[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160302154357.GF11670@leverpostej>
Date: Wed, 2 Mar 2016 15:43:58 +0000
From: Mark Rutland <mark.rutland@....com>
To: Andrey Ryabinin <aryabinin@...tuozzo.com>
Cc: linux-kernel@...r.kernel.org,
Alexander Potapenko <glider@...gle.com>,
Andrey Konovalov <adech.fo@...il.com>,
Dmitriy Vyukov <dvyukov@...gle.com>, will.deacon@....com,
catalin.marinas@....com, Andrew Morton <akpm@...ux-foundation.org>,
kasan-dev@...glegroups.com, Ingo Molnar <mingo@...nel.org>,
Peter Zijlstra <peterz@...radead.org>,
Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH 2/2] kasan: unpoison stack of idle task on cpu online
On Wed, Mar 02, 2016 at 06:27:49PM +0300, Andrey Ryabinin wrote:
> On 03/02/2016 05:50 PM, Mark Rutland wrote:
> > On Wed, Mar 02, 2016 at 04:51:59PM +0300, Andrey Ryabinin wrote:
[...]
> > Is all the above necessary?
> >
> > Surely we can just include <linux/smpboot.h> in mm/kasan/kasan.c?
>
> It is necessary. kernel/smpboot.h != include/linux/smpboot.h
Ah, I'd misread the patch. Sorry for the noise!
[...]
> >> + struct task_struct *tidle = idle_thread_get(cpu);
> >> + kasan_unpoison_shadow(task_stack_page(tidle), THREAD_SIZE);
> >
> > We never expect the stack to hit the end of the thread_info, so we can
> > start at task_stack_page(tidle) + 1, and avoid the shadow for
> > sizeof(struct thread_info).
> >
>
> I wouldn't bother, it's simpler to unpoison all. Size of struct thread_info is 32-bytes. That's 4-bytes of shadow.
> I don't think it matters whether you do memset of 2048 or 2044 bytes.
>
> > Do we do any poisoning of the thread_info structure in the thread_union?
>
> No, why would we poison it? It's absolutely valid memory and available for access.
For some reason I thought ASAN might poison gaps between struct
elements, or at least held open the option to. I guess inserting padding
would be an ABI issue, so it probably doesn't.
In the absence of that, I agree that always starting at
task_stack_page(t), and clearing the shadow for THREAD_SIZE bytes of
stack makes sense).
Thanks,
Mark.
Powered by blists - more mailing lists