[PATCH net-next] tcp: avoid dst->ops->check() call in tcp_v{4,6}_do_rcv()

If incoming skb dst matches the socket cached one,
there is no need to call again dst->ops->check().

Network layer already validated the skb dst for us,
usually from tcp_v{4,6}_early_demux().

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/ipv4/tcp_ipv4.c | 2 +-
 net/ipv6/tcp_ipv6.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 190e8a238876d966822531c74f17b896181e359f..b8e0562f68d7a1d02a5bb8c3b0aa0fe3b574d0dc 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c

@@ -1855,7 +1855,7 @@ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 
 		sock_rps_save_rxhash(sk, skb);
 		sk_mark_napi_id(sk, skb);
-		if (dst) {
+		if (dst && unlikely(dst != skb_dst(skb))) {
 			if (sk->sk_rx_dst_ifindex != skb->skb_iif ||
 			    !INDIRECT_CALL_1(dst->ops->check, ipv4_dst_check,
 					     dst, 0)) {

diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 164dceb842af767d93d0a2389564f543b2ba2fbf..074b83c9a551fafa52972a41506cdaf5299e24bd 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c

@@ -1596,7 +1596,7 @@ int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb)
 
 		sock_rps_save_rxhash(sk, skb);
 		sk_mark_napi_id(sk, skb);
-		if (dst) {
+		if (dst && unlikely(dst != skb_dst(skb))) {
 			if (sk->sk_rx_dst_ifindex != skb->skb_iif ||
 			    INDIRECT_CALL_1(dst->ops->check, ip6_dst_check,
 					    dst, sk->sk_rx_dst_cookie) == NULL) {

-- 
2.53.0.473.g4a7958ca14-goog

On Fri, Mar 6, 2026 at 7:43 AM Eric Dumazet [off-list ref] wrote:

If incoming skb dst matches the socket cached one,
there is no need to call again dst->ops->check().

Network layer already validated the skb dst for us,
usually from tcp_v{4,6}_early_demux().

Signed-off-by: Eric Dumazet <edumazet@google.com>

Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>

Hello:

This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski [off-list ref]:

On Fri,  6 Mar 2026 15:43:22 +0000 you wrote:

If incoming skb dst matches the socket cached one,
there is no need to call again dst->ops->check().

Network layer already validated the skb dst for us,
usually from tcp_v{4,6}_early_demux().

Signed-off-by: Eric Dumazet <edumazet@google.com>

[...]

Here is the summary with links:
  - [net-next] tcp: avoid dst->ops->check() call in tcp_v{4,6}_do_rcv()
    https://git.kernel.org/netdev/net-next/c/82f36517a13e

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