net: add annotations on hh->hh_len lockless accesses
Subsystems (3)
PATCH 1/1 net: add annotations on hh->hh_len lockless accesses
KCSAN reported a data-race [1]
While we can use READ_ONCE() on the read sides,
we need to make sure hh->hh_len is written last.
[1]
BUG: KCSAN: data-race in eth_header_cache / neigh_resolve_output
write to 0xffff8880b9dedcb8 of 4 bytes by task 29760 on cpu 0:
eth_header_cache+0xa9/0xd0 net/ethernet/eth.c:247 neigh_hh_init net/core/neighbour.c:1463 [inline] neigh_resolve_output net/core/neighbour.c:1480 [inline] neigh_resolve_output+0x415/0x470 net/core/neighbour.c:1470 neigh_output include/net/neighbour.h:511 [inline] ip6_finish_output2+0x7a2/0xec0 net/ipv6/ip6_output.c:116 __ip6_finish_output net/ipv6/ip6_output.c:142 [inline] __ip6_finish_output+0x2d7/0x330 net/ipv6/ip6_output.c:127 ip6_finish_output+0x41/0x160 net/ipv6/ip6_output.c:152 NF_HOOK_COND include/linux/netfilter.h:294 [inline] ip6_output+0xf2/0x280 net/ipv6/ip6_output.c:175 dst_output include/net/dst.h:436 [inline] NF_HOOK include/linux/netfilter.h:305 [inline] ndisc_send_skb+0x459/0x5f0 net/ipv6/ndisc.c:505 ndisc_send_ns+0x207/0x430 net/ipv6/ndisc.c:647 rt6_probe_deferred+0x98/0xf0 net/ipv6/route.c:615 process_one_work+0x3d4/0x890 kernel/workqueue.c:2269 worker_thread+0xa0/0x800 kernel/workqueue.c:2415 kthread+0x1d4/0x200 drivers/block/aoe/aoecmd.c:1253 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:352
read to 0xffff8880b9dedcb8 of 4 bytes by task 29572 on cpu 1:
neigh_resolve_output net/core/neighbour.c:1479 [inline] neigh_resolve_output+0x113/0x470 net/core/neighbour.c:1470 neigh_output include/net/neighbour.h:511 [inline] ip6_finish_output2+0x7a2/0xec0 net/ipv6/ip6_output.c:116 __ip6_finish_output net/ipv6/ip6_output.c:142 [inline] __ip6_finish_output+0x2d7/0x330 net/ipv6/ip6_output.c:127 ip6_finish_output+0x41/0x160 net/ipv6/ip6_output.c:152 NF_HOOK_COND include/linux/netfilter.h:294 [inline] ip6_output+0xf2/0x280 net/ipv6/ip6_output.c:175 dst_output include/net/dst.h:436 [inline] NF_HOOK include/linux/netfilter.h:305 [inline] ndisc_send_skb+0x459/0x5f0 net/ipv6/ndisc.c:505 ndisc_send_ns+0x207/0x430 net/ipv6/ndisc.c:647 rt6_probe_deferred+0x98/0xf0 net/ipv6/route.c:615 process_one_work+0x3d4/0x890 kernel/workqueue.c:2269 worker_thread+0xa0/0x800 kernel/workqueue.c:2415 kthread+0x1d4/0x200 drivers/block/aoe/aoecmd.c:1253 ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:352
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 29572 Comm: kworker/1:4 Not tainted 5.4.0-rc6+ #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Workqueue: events rt6_probe_deferred
Signed-off-by Eric Dumazet <[email protected]>
Reported-by syzbot <[email protected]>
--- drivers/firewire/net.c | 6 +++++- include/net/neighbour.h | 2 +- net/core/neighbour.c | 4 ++-- net/ethernet/eth.c | 7 ++++++- 4 files changed, 14 insertions(+), 5 deletions(-)
drivers/firewire/net.c +5 -1
@@ -250,7 +250,11 @@ static int fwnet_header_cache(const struct neighbour *neigh, h = (struct fwnet_header *)((u8 *)hh->hh_data +HH_DATA_OFF (sizeof (*h))); h->h_proto = type;memcpy (h->h_dest, neigh->ha, net->addr_len); - hh->hh_len =FWNET_HLEN ; + +/* Pairs with the READ_ONCE() in neigh_resolve_output(), +* neigh_hh_output() and neigh_update_hhs(). +*/ +smp_store_release (&hh->hh_len,FWNET_HLEN );return 0 ; }
include/net/neighbour.h +1 -1
@@ -468,7 +468,7 @@ static inline int neigh_hh_output(const struct hh_cache *hh, struct sk_buff *skb do { seq = read_seqbegin(&hh->hh_lock); - hh_len = hh->hh_len; + hh_len = READ_ONCE(hh->hh_len); if (likely(hh_len <= HH_DATA_MOD)) { hh_alen = HH_DATA_MOD;
net/core/neighbour.c +2 -2
@@ -1197,7 +1197,7 @@ static void neigh_update_hhs(struct neighbour *neigh)if (update) { hh = &neigh->hh; -if (hh->hh_len) { +if (READ_ONCE (hh->hh_len)) {write_seqlock_bh (&hh->hh_lock);update (hh, neigh->dev, neigh->ha);write_sequnlock_bh (&hh->hh_lock);
@@ -1476,7 +1476,7 @@ int neigh_resolve_output(struct neighbour *neigh, struct sk_buff *skb)struct net_device *dev = neigh->dev;unsigned int seq; -if (dev->header_ops->cache && !neigh->hh.hh_len) +if (dev->header_ops->cache && !READ_ONCE (neigh->hh.hh_len))neigh_hh_init (neigh); do {
net/ethernet/eth.c +6 -1
@@ -244,7 +244,12 @@ int eth_header_cache(const struct neighbour *neigh, struct hh_cache *hh, __be16 eth->h_proto = type;memcpy (eth->h_source, dev->dev_addr,ETH_ALEN );memcpy (eth->h_dest, neigh->ha,ETH_ALEN ); - hh->hh_len =ETH_HLEN ; + +/* Pairs with READ_ONCE() in neigh_resolve_output(), +* neigh_hh_output() and neigh_update_hhs(). +*/ +smp_store_release (&hh->hh_len,ETH_HLEN ); +return 0 ; }EXPORT_SYMBOL (eth_header_cache);
Signature · 1 line
2.24.0.432.g9d3f5f5b63-goog
From: Eric Dumazet <[email protected]>
Date: Thu, 7 Nov 2019 18:29:11 -0800
Quoted 8 lines
KCSAN reported a data-race [1]
While we can use
READ_ONCE()on the read sides,
we need to make surehh->hh_lenis written last.[1]
BUG: KCSAN: data-race in eth_header_cache / neigh_resolve_output
...
Quoted 2 lines
Signed-off-by Eric Dumazet <[email protected]>Reported-by syzbot <[email protected]>
Applied, thanks Eric.