[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <OSZP286MB140479B6DBDB0F13651A55F9959F9@OSZP286MB1404.JPNP286.PROD.OUTLOOK.COM>
Date: Thu, 4 Aug 2022 23:04:21 +0800
From: gfree.wind@...look.com
To: martin.lau@...ux.dev, daniel@...earbox.net,
john.fastabend@...il.com, song@...nel.org, yhs@...com,
kpsingh@...nel.org, davem@...emloft.net, edumazet@...gle.com,
kuba@...nel.org, pabeni@...hat.com, bpf@...r.kernel.org,
netdev@...r.kernel.org
Cc: gfree.wind@...look.com, Gao Feng <gfree.wind@...il.com>
Subject: [PATCH net] net: bpf: Use the protocol's set_rcvlowat behavior if there is one
From: Gao Feng <gfree.wind@...il.com>
The commit d1361840f8c5 ("tcp: fix SO_RCVLOWAT and RCVBUF autotuning")
add one new (struct proto_ops)->set_rcvlowat method so that a protocol
can override the default setsockopt(SO_RCVLOWAT) behavior.
The prior bpf codes don't check and invoke the protos's set_rcvlowat,
now correct it.
Signed-off-by: Gao Feng <gfree.wind@...il.com>
---
net/core/filter.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/net/core/filter.c b/net/core/filter.c
index 7950f7520765..beb6209897ab 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -5064,7 +5064,10 @@ static int _bpf_setsockopt(struct sock *sk, int level, int optname,
case SO_RCVLOWAT:
if (val < 0)
val = INT_MAX;
- WRITE_ONCE(sk->sk_rcvlowat, val ? : 1);
+ if (sk->sk_socket && sk->sk_socket->ops->set_rcvlowat)
+ ret = sk->sk_socket->ops->set_rcvlowat(sk, val);
+ else
+ WRITE_ONCE(sk->sk_rcvlowat, val ? : 1);
break;
case SO_MARK:
if (sk->sk_mark != val) {
--
2.25.1
Powered by blists - more mailing lists