From 895b5c9f206eb7d25dc1360a8ccfc5958895eb89 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Sun, 29 Sep 2019 20:54:03 +0200 Subject: netfilter: drop bridge nf reset from nf_reset commit 174e23810cd31 ("sk_buff: drop all skb extensions on free and skb scrubbing") made napi recycle always drop skb extensions. The additional skb_ext_del() that is performed via nf_reset on napi skb recycle is not needed anymore. Most nf_reset() calls in the stack are there so queued skb won't block 'rmmod nf_conntrack' indefinitely. This removes the skb_ext_del from nf_reset, and renames it to a more fitting nf_reset_ct(). In a few selected places, add a call to skb_ext_reset to make sure that no active extensions remain. I am submitting this for "net", because we're still early in the release cycle. The patch applies to net-next too, but I think the rename causes needless divergence between those trees. Suggested-by: Eric Dumazet Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/core/skbuff.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/skbuff.c') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 01d65206f4fb..529133611ea2 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -5120,7 +5120,7 @@ void skb_scrub_packet(struct sk_buff *skb, bool xnet) skb->ignore_df = 0; skb_dst_drop(skb); skb_ext_reset(skb); - nf_reset(skb); + nf_reset_ct(skb); nf_reset_trace(skb); #ifdef CONFIG_NET_SWITCHDEV -- cgit v1.2.3 From ebb3b78db7bf842270a46fd4fe7cc45c78fa5ed6 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 10 Oct 2019 20:17:44 -0700 Subject: tcp: annotate sk->sk_rcvbuf lockless reads For the sake of tcp_poll(), there are few places where we fetch sk->sk_rcvbuf while this field can change from IRQ or other cpu. We need to add READ_ONCE() annotations, and also make sure write sides use corresponding WRITE_ONCE() to avoid store-tearing. Note that other transports probably need similar fixes. Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 4 ++-- include/trace/events/sock.h | 2 +- net/core/filter.c | 3 ++- net/core/skbuff.c | 2 +- net/core/sock.c | 5 +++-- net/ipv4/tcp.c | 4 ++-- net/ipv4/tcp_input.c | 7 ++++--- 7 files changed, 15 insertions(+), 12 deletions(-) (limited to 'net/core/skbuff.c') diff --git a/include/net/tcp.h b/include/net/tcp.h index e1d08f69fd39..ab4eb5eb5d07 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -1380,14 +1380,14 @@ static inline int tcp_win_from_space(const struct sock *sk, int space) /* Note: caller must be prepared to deal with negative returns */ static inline int tcp_space(const struct sock *sk) { - return tcp_win_from_space(sk, sk->sk_rcvbuf - + return tcp_win_from_space(sk, READ_ONCE(sk->sk_rcvbuf) - READ_ONCE(sk->sk_backlog.len) - atomic_read(&sk->sk_rmem_alloc)); } static inline int tcp_full_space(const struct sock *sk) { - return tcp_win_from_space(sk, sk->sk_rcvbuf); + return tcp_win_from_space(sk, READ_ONCE(sk->sk_rcvbuf)); } extern void tcp_openreq_init_rwin(struct request_sock *req, diff --git a/include/trace/events/sock.h b/include/trace/events/sock.h index a0c4b8a30966..f720c32e7dfd 100644 --- a/include/trace/events/sock.h +++ b/include/trace/events/sock.h @@ -82,7 +82,7 @@ TRACE_EVENT(sock_rcvqueue_full, TP_fast_assign( __entry->rmem_alloc = atomic_read(&sk->sk_rmem_alloc); __entry->truesize = skb->truesize; - __entry->sk_rcvbuf = sk->sk_rcvbuf; + __entry->sk_rcvbuf = READ_ONCE(sk->sk_rcvbuf); ), TP_printk("rmem_alloc=%d truesize=%u sk_rcvbuf=%d", diff --git a/net/core/filter.c b/net/core/filter.c index a50c0b6846f2..7deceaeeed7b 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -4252,7 +4252,8 @@ BPF_CALL_5(bpf_setsockopt, struct bpf_sock_ops_kern *, bpf_sock, case SO_RCVBUF: val = min_t(u32, val, sysctl_rmem_max); sk->sk_userlocks |= SOCK_RCVBUF_LOCK; - sk->sk_rcvbuf = max_t(int, val * 2, SOCK_MIN_RCVBUF); + WRITE_ONCE(sk->sk_rcvbuf, + max_t(int, val * 2, SOCK_MIN_RCVBUF)); break; case SO_SNDBUF: val = min_t(u32, val, sysctl_wmem_max); diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 529133611ea2..8c178703467b 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -4415,7 +4415,7 @@ static void skb_set_err_queue(struct sk_buff *skb) int sock_queue_err_skb(struct sock *sk, struct sk_buff *skb) { if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >= - (unsigned int)sk->sk_rcvbuf) + (unsigned int)READ_ONCE(sk->sk_rcvbuf)) return -ENOMEM; skb_orphan(skb); diff --git a/net/core/sock.c b/net/core/sock.c index 2a053999df11..8c8f61e70141 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -831,7 +831,8 @@ set_rcvbuf: * returning the value we actually used in getsockopt * is the most desirable behavior. */ - sk->sk_rcvbuf = max_t(int, val * 2, SOCK_MIN_RCVBUF); + WRITE_ONCE(sk->sk_rcvbuf, + max_t(int, val * 2, SOCK_MIN_RCVBUF)); break; case SO_RCVBUFFORCE: @@ -3204,7 +3205,7 @@ void sk_get_meminfo(const struct sock *sk, u32 *mem) memset(mem, 0, sizeof(*mem) * SK_MEMINFO_VARS); mem[SK_MEMINFO_RMEM_ALLOC] = sk_rmem_alloc_get(sk); - mem[SK_MEMINFO_RCVBUF] = sk->sk_rcvbuf; + mem[SK_MEMINFO_RCVBUF] = READ_ONCE(sk->sk_rcvbuf); mem[SK_MEMINFO_WMEM_ALLOC] = sk_wmem_alloc_get(sk); mem[SK_MEMINFO_SNDBUF] = sk->sk_sndbuf; mem[SK_MEMINFO_FWD_ALLOC] = sk->sk_forward_alloc; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 577a8c6eef9f..bc0481aa6633 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -451,7 +451,7 @@ void tcp_init_sock(struct sock *sk) icsk->icsk_sync_mss = tcp_sync_mss; sk->sk_sndbuf = sock_net(sk)->ipv4.sysctl_tcp_wmem[1]; - sk->sk_rcvbuf = sock_net(sk)->ipv4.sysctl_tcp_rmem[1]; + WRITE_ONCE(sk->sk_rcvbuf, sock_net(sk)->ipv4.sysctl_tcp_rmem[1]); sk_sockets_allocated_inc(sk); sk->sk_route_forced_caps = NETIF_F_GSO; @@ -1711,7 +1711,7 @@ int tcp_set_rcvlowat(struct sock *sk, int val) val <<= 1; if (val > sk->sk_rcvbuf) { - sk->sk_rcvbuf = val; + WRITE_ONCE(sk->sk_rcvbuf, val); tcp_sk(sk)->window_clamp = tcp_win_from_space(sk, val); } return 0; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 16342e043ab3..6995df20710a 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -483,8 +483,9 @@ static void tcp_clamp_window(struct sock *sk) !(sk->sk_userlocks & SOCK_RCVBUF_LOCK) && !tcp_under_memory_pressure(sk) && sk_memory_allocated(sk) < sk_prot_mem_limits(sk, 0)) { - sk->sk_rcvbuf = min(atomic_read(&sk->sk_rmem_alloc), - net->ipv4.sysctl_tcp_rmem[2]); + WRITE_ONCE(sk->sk_rcvbuf, + min(atomic_read(&sk->sk_rmem_alloc), + net->ipv4.sysctl_tcp_rmem[2])); } if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf) tp->rcv_ssthresh = min(tp->window_clamp, 2U * tp->advmss); @@ -648,7 +649,7 @@ void tcp_rcv_space_adjust(struct sock *sk) rcvbuf = min_t(u64, rcvwin * rcvmem, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); if (rcvbuf > sk->sk_rcvbuf) { - sk->sk_rcvbuf = rcvbuf; + WRITE_ONCE(sk->sk_rcvbuf, rcvbuf); /* Make the window clamp follow along. */ tp->window_clamp = tcp_win_from_space(sk, rcvbuf); -- cgit v1.2.3 From dedc5a08da07874c6e0d411e7f39c5c2cf137014 Mon Sep 17 00:00:00 2001 From: Davide Caratti Date: Sat, 12 Oct 2019 13:55:06 +0200 Subject: net: avoid errors when trying to pop MLPS header on non-MPLS packets the following script: # tc qdisc add dev eth0 clsact # tc filter add dev eth0 egress matchall action mpls pop implicitly makes the kernel drop all packets transmitted by eth0, if they don't have a MPLS header. This behavior is uncommon: other encapsulations (like VLAN) just let the packet pass unmodified. Since the result of MPLS 'pop' operation would be the same regardless of the presence / absence of MPLS