[PATCH net] mptcp: fix race condition in mptcp_schedule_work()

[PATCH net] mptcp: fix race condition in mptcp_schedule_work()

[Eric Dumazet]

syzbot reported use-after-free in mptcp_schedule_work() [1]

Issue here is that mptcp_schedule_work() schedules a work,
then gets a refcount on sk->sk_refcnt if the work was scheduled.
This refcount will be released by mptcp_worker().

[A] if (schedule_work(...)) {
[B]     sock_hold(sk);
        return true;
    }

Problem is that mptcp_worker() can run immediately and complete before [B]

We need instead :

    sock_hold(sk);
    if (schedule_work(...))
        return true;
    sock_put(sk);

[1]
refcount_t: addition on 0; use-after-free.
 WARNING: CPU: 1 PID: 29 at lib/refcount.c:25 refcount_warn_saturate+0xfa/0x1d0 lib/refcount.c:25
Call Trace:
 <TASK>
 __refcount_add include/linux/refcount.h:-1 [inline]
  __refcount_inc include/linux/refcount.h:366 [inline]
  refcount_inc include/linux/refcount.h:383 [inline]
  sock_hold include/net/sock.h:816 [inline]
  mptcp_schedule_work+0x164/0x1a0 net/mptcp/protocol.c:943
  mptcp_tout_timer+0x21/0xa0 net/mptcp/protocol.c:2316
  call_timer_fn+0x17e/0x5f0 kernel/time/timer.c:1747
  expire_timers kernel/time/timer.c:1798 [inline]
  __run_timers kernel/time/timer.c:2372 [inline]
  __run_timer_base+0x648/0x970 kernel/time/timer.c:2384
  run_timer_base kernel/time/timer.c:2393 [inline]
  run_timer_softirq+0xb7/0x180 kernel/time/timer.c:2403
  handle_softirqs+0x22f/0x710 kernel/softirq.c:622
  __do_softirq kernel/softirq.c:656 [inline]
  run_ktimerd+0xcf/0x190 kernel/softirq.c:1138
  smpboot_thread_fn+0x542/0xa60 kernel/smpboot.c:160
  kthread+0x711/0x8a0 kernel/kthread.c:463
  ret_from_fork+0x4bc/0x870 arch/x86/kernel/process.c:158
  ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245

Fixes: 3b1d6210a957 ("mptcp: implement and use MPTCP-level retransmission")
Reported-by: syzbot+355158e7e301548a1424@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/6915b46f.050a0220.3565dc.0028.GAE@google.com/T/#u
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/mptcp/protocol.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 2d6b8de35c449374fec6dc935a459af0ef55b59a..e27e0fe2460f7de579b197e1c6c0e80fac10c1cd 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -935,14 +935,19 @@ static void mptcp_reset_rtx_timer(struct sock *sk)
 
 bool mptcp_schedule_work(struct sock *sk)
 {
-	if (inet_sk_state_load(sk) != TCP_CLOSE &&
-	    schedule_work(&mptcp_sk(sk)->work)) {
-		/* each subflow already holds a reference to the sk, and the
-		 * workqueue is invoked by a subflow, so sk can't go away here.
-		 */
-		sock_hold(sk);
+	if (inet_sk_state_load(sk) == TCP_CLOSE)
+		return false;
+
+	/* Get a reference on this socket, mptcp_worker() will release it.
+	 * As mptcp_worker() might complete before us, we can not avoid
+	 * a sock_hold()/sock_put() if schedule_work() returns false.
+	 */
+	sock_hold(sk);
+
+	if (schedule_work(&mptcp_sk(sk)->work))
 		return true;
-	}
+
+	sock_put(sk);
 	return false;
 }
 
-- 
2.51.2.1041.gc1ab5b90ca-goog

Re: [PATCH net] mptcp: fix race condition in mptcp_schedule_work()

[Matthieu Baerts]

Hi Eric,

(+cc MPTCP ML)

On 13/11/2025 11:39, Eric Dumazet wrote:
> syzbot reported use-after-free in mptcp_schedule_work() [1]
> 
> Issue here is that mptcp_schedule_work() schedules a work,
> then gets a refcount on sk->sk_refcnt if the work was scheduled.
> This refcount will be released by mptcp_worker().
> 
> [A] if (schedule_work(...)) {
> [B]     sock_hold(sk);
>         return true;
>     }
> 
> Problem is that mptcp_worker() can run immediately and complete before [B]
> 
> We need instead :
> 
>     sock_hold(sk);
>     if (schedule_work(...))
>         return true;
>     sock_put(sk);

Thank you for having released the syzbot issue with the fix! That's way
easier for us when the fix is provided with the bug report! :-D

The modifications look good to me:

Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>


This patch can be applied to -net directly, ideally with:

Cc: stable@vger.kernel.org

(Just for me to be able to track issues with the backports. If it cannot
be added, I can also track it "manually" if preferred.)

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

Re: [PATCH net] mptcp: fix race condition in mptcp_schedule_work()

[patchwork-bot+netdevbpf]

Hello:

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

On Thu, 13 Nov 2025 10:39:24 +0000 you wrote:
> syzbot reported use-after-free in mptcp_schedule_work() [1]
> 
> Issue here is that mptcp_schedule_work() schedules a work,
> then gets a refcount on sk->sk_refcnt if the work was scheduled.
> This refcount will be released by mptcp_worker().
> 
> [A] if (schedule_work(...)) {
> [B]     sock_hold(sk);
>         return true;
>     }
> 
> [...]

Here is the summary with links:
  - [net] mptcp: fix race condition in mptcp_schedule_work()
    https://git.kernel.org/netdev/net/c/035bca3f017e

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