linux-kernel - Re: /proc/diskstats buglet

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:	Tue, 26 Aug 2014 18:00:43 +0300
From:	Alexey Dobriyan <adobriyan@...il.com>
To:	Jens Axboe <axboe@...com>
Cc:	Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: /proc/diskstats buglet

On Tue, Aug 26, 2014 at 5:55 PM, Jens Axboe <axboe@...com> wrote:
> On 08/26/2014 08:47 AM, Alexey Dobriyan wrote:
>> Found and reproduced some time ago, almost forgot about :-)
>>
>> In part_round_stats_single(), ->stamp field is written but without
>> locking SMP-wise.
>>
>>     part->stamp = now;
>>
>> So, if two processes read /proc/diskstats, it is possible for "now -
>> part->stamp" value to become negative.
>>
>> And indeed this can happen:
>>
>> now 4294755500, ->stamp 4294755501
>> ------------[ cut here ]------------
>> WARNING: CPU: 0 PID: 1950 at block/blk-core.c:1229
>> part_round_stats_single+0xc0/0xd0()
>>  ...
>>  [<ffffffff811963f0>] part_round_stats_single+0xc0/0xd0
>>  [<ffffffff81196447>] part_round_stats+0x47/0x70
>>  [<ffffffff811a069d>] diskstats_show+0x8d/0x4b0
>>  ...
>>
>> Dunno how important used fields in /proc/diskstats but they can be
>> clearly bogus.
>
> Easiest fix is probably just to do the now - part->stamp math earlier,
> and ignore <= 0 instead of just now == part->stamp. I think that should
> be good enough for disk stats, and (most importantly), it would avoid
> the warning.
>
> Speaking of the warning, I don't see where that is. Where is it from?

I inserted the warning to be sure bug exists and I am not misreading the code.

  Alexey
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/