linux-kernel - Re: [PATCH] include/linux/kasan.h: Notice about 0 for kasan_[dis/en]able

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 May 2016 13:23:03 +0200
From:	Alexander Potapenko <glider@...gle.com>
To:	Chen Gang <chengang@...ndsoft.com.cn>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Andrey Ryabinin <aryabinin@...tuozzo.com>,
	Dmitriy Vyukov <dvyukov@...gle.com>,
	kasan-dev <kasan-dev@...glegroups.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Linux Memory Management List <linux-mm@...ck.org>,
	Chen Gang <gang.chen.5i5j@...il.com>
Subject: Re: [PATCH] include/linux/kasan.h: Notice about 0 for kasan_[dis/en]able_current()

On Mon, May 2, 2016 at 1:20 PM, Chen Gang <chengang@...ndsoft.com.cn> wrote:
> On 5/2/16 18:49, Alexander Potapenko wrote:
>> On Mon, May 2, 2016 at 7:35 AM,  <chengang@...ndsoft.com.cn> wrote:
>>>
>>> According to their comments and the kasan_depth's initialization, if
>>> kasan_depth is zero, it means disable. So kasan_depth need consider
>>> about the 0 overflow.
>>>
>>> Also remove useless comments for dummy kasan_slab_free().
>>>
>>> Signed-off-by: Chen Gang <gang.chen.5i5j@...il.com>
>>
>> Acked-by: Alexander Potapenko <glider@...gle.com>
Nacked-by: Alexander Potapenko <glider@...gle.com>
>>
>
> OK, thanks.
Well, on a second thought I take that back, there still might be problems.
I haven't noticed the other CL, and was too hasty reviewing this one.

As kasan_disable_current() and kasan_enable_current() always go
together, we need to prevent nested calls to them from breaking
everything.
If we ignore some calls to kasan_disable_current() to prevent
overflows, the pairing calls to kasan_enable_current() will bring
|current->kasan_depth| to an invalid state.

E.g. if I'm understanding your idea correctly, after the following
sequence of calls:
  kasan_disable_current();  // #1
  kasan_disable_current();  // #2
  kasan_enable_current();  // #3
  kasan_enable_current();  // #4

the value of |current->kasan_depth| will be 2, so a single subsequent
call to kasan_disable_current() won't disable KASAN.

I think we'd better add BUG checks to bail out if the value of
|current->kasan_depth| is too big or too small.

> Another patch thread is also related with this patch thread, please help
> check.
>
> And sorry, originally, I did not let the 2 patches in one patches set.
>
> Thanks.
> --
> Chen Gang (陈刚)
>
> Managing Natural Environments is the Duty of Human Beings.



-- 
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Matthew Scott Sucherman, Paul Terence Manicle
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg