Regression? 2.6.27-rc3 segfault on cold boot; not on warm boot.

[David Greaves]

Eric Sandeen wrote:
> David Greaves wrote:
>> Rafael J. Wysocki wrote:
>>> [Adding CCs]
>>>
>>> [The issue is probably present in 2.6.26 too]
>> Yes, below is a netconsole boot of 2.26.6.3
>>
>> I cold booted - lots of segfaults. I then did a ctrl-Alt-SysRq-S-U-B to a warm
>> boot. No segfaults
>> (lots of #### mark the split)
>>
>> I've just started 2.6.25.14 building for tomorrow to confirm how far back it
>> goes. (I'm 99% sure 2.6.25 is OK)
>>
>> David
> 
> (trimming cc's to xfs)
> 
> David, did you get anywhere with identifying this further?
> 
> Thanks,
> -Eric

Sorry, I didn't.

IIRC I spent a couple of weeks on it and found it too unpredictable to bisect.
I have been running 2.6.27.4 on that machine for a few months (checks... Nov 08)
and it's been good since then.

Linus suspected HW and we got into a habit of warming the machine up before
booting. That habit has long since gone and the last xfs hissy-fit was on 13 Sep 08.

David


-- 
"Don't worry, you'll be fine; I saw it work in a cartoon once..."