[Intel-gfx] [PATCH 1/4] drm/i915: promote FIFO underruns to DRM_ERROR
Paulo Zanoni
przanoni at gmail.com
Fri Sep 20 20:38:22 CEST 2013
2013/9/20 Ville Syrjälä <ville.syrjala at linux.intel.com>:
> On Thu, Sep 19, 2013 at 05:20:49PM -0300, Paulo Zanoni wrote:
>> 2013/9/19 Chris Wilson <chris at chris-wilson.co.uk>:
>> > On Thu, Sep 19, 2013 at 05:00:35PM -0300, Paulo Zanoni wrote:
>> >> From: Paulo Zanoni <paulo.r.zanoni at intel.com>
>> >>
>> >> Linus recently complained about screen corruption when coming out of
>> >> DPMS on his Haswell machine, and he also mentioned there were no error
>> >> messages on the log. I think I can reproduce this problem, and when it
>> >> happens I get a "FIFO underrun" message, but since it's just
>> >> DRM_DEBUG_DRIVER it's hard to notice. So promote underruns to error
>> >> messages because reports containing "I'm getting a FIFO underrun on
>> >> pipe A" are way much better than "I'm getting a screen corruption".
>> >
>> > Sadly FIFO underruns on quite a few platforms are expected and
>> > unfixable. Others they are just expected. We reduced the error level
>> > because they are too noisy, and we were not in a position where the
>> > underrun was actionable. As we do not yet have the underrun feedback in
>> > place, I do not think we are ready for the onslaught of errors.
>> >
>> > At minimum, you should only set the error message for the platforms you
>> > intend to fix.
>>
>> Hmmmm, yeah, I should have thought about this a little bit more. Sorry :(
>>
>> At least on ILK/SNB/IVB/HSW we have the code to stop flooding dmesg
>> and the code to ignore the expected underruns, so I guess a patch
>> touching just these platforms would be ok?
>
> I would limit it to HSW before my ILK/SNB/IVB conversion to HSW
> watermark code gets in. The current code apparently produces somewhat
> incorrect results since according to Imre it even fixes some kind of
> rare hang on ILK.
Oh, adding "if IS_HASWELL" checks to that code will make it so much uglier :(
I guess I'll wait your ILK/SNB/IVB watermark fixes first :)
>
> --
> Ville Syrjälä
> Intel OTC
--
Paulo Zanoni
More information about the Intel-gfx
mailing list