header(s) in the original packet, we can let skb_mpls_pop() return 0 when dealing with non-MPLS packets. For the OVS use-case, this is acceptable because __ovs_nla_copy_actions() already ensures that MPLS 'pop' operation only occurs with packets having an MPLS Ethernet type (and there are no other callers in current code, so the semantic change should be ok). v2: better documentation of use-cases for skb_mpls_pop(), thanks to Simon Horman Fixes: 2a2ea50870ba ("net: sched: add mpls manipulation actions to TC") Reviewed-by: Simon Horman Acked-by: John Hurley Signed-off-by: Davide Caratti Signed-off-by: David S. Miller --- net/core/skbuff.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/skbuff.c') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 8c178703467b..03b6809ebde4 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -5536,7 +5536,7 @@ int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto) int err; if (unlikely(!eth_p_mpls(skb->protocol))) - return -EINVAL; + return 0; err = skb_ensure_writable(skb, skb->mac_len + MPLS_HLEN); if (unlikely(err)) -- cgit v1.2.3 From fa4e0f8855fcba600e0be2575ee29c69166f74bd Mon Sep 17 00:00:00 2001 From: Davide Caratti Date: Sat, 12 Oct 2019 13:55:07 +0200 Subject: net/sched: fix corrupted L2 header with MPLS 'push' and 'pop' actions the following script: # tc qdisc add dev eth0 clsact # tc filter add dev eth0 egress protocol ip matchall \ > action mpls push protocol mpls_uc label 0x355aa bos 1 causes corruption of all IP packets transmitted by eth0. On TC egress, we can't rely on the value of skb->mac_len, because it's 0 and a MPLS 'push' operation will result in an overwrite of the first 4 octets in the packet L2 header (e.g. the Destination Address if eth0 is an Ethernet); the same error pattern is present also in the MPLS 'pop' operation. Fix this error in act_mpls data plane, computing 'mac_len' as the difference between the network header and the mac header (when not at TC ingress), and use it in MPLS 'push'/'pop' core functions. v2: unbreak 'make htmldocs' because of missing documentation of 'mac_len' in skb_mpls_pop(), reported by kbuild test robot CC: Lorenzo Bianconi Fixes: 2a2ea50870ba ("net: sched: add mpls manipulation actions to TC") Reviewed-by: Simon Horman Acked-by: John Hurley Signed-off-by: Davide Caratti Signed-off-by: David S. Miller --- include/linux/skbuff.h | 5 +++-- net/core/skbuff.c | 19 +++++++++++-------- net/openvswitch/actions.c | 5 +++-- net/sched/act_mpls.c | 12 ++++++++---- 4 files changed, 25 insertions(+), 16 deletions(-) (limited to 'net/core/skbuff.c') diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 4351577b14d7..7914fdaf4226 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -3510,8 +3510,9 @@ int skb_ensure_writable(struct sk_buff *skb, int write_len); int __skb_vlan_pop(struct sk_buff *skb, u16 *vlan_tci); int skb_vlan_pop(struct sk_buff *skb); int skb_vlan_push(struct sk_buff *skb, __be16 vlan_proto, u16 vlan_tci); -int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto); -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto); +int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, + int mac_len); +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len); int skb_mpls_update_lse(struct sk_buff *skb, __be32 mpls_lse); int skb_mpls_dec_ttl(struct sk_buff *skb); struct sk_buff *pskb_extract(struct sk_buff *skb, int off, int to_copy, diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 03b6809ebde4..867e61df00db 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -5477,12 +5477,14 @@ static void skb_mod_eth_type(struct sk_buff *skb, struct ethhdr *hdr, * @skb: buffer * @mpls_lse: MPLS label stack entry to push * @mpls_proto: ethertype of the new MPLS header (expects 0x8847 or 0x8848) + * @mac_len: