buffered and direct IO is ill-advised it does pose a thread for a data integrity, is unexpected and should be fixed. Fix this by deferring completion of asynchronous direct writes to a process context in the case that there are mapped pages to be found in the inode. Later before the completion in dio_complete() invalidate the pages in question. This ensures that after the completion the pages in the written area are either unmapped, or populated with up-to-date data. Also do the same for the iomap case which uses iomap_dio_complete() instead. This has a side effect of deferring the completion to a process context for every AIO DIO that happens on inode that has pages mapped. However since the consensus is that this is ill-advised practice the performance implication should not be a problem. This was based on proposal from Jeff Moyer, thanks! Reviewed-by: Jan Kara Reviewed-by: Darrick J. Wong Reviewed-by: Jeff Moyer Signed-off-by: Lukas Czerner Signed-off-by: Jens Axboe --- fs/direct-io.c | 49 +++++++++++++++++++++++++++++++++++++++++++------ fs/iomap.c | 29 ++++++++++++++++------------- mm/filemap.c | 10 ++++++++-- 3 files changed, 67 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/direct-io.c b/fs/direct-io.c index 5fa2211e49ae..62cf812ed0e5 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -229,6 +229,7 @@ static ssize_t dio_complete(struct dio *dio, ssize_t ret, bool is_async) { loff_t offset = dio->iocb->ki_pos; ssize_t transferred = 0; + int err; /* * AIO submission can race with bio completion to get here while @@ -258,8 +259,22 @@ static ssize_t dio_complete(struct dio *dio, ssize_t ret, bool is_async) if (ret == 0) ret = transferred; + /* + * Try again to invalidate clean pages which might have been cached by + * non-direct readahead, or faulted in by get_user_pages() if the source + * of the write was an mmap'ed region of the file we're writing. Either + * one is a pretty crazy thing to do, so we don't support it 100%. If + * this invalidation fails, tough, the write still worked... + */ + if (ret > 0 && dio->op == REQ_OP_WRITE && + dio->inode->i_mapping->nrpages) { + err = invalidate_inode_pages2_range(dio->inode->i_mapping, + offset >> PAGE_SHIFT, + (offset + ret - 1) >> PAGE_SHIFT); + WARN_ON_ONCE(err); + } + if (dio->end_io) { - int err; // XXX: ki_pos?? err = dio->end_io(dio->iocb, offset, ret, dio->private); @@ -304,6 +319,7 @@ static void dio_bio_end_aio(struct bio *bio) struct dio *dio = bio->bi_private; unsigned long remaining; unsigned long flags; + bool defer_completion = false; /* cleanup the bio */ dio_bio_complete(dio, bio); @@ -315,7 +331,19 @@ static void dio_bio_end_aio(struct bio *bio) spin_unlock_irqrestore(&dio->bio_lock, flags); if (remaining == 0) { - if (dio->result && dio->defer_completion) { + /* + * Defer completion when defer_completion is set or + * when the inode has pages mapped and this is AIO write. + * We need to invalidate those pages because there is a + * chance they contain stale data in the case buffered IO + * went in between AIO submission and completion into the + * same region. + */ + if (dio->result) + defer_completion = dio->defer_completion || + (dio->op == REQ_OP_WRITE && + dio->inode->i_mapping->nrpages); + if (defer_completion) { INIT_WORK(&dio->complete_work, dio_aio_complete_work); queue_work(dio->inode->i_sb->s_dio_done_wq, &dio->complete_work); @@ -1210,10 +1238,19 @@ do_blockdev_direct_IO(struct kiocb *iocb, struct inode *inode, * For AIO O_(D)SYNC writes we need to defer completions to a workqueue * so that we can call ->fsync. */ - if (dio->is_async && iov_iter_rw(iter) == WRITE && - ((iocb->ki_filp->f_flags & O_DSYNC) || - IS_SYNC(iocb->ki_filp->f_mapping->host))) { - retval = dio_set_defer_completion(dio); + if (dio->is_async && iov_iter_rw(iter) == WRITE) { + retval = 0; + if ((iocb->ki_filp->f_flags & O_DSYNC) || + IS_SYNC(iocb->ki_filp->f_mapping->host)) + retval = dio_set_defer_completion(dio); + else if (!dio->inode->i_sb->s_dio_done_wq) { + /* + * In case of AIO write racing with buffered read we + * need to defer completion. We can't decide this now, + * however the workqueue needs to be initialized here. + */ + retval = sb_init_dio_done_wq(dio->inode->i_sb); + } if (retval) { /* * We grab i_mutex only for reads so we don't have diff --git a/fs/iomap.c b/fs/iomap.c index 269b24a01f32..8194d30bdca0 100644 --- a/fs/iomap.c +++ b/fs/iomap.c @@ -713,8 +713,24 @@ struct iomap_dio { static ssize_t iomap_dio_complete(struct iomap_dio *dio) { struct kiocb *iocb = dio->iocb; + struct inode *inode = file_inode(iocb->ki_filp); ssize_t ret; + /* + * Try again to invalidate clean pages which might have been cached by + * non-direct readahead, or faulted in by get_user_pages() if the source + * of the write was an mmap'ed region of the file we're writing. Either + * one is a pretty crazy thing to do, so we don't support it 100%. If + * this invalidation fails, tough, the write still worked... + */ + if (!dio->error && + (dio->flags & IOMAP_DIO_WRITE) && inode->i_mapping->nrpages) { + ret = invalidate_inode_pages2_range(inode->i_mapping, + iocb->ki_pos >> PAGE_SHIFT, + (iocb->ki_pos + dio->size - 1) >> PAGE_SHIFT); + WARN_ON_ONCE(ret); + } + if (dio->end_io) { ret = dio->end_io(iocb, dio->error ? dio->error : dio->size, @@ -1042,19 +1058,6 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, ret = iomap_dio_complete(dio); - /* - * Try again to invalidate clean pages which might have been cached by - * non-direct readahead, or faulted in by get_user_pages() if the source - * of the write was an mmap'ed region of the file we're writing. Either - * one is a pretty crazy thing to do, so we don't support it 100%. If - * this invalidation fails, tough, the write still worked... - */ - if (iov_iter_rw(iter) == WRITE) { - int err = invalidate_inode_pages2_range(mapping, - start >> PAGE_SHIFT, end >> PAGE_SHIFT); - WARN_ON_ONCE(err); - } - return ret; out_free_dio: diff --git a/mm/filemap.c b/mm/filemap.c index 870971e20967..db250d0e0565 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2926,9 +2926,15 @@ generic_file_direct_write(struct kiocb *iocb, struct iov_iter *from) * we're writing. Either one is a pretty crazy thing to do, * so we don't support it 100%. If this invalidation * fails, tough, the write still worked... + * + * Most of the time we do not need this since dio_complete() will do + * the invalidation for us. However there are some file systems that + * do not end up with dio_complete() being called, so let's not break + * them by removing it completely */ - invalidate_inode_pages2_range(mapping, - pos >> PAGE_SHIFT, end); + if (mapping->nrpages) + invalidate_inode_pages2_range(mapping, + pos >> PAGE_SHIFT, end); if (written > 0) { pos += written; -- cgit v1.2.3 From 10201655b085df8e000822e496e5d4016a167a36 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Tue, 19 Sep 2017 07:15:35 -0500 Subject: gfs2: Fix debugfs glocks dump The switch to rhashtables (commit 88ffbf3e03) broke the debugfs glock dump (/sys/kernel/debug/gfs2//glocks) for dumps bigger than a single buffer: the right function for restarting an rhashtable iteration from the beginning of the hash table is rhashtable_walk_enter; rhashtable_walk_stop + rhashtable_walk_start will just resume from the current position. Signed-off-by: Andreas Gruenbacher Signed-off-by: Bob Peterson Cc: stable@vger.kernel.org # v4.3+ --- fs/gfs2/glock.c | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 98e845b7841b..11066d8647d2 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1945,13 +1945,9 @@ static void *gfs2_glock_seq_start(struct seq_file *seq, loff_t *pos) { struct gfs2_glock_iter *gi = seq->private; loff_t n = *pos; - int ret; - - if (gi->last_pos <= *pos) - n = (*pos - gi->last_pos); - ret = rhashtable_walk_start(&gi->hti); - if (ret) + rhashtable_walk_enter(&gl_hash_table, &gi->hti); + if (rhashtable_walk_start(&gi->hti) != 0) return NULL; do { @@ -1959,6 +1955,7 @@ static void *gfs2_glock_seq_start(struct seq_file *seq, loff_t *pos) } while (gi->gl && n--); gi->last_pos = *pos; + return gi->gl; } @@ -1970,6 +1967,7 @@ static void *gfs2_glock_seq_next(struct seq_file *seq, void *iter_ptr, (*pos)++; gi->last_pos = *pos; gfs2_glock_iter_next(gi); + return gi->gl; } @@ -1980,6 +1978,7 @@ static void gfs2_glock_seq_stop(struct seq_file *seq, void *iter_ptr) gi->gl = NULL; rhashtable_walk_stop(&gi->hti); + rhashtable_walk_exit(&gi->hti); } static int gfs2_glock_seq_show(struct seq_file *seq, void *iter_ptr) @@ -2042,12 +2041,10 @@ static int __gfs2_glocks_open(struct inode *inode, struct file *file, struct gfs2_glock_iter *gi = seq->private; gi->sdp = inode->i_private; - gi->last_pos = 0; seq->buf = kmalloc(GFS2_SEQ_GOODSIZE, GFP_KERNEL | __GFP_NOWARN); if (seq->buf) seq->size = GFS2_SEQ_GOODSIZE; gi->gl = NULL; - rhashtable_walk_enter(&gl_hash_table, &gi->hti); } return ret; } @@ -2063,7 +2060,6 @@ static int gfs2_glocks_release(struct inode *inode, struct file *file) struct gfs2_glock_iter *gi = seq->private; gi->gl = NULL; - rhashtable_walk_exit(&gi->hti); return seq_release_private(inode, file); } -- cgit v1.2.3 From cc6f77710a6de6210f9feda7cd53e2f5ee7a7e69 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 18 Sep 2017 09:41:16 -0700 Subject: xfs: don't unconditionally clear the reflink flag on zero-block files If we have speculative cow preallocations hanging around in the cow fork, don't let a truncate operation clear the reflink flag because if we do then there's a chance we'll forget to free those extents when we destroy the incore inode. Reported-by: Amir Goldstein Reviewed-by: Carlos Maiolino Reviewed-by: Christoph Hellwig Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_inode.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 5599dda4727a..4ec5b7f45401 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1624,10 +1624,12 @@ xfs_itruncate_extents( goto out; /* - * Clear the reflink flag if we truncated everything. + * Clear the reflink flag if there are no data fork blocks and + * there are no extents staged in the cow fork. */ - if (ip->i_d.di_nblocks == 0 && xfs_is_reflink_inode(ip)) { - ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + if (xfs_is_reflink_inode(ip) && ip->i_cnextents == 0) { + if (ip->i_d.di_nblocks == 0) + ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; xfs_inode_clear_cowblocks_tag(ip); } -- cgit v1.2.3 From 3af423b03435c81036fa710623d3ae92fbe346a3 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 18 Sep 2017 09:41:17 -0700 Subject: xfs: evict CoW fork extents when performing finsert/fcollapse When we perform an finsert/fcollapse operation, cancel all the CoW extents for the affected file offset range so that they don't end up pointing to the wrong blocks. Reported-by: Amir Goldstein Reviewed-by: Carlos Maiolino Reviewed-by: Christoph Hellwig Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_bmap_util.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index cd9a5400ba4f..bc6c6e10a969 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -1459,7 +1459,19 @@ xfs_shift_file_space( return error; /* - * The extent shiting code works on extent granularity. So, if + * Clean out anything hanging around in the cow fork now that + * we've flushed all the dirty data out to disk to avoid having + * CoW extents at the wrong offsets. + */ + if (xfs_is_reflink_inode(ip)) { + error = xfs_reflink_cancel_cow_range(ip, offset, NULLFILEOFF, + true); + if (error) + return error; + } + + /* + * The extent shifting code works on extent granularity. So, if * stop_fsb is not the starting block of extent, we need to split * the extent at stop_fsb. */ -- cgit v1.2.3 From e150dcd459e1b441eaf08f341a986f04e61bf3b8 Mon Sep 17 00:00:00 2001 From: Helge Deller Date: Mon, 18 Sep 2017 11:34:16 -0700 Subject: fs/xfs: Use %pS printk format for direct addresses Use the %pS instead of the %pF printk format specifier for printing symbols from direct addresses. This is needed for the ia64, ppc64 and parisc64 architectures. Signed-off-by: Helge Deller Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_error.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c index bd786a9ac2c3..eaf86f55b7f2 100644 --- a/fs/xfs/xfs_error.c +++ b/fs/xfs/xfs_error.c @@ -347,7 +347,7 @@ xfs_verifier_error( { struct xfs_mount *mp = bp->b_target->bt_mount; - xfs_alert(mp, "Metadata %s detected at %pF, %s block 0x%llx", + xfs_alert(mp, "Metadata %s detected at %pS, %s block 0x%llx", bp->b_error == -EFSBADCRC ? "CRC error" : "corruption", __return_address, bp->b_ops->name, bp->b_bn); -- cgit v1.2.3 From 64671bafbdd984535aa382bccadd91fbe7be0e80 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Mon, 18 Sep 2017 11:38:58 -0700 Subject: xfs: kill meaningless variable 'zero' In xfs_file_aio_write_checks(), variable 'zero' is there only to satisfy xfs_zero_eof(), the result of it is ignored. Now, with iomap_zero_range() based xfs_zero_eof(), we can safely pass NULL as the last param of it and kill 'zero'. Signed-off-by: Eryu Guan Reviewed-by: Carlos Maiolino Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_file.