Skip to contentSkip to conversation

tcp: update window_clamp when SO_RCVBUF is set

PATCH Applied Jakub Kicinski 3 participants 5 messages Apr 8 – Apr 13

Subsystems (4)
lore
Share & tools
lore

Jakub Kicinski

author maintainer
#0 lore raw

PATCH 1/1 tcp: update window_clamp when SO_RCVBUF is set

Commit under Fixes moved recomputing the window clamp to tcp_measure_rcv_mss() (when scaling_ratio changes). I suspect it missed the fact that we don't recompute the clamp when rcvbuf is set. Until scaling_ratio changes we are stuck with the old window clamp which may be based on the small initial buffer. scaling_ratio may never change.

Inspired by Eric's recent commit d1361840f8c5 ("tcp: fix SO_RCVLOWAT and RCVBUF autotuning") plumb the user action thru to TCP and have it update the clamp.

A smaller fix would be to just have tcp_rcvbuf_grow() adjust the clamp even if SOCK_RCVBUF_LOCK is set. But IIUC this is what we were trying to get away from in the first place.

Fixes a2cbb1603943 ("tcp: Update window clamping condition")

---
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]
CC: [email protected]

---
 include/linux/net.h | 1 +
 include/net/tcp.h   | 1 +
 net/core/sock.c     | 9 +++++++++
 net/ipv4/af_inet.c  | 1 +
 net/ipv4/tcp.c      | 5 +++++
 net/ipv6/af_inet6.c | 1 +
 6 files changed, 18 insertions(+)
include/linux/net.h +1 -0
index a8e818de95b3..ca6a7bc5c9ae 100644
@@ -223,6 +223,7 @@ struct proto_ops {
 	int		(*sendmsg_locked)(struct sock *sk, struct msghdr *msg,
 					  size_t size);
 	int		(*set_rcvlowat)(struct sock *sk, int val);
+	void		(*set_rcvbuf)(struct sock *sk, int val);
 };
 
 #define DECLARE_SOCKADDR(type, dst, src)	\
include/net/tcp.h +1 -0
index 6156d1d068e1..b9db447892dd 100644
@@ -516,6 +516,7 @@ void tcp_syn_ack_timeout(const struct request_sock *req);
 int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 		int flags);
 int tcp_set_rcvlowat(struct sock *sk, int val);
+void tcp_set_rcvbuf(struct sock *sk, int val);
 int tcp_set_window_clamp(struct sock *sk, int val);
 
 static inline void
net/core/sock.c +9 -0
index fdaf66e6dc18..f3a186376bc5 100644
@@ -973,6 +973,8 @@ EXPORT_SYMBOL(sock_set_keepalive);
 
 static void __sock_set_rcvbuf(struct sock *sk, int val)
 {
+	struct socket *sock = sk->sk_socket;
+
 	/* Ensure val * 2 fits into an int, to prevent max_t() from treating it
 	 * as a negative value.
 	 */
@@ -990,6 +992,13 @@ static void __sock_set_rcvbuf(struct sock *sk, int val)
 	 * we actually used in getsockopt is the most desirable behavior.
 	 */
 	WRITE_ONCE(sk->sk_rcvbuf, max_t(int, val * 2, SOCK_MIN_RCVBUF));
+
+	if (sock) {
+		const struct proto_ops *ops = READ_ONCE(sock->ops);
+
+		if (ops->set_rcvbuf)
+			ops->set_rcvbuf(sk, sk->sk_rcvbuf);
+	}
 }
 
 void sock_set_rcvbuf(struct sock *sk, int val)
net/ipv4/af_inet.c +1 -0
index f98e46ae3e30..0e62032e76b1 100644
@@ -1091,6 +1091,7 @@ const struct proto_ops inet_stream_ops = {
 	.compat_ioctl	   = inet_compat_ioctl,
 #endif
 	.set_rcvlowat	   = tcp_set_rcvlowat,
+	.set_rcvbuf	   = tcp_set_rcvbuf,
 };
 EXPORT_SYMBOL(inet_stream_ops);
 
net/ipv4/tcp.c +5 -0
index e57eaffc007a..1a494d18c5fd 100644
@@ -1858,6 +1858,11 @@ int tcp_set_rcvlowat(struct sock *sk, int val)
 	return 0;
 }
 