length of the MAC header * * Expects skb->data at mac header. * * Returns 0 on success, -errno otherwise. */ -int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto) +int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, + int mac_len) { struct mpls_shim_hdr *lse; int err; @@ -5499,15 +5501,15 @@ int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto) return err; if (!skb->inner_protocol) { - skb_set_inner_network_header(skb, skb->mac_len); + skb_set_inner_network_header(skb, mac_len); skb_set_inner_protocol(skb, skb->protocol); } skb_push(skb, MPLS_HLEN); memmove(skb_mac_header(skb) - MPLS_HLEN, skb_mac_header(skb), - skb->mac_len); + mac_len); skb_reset_mac_header(skb); - skb_set_network_header(skb, skb->mac_len); + skb_set_network_header(skb, mac_len); lse = mpls_hdr(skb); lse->label_stack_entry = mpls_lse; @@ -5526,29 +5528,30 @@ EXPORT_SYMBOL_GPL(skb_mpls_push); * * @skb: buffer * @next_proto: ethertype of header after popped MPLS header + * @mac_len: length of the MAC header * * Expects skb->data at mac header. * * Returns 0 on success, -errno otherwise. */ -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto) +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) { int err; if (unlikely(!eth_p_mpls(skb->protocol))) return 0; - err = skb_ensure_writable(skb, skb->mac_len + MPLS_HLEN); + err = skb_ensure_writable(skb, mac_len + MPLS_HLEN); if (unlikely(err)) return err; skb_postpull_rcsum(skb, mpls_hdr(skb), MPLS_HLEN); memmove(skb_mac_header(skb) + MPLS_HLEN, skb_mac_header(skb), - skb->mac_len); + mac_len); __skb_pull(skb, MPLS_HLEN); skb_reset_mac_header(skb); - skb_set_network_header(skb, skb->mac_len); + skb_set_network_header(skb, mac_len); if (skb->dev && skb->dev->type == ARPHRD_ETHER) { struct ethhdr *hdr; diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c index 3572e11b6f21..1c77f520f474 100644 --- a/net/openvswitch/actions.c +++ b/net/openvswitch/actions.c @@ -165,7 +165,8 @@ static int push_mpls(struct sk_buff *skb, struct sw_flow_key *key, { int err; - err = skb_mpls_push(skb, mpls->mpls_lse, mpls->mpls_ethertype); + err = skb_mpls_push(skb, mpls->mpls_lse, mpls->mpls_ethertype, + skb->mac_len); if (err) return err; @@ -178,7 +179,7 @@ static int pop_mpls(struct sk_buff *skb, struct sw_flow_key *key, { int err; - err = skb_mpls_pop(skb, ethertype); + err = skb_mpls_pop(skb, ethertype, skb->mac_len); if (err) return err; diff --git a/net/sched/act_mpls.c b/net/sched/act_mpls.c index e168df0e008a..4cf6c553bb0b 100644 --- a/net/sched/act_mpls.c +++ b/net/sched/act_mpls.c @@ -55,7 +55,7 @@ static int tcf_mpls_act(struct sk_buff *skb, const struct tc_action *a, struct tcf_mpls *m = to_mpls(a); struct tcf_mpls_params *p; __be32 new_lse; - int ret; + int ret, mac_len; tcf_lastuse_update(&m->tcf_tm); bstats_cpu_update(this_cpu_ptr(m->common.cpu_bstats), skb); @@ -63,8 +63,12 @@ static int tcf_mpls_act(struct sk_buff *skb, const struct tc_action *a, /* Ensure 'data' points at mac_header prior calling mpls manipulating * functions. */ - if (skb_at_tc_ingress(skb)) + if (skb_at_tc_ingress(skb)) { skb_push_rcsum(skb, skb->mac_len); + mac_len = skb->mac_len; + } else { + mac_len = skb_network_header(skb) - skb_mac_header(skb); + } ret = READ_ONCE(m->tcf_action); @@ -72,12 +76,12 @@ static int tcf_mpls_act(struct sk_buff *skb, const struct tc_action *a, switch (p->tcfm_action) { case TCA_MPLS_ACT_POP: - if (skb_mpls_pop(skb, p->tcfm_proto)) + if (skb_mpls_pop(skb, p->tcfm_proto, mac_len)) goto drop; break; case TCA_MPLS_ACT_PUSH: new_lse = tcf_mpls_get_lse(NULL, p, !eth_p_mpls(skb->protocol)); - if (skb_mpls_push(skb, new_lse, p->tcfm_proto)) + if (skb_mpls_push(skb, new_lse, p->tcfm_proto, mac_len)) goto drop; break; case TCA_MPLS_ACT_MODIFY: -- cgit v1.2.3