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index ebdd0bd2b261..261d83f1db76 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -377,8 +377,6 @@ restart: */ spin_lock(&ip->i_flags_lock); if (iocb->ki_pos > i_size_read(inode)) { - bool zero = false; - spin_unlock(&ip->i_flags_lock); if (!drained_dio) { if (*iolock == XFS_IOLOCK_SHARED) { @@ -399,7 +397,7 @@ restart: drained_dio = true; goto restart; } - error = xfs_zero_eof(ip, iocb->ki_pos, i_size_read(inode), &zero); + error = xfs_zero_eof(ip, iocb->ki_pos, i_size_read(inode), NULL); if (error) return error; } else -- cgit v1.2.3 From d20a5e3851969fa685f118a80e4df670255a4e8d Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Mon, 18 Sep 2017 11:39:23 -0700 Subject: xfs: report zeroed or not correctly in xfs_zero_range() The 'did_zero' param of xfs_zero_range() was not passed to iomap_zero_range() correctly. This was introduced by commit 7bb41db3ea16 ("xfs: handle 64-bit length in xfs_iozero"), and found by code inspection. Signed-off-by: Eryu Guan Reviewed-by: Carlos Maiolino Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 261d83f1db76..350b6d43ba23 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -58,7 +58,7 @@ xfs_zero_range( xfs_off_t count, bool *did_zero) { - return iomap_zero_range(VFS_I(ip), pos, count, NULL, &xfs_iomap_ops); + return iomap_zero_range(VFS_I(ip), pos, count, did_zero, &xfs_iomap_ops); } int -- cgit v1.2.3 From 1e6fa688bffc0ff419a4c3e78dbaf7aabfb55183 Mon Sep 17 00:00:00 2001 From: Kenjiro Nakayama Date: Mon, 18 Sep 2017 12:03:56 -0700 Subject: xfs: Output warning message when discard option was enabled even though the device does not support discard In order to using discard function, it is necessary that not only xfs is mounted with discard option, but also the discard function is supported by the device. Current code doesn't output any message when users mount with discard option on unsupported device, so it is difficult to notice that it was not enabled actually. This patch adds the warning message to notice that discard option is not enabled due to unsupported device when the filesystem is mounted. Changes in v2 (Suggested by Brian Foster): - Move the unsupported device check into xfs_fs_fill_super(). - Clear the discard flag when device is unsupported. Signed-off-by: Kenjiro Nakayama Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_super.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index c996f4ae4a5f..584cf2d573ba 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -1654,6 +1654,16 @@ xfs_fs_fill_super( "DAX and reflink have not been tested together!"); } + if (mp->m_flags & XFS_MOUNT_DISCARD) { + struct request_queue *q = bdev_get_queue(sb->s_bdev); + + if (!blk_queue_discard(q)) { + xfs_warn(mp, "mounting with \"discard\" option, but " + "the device does not support discard"); + mp->m_flags &= ~XFS_MOUNT_DISCARD; + } + } + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { if (mp->m_sb.sb_rblocks) { xfs_alert(mp, -- cgit v1.2.3 From 60915f83cd1e021a66fc1503a446aef5c772553a Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Mon, 18 Sep 2017 13:38:46 -0700 Subject: xfs: remove redundant re-initialization of total_nr_pages Variable total_nr_pages is being initialized and then updated with the same value, this latter assignment is redundant and can be removed. Cleans up clang build warning: Value stored to 'total_nr_pages' during its initialization is never read Signed-off-by: Colin Ian King Reviewed-by: Brian Foster Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_buf.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index da14658da310..2f97c12ca75e 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -1258,8 +1258,6 @@ xfs_buf_ioapply_map( int size; int offset; - total_nr_pages = bp->b_page_count; - /* skip the pages in the buffer before the start offset */ page_index = 0; offset = *buf_offset; -- cgit v1.2.3 From ff40adf7fbdff96860b1153332c0b1c7bab6e0c1 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Thu, 24 Aug 2017 18:19:48 -0600 Subject: Btrfs: use the new helper wbc_to_write_flags This updates btrfs to use the helper wbc_to_write_flags which has been applied in ext4/xfs/f2fs/block. Please note that, with this, btrfs's dirty pages written by a writeback job will carry the flag REQ_BACKGROUND, which is currently used by writeback-throttle to determine whether it should go to get a request or wait. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index d17783d70228..4ead6da5a645 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3471,8 +3471,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, unsigned int write_flags = 0; unsigned long nr_written = 0; - if (wbc->sync_mode == WB_SYNC_ALL) - write_flags = REQ_SYNC; + write_flags = wbc_to_write_flags(wbc); trace___extent_writepage(page, inode, wbc); @@ -3718,7 +3717,7 @@ static noinline_for_stack int write_one_eb(struct extent_buffer *eb, unsigned long i, num_pages; unsigned long bio_flags = 0; unsigned long start, end; - unsigned int write_flags = (epd->sync_io ? REQ_SYNC : 0) | REQ_META; + unsigned int write_flags = wbc_to_write_flags(wbc) | REQ_META; int ret = 0; clear_bit(EXTENT_BUFFER_WRITE_ERR, &eb->bflags); -- cgit v1.2.3 From 5f14efd3d437205143dcffcf776e0122eae1755a Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 23 Aug 2017 12:15:09 -0600 Subject: Btrfs: do not reset bio->bi_ops while writing bio flush_epd_write_bio() sets bio->bi_opf by itself to honor REQ_SYNC, but it's not needed at all since bio->bi_opf has set up properly in both __extent_writepage() and write_one_eb(), and in the case of write_one_eb(), it also sets REQ_META, which we will lose in flush_epd_write_bio(). This remove this unnecessary bio->bi_opf setting. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 4ead6da5a645..3738d245518c 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4062,9 +4062,6 @@ static void flush_epd_write_bio(struct extent_page_data *epd) if (epd->bio) { int ret; - bio_set_op_attrs(epd->bio, REQ_OP_WRITE, - epd->sync_io ? REQ_SYNC : 0); - ret = submit_one_bio(epd->bio, 0, epd->bio_flags); BUG_ON(ret < 0); /* -ENOMEM */ epd->bio = NULL; -- cgit v1.2.3 From bea7eafdbda3ba1d4b2ccb9cca829eefb7989bb9 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Tue, 22 Aug 2017 23:46:00 -0700 Subject: Btrfs: fix incorrect {node,sector}size endianness from BTRFS_IOC_FS_INFO fs_info->super_copy->{node,sector}size are little-endian, but the ioctl should return the values in native endianness. Use the cached values in btrfs_fs_info instead. Found with sparse. Fixes: 80a773fbfc2d ("btrfs: retrieve more info from FS_INFO ioctl") Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index ae8fbf9d3de2..cf1c2ee030dd 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -2769,9 +2769,9 @@ static long btrfs_ioctl_fs_info(struct btrfs_fs_info *fs_info, } mutex_unlock(&fs_devices->device_list_mutex); - fi_args->nodesize = fs_info->super_copy->nodesize; - fi_args->sectorsize = fs_info->super_copy->sectorsize; - fi_args->clone_alignment = fs_info->super_copy->sectorsize; + fi_args->nodesize = fs_info->nodesize; + fi_args->sectorsize = fs_info->sectorsize; + fi_args->clone_alignment = fs_info->sectorsize; if (copy_to_user(arg, fi_args, sizeof(*fi_args))) ret = -EFAULT; -- cgit v1.2.3 From 63d71450c8d817649a79e37d685523f988b9cc98 Mon Sep 17 00:00:00 2001 From: Naohiro Aota Date: Fri, 1 Sep 2017 17:58:47 +0900 Subject: btrfs: clear ordered flag on cleaning up ordered extents Commit 524272607e88 ("btrfs: Handle delalloc error correctly to avoid ordered extent hang") introduced btrfs_cleanup_ordered_extents() to cleanup submitted ordered extents. However, it does not clear the ordered bit (Private2) of corresponding pages. Thus, the following BUG occurs from free_pages_check_bad() (on btrfs/125 with nospace_cache). BUG: Bad page state in process btrfs pfn:3fa787 page:ffffdf2acfe9e1c0 count:0 mapcount:0 mapping: (null) index:0xd flags: 0x8000000000002008(uptodate|private_2) raw: 8000000000002008 0000000000000000 000000000000000d 00000000ffffffff raw: ffffdf2acf5c1b20 ffffb443802238b0 0000000000000000 0000000000000000 page dumped because: PAGE_FLAGS_CHECK_AT_FREE flag(s) set bad because of flags: 0x2000(private_2) This patch clears the flag same as other places calling btrfs_dec_test_ordered_pending() for every page in the specified range. Fixes: 524272607e88 ("btrfs: Handle delalloc error correctly to avoid ordered extent hang") Cc: # 4.12 Signed-off-by: Naohiro Aota Reviewed-by: Qu Wenruo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/inode.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d184a46e46c4..455c0f22fe2d 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -135,6 +135,18 @@ static inline void btrfs_cleanup_ordered_extents(struct inode *inode, const u64 offset, const u64 bytes) { + unsigned long index = offset >> PAGE_SHIFT; + unsigned long end_index = (offset + bytes - 1) >> PAGE_SHIFT; + struct page *page; + + while (index <= end_index) { + page = find_get_page(inode->i_mapping, index); + index++; + if (!page) + continue; + ClearPagePrivate2(page); + put_page(page); + } return __endio_write_update_ordered(inode, offset + PAGE_SIZE, bytes - PAGE_SIZE, false); } -- cgit v1.2.3 From 67c003f90fd68062d92a7ffade36f9b2a9098bd8 Mon Sep 17 00:00:00 2001 From: Naohiro Aota Date: Fri, 1 Sep 2017 17:59:07 +0900 Subject: btrfs: finish ordered extent cleaning if no progress is found __endio_write_update_ordered() repeats the search until it reaches the end of the specified range. This works well with direct IO path, because before the function is called, it's ensured that there are ordered extents filling whole the range. It's not the case, however, when it's called from run_delalloc_range(): it is possible to have error in the midle of the loop in e.g. run_delalloc_nocow(), so that there exisits the range not covered by any ordered extents. By cleaning such "uncomplete" range, __endio_write_update_ordered() stucks at offset where there're no ordered extents. Since the ordered extents are created from head to tail, we can stop the search if there are no offset progress. Fixes: 524272607e88 ("btrfs: Handle delalloc error correctly to avoid ordered extent hang") Cc: # 4.12 Signed-off-by: Naohiro Aota Reviewed-by: Qu Wenruo Reviewed-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/inode.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 455c0f22fe2d..f78c5640c6dc 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -8396,6 +8396,7 @@ static void __endio_write_update_ordered(struct inode *inode, btrfs_work_func_t func; u64 ordered_offset = offset; u64 ordered_bytes = bytes; + u64 last_offset; int ret; if (btrfs_is_free_space_inode(BTRFS_I(inode))) { @@ -8407,6 +8408,7 @@ static void __endio_write_update_ordered(struct inode *inode, } again: + last_offset = ordered_offset; ret = btrfs_dec_test_first_ordered_pending(inode, &ordered, &ordered_offset, ordered_bytes, @@ -8417,6 +8419,12 @@ again: btrfs_init_work(&ordered->work, func, finish_ordered_fn, NULL, NULL); btrfs_queue_work(wq, &ordered->work); out_test: + /* + * If btrfs_dec_test_ordered_pending does not find any ordered extent + * in the range, we can exit. + */ + if (ordered_offset == last_offset) + return; /* * our bio might span multiple ordered extents. If we haven't * completed the accounting for the whole dio, go back and try again -- cgit v1.2.3 From bb166d7207432d3c7d10c45dc052f12ba3a2121d Mon Sep 17 00:00:00 2001 From: Naohiro Aota Date: Fri, 25 Aug 2017 14:15:14 +0900 Subject: btrfs: fix NULL pointer dereference from free_reloc_roots() __del_reloc_root should be called before freeing up reloc_root->node. If not, calling __del_reloc_root() dereference reloc_root->node, causing the system BUG. Fixes: 6bdf131fac23 ("Btrfs: don't leak reloc root nodes on error") Cc: # 4.9 Signed-off-by: Naohiro Aota Reviewed-by: Nikolay Borisov Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/relocation.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 3a49a3c2fca4..9841faef08ea 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2400,11 +2400,11 @@ void free_reloc_roots(struct list_head *list) while (!list_empty(list)) { reloc_root = list_entry(list->next, struct btrfs_root, root_list); + __del_reloc_root(reloc_root); free_extent_buffer(reloc_root->node); free_extent_buffer(reloc_root->commit_root); reloc_root->node = NULL; reloc_root->commit_root = NULL; - __del_reloc_root(reloc_root); } } -- cgit v1.2.3 From ca6842bf01dc1ad41195eac1e343b4f08c496ba8 Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Fri, 22 Jan 2016 09:13:25 +0900 Subject: Btrfs: send: fix error number for unknown inode types ENOTSUPP should not be returned to the user program. (cf. include/linux/errno.h) Therefore, EOPNOTSUPP is used instead of ENOTSUPP. Signed-off-by: Tsutomu Itoh Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/send.