| Message ID | d47d53e6f8ee7a11228ca2f025d6243cc04b77f3.1689691004.git.pabeni@redhat.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 0558e1674598ec9029c1d3bceb787c8340272b51 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [v2,net-next] udp: use indirect call wrapper for data ready() | expand |
Paolo Abeni wrote: > In most cases UDP sockets use the default data ready callback. > Leverage the indirect call wrapper for such callback to avoid an > indirect call in fastpath. > > The above gives small but measurable performance gain under UDP flood. > > Signed-off-by: Paolo Abeni <pabeni@redhat.com> Reviewed-by: Willem de Bruijn <willemb@google.com> > --- > v1 -> v2: > - do not introduce the specific helper (Willem) I was just about to Ack v1. Did not mean to request a respin if no one else spoke up. But thanks for humoring me :) > --- > net/ipv4/udp.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c > index 42a96b3547c9..8c3ebd95f5b9 100644 > --- a/net/ipv4/udp.c > +++ b/net/ipv4/udp.c > @@ -1553,7 +1553,7 @@ int __udp_enqueue_schedule_skb(struct sock *sk, struct sk_buff *skb) > spin_unlock(&list->lock); > > if (!sock_flag(sk, SOCK_DEAD)) > - sk->sk_data_ready(sk); > + INDIRECT_CALL_1(sk->sk_data_ready, sock_def_readable, sk); > > busylock_release(busy); > return 0; > -- > 2.41.0 >
Hello: This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <kuba@kernel.org>: On Tue, 18 Jul 2023 16:38:09 +0200 you wrote: > In most cases UDP sockets use the default data ready callback. > Leverage the indirect call wrapper for such callback to avoid an > indirect call in fastpath. > > The above gives small but measurable performance gain under UDP flood. > > Signed-off-by: Paolo Abeni <pabeni@redhat.com> > > [...] Here is the summary with links: - [v2,net-next] udp: use indirect call wrapper for data ready() https://git.kernel.org/netdev/net-next/c/0558e1674598 You are awesome, thank you!
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 42a96b3547c9..8c3ebd95f5b9 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -1553,7 +1553,7 @@ int __udp_enqueue_schedule_skb(struct sock *sk, struct sk_buff *skb) spin_unlock(&list->lock); if (!sock_flag(sk, SOCK_DEAD)) - sk->sk_data_ready(sk); + INDIRECT_CALL_1(sk->sk_data_ready, sock_def_readable, sk); busylock_release(busy); return 0;
In most cases UDP sockets use the default data ready callback. Leverage the indirect call wrapper for such callback to avoid an indirect call in fastpath. The above gives small but measurable performance gain under UDP flood. Signed-off-by: Paolo Abeni <pabeni@redhat.com> --- v1 -> v2: - do not introduce the specific helper (Willem) --- net/ipv4/udp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)