Re: [BUG] signal: sighand unprotected when accessed by /proc

From: Steven Rostedt <rostedt@goodmis.org>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
	LKML <linux-kernel@vger.kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Peter Zijlstra <peterz@infradead.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Ingo Molnar <mingo@kernel.org>,
	Clark Williams <williams@redhat.com>,
	"Luis Claudio R. Goncalves" <lclaudio@uudg.org>
Subject: Re: [BUG] signal: sighand unprotected when accessed by /proc
Date: Tue, 3 Jun 2014 16:25:14 -0400	[thread overview]
Message-ID: <20140603162514.7b11f3ea@gandalf.local.home> (raw)
In-Reply-To: <20140603200938.GA3384@redhat.com>

On Tue, 3 Jun 2014 22:09:38 +0200
Oleg Nesterov <oleg@redhat.com> wrote:

> > Now, if that lock is released and reused (I didn't trace other tasks
> > allocating these locks), it reinitializes the lock->wait_list.
> 
> How? From where? This should be done by sighand_ctor() only?

This looks definitely like an -rt only bug and it's an obvious one at
that :-p

Looking in mm/slub.c: slab_alloc_node() we have this:

        if (unlikely(gfpflags & __GFP_ZERO) && object)
                memset(object, 0, s->object_size);
#ifdef CONFIG_PREEMPT_RT_FULL
        if (unlikely(s->ctor) && object)
                s->ctor(object);
#endif  

        slab_post_alloc_hook(s, gfpflags, object);

        return object;
}

We call the ctor on the object when it is allocated, not when the page
is created.

Doh! I guess we now know why we shouldn't do that.

-- Steve