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 8f1d3d6e7087..43430e6c99aa 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -2640,7 +2640,7 @@ static int send_create_inode(struct send_ctx *sctx, u64 ino) } else { btrfs_warn(sctx->send_root->fs_info, "unexpected inode type %o", (int)(mode & S_IFMT)); - ret = -ENOTSUPP; + ret = -EOPNOTSUPP; goto out; } -- cgit v1.2.3 From 6d6d282932d1a609e60dc4467677e0e863682f57 Mon Sep 17 00:00:00 2001 From: satoru takeuchi Date: Tue, 12 Sep 2017 22:42:52 +0900 Subject: btrfs: prevent to set invalid default subvolid `btrfs sub set-default` succeeds to set an ID which isn't corresponding to any fs/file tree. If such the bad ID is set to a filesystem, we can't mount this filesystem without specifying `subvol` or `subvolid` mount options. Fixes: 6ef5ed0d386b ("Btrfs: add ioctl and incompat flag to set the default mount subvol") Cc: Signed-off-by: Satoru Takeuchi Reviewed-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index cf1c2ee030dd..d4a77993e52f 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -4057,6 +4057,10 @@ static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp) ret = PTR_ERR(new_root); goto out; } + if (!is_fstree(new_root->objectid)) { + ret = -ENOENT; + goto out; + } path = btrfs_alloc_path(); if (!path) { -- cgit v1.2.3 From 78ad4ce014d025f41b8dde3a81876832ead643cf Mon Sep 17 00:00:00 2001 From: Naohiro Aota Date: Fri, 8 Sep 2017 17:48:55 +0900 Subject: btrfs: propagate error to btrfs_cmp_data_prepare caller btrfs_cmp_data_prepare() (almost) always returns 0 i.e. ignoring errors from gather_extent_pages(). While the pages are freed by btrfs_cmp_data_free(), cmp->num_pages still has > 0. Then, btrfs_extent_same() try to access the already freed pages causing faults (or violates PageLocked assertion). This patch just return the error as is so that the caller stop the process. Signed-off-by: Naohiro Aota Fixes: f441460202cb ("btrfs: fix deadlock with extent-same and readpage") Cc: # 4.2 Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index d4a77993e52f..802df5755cd3 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -3028,7 +3028,7 @@ static int btrfs_cmp_data_prepare(struct inode *src, u64 loff, out: if (ret) btrfs_cmp_data_free(cmp); - return 0; + return ret; } static int btrfs_cmp_data(u64 len, struct cmp_pages *cmp) -- cgit v1.2.3 From c2faff790ccd11ea5be8e3ca99713f116fcd6030 Mon Sep 17 00:00:00 2001 From: "Misono, Tomohiro" Date: Wed, 30 Aug 2017 16:33:16 +0900 Subject: btrfs: remove BTRFS_FS_QUOTA_DISABLING flag Currently, "btrfs quota enable" would fail after "btrfs quota disable" on the first time with syslog output "qgroup_rescan_init failed with -22", but it would succeed on the second time. When "quota disable" is called, BTRFS_FS_QUOTA_DISABLING flag bit will be set in fs_info->flags in btrfs_quota_disable(), but it will not be droppd in btrfs_run_qgroups() (which is called in btrfs_commit_transaction()) because quota_root has already been freed. If "quota enable" is called after that, both BTRFS_FS_QUOTA_DISABLING and BTRFS_FS_QUOTA_ENABLED flag would be dropped in the btrfs_run_qgroups() since quota_root is not NULL. This leads to the failure of "quota enable" on the first time. BTRFS_FS_QUOTA_DISABLING flag is not used outside of "quota disable" context and is equivalent to whether quota_root is NULL or not. btrfs_run_qgroups() checks whether quota_root is NULL or not in the first place. So, let's remove BTRFS_FS_QUOTA_DISABLING flag. Signed-off-by: Tomohiro Misono Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 1 - fs/btrfs/qgroup.c | 4 ---- 2 files changed, 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 2add002662f4..b7ccfcc01732 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -708,7 +708,6 @@ struct btrfs_delayed_root; #define BTRFS_FS_OPEN 5 #define BTRFS_FS_QUOTA_ENABLED 6 #define BTRFS_FS_QUOTA_ENABLING 7 -#define BTRFS_FS_QUOTA_DISABLING 8 #define BTRFS_FS_UPDATE_UUID_TREE_GEN 9 #define BTRFS_FS_CREATING_FREE_SPACE_TREE 10 #define BTRFS_FS_BTREE_ERR 11 diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 5c8b61c86e61..770f667269f5 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -807,7 +807,6 @@ static int btrfs_clean_quota_tree(struct btrfs_trans_handle *trans, } ret = 0; out: - set_bit(BTRFS_FS_QUOTA_DISABLING, &root->fs_info->flags); btrfs_free_path(path); return ret; } @@ -953,7 +952,6 @@ int btrfs_quota_disable(struct btrfs_trans_handle *trans, if (!fs_info->quota_root) goto out; clear_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags); - set_bit(BTRFS_FS_QUOTA_DISABLING, &fs_info->flags); btrfs_qgroup_wait_for_completion(fs_info, false); spin_lock(&fs_info->qgroup_lock); quota_root = fs_info->quota_root; @@ -2086,8 +2084,6 @@ int btrfs_run_qgroups(struct btrfs_trans_handle *trans, if (test_and_clear_bit(BTRFS_FS_QUOTA_ENABLING, &fs_info->flags)) set_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags); - if (test_and_clear_bit(BTRFS_FS_QUOTA_DISABLING, &fs_info->flags)) - clear_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags); spin_lock(&fs_info->qgroup_lock); while (!list_empty(&fs_info->dirty_qgroups)) { -- cgit v1.2.3 From fed3b381145e2e7c66b0b3f640851e1633ebd07f Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 13 Sep 2017 12:25:21 -0600 Subject: Btrfs: do not backup tree roots when fsync It doesn't make sense to backup tree roots when doing fsync, since during fsync those tree roots have not been consistent on disk. Signed-off-by: Liu Bo Reviewed-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 27d458640536..0f2271815eb6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3641,7 +3641,14 @@ int write_all_supers(struct btrfs_fs_info *fs_info, int max_mirrors) u64 flags; do_barriers = !btrfs_test_opt(fs_info, NOBARRIER); - backup_super_roots(fs_info); + + /* + * max_mirrors == 0 indicates we're from commit_transaction, + * not from fsync where the tree roots in fs_info have not + * been consistent on disk. + */ + if (max_mirrors == 0) + backup_super_roots(fs_info); sb = fs_info->super_for_commit; dev_item = &sb->dev_item; -- cgit v1.2.3 From bd7d63c2ceaf737eeb21630a2b62fc5fe34dba29 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 19 Sep 2017 17:50:09 -0600 Subject: Btrfs: use btrfs_op instead of bio_op in __btrfs_map_block This seems to be a leftover of commit cf8cddd38bab ("btrfs: don't abuse REQ_OP_* flags for btrfs_map_block"). It should use btrfs_op() helper to provide one of 'enum btrfs_map_op' types. Fixes: cf8cddd38bab ("btrfs: don't abuse REQ_OP_* flags for btrfs_map_block") Signed-off-by: Liu Bo Reviewed-by: Satoru Takeuchi Reviewed-by: David Sterba 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 d024f1b07282..6a72f88f77b6 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -6166,7 +6166,7 @@ int btrfs_map_bio(struct btrfs_fs_info *fs_info, struct bio *bio, map_length = length; btrfs_bio_counter_inc_blocked(fs_info); - ret = __btrfs_map_block(fs_info, bio_op(bio), logical, + ret = __btrfs_map_block(fs_info, btrfs_op(bio), logical, &map_length, &bbio, mirror_num, 1); if (ret) { btrfs_bio_counter_dec(fs_info); -- cgit v1.2.3 From cf1167d5c1abf3bc42b2a1562bfa7937c05337e2 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 20 Sep 2017 17:50:18 -0600 Subject: Btrfs: fix kernel oops while reading compressed data The kernel oops happens at kernel BUG at fs/btrfs/extent_io.c:2104! ... RIP: clean_io_failure+0x263/0x2a0 [btrfs] It's showing that read-repair code is using an improper mirror index. This is due to the fact that compression read's endio hasn't recorded the failed mirror index in %cb->orig_bio. With this, btrfs's read-repair can work properly on reading compressed data. Signed-off-by: Liu Bo Reported-by: Paul Jones Tested-by: Paul Jones Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/compression.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 883ecc58fd0d..c6aa53c4c102 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -107,6 +107,7 @@ static void end_compressed_bio_read(struct bio *bio) struct inode *inode; struct page *page; unsigned long index; + unsigned int mirror = btrfs_io_bio(bio)->mirror_num; int ret; if (bio->bi_status) @@ -118,6 +119,14 @@ static void end_compressed_bio_read(struct bio *bio) if (!refcount_dec_and_test(&cb->pending_bios)) goto out; + /* + * Record the correct mirror_num in cb->orig_bio so that + * read-repair can work properly. + */ + ASSERT(btrfs_io_bio(cb->orig_bio)); + btrfs_io_bio(cb->orig_bio)->mirror_num = mirror; + cb->mirror_num = mirror; + inode = cb->inode; ret = check_compressed_csum(BTRFS_I(inode), cb, (u64)bio->bi_iter.bi_sector << 9); -- cgit v1.2.3 From e6311f240c946788131ba2b97e14f37312688072 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 20 Sep 2017 17:50:19 -0600 Subject: Btrfs: skip checksum when reading compressed data if some IO have failed Currently even if the underlying disk reports failure on IO, compressed read endio still gets to verify checksum and reports it as a checksum error. In fact, if some IO have failed during reading a compressed data extent , there's no way the checksum could match, therefore, we can skip that in order to return error quickly to the upper layer. Please note that we need to do this after recording the failed mirror index so that read-repair in the upper layer's endio can work properly. Signed-off-by: Liu Bo Tested-by: Paul Jones Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/compression.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index c6aa53c4c102..fc31af98a41b 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -108,7 +108,7 @@ static void end_compressed_bio_read(struct bio *bio) struct page *page; unsigned long index; unsigned int mirror = btrfs_io_bio(bio)->mirror_num; - int ret; + int ret = 0; if (bio->bi_status) cb->errors = 1; @@ -127,6 +127,13 @@ static void end_compressed_bio_read(struct bio *bio) btrfs_io_bio(cb->orig_bio)->mirror_num = mirror; cb->mirror_num = mirror; + /* + * Some IO in this cb have failed, just skip checksum as there + * is no way it could be correct. + */ + if (cb->errors == 1) + goto csum_failed; + inode = cb->inode; ret = check_compressed_csum(BTRFS_I(inode), cb, (u64)bio->bi_iter.bi_sector << 9); -- cgit v1.2.3 From 36b96fdc6b2dc6f4a0fedc563fa7508c91b90a10 Mon Sep 17 00:00:00 2001 From: Sargun Dhillon Date: Sun, 17 Sep 2017 09:02:29 +0000 Subject: btrfs: Report error on removing qgroup if del_qgroup_item fails Previously, we were calling del_qgroup_item, and ignoring the return code resulting in a potential to have divergent in-memory state without an error. Perhaps, it makes sense to handle this error code, and put the filesystem into a read only, or similar state. This patch only adds reporting of the error if the error is fatal, (any error other than qgroup not found). Signed-off-by: Sargun Dhillon Reviewed-by: Qu Wenruo Signed-off-by: David Sterba --- fs/btrfs/qgroup.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/btrfs/qgroup.c b/fs/btrfs/qgroup.c index 770f667269f5..e172d4843eae 100644 --- a/fs/btrfs/qgroup.c +++ b/fs/btrfs/qgroup.c @@ -1305,6 +1305,8 @@ int btrfs_remove_qgroup(struct btrfs_trans_handle *trans, } } ret = del_qgroup_item(trans, quota_root, qgroupid); + if (ret && ret != -ENOENT) + goto out; while (!list_empty(&qgroup->groups)) { list = list_first_entry(&qgroup->groups, -- cgit v1.2.3 From 99c4e3b96c797f047be4e6b7c03cfca01959f146 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 15 Sep 2017 15:06:51 -0600 Subject: Btrfs: fix unexpected result when dio reading corrupted blocks commit 4246a0b63bd8 ("block: add a bi_error field to struct bio") changed the logic of how dio read endio reports errors. For single stripe dio read, %bio->bi_status reflects the error before verifying checksum, and now we're updating it when data block matches with its checksum, while in the mismatching case, %bio->bi_status is not updated to relfect that. When some blocks in a file have been corrupted on disk, reading such a file ends up with 1) checksum errors are reported in kernel log 2) read(2) returns successfully with some content being 0x01. In order to fix it, we need to report its checksum mismatch error to the upper layer (dio layer in this case) as well. Fixes: 4246a0b63bd8 ("block: add a bi_error field to struct bio") Signed-off-by: Liu Bo Reported-by: Goffredo Baroncelli Tested-by: Goffredo Baroncelli Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/inode.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f78c5640c6dc..c242d0230db9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -8366,11 +8366,8 @@ static void btrfs_endio_direct_read(struct bio *bio) struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); blk_status_t err = bio->bi_status; - if (dip->flags & BTRFS_DIO_ORIG_BIO_SUBMITTED) { + if (dip->flags & BTRFS_DIO_ORIG_BIO_SUBMITTED) err = btrfs_subio_endio_read(inode, io_bio, err); - if (!err) - bio->bi_status = 0; - } unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset, dip->logical_offset + dip->bytes - 1); @@ -8378,7 +8375,7 @@ static void btrfs_endio_direct_read(struct bio *bio) kfree(dip); - dio_bio->bi_status = bio->bi_status; + dio_bio->bi_status = err; dio_end_io(dio_bio); if (io_bio->end_io) -- cgit v1.2.3 From 8c6c592831a09a28428448e68fb08c6bbb8b9b8b Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 29 Aug 2017 10:11:39 -0400 Subject: btrfs: log csums for all modified extents Amir reported a bug discovered by his cleaned up version of my dm-log-writes xfstests where we were missing csums at certain replay points. This is because fsx was doing an msync(), which essentially fsync()'s a specific range of a file. We will log all modified extents, but only search for the checksums in the range we are being asked to sync. We cannot simply log the extents in the range we're being asked because we are logging the inode item as it is currently, which if it has had a i_size update before the msync means we will miss extents when replaying. We could possibly get around this by marking the inode with the transaction that extended the i_size to see if we have this case, but this would be racy and we'd have to lock the whole range of the inode to make sure we didn't have an ordered extent outside of our range that was in the middle of completing. Fix this simply by keeping track of the modified extents range and logging the csums for the entire range of extents that we are logging. This makes the xfstest pass. Reported-by: Amir Goldstein Signed-off-by: Josef Bacik Signed-off-by: David Sterba --- fs/btrfs/tree-log.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index ad7f4bab640b..c800d067fcbf 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -4181,6 +4181,7 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, struct extent_map *em, *n; struct list_head extents; struct extent_map_tree *tree = &inode->extent_tree; + u64 logged_start, logged_end; u64 test_gen; int ret = 0; int num = 0; @@ -4190,10 +4191,11 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, down_write(&inode->dio_sem); write_lock(&tree->lock); test_gen = root->fs_info->last_trans_committed; + logged_start = start; + logged_end = end; list_for_each_entry_safe(em, n, &tree->modified_extents, list) { list_del_init(&em->list); - /* * Just an arbitrary number, this can be really CPU intensive * once we start getting a lot of extents, and really once we @@ -4208,6 +4210,12 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, if (em->generation <= test_gen) continue; + + if (em->start < logged_start) + logged_start = em->start; + if ((em->start + em->len - 1) > logged_end) + logged_end = em->start + em->len - 1; + /* Need a ref to keep it from getting evicted from cache */ refcount_inc(&em->refs); set_bit(EXTENT_FLAG_LOGGING, &em->flags); @@ -4216,7 +4224,7 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans, } list_sort(NULL, &extents, extent_cmp); - btrfs_get_logged_extents(inode, logged_list, start, end); + btrfs_get_logged_extents(inode, logged_list, logged_start, logged_end); /* * Some ordered extents started by fsync might have completed * before we could collect them into the list logged_list, which -- cgit v1.2.3 From 6851a3db7e224bbb85e23b3c64a506c9e0904382 Mon Sep 17 00:00:00 2001 From: Ross Zwisler Date: Mon, 18 Sep 2017 14:46:03 -0700 Subject: xfs: validate bdev support for DAX inode flag Currently only the blocksize is checked, but we should really be calling bdev_dax_supported() which also tests to make sure we can get a struct dax_device and that the dax_direct_access() path is working. This is the same check that we do for the "-o dax" mount option in xfs_fs_fill_super(). This does not fix the race issues that caused the XFS DAX inode option to be disabled, so that option will still be disabled. If/when we re-enable it, though, I think we will want this issue to have been fixed. I also do think that we want to fix this in stable kernels. Signed-off-by: Ross Zwisler CC: stable@vger.kernel.org Reviewed-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_ioctl.