[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87frn5vjul.ffs@tglx>
Date: Mon, 02 Dec 2024 22:46:42 +0100
From: Thomas Gleixner <tglx@...utronix.de>
To: Ilya Leoshkevich <iii@...ux.ibm.com>
Cc: LKML <linux-kernel@...r.kernel.org>, Frederic Weisbecker
<frederic@...nel.org>, Oleg Nesterov <oleg@...hat.com>, Heiko Carstens
<hca@...ux.ibm.com>, Vasily Gorbik <gor@...ux.ibm.com>, Alexander Gordeev
<agordeev@...ux.ibm.com>, Richard Henderson
<richard.henderson@...aro.org>, Ilya Leoshkevich <iii@...ux.ibm.com>
Subject: Re: [PATCH] posix-timers: Fix a race with __exit_signal()
On Mon, Dec 02 2024 at 16:54, Ilya Leoshkevich wrote:
> If a thread exit happens simultaneously with a POSIX timer signal, this
> POSIX timer may end up being erroneously stopped. This problem may be
> reproduced with a small C program that creates a POSIX timer and
> constantly respawns threads, e.g. [1].
AFAICT this has been fixed with commit:
63dffecfba3e ("posix-timers: Target group sigqueue to current task only if not exiting")
Are you sure, that you have this commit in your testing? It's in rc1.
Thanks,
tglx
Powered by blists - more mailing lists