linux-kernel - Re: [patch V2 3/3] signal: Allow tasks to cache one sigqueue struct

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <871rck2hze.fsf@nanos.tec.linutronix.de>
Date:   Fri, 12 Mar 2021 22:13:25 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Oleg Nesterov <oleg@...hat.com>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Ingo Molnar <mingo@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Juri Lelli <juri.lelli@...hat.com>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
        Daniel Bristot de Oliveira <bristot@...hat.com>,
        Matt Fleming <matt@...eblueprint.co.uk>,
        "Eric W. Biederman" <ebiederm@...ssion.com>
Subject: Re: [patch V2 3/3] signal: Allow tasks to cache one sigqueue struct

On Fri, Mar 12 2021 at 20:26, Thomas Gleixner wrote:
> On Fri, Mar 12 2021 at 17:11, Oleg Nesterov wrote:
>> On 03/11, Thomas Gleixner wrote:
>>>
>>> @@ -456,7 +460,12 @@ static void __sigqueue_free(struct sigqu
>>>  		return;
>>>  	if (atomic_dec_and_test(&q->user->sigpending))
>>>  		free_uid(q->user);
>>> -	kmem_cache_free(sigqueue_cachep, q);
>>> +
>>> +	/* Cache one sigqueue per task */
>>> +	if (!current->sigqueue_cache)
>>> +		current->sigqueue_cache = q;
>>> +	else
>>> +		kmem_cache_free(sigqueue_cachep, q);
>>>  }
>>
>> This doesn't look right, note that __exit_signal() does
>> flush_sigqueue(&sig->shared_pending) at the end, after exit_task_sighand()
>> was already called.
>>
>> I'd suggest to not add the new exit_task_sighand() helper and simply free
>> current->sigqueue_cache at the end of __exit_signal().
>
> Ooops. Thanks for spotting this!

Hrm.

The task which is released is obviously not current, so even if there
are still sigqueues in shared_pending then they wont end up in the
released tasks sigqueue_cache. They can only ever end up in
current->sigqueue_cache.

But that brings my memory back why I had cmpxchg() in the original
version. This code runs without current->sighand->siglock held.

So we need READ/WRITE_ONCE() for that on both sides which is sufficient.

Thanks,

        tglx