[Intel-gfx] [PATCH] drm/i915: Record batch buffer following GPU error

[2points.mailinglist at gmx.org]

Hi Chris,

thank you for fixing that. Didn't run into any problems with the new patch.

Maybe the output can help along to pinpoint one of the/the bug that has 
apparently been haunting the 855 series with random freezes ever since 2.6.28. 
Will post relevant captures to appropriate bug report on the FD tracker. 
Probably #21826, in case you're interested.

Best regards,
Moritz

On Thursday 11 February 2010 15:19:58 you wrote:
> On Wed, 10 Feb 2010 22:33:40 +0100, 2points.mailinglist at gmx.org wrote:
> > Hi Chris,
> >
> > tried to gather some debugging output with this patch, but came upon the
> > following problem: The moment an error is to be recorded, the kernel
> > panics. From what I gather from the backtrace, it seems to crash in
> > clone_obj, somewhere along kmap_atomic, probably.
> 
> Thanks for testing, AFAICS this was probably triggered by my broken use of
> kunmap_atomic(). I've posted the updated patch and I would be grateful if
> you could try testing it again -- and sorry for the whatever it is that is
> causing the error in the first place. :|
> -ickle
>