diff mbox series

[net] tcp: do not mess with cloned skbs in tcp_add_backlog()

Message ID 20210119164900.766957-1-eric.dumazet@gmail.com (mailing list archive)
State Accepted
Commit b160c28548bc0a87cbd16d5af6d3edcfd70b8c9a
Delegated to: Netdev Maintainers
Headers show
Series [net] tcp: do not mess with cloned skbs in tcp_add_backlog() | expand

Checks

Context Check Description
netdev/cover_letter success Link
netdev/fixes_present success Link
netdev/patch_count success Link
netdev/tree_selection success Clearly marked for net
netdev/subject_prefix success Link
netdev/cc_maintainers warning 2 maintainers not CCed: yoshfuji@linux-ipv6.org ncardwell@google.com
netdev/source_inline success Was 0 now: 0
netdev/verify_signedoff success Link
netdev/module_param success Was 0 now: 0
netdev/build_32bit success Errors and warnings before: 15 this patch: 15
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/verify_fixes success Link
netdev/checkpatch warning WARNING: Non-standard signature: Bisected-by:
netdev/build_allmodconfig_warn success Errors and warnings before: 15 this patch: 15
netdev/header_inline success Link
netdev/stable success Stable not CCed

Commit Message

Eric Dumazet Jan. 19, 2021, 4:49 p.m. UTC 
From: Eric Dumazet <edumazet@google.com>

Heiner Kallweit reported that some skbs were sent with
the following invalid GSO properties :
- gso_size > 0
- gso_type == 0

This was triggerring a WARN_ON_ONCE() in rtl8169_tso_csum_v2.

Juerg Haefliger was able to reproduce a similar issue using
a lan78xx NIC and a workload mixing TCP incoming traffic
and forwarded packets.

The problem is that tcp_add_backlog() is writing
over gso_segs and gso_size even if the incoming packet will not
be coalesced to the backlog tail packet.

While skb_try_coalesce() would bail out if tail packet is cloned,
this overwriting would lead to corruptions of other packets
cooked by lan78xx, sharing a common super-packet.

The strategy used by lan78xx is to use a big skb, and split
it into all received packets using skb_clone() to avoid copies.
The drawback of this strategy is that all the small skb share a common
struct skb_shared_info.

This patch rewrites TCP gso_size/gso_segs handling to only
happen on the tail skb, since skb_try_coalesce() made sure
it was not cloned.

Fixes: 4f693b55c3d2 ("tcp: implement coalescing on backlog queue")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Bisected-by: Juerg Haefliger <juergh@canonical.com>
Tested-by: Juerg Haefliger <juergh@canonical.com>
Reported-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=209423
---
 net/ipv4/tcp_ipv4.c | 25 +++++++++++++------------
 1 file changed, 13 insertions(+), 12 deletions(-)

Comments

patchwork-bot+netdevbpf@kernel.org Jan. 20, 2021, 2:10 a.m. UTC | #1 
Hello:

This patch was applied to netdev/net.git (refs/heads/master):

On Tue, 19 Jan 2021 08:49:00 -0800 you wrote:
> From: Eric Dumazet <edumazet@google.com>
> 
> Heiner Kallweit reported that some skbs were sent with
> the following invalid GSO properties :
> - gso_size > 0
> - gso_type == 0
> 
> [...]

Here is the summary with links:
  - [net] tcp: do not mess with cloned skbs in tcp_add_backlog()
    https://git.kernel.org/netdev/net/c/b160c28548bc

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
diff mbox series

Patch

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 58207c7769d05693b650e3c93e4ef405a5d4b23a..4e82745d336fc3fb0d9ce8c92aaeb39702f64b8a 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1760,6 +1760,7 @@  int tcp_v4_early_demux(struct sk_buff *skb)
 bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 {
 	u32 limit = READ_ONCE(sk->sk_rcvbuf) + READ_ONCE(sk->sk_sndbuf);
+	u32 tail_gso_size, tail_gso_segs;
 	struct skb_shared_info *shinfo;
 	const struct tcphdr *th;
 	struct tcphdr *thtail;
@@ -1767,6 +1768,7 @@  bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 	unsigned int hdrlen;
 	bool fragstolen;
 	u32 gso_segs;
+	u32 gso_size;
 	int delta;
 
 	/* In case all data was pulled from skb frags (in __pskb_pull_tail()),
@@ -1792,13 +1794,6 @@  bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 	 */
 	th = (const struct tcphdr *)skb->data;
 	hdrlen = th->doff * 4;
-	shinfo = skb_shinfo(skb);
-
-	if (!shinfo->gso_size)
-		shinfo->gso_size = skb->len - hdrlen;
-
-	if (!shinfo->gso_segs)
-		shinfo->gso_segs = 1;
 
 	tail = sk->sk_backlog.tail;
 	if (!tail)
@@ -1821,6 +1816,15 @@  bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 		goto no_coalesce;
 
 	__skb_pull(skb, hdrlen);
+
+	shinfo = skb_shinfo(skb);
+	gso_size = shinfo->gso_size ?: skb->len;
+	gso_segs = shinfo->gso_segs ?: 1;
+
+	shinfo = skb_shinfo(tail);
+	tail_gso_size = shinfo->gso_size ?: (tail->len - hdrlen);
+	tail_gso_segs = shinfo->gso_segs ?: 1;
+
 	if (skb_try_coalesce(tail, skb, &fragstolen, &delta)) {
 		TCP_SKB_CB(tail)->end_seq = TCP_SKB_CB(skb)->end_seq;
 
@@ -1847,11 +1851,8 @@  bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 		}
 
 		/* Not as strict as GRO. We only need to carry mss max value */
-		skb_shinfo(tail)->gso_size = max(shinfo->gso_size,
-						 skb_shinfo(tail)->gso_size);
-
-		gso_segs = skb_shinfo(tail)->gso_segs + shinfo->gso_segs;
-		skb_shinfo(tail)->gso_segs = min_t(u32, gso_segs, 0xFFFF);
+		shinfo->gso_size = max(gso_size, tail_gso_size);
+		shinfo->gso_segs = min_t(u32, gso_segs + tail_gso_segs, 0xFFFF);
 
 		sk->sk_backlog.len += delta;
 		__NET_INC_STATS(sock_net(sk),