[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABBYNZKnCyLkRKH=DFZbDSL=m0O5PUVkQjtiB0xpCZM7v78HmQ@mail.gmail.com>
Date: Thu, 17 Nov 2022 13:04:12 -0800
From: Luiz Augusto von Dentz <luiz.dentz@...il.com>
To: Thomas Gleixner <tglx@...utronix.de>
Cc: Hillf Danton <hdanton@...a.com>,
syzbot <syzbot+6fb78d577e89e69602f9@...kaller.appspotmail.com>,
linux-kernel@...r.kernel.org, pbonzini@...hat.com,
syzkaller-bugs@...glegroups.com,
Steven Rostedt <rosted@...dmis.org>,
Marcel Holtmann <marcel@...tmann.org>
Subject: Re: [syzbot] WARNING in call_timer_fn
Hi Thomas,
On Thu, Nov 17, 2022 at 8:06 AM Thomas Gleixner <tglx@...utronix.de> wrote:
>
> On Thu, Nov 17 2022 at 20:55, Hillf Danton wrote:
> > On 17 Nov 2022 12:54:28 +0100 Thomas Gleixner <tglx@...utronix.de>
> >>
> >> The work has been canceled already before in the same function and there
> >> are some more delayed works which can trigger this.
> >>
> >> So no, this whole close_sync() function is prone to teardown races and
> >> just slapping a single cancel here without deeper analysis does not cut
> >> it.
> >
> > Agree.
> >
> > A set of sync cancelations can do the job, given what is defined in struct
> > hci_dev wrt workqueue.
>
> It's only part of the solution because you also have to prevent that
> work is queued from other parts of the code....
I thought we would have something similar to shutdown_timer (e.g.
shutdown_delayed_work) so we can safely free its object/struct, at
least that was the impression I got when discussing with Steven.
--
Luiz Augusto von Dentz
Powered by blists - more mailing lists