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index 5049e8ab6e30..aa75389be8cf 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -1088,6 +1088,7 @@ xfs_ioctl_setattr_dax_invalidate( int *join_flags) { struct inode *inode = VFS_I(ip); + struct super_block *sb = inode->i_sb; int error; *join_flags = 0; @@ -1100,7 +1101,7 @@ xfs_ioctl_setattr_dax_invalidate( if (fa->fsx_xflags & FS_XFLAG_DAX) { if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode))) return -EINVAL; - if (ip->i_mount->m_sb.sb_blocksize != PAGE_SIZE) + if (bdev_dax_supported(sb, sb->s_blocksize) < 0) return -EINVAL; } -- cgit v1.2.3 From 546e7be8244dc050effef0555df5b8d94d10dafc Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Fri, 22 Sep 2017 11:47:33 -0700 Subject: iomap_dio_rw: Allocate AIO completion queue before submitting dio Executing xfs/104 test in a loop on Linux-v4.13 kernel on a ppc64 machine can cause the following NULL pointer dereference, .queue_work_on+0x4c/0x80 .iomap_dio_bio_end_io+0xbc/0x1f0 .bio_endio+0x118/0x1f0 .blk_update_request+0xd0/0x470 .blk_mq_end_request+0x24/0xc0 .lo_complete_rq+0x40/0xe0 .__blk_mq_complete_request_remote+0x28/0x40 .flush_smp_call_function_queue+0xc4/0x1e0 .smp_ipi_demux_relaxed+0x8c/0x100 .icp_hv_ipi_action+0x54/0xa0 .__handle_irq_event_percpu+0x84/0x2c0 .handle_irq_event_percpu+0x28/0x80 .handle_percpu_irq+0x78/0xc0 .generic_handle_irq+0x40/0x70 .__do_irq+0x88/0x200 .call_do_irq+0x14/0x24 .do_IRQ+0x84/0x130 This occurs due to the following sequence of events, 1. Allocate dio for Direct I/O write. 2. Invoke iomap_apply() until iov_iter_count() bytes have been submitted. - Assume that we have submitted atleast one bio. Hence iomap_dio->ref value will be >= 2. - If during the second iteration, iomap_apply() ends up returning -ENOSPC, we would break out of the loop and since the 'ret' value is a negative number we end up not allocating memory for super_block->s_dio_done_wq. 3. Meanwhile, iomap_dio_bio_end_io() is invoked for bios that have been submitted and here the code ends up dereferencing the NULL pointer stored at super_block->s_dio_done_wq. This commit fixes the bug by allocating memory for super_block->s_dio_done_wq before iomap_apply() is invoked. Reported-by: Eryu Guan Reviewed-by: Christoph Hellwig Tested-by: Eryu Guan Signed-off-by: Chandan Rajendra Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/iomap.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/iomap.c b/fs/iomap.c index 269b24a01f32..d4f526a3f5b2 100644 --- a/fs/iomap.c +++ b/fs/iomap.c @@ -993,6 +993,13 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, WARN_ON_ONCE(ret); ret = 0; + if (iov_iter_rw(iter) == WRITE && !is_sync_kiocb(iocb) && + !inode->i_sb->s_dio_done_wq) { + ret = sb_init_dio_done_wq(inode->i_sb); + if (ret < 0) + goto out_free_dio; + } + inode_dio_begin(inode); blk_start_plug(&plug); @@ -1015,13 +1022,6 @@ iomap_dio_rw(struct kiocb *iocb, struct iov_iter *iter, if (ret < 0) iomap_dio_set_error(dio, ret); - if (ret >= 0 && iov_iter_rw(iter) == WRITE && !is_sync_kiocb(iocb) && - !inode->i_sb->s_dio_done_wq) { - ret = sb_init_dio_done_wq(inode->i_sb); - if (ret < 0) - iomap_dio_set_error(dio, ret); - } - if (!atomic_dec_and_test(&dio->ref)) { if (!is_sync_kiocb(iocb)) return -EIOCBQUEUED; -- cgit v1.2.3 From ee70daaba82d70766d0723b743d9fdeb3b06102a Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Thu, 21 Sep 2017 11:26:18 -0700 Subject: xfs: update i_size after unwritten conversion in dio completion Since commit d531d91d6990 ("xfs: always use unwritten extents for direct I/O writes"), we start allocating unwritten extents for all direct writes to allow appending aio in XFS. But for dio writes that could extend file size we update the in-core inode size first, then convert the unwritten extents to real allocations at dio completion time in xfs_dio_write_end_io(). Thus a racing direct read could see the new i_size and find the unwritten extents first and read zeros instead of actual data, if the direct writer also takes a shared iolock. Fix it by updating the in-core inode size after the unwritten extent conversion. To do this, introduce a new boolean argument to xfs_iomap_write_unwritten() to tell if we want to update in-core i_size or not. Suggested-by: Brian Foster Reviewed-by: Brian Foster Signed-off-by: Eryu Guan Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_aops.c | 3 ++- fs/xfs/xfs_file.c | 33 +++++++++++++++++++-------------- fs/xfs/xfs_iomap.c | 7 +++++-- fs/xfs/xfs_iomap.h | 2 +- fs/xfs/xfs_pnfs.c | 2 +- 5 files changed, 28 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 29172609f2a3..f18e5932aec4 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -343,7 +343,8 @@ xfs_end_io( error = xfs_reflink_end_cow(ip, offset, size); break; case XFS_IO_UNWRITTEN: - error = xfs_iomap_write_unwritten(ip, offset, size); + /* writeback should never update isize */ + error = xfs_iomap_write_unwritten(ip, offset, size, false); break; default: ASSERT(!xfs_ioend_is_append(ioend) || ioend->io_append_trans); diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index 350b6d43ba23..309e26c9dddb 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -434,7 +434,6 @@ xfs_dio_write_end_io( struct inode *inode = file_inode(iocb->ki_filp); struct xfs_inode *ip = XFS_I(inode); loff_t offset = iocb->ki_pos; - bool update_size = false; int error = 0; trace_xfs_end_io_direct_write(ip, offset, size); @@ -445,6 +444,21 @@ xfs_dio_write_end_io( if (size <= 0) return size; + if (flags & IOMAP_DIO_COW) { + error = xfs_reflink_end_cow(ip, offset, size); + if (error) + return error; + } + + /* + * Unwritten conversion updates the in-core isize after extent + * conversion but before updating the on-disk size. Updating isize any + * earlier allows a racing dio read to find unwritten extents before + * they are converted. + */ + if (flags & IOMAP_DIO_UNWRITTEN) + return xfs_iomap_write_unwritten(ip, offset, size, true); + /* * We need to update the in-core inode size here so that we don't end up * with the on-disk inode size being outside the in-core inode size. We @@ -459,20 +473,11 @@ xfs_dio_write_end_io( spin_lock(&ip->i_flags_lock); if (offset + size > i_size_read(inode)) { i_size_write(inode, offset + size); - update_size = true; - } - spin_unlock(&ip->i_flags_lock); - - if (flags & IOMAP_DIO_COW) { - error = xfs_reflink_end_cow(ip, offset, size); - if (error) - return error; - } - - if (flags & IOMAP_DIO_UNWRITTEN) - error = xfs_iomap_write_unwritten(ip, offset, size); - else if (update_size) + spin_unlock(&ip->i_flags_lock); error = xfs_setfilesize(ip, offset, size); + } else { + spin_unlock(&ip->i_flags_lock); + } return error; } diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index a1909bc064e9..f179bdf1644d 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -829,7 +829,8 @@ int xfs_iomap_write_unwritten( xfs_inode_t *ip, xfs_off_t offset, - xfs_off_t count) + xfs_off_t count, + bool update_isize) { xfs_mount_t *mp = ip->i_mount; xfs_fileoff_t offset_fsb; @@ -840,6 +841,7 @@ xfs_iomap_write_unwritten( xfs_trans_t *tp; xfs_bmbt_irec_t imap; struct xfs_defer_ops dfops; + struct inode *inode = VFS_I(ip); xfs_fsize_t i_size; uint resblks; int error; @@ -899,7 +901,8 @@ xfs_iomap_write_unwritten( i_size = XFS_FSB_TO_B(mp, offset_fsb + count_fsb); if (i_size > offset + count) i_size = offset + count; - + if (update_isize && i_size > i_size_read(inode)) + i_size_write(inode, i_size); i_size = xfs_new_eof(ip, i_size); if (i_size) { ip->i_d.di_size = i_size; diff --git a/fs/xfs/xfs_iomap.h b/fs/xfs/xfs_iomap.h index 00db3ecea084..ee535065c5d0 100644 --- a/fs/xfs/xfs_iomap.h +++ b/fs/xfs/xfs_iomap.h @@ -27,7 +27,7 @@ int xfs_iomap_write_direct(struct xfs_inode *, xfs_off_t, size_t, struct xfs_bmbt_irec *, int); int xfs_iomap_write_allocate(struct xfs_inode *, int, xfs_off_t, struct xfs_bmbt_irec *); -int xfs_iomap_write_unwritten(struct xfs_inode *, xfs_off_t, xfs_off_t); +int xfs_iomap_write_unwritten(struct xfs_inode *, xfs_off_t, xfs_off_t, bool); void xfs_bmbt_to_iomap(struct xfs_inode *, struct iomap *, struct xfs_bmbt_irec *); diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c index 2f2dc3c09ad0..4246876df7b7 100644 --- a/fs/xfs/xfs_pnfs.c +++ b/fs/xfs/xfs_pnfs.c @@ -274,7 +274,7 @@ xfs_fs_commit_blocks( (end - 1) >> PAGE_SHIFT); WARN_ON_ONCE(error); - error = xfs_iomap_write_unwritten(ip, start, length); + error = xfs_iomap_write_unwritten(ip, start, length, false); if (error) goto out_drop_iolock; } -- cgit v1.2.3 From 9789dd9e1d939232e8ff4c50ef8e75aa6781b3fb Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 18 Sep 2017 09:42:09 -0700 Subject: xfs: perag initialization should only touch m_ag_max_usable for AG 0 We call __xfs_ag_resv_init to make a per-AG reservation for each AG. This makes the reservation per-AG, not per-filesystem. Therefore, it is incorrect to adjust m_ag_max_usable for each AG. Adjust it only when we're reserving AG 0's blocks so that we only do it once per fs. Signed-off-by: Darrick J. Wong Reviewed-by: Brian Foster --- fs/xfs/libxfs/xfs_ag_resv.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c index b008ff3250eb..df3e600835e8 100644 --- a/fs/xfs/libxfs/xfs_ag_resv.c +++ b/fs/xfs/libxfs/xfs_ag_resv.c @@ -156,7 +156,8 @@ __xfs_ag_resv_free( trace_xfs_ag_resv_free(pag, type, 0); resv = xfs_perag_resv(pag, type); - pag->pag_mount->m_ag_max_usable += resv->ar_asked; + if (pag->pag_agno == 0) + pag->pag_mount->m_ag_max_usable += resv->ar_asked; /* * AGFL blocks are always considered "free", so whatever * was reserved at mount time must be given back at umount. @@ -216,7 +217,14 @@ __xfs_ag_resv_init( return error; } - mp->m_ag_max_usable -= ask; + /* + * Reduce the maximum per-AG allocation length by however much we're + * trying to reserve for an AG. Since this is a filesystem-wide + * counter, we only make the adjustment for AG 0. This assumes that + * there aren't any AGs hungrier for per-AG reservation than AG 0. + */ + if (pag->pag_agno == 0) + mp->m_ag_max_usable -= ask; resv = xfs_perag_resv(pag, type); resv->ar_asked = ask; -- cgit v1.2.3 From 842f6e9f786226c58fcbd5ef80eadca72fdfe652 Mon Sep 17 00:00:00 2001 From: Carlos Maiolino Date: Fri, 22 Sep 2017 11:47:46 -0700 Subject: xfs: Capture state of the right inode in xfs_iflush_done My previous patch: d3a304b6292168b83b45d624784f973fdc1ca674 check for XFS_LI_FAILED flag xfs_iflush done, so the failed item can be properly resubmitted. In the loop scanning other inodes being completed, it should check the current item for the XFS_LI_FAILED, and not the initial one. The state of the initial inode is checked after the loop ends Kudos to Eric for catching this. Signed-off-by: Carlos Maiolino Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_inode_item.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 6d0f74ec31e8..a705f34b58fa 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -745,7 +745,7 @@ xfs_iflush_done( */ iip = INODE_ITEM(blip); if ((iip->ili_logged && blip->li_lsn == iip->ili_flush_lsn) || - lip->li_flags & XFS_LI_FAILED) + (blip->li_flags & XFS_LI_FAILED)) need_ail++; blip = next; -- cgit v1.2.3 From 5e5c943c1f257c2b3424fc3f8a7b18570152dab3 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 18 Sep 2017 09:41:17 -0700 Subject: xfs: revert "xfs: factor rmap btree size into the indlen calculations" In commit fd26a88093ba we added a worst case estimate for rmapbt blocks needed to satisfy the block mapping request. Since then, we added the ability to reserve enough space in each AG such that we should never run out of blocks to grow the rmapbt, which makes this calculation unnecessary. Revert the commit because it makes the extra delalloc indlen accounting unnecessary and incorrect. Reported-by: Eryu Guan Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Darrick J. Wong --- fs/xfs/libxfs/xfs_bmap.c | 17 ++--------------- 1 file changed, 2 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index 459f4b4f08fe..044a363119be 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -49,7 +49,6 @@ #include "xfs_rmap.h" #include "xfs_ag_resv.h" #include "xfs_refcount.h" -#include "xfs_rmap_btree.h" #include "xfs_icache.h" @@ -192,12 +191,8 @@ xfs_bmap_worst_indlen( int maxrecs; /* maximum record count at this level */ xfs_mount_t *mp; /* mount structure */ xfs_filblks_t rval; /* return value */ - xfs_filblks_t orig_len; mp = ip->i_mount; - - /* Calculate the worst-case size of the bmbt. */ - orig_len = len; maxrecs = mp->m_bmap_dmxr[0]; for (level = 0, rval = 0; level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK); @@ -205,20 +200,12 @@ xfs_bmap_worst_indlen( len += maxrecs - 1; do_div(len, maxrecs); rval += len; - if (len == 1) { - rval += XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - + if (len == 1) + return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - level - 1; - break; - } if (level == 0) maxrecs = mp->m_bmap_dmxr[1]; } - - /* Calculate the worst-case size of the rmapbt. */ - if (xfs_sb_version_hasrmapbt(&mp->m_sb)) - rval += 1 + xfs_rmapbt_calc_size(mp, orig_len) + - mp->m_rmap_maxlevels; - return rval; } -- cgit v1.2.3 From fc46820b27a2d9a46f7e90c9ceb4a64a1bc5fab8 Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher Date: Mon, 25 Sep 2017 12:23:03 +0200 Subject: vfs: Return -ENXIO for negative SEEK_HOLE / SEEK_DATA offsets In generic_file_llseek_size, return -ENXIO for negative offsets as well as offsets beyond EOF. This affects filesystems which don't implement SEEK_HOLE / SEEK_DATA internally, possibly because they don't support holes. Fixes xfstest generic/448. Signed-off-by: Andreas Gruenbacher Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- fs/read_write.