linux-kernel - Re: 2.6.29 regression: ATA bus errors on resume

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]

Message-ID: <49DA5A83.2070002@kernel.org>
Date:	Mon, 06 Apr 2009 12:39:47 -0700
From:	Tejun Heo <tj@...nel.org>
To:	Niel Lambrechts <niel.lambrechts@...il.com>
CC:	"linux.kernel" <linux-kernel@...r.kernel.org>
Subject: Re: 2.6.29 regression: ATA bus errors on resume

Hello,

Niel Lambrechts wrote:
> On 04/06/2009 12:09 PM, Tejun Heo wrote:
>>> Will the fix naturally make its way into the mainline kernel, or is
>>> there any extra debugging/testing I can help with?
>>>     
>> Well, the problem is the debug patch doesn't actually do anything
>> other than printing out messages.  It could be that the problem is
>> timing dependent (which is likely anyway).  You still can reporduce
>> the problem with the patch, right?
>>   
> Heh? You provided two patches, with the last one you said:

Yeah, the second one actually only added printks to see whether that's
the case.  No behavior change.

>> Strange.  Maybe IO commands are getting through while the sdev is
>> still in quiesce state?  Can you please repeat the test with the
>> attached patch?
>
> With the latter, I have not encountered the original problem i.e. any
> severe EXT4 corruption again, not in 2.6.29 and not in 2.6.29.1.

Eh... so, we're definitely seeing something which is dependent on
timing.

> Do I also need to try the last patch without any debugging messages?

Then there will be nothing left.  :-) Can you please try the attached
patch?  It's still only debug messages but lighter; hopefully, it
won't mask the problem.

Thanks.

-- 
tejun

View attachment "libata-eh-debug-2.patch" of type "text/x-patch" (4378 bytes)