Bug#498703: more info

["david manyé" <dmanye@gmail.com>]

2008/9/15 david manyé <dmanye@gmail.com>:
> 2008/9/13 Brice Goglin <Brice.Goglin@ens-lyon.org>:
>> david manyé wrote:
>>> hello again,
>>>
>>> i attach some related files:
>>>
>>> script.d122.ko.gz contains the output of
>>> /usr/share/bug/xserver-xorg-video-intel/script when xwindow crashed.
>>> Xorg.0.log.d122.ko.gz contains /var/log/Xorg.0.log also when xwindow crashed.
>>>
>>
>> When X fails to start, do you have gdm or so trying to restart it a
>> couple times?
>
> yes, maybe three times.
>
>> The log of the failure says:
>> (WW) intel(0): ESR is 0x00000010, page table error
>> (WW) intel(0): PGTBL_ER is 0x00000021
>> (WW) intel(0): PRB0_HEAD (0x00004ac4) and PRB0_TAIL (0x00000000)
>> indicate ring buffer not flushed
>> (WW) intel(0): Existing errors found in hardware state.
>> But these errors may come from the previous failure to start, leaving
>> the hardware in a bad state. So it would be nice if you could catch the
>> log of the first failure to start after a reboot.
>>
>
> if i'm not wrong, this is what the attached files whose name contains
> the string "ko" contains...
>
>> Otherwise, you're getting a GPU lockup, which can be caused by many
>> reasons, and we have pretty much no clue to debug it.
>
> in this case, what can be done? these machines also boot *without
> problems* a propietary operating system ...
>
>>
>> Brice
>>
>>
>
oops, i forgot to send the previous response to the bug number...

i've booted 24 computers and got a failure ratio of ~20%. some
machines got different PGTBL: 0x00000011, 0x00000021 and 0x00000039.
don't know if this could be useful...