linux-kernel - Re: [BUG] Use of probe_kernel_address() in task_rcu_dereference() without checking return value

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]

Date:   Mon, 2 Sep 2019 18:20:36 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Oleg Nesterov <oleg@...hat.com>
Cc:     "Eric W. Biederman" <ebiederm@...ssion.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Russell King - ARM Linux admin <linux@...linux.org.uk>,
        Chris Metcalf <cmetcalf@...hip.com>,
        Christoph Lameter <cl@...ux.com>,
        Kirill Tkhai <tkhai@...dex.ru>, Mike Galbraith <efault@....de>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...nel.org>,
        Linux List Kernel Mailing <linux-kernel@...r.kernel.org>,
        Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
Subject: Re: [BUG] Use of probe_kernel_address() in task_rcu_dereference()
 without checking return value

On Mon, Sep 02, 2019 at 04:44:24PM +0200, Oleg Nesterov wrote:

> speaking of the users of task_rcu_dereference(), membarrier_global_expedited()
> does
> 
> 		rcu_read_lock();
> 		p = task_rcu_dereference(&cpu_rq(cpu)->curr);
> 		if (p && p->mm && (atomic_read(&p->mm->membarrier_state) &
> 				   MEMBARRIER_STATE_GLOBAL_EXPEDITED)) {
> 
> This asks for READ_ONCE, but this is minor. Why can't p->mm be freed?
> 
> I guess it is fine to read the garbage from &p->mm->membarrier_state if we race
> with the exiting task, but in theory this looks unsafe if CONFIG_DEBUG_PAGEALLOC.
> 
> Another possible user of probe_slab_address() or I am totally confused?

You're quite right; that's busted.

Due to the lack of READ_ONCE() on p->mm, the above can in fact turn into
a NULL deref when we hit do_exit() around exit_mm(). The first p->mm
read is before and sees !NULL, the second is after and does observe
NULL, and *bang*.

I suppose it wants to be something like:

	mm = READ_ONCE(p->mm);
	if (mm && probe_address())

(I'm not sure _slab_ is a useful part of the name; it should work on
kernel memory irrespective of the allocator)

If it got freed, that CPU already just did something that implies
smp_mb() so we're good. So whatever garbage gets read, is fine. Either
we do a superfluous IPI or not is immaterial.