+void tcp_set_rcvbuf(struct sock *sk, int val)
+{
+	tcp_set_window_clamp(sk, tcp_win_from_space(sk, val));
+}
+
 #ifdef CONFIG_MMU
 static const struct vm_operations_struct tcp_vm_ops = {
 };
net/ipv6/af_inet6.c +1 -0
index ee341a8254bf..0a88b376141d 100644
@@ -690,6 +690,7 @@ const struct proto_ops inet6_stream_ops = {
 	.compat_ioctl	   = inet6_compat_ioctl,
 #endif
 	.set_rcvlowat	   = tcp_set_rcvlowat,
+	.set_rcvbuf	   = tcp_set_rcvbuf,
 };
 EXPORT_SYMBOL_GPL(inet6_stream_ops);
 
Signature · 1 line
2.53.0
author maintainer
#1 lore raw
Quoted 7 lines · On Tue, 7 Apr 2026 17:14:38 -0700 Jakub Kicinski wrote:

Commit under Fixes moved recomputing the window clamp to tcp_measure_rcv_mss() (when scaling_ratio changes). I suspect it missed the fact that we don't recompute the clamp when rcvbuf is set. Until scaling_ratio changes we are stuck with the old window clamp which may be based on the small initial buffer. scaling_ratio may never change.

Quoted 8 lines

Inspired by Eric's recent commit d1361840f8c5 ("tcp: fix SO_RCVLOWAT and RCVBUF autotuning") plumb the user action thru to TCP and have it update the clamp.

A smaller fix would be to just have tcp_rcvbuf_grow() adjust the clamp even if SOCK_RCVBUF_LOCK is set. But IIUC this is what we were trying to get away from in the first place.

Hi Eric, any thoughts?
I always assume you are displeased if you don't reply within 8 hours :)

I should say that everyone has obviously discouraged the team that run into this from using SO_RCVBUF. I'm fascinated by how they decided that it helps since it clearly doesn't work. AI sure makes it easy for people to "try things". Sigh.

#2 lore raw
Quoted 13 lines · Jakub Kicinski
Quoted 11 lines · On Tue, 7 Apr 2026 17:14:38 -0700 Jakub Kicinski wrote:

Commit under Fixes moved recomputing the window clamp to tcp_measure_rcv_mss() (when scaling_ratio changes). I suspect it missed the fact that we don't recompute the clamp when rcvbuf is set. Until scaling_ratio changes we are stuck with the old window clamp which may be based on the small initial buffer. scaling_ratio may never change.

Inspired by Eric's recent commit d1361840f8c5 ("tcp: fix SO_RCVLOWAT and RCVBUF autotuning") plumb the user action thru to TCP and have it update the clamp.

Quoted 8 lines
Quoted 4 lines

A smaller fix would be to just have tcp_rcvbuf_grow() adjust the clamp even if SOCK_RCVBUF_LOCK is set. But IIUC this is what we were trying to get away from in the first place.

Hi Eric, any thoughts?
I always assume you are displeased if you don't reply within 8 hours :)

Not at all, I simply missed this patch. Too many emails to triage.

I will take a look asap.

Quoted 4 lines

I should say that everyone has obviously discouraged the team that run into this from using SO_RCVBUF. I'm fascinated by how they decided that it helps since it clearly doesn't work. AI sure makes it easy for people to "try things". Sigh.

Reviewed-by on v1 1/1
#3 lore raw
In reply to Jakub Kicinski
Quoted 12 lines · Jakub Kicinski

Commit under Fixes moved recomputing the window clamp to tcp_measure_rcv_mss() (when scaling_ratio changes). I suspect it missed the fact that we don't recompute the clamp when rcvbuf is set. Until scaling_ratio changes we are stuck with the old window clamp which may be based on the small initial buffer. scaling_ratio may never change.

Inspired by Eric's recent commit d1361840f8c5 ("tcp: fix SO_RCVLOWAT and RCVBUF autotuning") plumb the user action thru to TCP and have it update the clamp.

Quoted 7 lines

A smaller fix would be to just have tcp_rcvbuf_grow() adjust the clamp even if SOCK_RCVBUF_LOCK is set. But IIUC this is what we were trying to get away from in the first place.

Fixes a2cbb1603943 ("tcp: Update window clamping condition")
#4 lore raw
In reply to Jakub Kicinski
This patch was applied to netdev/net-next.git (main) by Paolo Abeni <[email protected]>:

Hello:

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

Quoted 8 lines · On Tue, 7 Apr 2026 17:14:38 -0700 you wrote:

Commit under Fixes moved recomputing the window clamp to tcp_measure_rcv_mss() (when scaling_ratio changes). I suspect it missed the fact that we don't recompute the clamp when rcvbuf is set. Until scaling_ratio changes we are stuck with the old window clamp which may be based on the small initial buffer. scaling_ratio may never change.

[...]

Here is the summary with links:
- [net] tcp: update window_clamp when SO_RCVBUF is set

    https://git.kernel.org/netdev/net-next/c/b025461303d8

You are awesome, thank you!

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