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/read_write.c b/fs/read_write.c index a2b9a47235c5..f0d4b16873e8 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -112,7 +112,7 @@ generic_file_llseek_size(struct file *file, loff_t offset, int whence, * In the generic case the entire file is data, so as long as * offset isn't at the end of the file then the offset is data. */ - if (offset >= eof) + if ((unsigned long long)offset >= eof) return -ENXIO; break; case SEEK_HOLE: @@ -120,7 +120,7 @@ generic_file_llseek_size(struct file *file, loff_t offset, int whence, * There is a virtual hole at the end of the file, so as long as * offset isn't i_size or larger, return i_size. */ - if (offset >= eof) + if ((unsigned long long)offset >= eof) return -ENXIO; offset = eof; break; -- cgit v1.2.3 From 4c6bb69663b3a3f2db8f488356e96acb5460f25f Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 26 Sep 2017 10:36:05 +0200 Subject: quota: Fix quota corruption with generic/232 test Eric has reported that since commit d2faa415166b "quota: Do not acquire dqio_sem for dquot overwrites in v2 format" test generic/232 occasionally fails due to quota information being incorrect. Indeed that commit was too eager to remove dqio_sem completely from the path that just overwrites quota structure with updated information. Although that is innocent on its own, another process that inserts new quota structure to the same block can perform read-modify-write cycle of that block thus effectively discarding quota information update if they race in a wrong way. Fix the problem by acquiring dqio_sem for reading for overwrites of quota structure. Note that it *is* possible to completely avoid taking dqio_sem in the overwrite path however that will require modifying path inserting / deleting quota structures to avoid RMW cycles of the full block and for now it is not clear whether it is worth the hassle. Fixes: d2faa415166b2883428efa92f451774ef44373ac Reported-and-tested-by: Eric Whitney Signed-off-by: Jan Kara --- fs/quota/quota_v2.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/quota/quota_v2.c b/fs/quota/quota_v2.c index c0187cda2c1e..a73e5b34db41 100644 --- a/fs/quota/quota_v2.c +++ b/fs/quota/quota_v2.c @@ -328,12 +328,16 @@ static int v2_write_dquot(struct dquot *dquot) if (!dquot->dq_off) { alloc = true; down_write(&dqopt->dqio_sem); + } else { + down_read(&dqopt->dqio_sem); } ret = qtree_write_dquot( sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv, dquot); if (alloc) up_write(&dqopt->dqio_sem); + else + up_read(&dqopt->dqio_sem); return ret; } -- cgit v1.2.3 From 1593baab910da72480d651ea7bf2ce6e3a25a484 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Sep 2017 18:09:26 +0200 Subject: sched/debug: Implement consistent task-state printing Currently get_task_state() and task_state_to_char() report different states, create a number of common helpers and unify the reported state space. Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar --- fs/proc/array.c | 15 ++------------- include/linux/sched.h | 26 +++++++++++++++++++------- 2 files changed, 21 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index 525157ca25cb..01196d3ad452 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -130,19 +130,8 @@ static const char * const task_state_array[] = { static inline const char *get_task_state(struct task_struct *tsk) { - unsigned int state = (tsk->state | tsk->exit_state) & TASK_REPORT; - - /* - * Parked tasks do not run; they sit in __kthread_parkme(). - * Without this check, we would report them as running, which is - * clearly wrong, so we report them as sleeping instead. - */ - if (tsk->state == TASK_PARKED) - state = TASK_INTERRUPTIBLE; - - BUILD_BUG_ON(1 + ilog2(TASK_REPORT) != ARRAY_SIZE(task_state_array)-1); - - return task_state_array[fls(state)]; + BUILD_BUG_ON(1 + ilog2(TASK_REPORT) != ARRAY_SIZE(task_state_array) - 1); + return task_state_array[__get_task_state(tsk)]; } static inline int get_task_umask(struct task_struct *tsk) diff --git a/include/linux/sched.h b/include/linux/sched.h index 92fb8dd5a9e4..163a0b738908 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1243,17 +1243,29 @@ static inline pid_t task_pgrp_nr(struct task_struct *tsk) return task_pgrp_nr_ns(tsk, &init_pid_ns); } -static inline char task_state_to_char(struct task_struct *task) +static inline unsigned int __get_task_state(struct task_struct *tsk) { - const char stat_nam[] = TASK_STATE_TO_CHAR_STR; - unsigned long state = task->state; + unsigned int tsk_state = READ_ONCE(tsk->state); + unsigned int state = (tsk_state | tsk->exit_state) & TASK_REPORT; - state = state ? __ffs(state) + 1 : 0; + if (tsk_state == TASK_PARKED) + state = TASK_INTERRUPTIBLE; - /* Make sure the string lines up properly with the number of task states: */ - BUILD_BUG_ON(sizeof(TASK_STATE_TO_CHAR_STR)-1 != ilog2(TASK_STATE_MAX)+1); + return fls(state); +} + +static inline char __task_state_to_char(unsigned int state) +{ + static const char state_char[] = "RSDTtXZ"; + + BUILD_BUG_ON(1 + ilog2(TASK_REPORT) != sizeof(state_char) - 2); - return state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?'; + return state_char[state]; +} + +static inline char task_state_to_char(struct task_struct *tsk) +{ + return __task_state_to_char(__get_task_state(tsk)); } /** -- cgit v1.2.3 From 06eb61844d841d0032a9950ce7f8e783ee49c0d0 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Sep 2017 18:30:40 +0200 Subject: sched/debug: Add explicit TASK_IDLE printing Markus reported that kthreads that idle using TASK_IDLE instead of TASK_INTERRUPTIBLE are reported in as TASK_UNINTERRUPTIBLE and things like htop mark those red. This is undesirable, so add an explicit state for TASK_IDLE. Reported-by: Markus Trippelsdorf Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar --- fs/proc/array.c | 21 +++++++++++++-------- include/linux/sched.h | 12 ++++++++++-- include/trace/events/sched.h | 7 ++++--- 3 files changed, 27 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index 01196d3ad452..a120a4549d48 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -119,18 +119,23 @@ static inline void task_name(struct seq_file *m, struct task_struct *p) * simple bit tests. */ static const char * const task_state_array[] = { - "R (running)", /* 0 */ - "S (sleeping)", /* 1 */ - "D (disk sleep)", /* 2 */ - "T (stopped)", /* 4 */ - "t (tracing stop)", /* 8 */ - "X (dead)", /* 16 */ - "Z (zombie)", /* 32 */ + + /* states in TASK_REPORT: */ + "R (running)", /* 0x00 */ + "S (sleeping)", /* 0x01 */ + "D (disk sleep)", /* 0x02 */ + "T (stopped)", /* 0x04 */ + "t (tracing stop)", /* 0x08 */ + "X (dead)", /* 0x10 */ + "Z (zombie)", /* 0x20 */ + + /* states beyond TASK_REPORT: */ + "I (idle)", /* 0x40 */ }; static inline const char *get_task_state(struct task_struct *tsk) { - BUILD_BUG_ON(1 + ilog2(TASK_REPORT) != ARRAY_SIZE(task_state_array) - 1); + BUILD_BUG_ON(1 + ilog2(TASK_REPORT_MAX) != ARRAY_SIZE(task_state_array)); return task_state_array[__get_task_state(tsk)]; } diff --git a/include/linux/sched.h b/include/linux/sched.h index bc7807933415..286fc1117046 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1241,22 +1241,30 @@ static inline pid_t task_pgrp_nr(struct task_struct *tsk) return task_pgrp_nr_ns(tsk, &init_pid_ns); } +#define TASK_REPORT_IDLE (TASK_REPORT + 1) +#define TASK_REPORT_MAX (TASK_REPORT_IDLE << 1) + static inline unsigned int __get_task_state(struct task_struct *tsk) { unsigned int tsk_state = READ_ONCE(tsk->state); unsigned int state = (tsk_state | tsk->exit_state) & TASK_REPORT; + BUILD_BUG_ON_NOT_POWER_OF_2(TASK_REPORT_MAX); + if (tsk_state == TASK_PARKED) state = TASK_INTERRUPTIBLE; + if (tsk_state == TASK_IDLE) + state = TASK_REPORT_IDLE; + return fls(state); } static inline char __task_state_to_char(unsigned int state) { - static const char state_char[] = "RSDTtXZ"; + static const char state_char[] = "RSDTtXZI"; - BUILD_BUG_ON(1 + ilog2(TASK_REPORT) != sizeof(state_char) - 2); + BUILD_BUG_ON(1 + ilog2(TASK_REPORT_MAX) != sizeof(state_char) - 1); return state_char[state]; } diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h index c63e20c9ef24..b371ef8206e1 100644 --- a/include/trace/events/sched.h +++ b/include/trace/events/sched.h @@ -156,10 +156,11 @@ TRACE_EVENT(sched_switch, TP_printk("prev_comm=%s prev_pid=%d prev_prio=%d prev_state=%s%s ==> next_comm=%s next_pid=%d next_prio=%d", __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, - (__entry->prev_state & TASK_REPORT) ? - __print_flags(__entry->prev_state & TASK_REPORT, "|", + (__entry->prev_state & (TASK_REPORT_MAX - 1)) ? + __print_flags(__entry->prev_state & (TASK_REPORT_MAX - 1), "|", { 0x01, "S" }, { 0x02, "D" }, { 0x04, "T" }, - { 0x08, "t" }, { 0x10, "X" }, { 0x20, "Z" }) : + { 0x08, "t" }, { 0x10, "X" }, { 0x20, "Z" }, + { 0x40, "I" }) : "R", __entry->prev_state & TASK_STATE_MAX ? "+" : "", -- cgit v1.2.3 From 8ef9925b02c23e3838d5e593c5cf37984141150f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 22 Sep 2017 18:37:28 +0200 Subject: sched/debug: Add explicit TASK_PARKED printing Currently TASK_PARKED is masqueraded as TASK_INTERRUPTIBLE, give it its own print state because it will not in fact get woken by regular wakeups and is a long-term state. This requires moving TASK_PARKED into the TASK_REPORT mask, and since that latter needs to be a contiguous bitmask, we need to shuffle the bits around a bit. Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar --- fs/proc/array.c | 3 ++- include/linux/sched.h | 16 +++++++--------- include/trace/events/sched.h | 2 +- 3 files changed, 10 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/proc/array.c b/fs/proc/array.c index a120a4549d48..77a8eacbe032 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -128,9 +128,10 @@ static const char * const task_state_array[] = { "t (tracing stop)", /* 0x08 */ "X (dead)", /* 0x10 */ "Z (zombie)", /* 0x20 */ + "P (parked)", /* 0x40 */ /* states beyond TASK_REPORT: */ - "I (idle)", /* 0x40 */ + "I (idle)", /* 0x80 */ }; static inline const char *get_task_state(struct task_struct *tsk) diff --git a/include/linux/sched.h b/include/linux/sched.h index 286fc1117046..26a7df4e558c 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -75,10 +75,10 @@ struct task_group; #define EXIT_ZOMBIE 0x0020 #define EXIT_TRACE (EXIT_ZOMBIE | EXIT_DEAD) /* Used in tsk->state again: */ -#define TASK_DEAD 0x0040 -#define TASK_WAKEKILL 0x0080 -#define TASK_WAKING 0x0100 -#define TASK_PARKED 0x0200 +#define TASK_PARKED 0x0040 +#define TASK_DEAD 0x0080 +#define TASK_WAKEKILL 0x0100 +#define TASK_WAKING 0x0200 #define TASK_NOLOAD 0x0400 #define TASK_NEW 0x0800 #define TASK_STATE_MAX 0x1000 @@ -97,7 +97,8 @@ struct task_group; /* get_task_state(): */ #define TASK_REPORT (TASK_RUNNING | TASK_INTERRUPTIBLE | \ TASK_UNINTERRUPTIBLE | __TASK_STOPPED | \ - __TASK_TRACED | EXIT_DEAD | EXIT_ZOMBIE) + __TASK_TRACED | EXIT_DEAD | EXIT_ZOMBIE | \ + TASK_PARKED) #define task_is_traced(task) ((task->state & __TASK_TRACED) != 0) @@ -1251,9 +1252,6 @@ static inline unsigned int __get_task_state(struct task_struct *tsk) BUILD_BUG_ON_NOT_POWER_OF_2(TASK_REPORT_MAX); - if (tsk_state == TASK_PARKED) - state = TASK_INTERRUPTIBLE; - if (tsk_state == TASK_IDLE) state = TASK_REPORT_IDLE; @@ -1262,7 +1260,7 @@ static inline unsigned int __get_task_state(struct task_struct *tsk) static inline char __task_state_to_char(unsigned int state) { - static const char state_char[] = "RSDTtXZI"; + static const char state_char[] = "RSDTtXZPI"; BUILD_BUG_ON(1 + ilog2(TASK_REPORT_MAX) != sizeof(state_char) - 1); diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h index b371ef8206e1..3c8b7f625670 100644 --- a/include/trace/events/sched.h +++ b/include/trace/events/sched.h @@ -160,7 +160,7 @@ TRACE_EVENT(sched_switch, __print_flags(__entry->prev_state & (TASK_REPORT_MAX - 1), "|", { 0x01, "S" }, { 0x02, "D" }, { 0x04, "T" }, { 0x08, "t" }, { 0x10, "X" }, { 0x20, "Z" }, - { 0x40, "I" }) : + { 0x40, "P" }, { 0x80, "I" }) : "R", __entry->prev_state & TASK_STATE_MAX ? "+" : "", -- cgit v1.2.3 From 9f4057fc937f200f000dbc378c5c3e37d45e31dc Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 22 Sep 2017 09:26:57 +0800 Subject: ceph: properly queue cap snap for newly created snap realm commit 3ae0bebc "ceph: queue cap snap only when snap realm's context changes" introduced a regression: we may not call queue_realm_cap_snaps() for newly created snap realm. This regression allows unflushed snapshot data to be overwritten. Link: http://tracker.ceph.com/issues/21483 Signed-off-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov --- fs/ceph/snap.