Skip to contentSkip to conversation

tcp: fix tcp_grow_skb() vs tstamps

PATCH Applied Eric Dumazet 3 participants 3 messages Apr 25, 2024 – Apr 26, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tcp: fix tcp_grow_skb() vs tstamps

I forgot to call tcp_skb_collapse_tstamp() in the case we consume the second skb in write queue.

Neal suggested to create a common helper used by tcp_mtu_probe() and tcp_grow_skb().

Fixes 8ee602c63520 ("tcp: try to send bigger TSO packets")
Cc Neal Cardwell <[email protected]>
---
 net/ipv4/tcp_output.c | 33 +++++++++++++++++++--------------
 1 file changed, 19 insertions(+), 14 deletions(-)
net/ipv4/tcp_output.c +19 -14
index ce59e4499b66dfd2fe00c255c11187e3c08c5806..ef25556c48139c1904ac3fe6e484676ba10ba1bf 100644
@@ -2403,6 +2403,21 @@ static int tcp_clone_payload(struct sock *sk, struct sk_buff *to,
 	return 0;
 }
 
+/* tcp_mtu_probe() and tcp_grow_skb() can both eat an skb (src) if
+ * all its payload was moved to another one (dst).
+ * Make sure to transfer tcp_flags, eor, and tstamp.
+ */
+static void tcp_eat_one_skb(struct sock *sk,
+			    struct sk_buff *dst,
+			    struct sk_buff *src)
+{
+	TCP_SKB_CB(dst)->tcp_flags |= TCP_SKB_CB(src)->tcp_flags;
+	TCP_SKB_CB(dst)->eor = TCP_SKB_CB(src)->eor;
+	tcp_skb_collapse_tstamp(dst, src);
+	tcp_unlink_write_queue(src, sk);
+	tcp_wmem_free_skb(sk, src);
+}
+
 /* Create a new MTU probe if we are ready.
  * MTU probe is regularly attempting to increase the path MTU by
  * deliberately sending larger packets.  This discovers routing
@@ -2508,16 +2523,7 @@ static int tcp_mtu_probe(struct sock *sk)
 		copy = min_t(int, skb->len, probe_size - len);
 
 		if (skb->len <= copy) {
-			/* We've eaten all the data from this skb.
-			 * Throw it away. */
-			TCP_SKB_CB(nskb)->tcp_flags |= TCP_SKB_CB(skb)->tcp_flags;
-			/* If this is the last SKB we copy and eor is set
-			 * we need to propagate it to the new skb.
-			 */
-			TCP_SKB_CB(nskb)->eor = TCP_SKB_CB(skb)->eor;
-			tcp_skb_collapse_tstamp(nskb, skb);
-			tcp_unlink_write_queue(skb, sk);
-			tcp_wmem_free_skb(sk, skb);
+			tcp_eat_one_skb(sk, nskb, skb);
 		} else {
 			TCP_SKB_CB(nskb)->tcp_flags |= TCP_SKB_CB(skb)->tcp_flags &
 						   ~(TCPHDR_FIN|TCPHDR_PSH);
@@ -2705,11 +2711,10 @@ static void tcp_grow_skb(struct sock *sk, struct sk_buff *skb, int amount)
 	TCP_SKB_CB(next_skb)->seq += nlen;
 
 	if (!next_skb->len) {
+		/* In case FIN is set, we need to update end_seq */
 		TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(next_skb)->end_seq;
-		TCP_SKB_CB(skb)->eor = TCP_SKB_CB(next_skb)->eor;
-		TCP_SKB_CB(skb)->tcp_flags |= TCP_SKB_CB(next_skb)->tcp_flags;
-		tcp_unlink_write_queue(next_skb, sk);
-		tcp_wmem_free_skb(sk, next_skb);
+
+		tcp_eat_one_skb(sk, skb, next_skb);
 	}
 }
 
Signature · 1 line
2.44.0.769.g3c40516874-goog
maintainer Acked-by on v1 1/1
#1 lore raw
Quoted 11 lines · Eric Dumazet

I forgot to call tcp_skb_collapse_tstamp() in the case we consume the second skb in write queue.

Neal suggested to create a common helper used by tcp_mtu_probe() and tcp_grow_skb().

Fixes 8ee602c63520 ("tcp: try to send bigger TSO packets")
Cc Neal Cardwell <[email protected]>

---

Great. Thanks, Eric!

neal

#2 lore raw
In reply to Eric Dumazet
This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <[email protected]>:

Quoted 11 lines · On Thu, 25 Apr 2024 19:34:50 +0000 you wrote:

I forgot to call tcp_skb_collapse_tstamp() in the case we consume the second skb in write queue.

Neal suggested to create a common helper used by tcp_mtu_probe() and tcp_grow_skb().

Fixes 8ee602c63520 ("tcp: try to send bigger TSO packets")
Cc Neal Cardwell <[email protected]>

[...]

Here is the summary with links:
- [net-next] tcp: fix tcp_grow_skb() vs tstamps

    https://git.kernel.org/netdev/net-next/c/1bede0a12d3a

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html