Re: [BISECTED, REGRESSION] Successful resume from suspend but freezes after I/O

[Yinghai Lu <yinghai@kernel.org>]

On Mon, Dec 21, 2009 at 12:54 AM, Volker Lanz <vl@fidra.de> wrote:
> On Tuesday 08 December 2009 02:13:20 Volker Lanz wrote:
>> On Tuesday 08 December 2009 00:05:11 Yinghai Lu wrote:
>> > the updated e820 table is the same to the one for 2.6.29...
>> >
>> > so could be some other patch cause the problem.
>> >
>> > please check if revert
>> > 2547089ca2db132e307ef68848ba029a8ec2f341
>> > could help.
>>
>> I reverted to one commit before that, which would be
>> eaa959df299157e2640fcb3321537501b6afd9e6
>> if I am not mistaken. I built and booted that kernel and could reproduce
>>  the freeze with it (TBH this was not surprising to me, because I
>>  originally pinned down the problem by bisecting the commits and this
>>  commit is after the one bisecting found. Am I missing a detail here?)
>>
>> But strangely enough the freeze this time produced an oops before it locked
>> the machine. I couldn't save it to a file but took a picture. The
>>  screenshot is here:
>> http://banane.volker-lanz.de/08122009013.jpg
>> The oops is _not _reproducable. After a reboot with this same kernel, a
>> suspend/resume cycle and some I/O the machine locks again without showing
>>  the oops.
>>
>> Volker
>>
>
> Any news on this problem? Anything else I can do to help debug the situation?
> The prospect of being confined to distros with kernels < 2.6.29 for the next
> couple of years on this machine worries me, obviously.
>

as 6d794 said that is used fix exposed bug.
...
please checkout out

c61cf4cfe7c73c7aa62dde3ff82cd475b9c41481

and boot with debug.

YH