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 1ffc8b426c1c..7fc0b850c352 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -374,12 +374,10 @@ static int build_snap_context(struct ceph_snap_realm *realm, realm->ino, realm, snapc, snapc->seq, (unsigned int) snapc->num_snaps); - if (realm->cached_context) { - ceph_put_snap_context(realm->cached_context); - /* queue realm for cap_snap creation */ - list_add_tail(&realm->dirty_item, dirty_realms); - } + ceph_put_snap_context(realm->cached_context); realm->cached_context = snapc; + /* queue realm for cap_snap creation */ + list_add_tail(&realm->dirty_item, dirty_realms); return 0; fail: -- cgit v1.2.3 From 38f340ccdf9ed5f1350505b46c5689d015967057 Mon Sep 17 00:00:00 2001 From: "Yan, Zheng" Date: Fri, 22 Sep 2017 11:41:06 +0800 Subject: ceph: fix __choose_mds() for LSSNAP request previous commit 5d37ca14 "ceph: send LSSNAP request to auth mds of directory inode" is buggy. It makes __choose_mds() choose mds base on hash of '.snap' dentry. Signed-off-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov --- fs/ceph/mds_client.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 84edfc60d87a..f23c820daaed 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -734,12 +734,13 @@ static int __choose_mds(struct ceph_mds_client *mdsc, inode = req->r_inode; ihold(inode); } else { - /* req->r_dentry is non-null for LSSNAP request. - * fall-thru */ - WARN_ON_ONCE(!req->r_dentry); + /* req->r_dentry is non-null for LSSNAP request */ + rcu_read_lock(); + inode = get_nonsnap_parent(req->r_dentry); + rcu_read_unlock(); + dout("__choose_mds using snapdir's parent %p\n", inode); } - } - if (!inode && req->r_dentry) { + } else if (req->r_dentry) { /* ignore race with rename; old or new d_parent is okay */ struct dentry *parent; struct inode *dir; -- cgit v1.2.3 From 384632e67e0829deb8015ee6ad916b180049d252 Mon Sep 17 00:00:00 2001 From: Andrea Arcangeli Date: Tue, 3 Oct 2017 16:15:38 -0700 Subject: userfaultfd: non-cooperative: fix fork use after free When reading the event from the uffd, we put it on a temporary fork_event list to detect if we can still access it after releasing and retaking the event_wqh.lock. If fork aborts and removes the event from the fork_event all is fine as long as we're still in the userfault read context and fork_event head is still alive. We've to put the event allocated in the fork kernel stack, back from fork_event list-head to the event_wqh head, before returning from userfaultfd_ctx_read, because the fork_event head lifetime is limited to the userfaultfd_ctx_read stack lifetime. Forgetting to move the event back to its event_wqh place then results in __remove_wait_queue(&ctx->event_wqh, &ewq->wq); in userfaultfd_event_wait_completion to remove it from a head that has been already freed from the reader stack. This could only happen if resolve_userfault_fork failed (for example if there are no file descriptors available to allocate the fork uffd). If it succeeded it was put back correctly. Furthermore, after find_userfault_evt receives a fork event, the forked userfault context in fork_nctx and uwq->msg.arg.reserved.reserved1 can be released by the fork thread as soon as the event_wqh.lock is released. Taking a reference on the fork_nctx before dropping the lock prevents an use after free in resolve_userfault_fork(). If the fork side aborted and it already released everything, we still try to succeed resolve_userfault_fork(), if possible. Fixes: 893e26e61d04eac9 ("userfaultfd: non-cooperative: Add fork() event") Link: http://lkml.kernel.org/r/20170920180413.26713-1-aarcange@redhat.com Signed-off-by: Andrea Arcangeli Reported-by: Mark Rutland Tested-by: Mark Rutland Cc: Pavel Emelyanov Cc: Mike Rapoport Cc: "Dr. David Alan Gilbert" Cc: Mike Kravetz Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/userfaultfd.c | 66 +++++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 56 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index ef4b48d1ea42..1c713fd5b3e6 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -588,6 +588,12 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx, break; if (ACCESS_ONCE(ctx->released) || fatal_signal_pending(current)) { + /* + * &ewq->wq may be queued in fork_event, but + * __remove_wait_queue ignores the head + * parameter. It would be a problem if it + * didn't. + */ __remove_wait_queue(&ctx->event_wqh, &ewq->wq); if (ewq->msg.event == UFFD_EVENT_FORK) { struct userfaultfd_ctx *new; @@ -1061,6 +1067,12 @@ static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait, (unsigned long) uwq->msg.arg.reserved.reserved1; list_move(&uwq->wq.entry, &fork_event); + /* + * fork_nctx can be freed as soon as + * we drop the lock, unless we take a + * reference on it. + */ + userfaultfd_ctx_get(fork_nctx); spin_unlock(&ctx->event_wqh.lock); ret = 0; break; @@ -1091,19 +1103,53 @@ static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait, if (!ret && msg->event == UFFD_EVENT_FORK) { ret = resolve_userfault_fork(ctx, fork_nctx, msg); + spin_lock(&ctx->event_wqh.lock); + if (!list_empty(&fork_event)) { + /* + * The fork thread didn't abort, so we can + * drop the temporary refcount. + */ + userfaultfd_ctx_put(fork_nctx); + + uwq = list_first_entry(&fork_event, + typeof(*uwq), + wq.entry); + /* + * If fork_event list wasn't empty and in turn + * the event wasn't already released by fork + * (the event is allocated on fork kernel + * stack), put the event back to its place in + * the event_wq. fork_event head will be freed + * as soon as we return so the event cannot + * stay queued there no matter the current + * "ret" value. + */ + list_del(&uwq->wq.entry); + __add_wait_queue(&ctx->event_wqh, &uwq->wq); - if (!ret) { - spin_lock(&ctx->event_wqh.lock); - if (!list_empty(&fork_event)) { - uwq = list_first_entry(&fork_event, - typeof(*uwq), - wq.entry); - list_del(&uwq->wq.entry); - __add_wait_queue(&ctx->event_wqh, &uwq->wq); + /* + * Leave the event in the waitqueue and report + * error to userland if we failed to resolve + * the userfault fork. + */ + if (likely(!ret)) userfaultfd_event_complete(ctx, uwq); - } - spin_unlock(&ctx->event_wqh.lock); + } else { + /* + * Here the fork thread aborted and the + * refcount from the fork thread on fork_nctx + * has already been released. We still hold + * the reference we took before releasing the + * lock above. If resolve_userfault_fork + * failed we've to drop it because the + * fork_nctx has to be freed in such case. If + * it succeeded we'll hold it because the new + * uffd references it. + */ + if (ret) + userfaultfd_ctx_put(fork_nctx); } + spin_unlock(&ctx->event_wqh.lock); } return ret; -- cgit v1.2.3 From c2315c187fa0d3ab363fdebe22718170b40473e3 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:42 -0700 Subject: exec: load_script: kill the onstack interp[BINPRM_BUF_SIZE] array Patch series "exec: binfmt_misc: fix use-after-free, kill iname[BINPRM_BUF_SIZE]". It looks like this code was always wrong, then commit 948b701a607f ("binfmt_misc: add persistent opened binary handler for containers") added more problems. This patch (of 6): load_script() can simply use i_name instead, it points into bprm->buf[] and nobody can change this memory until we call prepare_binprm(). The only complication is that we need to also change the signature of bprm_change_interp() but this change looks good too. While at it, do whitespace/style cleanups. NOTE: the real motivation for this change is that people want to increase BINPRM_BUF_SIZE, we need to change load_misc_binary() too but this looks more complicated because afaics it is very buggy. Link: http://lkml.kernel.org/r/20170918163446.GA26793@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Travis Gummels Cc: Ben Woodard Cc: Jim Foraker Cc: Cc: Al Viro Cc: James Bottomley Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_script.c | 17 +++++++++-------- fs/exec.c | 2 +- include/linux/binfmts.h | 2 +- 3 files changed, 11 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_script.c b/fs/binfmt_script.c index afdf4e3cafc2..7cde3f46ad26 100644 --- a/fs/binfmt_script.c +++ b/fs/binfmt_script.c @@ -19,7 +19,6 @@ static int load_script(struct linux_binprm *bprm) const char *i_arg, *i_name; char *cp; struct file *file; - char interp[BINPRM_BUF_SIZE]; int retval; if ((bprm->buf[0] != '#') || (bprm->buf[1] != '!')) @@ -55,7 +54,7 @@ static int load_script(struct linux_binprm *bprm) break; } for (cp = bprm->buf+2; (*cp == ' ') || (*cp == '\t'); cp++); - if (*cp == '\0') + if (*cp == '\0') return -ENOEXEC; /* No interpreter name found */ i_name = cp; i_arg = NULL; @@ -65,7 +64,6 @@ static int load_script(struct linux_binprm *bprm) *cp++ = '\0'; if (*cp) i_arg = cp; - strcpy (interp, i_name); /* * OK, we've parsed out the interpreter name and * (optional) argument. @@ -80,24 +78,27 @@ static int load_script(struct linux_binprm *bprm) if (retval) return retval; retval = copy_strings_kernel(1, &bprm->interp, bprm); - if (retval < 0) return retval; + if (retval < 0) + return retval; bprm->argc++; if (i_arg) { retval = copy_strings_kernel(1, &i_arg, bprm); - if (retval < 0) return retval; + if (retval < 0) + return retval; bprm->argc++; } retval = copy_strings_kernel(1, &i_name, bprm); - if (retval) return retval; + if (retval) + return retval; bprm->argc++; - retval = bprm_change_interp(interp, bprm); + retval = bprm_change_interp(i_name, bprm); if (retval < 0) return retval; /* * OK, now restart the process with the interpreter's dentry. */ - file = open_exec(interp); + file = open_exec(i_name); if (IS_ERR(file)) return PTR_ERR(file); diff --git a/fs/exec.c b/fs/exec.c index ac34d9724684..5470d3c1892a 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1410,7 +1410,7 @@ static void free_bprm(struct linux_binprm *bprm) kfree(bprm); } -int bprm_change_interp(char *interp, struct linux_binprm *bprm) +int bprm_change_interp(const char *interp, struct linux_binprm *bprm) { /* If a binfmt changed the interp, free it first. */ if (bprm->interp != bprm->filename) diff --git a/include/linux/binfmts.h b/include/linux/binfmts.h index fb44d6180ca0..18d05b5491f3 100644 --- a/include/linux/binfmts.h +++ b/include/linux/binfmts.h @@ -131,7 +131,7 @@ extern int setup_arg_pages(struct linux_binprm * bprm, 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 bprm_change_interp(const char *interp, struct linux_binprm *bprm); extern int copy_strings_kernel(int argc, const char *const *argv, struct linux_binprm *bprm); extern int prepare_bprm_creds(struct linux_binprm *bprm); -- cgit v1.2.3 From baba1b29731c79d605100087b8f02f9e1cf5a344 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:45 -0700 Subject: exec: binfmt_misc: don't nullify Node->dentry in kill_node() kill_node() nullifies/checks Node->dentry to avoid double free. This complicates the next changes and this is very confusing: - we do not need to check dentry != NULL under entries_lock, kill_node() is always called under inode_lock(d_inode(root)) and we rely on this inode_lock() anyway, without this lock the MISC_FMT_OPEN_FILE cleanup could race with itself. - if kill_inode() was already called and ->dentry == NULL we should not even try to close e->interp_file. We can change bm_entry_write() to simply check !list_empty(list) before kill_node. Again, we rely on inode_lock(), in particular it saves us from the race with bm_status_write(), another caller of kill_node(). Link: http://lkml.kernel.org/r/20170922143641.GA17210@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Al Viro Cc: Ben Woodard Cc: James Bottomley Cc: Jim Foraker Cc: Cc: Travis Gummels Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 22 +++++++++------------- 1 file changed, 9 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index ce7181ea60fa..6451e7520e05 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -603,11 +603,7 @@ static void kill_node(Node *e) struct dentry *dentry; write_lock(&entries_lock); - dentry = e->dentry; - if (dentry) { - list_del_init(&e->list); - e->dentry = NULL; - } + list_del_init(&e->list); write_unlock(&entries_lock); if ((e->flags & MISC_FMT_OPEN_FILE) && e->interp_file) { @@ -615,12 +611,11 @@ static void kill_node(Node *e) e->interp_file = NULL; } - if (dentry) { - drop_nlink(d_inode(dentry)); - d_drop(dentry); - dput(dentry); - simple_release_fs(&bm_mnt, &entry_count); - } + dentry = e->dentry; + drop_nlink(d_inode(dentry)); + d_drop(dentry); + dput(dentry); + simple_release_fs(&bm_mnt, &entry_count); } /* / */ @@ -665,7 +660,8 @@ static ssize_t bm_entry_write(struct file *file, const char __user *buffer, root = file_inode(file)->i_sb->s_root; inode_lock(d_inode(root)); - kill_node(e); + if (!list_empty(&e->list)) + kill_node(e); inode_unlock(d_inode(root)); break; @@ -794,7 +790,7 @@ static ssize_t bm_status_write(struct file *file, const char __user *buffer, inode_lock(d_inode(root)); while (!list_empty(&entries)) - kill_node(list_entry(entries.next, Node, list)); + kill_node(list_first_entry(&entries, Node, list)); inode_unlock(d_inode(root)); break; -- cgit v1.2.3 From 83f918274e4b841d6fb817861ea0c896fba0c179 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:48 -0700 Subject: exec: binfmt_misc: shift filp_close(interp_file) from kill_node() to bm_evict_inode() To ensure that load_misc_binary() can't use the partially destroyed Node, see also the next patch. The current logic looks wrong in any case, once we close interp_file it doesn't make any sense to delay kfree(inode->i_private), this Node is no longer valid. Even if the MISC_FMT_OPEN_FILE/interp_file checks were not racy (they are), load_misc_binary() should not try to reopen ->interpreter if MISC_FMT_OPEN_FILE is set but ->interp_file is NULL. And I can't understand why do we use filp_close(), not fput(). Link: http://lkml.kernel.org/r/20170922143644.GA17216@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Al Viro Cc: Ben Woodard Cc: James Bottomley Cc: Jim Foraker Cc: Cc: Travis Gummels Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index 6451e7520e05..203598ccb40a 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -594,8 +594,13 @@ static struct inode *bm_get_inode(struct super_block *sb, int mode) static void bm_evict_inode(struct inode *inode) { + Node *e = inode->i_private; + + if ((e->flags & MISC_FMT_OPEN_FILE) && e->interp_file) + filp_close(e->interp_file, NULL); + clear_inode(inode); - kfree(inode->i_private); + kfree(e); } static