[PATCH net-next] tcp: only release congestion control if it has been initialized

[PATCH net-next] tcp: only release congestion control if it has been initialized

[Pengcheng Yang]

Currently, when cleaning up congestion control, we always call the
release regardless of whether it has been initialized. There is no
need to release when closing TCP_LISTEN and TCP_CLOSE (close
immediately after socket()).

In this case, tcp_cdg calls kfree(NULL) in release without causing
an exception, but for some customized ca, this could lead to
unexpected exceptions. We need to ensure that init and release are
called in pairs.

Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
---
 net/ipv4/tcp.c      | 2 +-
 net/ipv4/tcp_cong.c | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 4f77bd862e95..71005daf9102 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -3335,7 +3335,7 @@ int tcp_disconnect(struct sock *sk, int flags)
 	tp->window_clamp = 0;
 	tp->delivered = 0;
 	tp->delivered_ce = 0;
-	if (icsk->icsk_ca_ops->release)
+	if (icsk->icsk_ca_initialized && icsk->icsk_ca_ops->release)
 		icsk->icsk_ca_ops->release(sk);
 	memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
 	icsk->icsk_ca_initialized = 0;
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
index 0306d257fa64..df758adbb445 100644
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -270,8 +270,9 @@ void tcp_cleanup_congestion_control(struct sock *sk)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 
-	if (icsk->icsk_ca_ops->release)
+	if (icsk->icsk_ca_initialized && icsk->icsk_ca_ops->release)
 		icsk->icsk_ca_ops->release(sk);
+	icsk->icsk_ca_initialized = 0;
 	bpf_module_put(icsk->icsk_ca_ops, icsk->icsk_ca_ops->owner);
 }
 
-- 
2.38.1

Re: [PATCH net-next] tcp: only release congestion control if it has been initialized

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Fri, 25 Oct 2024 16:45:44 +0800 you wrote:
> Currently, when cleaning up congestion control, we always call the
> release regardless of whether it has been initialized. There is no
> need to release when closing TCP_LISTEN and TCP_CLOSE (close
> immediately after socket()).
> 
> In this case, tcp_cdg calls kfree(NULL) in release without causing
> an exception, but for some customized ca, this could lead to
> unexpected exceptions. We need to ensure that init and release are
> called in pairs.
> 
> [...]

Here is the summary with links:
  - [net-next] tcp: only release congestion control if it has been initialized
    https://git.kernel.org/netdev/net-next/c/cf44bd08cdee

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