| Message ID | 20240411082530.907113-1-edumazet@google.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | d13b05962369bade0fb8e2d55af56492f866cd36 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next] tcp: small optimization when TCP_TW_SYN is processed | expand |
Hello: This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <kuba@kernel.org>: On Thu, 11 Apr 2024 08:25:29 +0000 you wrote: > When TCP_TW_SYN is processed, we perform a lookup to find > a listener and jump back in tcp_v6_rcv() and tcp_v4_rcv() > > Paolo suggested that we do not have to check if the > found socket is a TIME_WAIT or NEW_SYN_RECV one. > > Suggested-by: Paolo Abeni <pabeni@redhat.com> > Link: https://lore.kernel.org/netdev/68085c8a84538cacaac991415e4ccc72f45e76c2.camel@redhat.com/ > Signed-off-by: Eric Dumazet <edumazet@google.com> > Cc: Neal Cardwell <ncardwell@google.com> > > [...] Here is the summary with links: - [net-next] tcp: small optimization when TCP_TW_SYN is processed https://git.kernel.org/netdev/net-next/c/d13b05962369 You are awesome, thank you!
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 1e650ec71d2fe5198b9dad9e6ea9c5eaf868277f..88c83ac4212957f19efad0f967952d2502bdbc7f 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2205,7 +2205,6 @@ int tcp_v4_rcv(struct sk_buff *skb) if (!sk) goto no_tcp_socket; -process: if (sk->sk_state == TCP_TIME_WAIT) goto do_time_wait; @@ -2285,6 +2284,7 @@ int tcp_v4_rcv(struct sk_buff *skb) } } +process: if (static_branch_unlikely(&ip4_min_ttl)) { /* min_ttl can be changed concurrently from do_ip_setsockopt() */ if (unlikely(iph->ttl < READ_ONCE(inet_sk(sk)->min_ttl))) { diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 3aa9da5c9a669d2754b421cfb704ad28def5a748..bb7c3caf4f8536dabdcb3dbe7c90aff9c8985c90 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -1794,7 +1794,6 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb) if (!sk) goto no_tcp_socket; -process: if (sk->sk_state == TCP_TIME_WAIT) goto do_time_wait; @@ -1871,6 +1870,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb) } } +process: if (static_branch_unlikely(&ip6_min_hopcount)) { /* min_hopcount can be changed concurrently from do_ipv6_setsockopt() */ if (unlikely(hdr->hop_limit < READ_ONCE(tcp_inet6_sk(sk)->min_hopcount))) {
When TCP_TW_SYN is processed, we perform a lookup to find a listener and jump back in tcp_v6_rcv() and tcp_v4_rcv() Paolo suggested that we do not have to check if the found socket is a TIME_WAIT or NEW_SYN_RECV one. Suggested-by: Paolo Abeni <pabeni@redhat.com> Link: https://lore.kernel.org/netdev/68085c8a84538cacaac991415e4ccc72f45e76c2.camel@redhat.com/ Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Neal Cardwell <ncardwell@google.com> --- net/ipv4/tcp_ipv4.c | 2 +- net/ipv6/tcp_ipv6.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)