void kill_node(Node *e) @@ -606,11 +611,6 @@ static void kill_node(Node *e) list_del_init(&e->list); write_unlock(&entries_lock); - if ((e->flags & MISC_FMT_OPEN_FILE) && e->interp_file) { - filp_close(e->interp_file, NULL); - e->interp_file = NULL; - } - dentry = e->dentry; drop_nlink(d_inode(dentry)); d_drop(dentry); -- cgit v1.2.3 From eb23aa0317eb1f08e8d9d36b8753d42f03b32764 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:51 -0700 Subject: exec: binfmt_misc: remove the confusing e->interp_file != NULL checks If MISC_FMT_OPEN_FILE flag is set e->interp_file must be valid or we have a bug which should not be silently ignored. Link: http://lkml.kernel.org/r/20170922143647.GA17222@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Al Viro Cc: Ben Woodard Cc: James Bottomley Cc: Jim Foraker Cc: Cc: Travis Gummels Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index 203598ccb40a..babfe7bd56f0 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -205,7 +205,7 @@ static int load_misc_binary(struct linux_binprm *bprm) if (retval < 0) goto error; - if (fmt->flags & MISC_FMT_OPEN_FILE && fmt->interp_file) { + if (fmt->flags & MISC_FMT_OPEN_FILE) { interp_file = filp_clone_open(fmt->interp_file); if (!IS_ERR(interp_file)) deny_write_access(interp_file); @@ -596,7 +596,7 @@ static void bm_evict_inode(struct inode *inode) { Node *e = inode->i_private; - if ((e->flags & MISC_FMT_OPEN_FILE) && e->interp_file) + if (e->flags & MISC_FMT_OPEN_FILE) filp_close(e->interp_file, NULL); clear_inode(inode); -- cgit v1.2.3 From 43a4f2619038002f48c78698c42c05692d4b4eb2 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:55 -0700 Subject: exec: binfmt_misc: fix race between load_misc_binary() and kill_node() load_misc_binary() makes a local copy of fmt->interpreter under entries_lock to avoid the race with kill_node() but this is not enough; the whole Node can be freed after we drop entries_lock, not only the ->interpreter string. Add dget/dput(fmt->dentry) to ensure bm_evict_inode() can't destroy/free this Node. Link: http://lkml.kernel.org/r/20170922143650.GA17227@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Al Viro Cc: Ben Woodard Cc: James Bottomley Cc: Jim Foraker Cc: Travis Gummels Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index babfe7bd56f0..f5f8c2541790 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -138,20 +138,23 @@ static int load_misc_binary(struct linux_binprm *bprm) retval = -ENOEXEC; if (!enabled) - goto ret; + return retval; /* to keep locking time low, we copy the interpreter string */ read_lock(&entries_lock); fmt = check_file(bprm); - if (fmt) + if (fmt) { + dget(fmt->dentry); strlcpy(iname, fmt->interpreter, BINPRM_BUF_SIZE); + } read_unlock(&entries_lock); if (!fmt) - goto ret; + return retval; /* Need to be able to load the file after exec */ + retval = -ENOENT; if (bprm->interp_flags & BINPRM_FLAGS_PATH_INACCESSIBLE) - return -ENOENT; + goto ret; if (!(fmt->flags & MISC_FMT_PRESERVE_ARGV0)) { retval = remove_arg_zero(bprm); @@ -238,6 +241,7 @@ static int load_misc_binary(struct linux_binprm *bprm) goto error; ret: + dput(fmt->dentry); return retval; error: if (fd_binary > 0) -- cgit v1.2.3 From 50097f74934e3ec8fb1e6f3087568b958972817d Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Tue, 3 Oct 2017 16:15:58 -0700 Subject: exec: binfmt_misc: kill the onstack iname[BINPRM_BUF_SIZE] array After the previous change "fmt" can't go away, we can kill iname/iname_addr and use fmt->interpreter. Link: http://lkml.kernel.org/r/20170922143653.GA17232@redhat.com Signed-off-by: Oleg Nesterov Acked-by: Kees Cook Cc: Al Viro Cc: Ben Woodard Cc: James Bottomley Cc: Jim Foraker Cc: Cc: Travis Gummels Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/binfmt_misc.c | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index f5f8c2541790..2a46762def31 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -54,7 +54,7 @@ typedef struct { int size; /* size of magic/mask */ char *magic; /* magic or filename extension */ char *mask; /* mask, NULL for exact match */ - char *interpreter; /* filename of interpreter */ + const char *interpreter; /* filename of interpreter */ char *name; struct dentry *dentry; struct file *interp_file; @@ -131,8 +131,6 @@ static int load_misc_binary(struct linux_binprm *bprm) { Node *fmt; struct file *interp_file = NULL; - char iname[BINPRM_BUF_SIZE]; - const char *iname_addr = iname; int retval; int fd_binary = -1; @@ -143,10 +141,8 @@ static int load_misc_binary(struct linux_binprm *bprm) /* to keep locking time low, we copy the interpreter string */ read_lock(&entries_lock); fmt = check_file(bprm); - if (fmt) { + if (fmt) dget(fmt->dentry); - strlcpy(iname, fmt->interpreter, BINPRM_BUF_SIZE); - } read_unlock(&entries_lock); if (!fmt) return retval; @@ -198,13 +194,13 @@ static int load_misc_binary(struct linux_binprm *bprm) bprm->argc++; /* add the interp as argv[0] */ - retval = copy_strings_kernel(1, &iname_addr, bprm); + retval = copy_strings_kernel(1, &fmt->interpreter, bprm); if (retval < 0) goto error; bprm->argc++; /* Update interp in case binfmt_script needs it. */ - retval = bprm_change_interp(iname, bprm); + retval = bprm_change_interp(fmt->interpreter, bprm); if (retval < 0) goto error; @@ -213,7 +209,7 @@ static int load_misc_binary(struct linux_binprm *bprm) if (!IS_ERR(interp_file)) deny_write_access(interp_file); } else { - interp_file = open_exec(iname); + interp_file = open_exec(fmt->interpreter); } retval = PTR_ERR(interp_file); if (IS_ERR(interp_file)) -- cgit v1.2.3 From 52bfcdd7adbc26639bc7b2356ab9a3f5dad68ad6 Mon Sep 17 00:00:00 2001 From: "Darrick J. Wong" Date: Mon, 18 Sep 2017 09:41:18 -0700 Subject: xfs: always swap the cow forks when swapping extents Since the CoW fork exists as a secondary data structure to the data fork, we must always swap cow forks during swapext. We also need to swap the extent counts and reset the cowblocks tags. Reviewed-by: Brian Foster Reviewed-by: Christoph Hellwig Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_bmap_util.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index bc6c6e10a969..e9db7fc95b70 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -2122,11 +2122,31 @@ xfs_swap_extents( ip->i_d.di_flags2 |= tip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK; tip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; tip->i_d.di_flags2 |= f & XFS_DIFLAG2_REFLINK; + } + + /* Swap the cow forks. */ + if (xfs_sb_version_hasreflink(&mp->m_sb)) { + xfs_extnum_t extnum; + + ASSERT(ip->i_cformat == XFS_DINODE_FMT_EXTENTS); + ASSERT(tip->i_cformat == XFS_DINODE_FMT_EXTENTS); + + extnum = ip->i_cnextents; + ip->i_cnextents = tip->i_cnextents; + tip->i_cnextents = extnum; + cowfp = ip->i_cowfp; ip->i_cowfp = tip->i_cowfp; tip->i_cowfp = cowfp; - xfs_inode_set_cowblocks_tag(ip); - xfs_inode_set_cowblocks_tag(tip); + + if (ip->i_cowfp && ip->i_cnextents) + xfs_inode_set_cowblocks_tag(ip); + else + xfs_inode_clear_cowblocks_tag(ip); + if (tip->i_cowfp && tip->i_cnextents) + xfs_inode_set_cowblocks_tag(tip); + else + xfs_inode_clear_cowblocks_tag(tip); } xfs_trans_log_inode(tp, ip, src_log_flags); -- cgit v1.2.3 From e12199f85d0ad1b04ce6c425ad93cd847fe930bb Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 3 Oct 2017 08:58:33 -0700 Subject: xfs: handle racy AIO in xfs_reflink_end_cow If we got two AIO writes into a COW area the second one might not have any COW extents left to convert. Handle that case gracefully instead of triggering an assert or accessing beyond the bounds of the extent list. Signed-off-by: Christoph Hellwig Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong --- fs/xfs/xfs_reflink.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c index 3246815c24d6..37e603bf1591 100644 --- a/fs/xfs/xfs_reflink.c +++ b/fs/xfs/xfs_reflink.c @@ -736,7 +736,13 @@ xfs_reflink_end_cow( /* If there is a hole at end_fsb - 1 go to the previous extent */ if (!xfs_iext_lookup_extent(ip, ifp, end_fsb - 1, &idx, &got) || got.br_startoff > end_fsb) { - ASSERT(idx > 0); + /* + * In case of racing, overlapping AIO writes no COW extents + * might be left by the time I/O completes for the loser of + * the race. In that case we are done. + */ + if (idx <= 0) + goto out_cancel; xfs_iext_get_extent(ifp, --idx, &got); } @@ -809,6 +815,7 @@ next_extent: out_defer: xfs_defer_cancel(&dfops); +out_cancel: xfs_trans_cancel(tp); xfs_iunlock(ip, XFS_ILOCK_EXCL); out: -- cgit v1.2.3 From 57e7ba04d422c3d41c8426380303ec9b7533ded9 Mon Sep 17 00:00:00 2001 From: Casey Schaufler Date: Tue, 19 Sep 2017 09:39:08 -0700 Subject: lsm: fix smack_inode_removexattr and xattr_getsecurity memleak security_inode_getsecurity() provides the text string value of a security attribute. It does not provide a "secctx". The code in xattr_getsecurity() that calls security_inode_getsecurity() and then calls security_release_secctx() happened to work because SElinux and Smack treat the attribute and the secctx the same way. It fails for cap_inode_getsecurity(), because that module has no secctx that ever needs releasing. It turns out that Smack is the one that's doing things wrong by not allocating memory when instructed to do so by the "alloc" parameter. The fix is simple enough. Change the security_release_secctx() to kfree() because it isn't a secctx being returned by security_inode_getsecurity(). Change Smack to allocate the string when told to do so. Note: this also fixes memory leaks for LSMs which implement inode_getsecurity but not release_secctx, such as capabilities. Signed-off-by: Casey Schaufler Reported-by: Konstantin Khlebnikov Cc: stable@vger.kernel.org Signed-off-by: James Morris --- fs/xattr.c | 2 +- security/smack/smack_lsm.c | 55 +++++++++++++++++++++------------------------- 2 files changed, 26 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/xattr.c b/fs/xattr.c index 4424f7fecf14..61cd28ba25f3 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -250,7 +250,7 @@ xattr_getsecurity(struct inode *inode, const char *name, void *value, } memcpy(value, buffer, len); out: - security_release_secctx(buffer, len); + kfree(buffer); out_noalloc: return len; } diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c index 319add31b4a4..286171a16ed2 100644 --- a/security/smack/smack_lsm.c +++ b/security/smack/smack_lsm.c @@ -1473,7 +1473,7 @@ static int smack_inode_removexattr(struct dentry *dentry, const char *name) * @inode: the object * @name: attribute name * @buffer: where to put the result - * @alloc: unused + * @alloc: duplicate memory * * Returns the size of the attribute or an error code */ @@ -1486,43 +1486,38 @@ static int smack_inode_getsecurity(struct inode *inode, struct super_block *sbp; struct inode *ip = (struct inode *)inode; struct smack_known *isp; - int ilen; - int rc = 0; - if (strcmp(name, XATTR_SMACK_SUFFIX) == 0) { + if (strcmp(name, XATTR_SMACK_SUFFIX) == 0) isp = smk_of_inode(inode); - ilen = strlen(isp->smk_known); - *buffer = isp->smk_known; - return ilen; - } + else { + /* + * The rest of the Smack xattrs are only on sockets. + */ + sbp = ip->i_sb; + if (sbp->s_magic != SOCKFS_MAGIC) + return -EOPNOTSUPP; - /* - * The rest of the Smack xattrs are only on sockets. - */ - sbp = ip->i_sb; - if (sbp->s_magic != SOCKFS_MAGIC) - return -EOPNOTSUPP; + sock = SOCKET_I(ip); + if (sock == NULL || sock->sk == NULL) + return -EOPNOTSUPP; - sock = SOCKET_I(ip); - if (sock == NULL || sock->sk == NULL) - return -EOPNOTSUPP; - - ssp = sock->sk->sk_security; + ssp = sock->sk->sk_security; - if (strcmp(name, XATTR_SMACK_IPIN) == 0) - isp = ssp->smk_in; - else if (strcmp(name, XATTR_SMACK_IPOUT) == 0) - isp = ssp->smk_out; - else - return -EOPNOTSUPP; + if (strcmp(name, XATTR_SMACK_IPIN) == 0) + isp = ssp->smk_in; + else if (strcmp(name, XATTR_SMACK_IPOUT) == 0) + isp = ssp->smk_out; + else + return -EOPNOTSUPP; + } - ilen = strlen(isp->smk_known); - if (rc == 0) { - *buffer = isp->smk_known; - rc = ilen; + if (alloc) { + *buffer = kstrdup(isp->smk_known, GFP_KERNEL); + if (*buffer == NULL) + return -ENOMEM; } - return rc; + return strlen(isp->smk_known); } -- cgit v1.2.3 From 2d8ce70a08fe033c904115d59276ad86adeaa337 Mon Sep 17 00:00:00 2001 From: Goffredo Baroncelli Date: Tue, 3 Oct 2017 19:31:10 +0200 Subject: btrfs: avoid overflow when sector_t is 32 bit Jean-Denis Girard noticed commit c821e7f3 "pass bytes to btrfs_bio_alloc" (https://patchwork.kernel.org/patch/9763081/) introduces a regression on 32 bit machines. When CONFIG_LBDAF is _not_ defined (CONFIG_LBDAF == Support for large (2TB+) block devices and files) sector_t is 32 bit on 32bit machines. In the function submit_extent_page, 'sector' (which is sector_t type) is multiplied by 512 to convert it from sectors to bytes, leading to an overflow when the disk is bigger than 4GB (!). I added a cast to u64 to avoid overflow. Fixes: c821e7f3 ("btrfs: pass bytes to btrfs_bio_alloc") CC: stable@vger.kernel.org # 4.13+ Signed-off-by: Goffredo Baroncelli Tested-by: Jean-Denis Girard Reviewed-by: David Sterba Signed-off-by: David Sterba --- 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 3738d245518c..8dfa181b1a92 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2801,7 +2801,7 @@ static int submit_extent_page(unsigned int opf, struct extent_io_tree *tree, } } - bio = btrfs_bio_alloc(bdev, sector << 9); + bio = btrfs_bio_alloc(bdev, (u64)sector << 9); bio_add_page(bio, page, page_size, offset); bio->bi_end_io = end_io_func; bio->bi_private = tree; -- cgit v1.2.3 From 69ad59767d094752c23c0fc180a79532fde073d0 Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Wed, 4 Oct 2017 11:05:17 +0900 Subject: Btrfs: fix overlap of fs_info::flags values Because the values of BTRFS_FS_EXCL_OP and BTRFS_FS_QUOTA_OVERRIDE overlap, we should change the value. First, BTRFS_FS_EXCL_OP was set to 14. commit 171938e52807 ("btrfs: track exclusive filesystem operation in flags") Next, the value of BTRFS_FS_QUOTA_OVERRIDE was set to 14. commit f29efe292198 ("btrfs: add quota override flag to enable quota override for CAP_SYS_RESOURCE") As a result, the value 14 overlapped, by accident. This problem is solved by defining the value of BTRFS_FS_EXCL_OP as 16, the flags are internal. Fixes: f29efe292198 ("btrfs: add quota override flag to enable quota override for CAP_SYS_RESOURCE") CC: stable@vger.kernel.org # 4.13+ Signed-off-by: Tsutomu Itoh Reviewed-by: David Sterba [ minimize the change, update only BTRFS_FS_EXCL_OP ] Signed-off-by: David Sterba --- fs/btrfs/ctree.