Re: [PATCH 1/8] perf bench futex: Retry futex_wait() when interrupted by a signal

[Namhyung Kim <namhyung@kernel.org>]

Hello,

On Sat, Sep 26, 2026 at 07:13:47PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] perf: deadlock in futex wake-parallel due to missed one-shot futex_wake()
> --
> 
> commit 4425d175409b64dbe77117f655f0ffa4eca2b864
> Author: Michal Pluta <michalpl2003@gmail.com>
> 
> perf bench futex: Retry futex_wait() when interrupted by a signal
> 
> The worker threads in 'futex wake' and 'futex wake-parallel' are meant
> to go back to sleep when a signal interrupts futex_wait(), but the check
> never works. futex_wait() returns -1 and sets errno, so comparing its
> return value with EINTR is always true and the loop ends at the first
> interruption.
> 
> > diff --git a/tools/perf/bench/futex-wake-parallel.c b/tools/perf/bench/futex-wake-parallel.c
> > index 6aede7c46b337..a089d8ee4b023 100644
> > --- a/tools/perf/bench/futex-wake-parallel.c
> > +++ b/tools/perf/bench/futex-wake-parallel.c
> > @@ -139,10 +139,9 @@ static void *blocked_workerfn(void *arg __maybe_unused)
> >  	cond_wait(&thread_worker, &thread_lock);
> >  	mutex_unlock(&thread_lock);
> >  
> > -	while (1) { /* handle spurious wakeups */
> > -		if (futex_wait(&futex, 0, NULL, futex_flag) != EINTR)
> > -			break;
> > -	}
> > +	/* Retry if the wait was interrupted by a signal. */
> > +	while (futex_wait(&futex, 0, NULL, futex_flag) == -1 && errno == EINTR)
> > +		;
> 
> [Severity: Medium]
> Can this retry loop cause a deadlock if a signal interrupts the wait?
> 
> If blocked_workerfn() receives a signal, futex_wait() returns -1 and
> errno == EINTR. The thread then temporarily leaves the futex wait queue before
> looping back to call futex_wait() again.
> 
> If waking_workerfn() in tools/perf/bench/futex-wake-parallel.c happens to
> execute its single futex_wake() call during this unqueued window:
> 
> tools/perf/bench/futex-wake-parallel.c:waking_workerfn() {
>     ...
>     waker->nwoken = futex_wake(&futex, nwakes, futex_flag);
>     if (waker->nwoken != nwakes)
>         warnx("couldn't wakeup all tasks (%d/%d)",
>     ...
> }
> 
> Will the waker thread miss the unqueued worker and exit, leaving the worker to
> block forever on its next futex_wait() and causing the main thread to hang in
> pthread_join()?

Looks like this patch discovers an existing bug.

Probably you can add a logic at the end of wakeup_threads() to check
each waker's nwoken and call futex_wake() additionally if needed.

Thanks,
Namhyung