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index b7ccfcc01732..aff3248beb90 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -721,7 +721,7 @@ struct btrfs_delayed_root; * Indicate that a whole-filesystem exclusive operation is running * (device replace, resize, device add/delete, balance) */ -#define BTRFS_FS_EXCL_OP 14 +#define BTRFS_FS_EXCL_OP 16 struct btrfs_fs_info { u8 fsid[BTRFS_FSID_SIZE]; -- cgit v1.2.3 From 954c736f865d6c0c68ae4263a2f3502ee7c447a3 Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Mon, 18 Sep 2017 22:53:43 +0300 Subject: ovl: fix may_write_real() for overlayfs directories Overlayfs directory file_inode() is the overlay inode whether the real inode is upper or lower. This fixes a regression in xfstest generic/158. Fixes: 7c6893e3c9ab ("ovl: don't allow writing ioctl on lower layer") Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/namespace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 54059b142d6b..3b601f115b6c 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -468,7 +468,9 @@ static inline int may_write_real(struct file *file) /* File refers to upper, writable layer? */ upperdentry = d_real(dentry, NULL, 0, D_REAL_UPPER); - if (upperdentry && file_inode(file) == d_inode(upperdentry)) + if (upperdentry && + (file_inode(file) == d_inode(upperdentry) || + file_inode(file) == d_inode(dentry))) return 0; /* Lower layer: can't write to real file, sorry... */ -- cgit v1.2.3 From e0082a0f04c432cb6d7128ef60d8e425e45ce025 Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Sun, 24 Sep 2017 13:01:35 +0300 Subject: ovl: fix error value printed in ovl_lookup_index() Fixes: 359f392ca53e ("ovl: lookup index entry for copy up origin") Cc: # v4.13 Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/overlayfs/namei.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/overlayfs/namei.c b/fs/overlayfs/namei.c index c3addd1114f1..654bea1a5ac9 100644 --- a/fs/overlayfs/namei.c +++ b/fs/overlayfs/namei.c @@ -506,6 +506,7 @@ static struct dentry *ovl_lookup_index(struct dentry *dentry, index = lookup_one_len_unlocked(name.name, ofs->indexdir, name.len); if (IS_ERR(index)) { + err = PTR_ERR(index); pr_warn_ratelimited("overlayfs: failed inode index lookup (ino=%lu, key=%*s, err=%i);\n" "overlayfs: mount with '-o index=off' to disable inodes index.\n", d_inode(origin)->i_ino, name.len, name.name, -- cgit v1.2.3 From 9f4ec904dbd4eb1a2db10d5e7dc16eae386fe64d Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Sun, 24 Sep 2017 17:36:26 +0300 Subject: ovl: fix dput() of ERR_PTR in ovl_cleanup_index() Fixes: caf70cb2ba5d ("ovl: cleanup orphan index entries") Cc: # v4.13 Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/overlayfs/util.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c index 117794582f9f..7ae33d225a67 100644 --- a/fs/overlayfs/util.c +++ b/fs/overlayfs/util.c @@ -430,7 +430,7 @@ void ovl_inuse_unlock(struct dentry *dentry) } } -/* Called must hold OVL_I(inode)->oi_lock */ +/* Caller must hold OVL_I(inode)->lock */ static void ovl_cleanup_index(struct dentry *dentry) { struct inode *dir = ovl_indexdir(dentry->d_sb)->d_inode; @@ -469,6 +469,9 @@ static void ovl_cleanup_index(struct dentry *dentry) err = PTR_ERR(index); if (!IS_ERR(index)) err = ovl_cleanup(dir, index); + else + index = NULL; + inode_unlock(dir); if (err) goto fail; -- cgit v1.2.3 From dc7ab6773e8171e07f16fd0df0c5eea28c899503 Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Sun, 24 Sep 2017 22:19:10 +0300 Subject: ovl: fix dentry leak in ovl_indexdir_cleanup() index dentry was not released when breaking out of the loop due to index verification error. Fixes: 415543d5c64f ("ovl: cleanup bad and stale index entries on mount") Cc: # v4.13 Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/overlayfs/readdir.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c index 62e9b22a2077..0f85ee9c3268 100644 --- a/fs/overlayfs/readdir.c +++ b/fs/overlayfs/readdir.c @@ -988,6 +988,7 @@ int ovl_indexdir_cleanup(struct dentry *dentry, struct vfsmount *mnt, struct path *lowerstack, unsigned int numlower) { int err; + struct dentry *index = NULL; struct inode *dir = dentry->d_inode; struct path path = { .mnt = mnt, .dentry = dentry }; LIST_HEAD(list); @@ -1007,8 +1008,6 @@ int ovl_indexdir_cleanup(struct dentry *dentry, struct vfsmount *mnt, inode_lock_nested(dir, I_MUTEX_PARENT); list_for_each_entry(p, &list, l_node) { - struct dentry *index; - if (p->name[0] == '.') { if (p->len == 1) continue; @@ -1018,6 +1017,7 @@ int ovl_indexdir_cleanup(struct dentry *dentry, struct vfsmount *mnt, index = lookup_one_len(p->name, dentry, p->len); if (IS_ERR(index)) { err = PTR_ERR(index); + index = NULL; break; } err = ovl_verify_index(index, lowerstack, numlower); @@ -1029,7 +1029,9 @@ int ovl_indexdir_cleanup(struct dentry *dentry, struct vfsmount *mnt, break; } dput(index); + index = NULL; } + dput(index); inode_unlock(dir); out: ovl_cache_free(&list); -- cgit v1.2.3 From 5820dc0888d302ac05f8b91ffdf7e4e53b4fbf53 Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Mon, 25 Sep 2017 16:39:55 +0300 Subject: ovl: fix missing unlock_rename() in ovl_do_copy_up() Use the ovl_lock_rename_workdir() helper which requires unlock_rename() only on lock success. Fixes: ("fd210b7d67ee ovl: move copy up lock out") Cc: # v4.13 Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/overlayfs/copy_up.c | 6 ++---- fs/overlayfs/dir.c | 20 -------------------- fs/overlayfs/overlayfs.h | 1 + fs/overlayfs/util.c | 19 +++++++++++++++++++ 4 files changed, 22 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index aad97b30d5e6..c441f9387a1b 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -561,10 +561,8 @@ static int ovl_do_copy_up(struct ovl_copy_up_ctx *c) c->tmpfile = true; err = ovl_copy_up_locked(c); } else { - err = -EIO; - if (lock_rename(c->workdir, c->destdir) != NULL) { - pr_err("overlayfs: failed to lock workdir+upperdir\n"); - } else { + err = ovl_lock_rename_workdir(c->workdir, c->destdir); + if (!err) { err = ovl_copy_up_locked(c); unlock_rename(c->workdir, c->destdir); } diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c index 3309b1912241..cc961a3bd3bd 100644 --- a/fs/overlayfs/dir.c +++ b/fs/overlayfs/dir.c @@ -216,26 +216,6 @@ out_unlock: return err; } -static int ovl_lock_rename_workdir(struct dentry *workdir, - struct dentry *upperdir) -{ - /* Workdir should not be the same as upperdir */ - if (workdir == upperdir) - goto err; - - /* Workdir should not be subdir of upperdir and vice versa */ - if (lock_rename(workdir, upperdir) != NULL) - goto err_unlock; - - return 0; - -err_unlock: - unlock_rename(workdir, upperdir); -err: - pr_err("overlayfs: failed to lock workdir+upperdir\n"); - return -EIO; -} - static struct dentry *ovl_clear_empty(struct dentry *dentry, struct list_head *list) { diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h index d4e8c1a08fb0..c706a6f99928 100644 --- a/fs/overlayfs/overlayfs.h +++ b/fs/overlayfs/overlayfs.h @@ -235,6 +235,7 @@ bool ovl_inuse_trylock(struct dentry *dentry); void ovl_inuse_unlock(struct dentry *dentry); int ovl_nlink_start(struct dentry *dentry, bool *locked); void ovl_nlink_end(struct dentry *dentry, bool locked); +int ovl_lock_rename_workdir(struct dentry *workdir, struct dentry *upperdir); static inline bool ovl_is_impuredir(struct dentry *dentry) { diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c index 7ae33d225a67..b9b239fa5cfd 100644 --- a/fs/overlayfs/util.c +++ b/fs/overlayfs/util.c @@ -560,3 +560,22 @@ void ovl_nlink_end(struct dentry *dentry, bool locked) mutex_unlock(&OVL_I(d_inode(dentry))->lock); } } + +int ovl_lock_rename_workdir(struct dentry *workdir, struct dentry *upperdir) +{ + /* Workdir should not be the same as upperdir */ + if (workdir == upperdir) + goto err; + + /* Workdir should not be subdir of upperdir and vice versa */ + if (lock_rename(workdir, upperdir) != NULL) + goto err_unlock; + + return 0; + +err_unlock: + unlock_rename(workdir, upperdir); +err: + pr_err("overlayfs: failed to lock workdir+upperdir\n"); + return -EIO; +} -- cgit v1.2.3 From 85fdee1eef1a9e48ad5716916677e0c5fbc781e3 Mon Sep 17 00:00:00 2001 From: Amir Goldstein Date: Fri, 29 Sep 2017 10:21:21 +0300 Subject: ovl: fix regression caused by exclusive upper/work dir protection Enforcing exclusive ownership on upper/work dirs caused a docker regression: https://github.com/moby/moby/issues/34672. Euan spotted the regression and pointed to the offending commit. Vivek has brought the regression to my attention and provided this reproducer: Terminal 1: mount -t overlay -o workdir=work,lowerdir=lower,upperdir=upper none merged/ Terminal 2: unshare -m Terminal 1: umount merged mount -t overlay -o workdir=work,lowerdir=lower,upperdir=upper none merged/ mount: /root/overlay-testing/merged: none already mounted or mount point busy To fix the regression, I replaced the error with an alarming warning. With index feature enabled, mount does fail, but logs a suggestion to override exclusive dir protection by disabling index. Note that index=off mount does take the inuse locks, so a concurrent index=off will issue the warning and a concurrent index=on mount will fail. Documentation was updated to reflect this change. Fixes: 2cac0c00a6cd ("ovl: get exclusive ownership on upper/work dirs") Cc: # v4.13 Reported-by: Euan Kemp Reported-by: Vivek Goyal Signed-off-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- Documentation/filesystems/overlayfs.txt | 5 ++++- fs/overlayfs/ovl_entry.h | 3 +++ fs/overlayfs/super.c | 27 +++++++++++++++++++-------- 3 files changed, 26 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/Documentation/filesystems/overlayfs.txt b/Documentation/filesystems/overlayfs.txt index 36f528a7fdd6..8caa60734647 100644 --- a/Documentation/filesystems/overlayfs.txt +++ b/Documentation/filesystems/overlayfs.txt @@ -210,8 +210,11 @@ path as another overlay mount and it may use a lower layer path that is beneath or above the path of another overlay lower layer path. Using an upper layer path and/or a workdir path that are already used by -another overlay mount is not allowed and will fail with EBUSY. Using +another overlay mount is not allowed and may fail with EBUSY. Using partially overlapping paths is not allowed but will not fail with EBUSY. +If files are accessed from two overlayfs mounts which share or overlap the +upper layer and/or workdir path the behavior of the overlay is undefined, +though it will not result in a crash or deadlock. Mounting an overlay using an upper layer path, where the upper layer path was previously used by another mounted overlay in combination with a diff --git a/fs/overlayfs/ovl_entry.h b/fs/overlayfs/ovl_entry.h index 878a750986dd..25d9b5adcd42 100644 --- a/fs/overlayfs/ovl_entry.h +++ b/fs/overlayfs/ovl_entry.h @@ -37,6 +37,9 @@ struct ovl_fs { bool noxattr; /* sb common to all layers */ struct super_block *same_sb; + /* Did we take the inuse lock? */ + bool upperdir_locked; + bool workdir_locked; }; /* private information held for every overlayfs dentry */ diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c index fd5ea4facc62..092d150643c1 100644 --- a/fs/overlayfs/super.c +++ b/fs/overlayfs/super.c @@ -211,9 +211,10 @@ static void ovl_put_super(struct super_block *sb) dput(ufs->indexdir); dput(ufs->workdir); - ovl_inuse_unlock(ufs->workbasedir); + if (ufs->workdir_locked) + ovl_inuse_unlock(ufs->workbasedir); dput(ufs->workbasedir); - if (ufs->upper_mnt) + if (ufs->upper_mnt && ufs->upperdir_locked) ovl_inuse_unlock(ufs->upper_mnt->mnt_root); mntput(ufs->upper_mnt); for (i = 0; i < ufs->numlower; i++) @@ -881,9 +882,13 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent) goto out_put_upperpath; err = -EBUSY; - if (!ovl_inuse_trylock(upperpath.dentry)) { - pr_err("overlayfs: upperdir is in-use by another mount\n"); + if (ovl_inuse_trylock(upperpath.dentry)) { + ufs->upperdir_locked = true; + } else if (ufs->config.index) { + pr_err("overlayfs: upperdir is in-use by another mount, mount with '-o index=off' to override exclusive upperdir protection.\n"); goto out_put_upperpath; + } else { + pr_warn("overlayfs: upperdir is in-use by another mount, accessing files from both mounts will result in undefined behavior.\n"); } err = ovl_mount_dir(ufs->config.workdir, &workpath); @@ -901,9 +906,13 @@ static int ovl_fill_super(struct super_block *sb, void *data, int silent) } err = -EBUSY; - if (!ovl_inuse_trylock(workpath.dentry)) { - pr_err("overlayfs: workdir is in-use by another mount\n"); + if (ovl_inuse_trylock(workpath.dentry)) { + ufs->workdir_locked = true; + } else if (ufs->config.index) { + pr_err("overlayfs: workdir is in-use by another mount, mount with '-o index=off' to override exclusive workdir protection.\n"); goto out_put_workpath; + } else { + pr_warn("overlayfs: workdir is in-use by another mount, accessing files from both mounts will result in undefined behavior.\n"); } ufs->workbasedir = workpath.dentry; @@ -1156,11 +1165,13 @@ out_put_lowerpath: out_free_lowertmp: kfree(lowertmp); out_unlock_workdentry: - ovl_inuse_unlock(workpath.dentry); + if (ufs->workdir_locked) + ovl_inuse_unlock(workpath.dentry); out_put_workpath: path_put(&workpath); out_unlock_upperdentry: - ovl_inuse_unlock(upperpath.dentry); + if (ufs->upperdir_locked) + ovl_inuse_unlock(upperpath.dentry); out_put_upperpath: path_put(&upperpath); out_free_config: -- cgit v1.2.3