Flickering with page-flipping on Acer Iconia W500 (AMD C-50 APU)

[Felix Kuehling]

On 12-02-22 11:20 AM, Felix Kuehling wrote:
> On 12-02-21 07:49 PM, Mario Kleiner wrote:
>> On 02/21/2012 09:07 PM, Alex Deucher wrote:
> [snip]
>>> The fix looks ok to me.  Mario any thoughts?
>>>
>>> Reviewed-by: Alex Deucher<alexdeucher at gmail.com>
>>>
>> Hi,
>>
>> the fix looks ok to me for that device, but could we make it
>> conditional on the AMD C-50 APU and similar pieces? It is the right
>> thing to do for that gpu, but for regular desktop gpus it is too
>> pessimistic if it defers the pageflip timestamping and completion
>> event for an already completed flip:
>>
>> 1. Makes the timestamps 1 refresh too late, causing timing sensitive
>> software like mine to detect false positives -- reporting skipped
>> frames were there weren't any. Not as bad as missing a really skipped
>> frame, but still not great.
> Agreed. I was going to perform some more experiments on other hardware
> to determine what the right threshold is for different hardware
> generations. I hope I'll get to that this week.

I have a final version of my patch including an explanation of the
observations it's based on (attached). It's against current drm-next
from git://people.freedesktop.org/~airlied/linux.

>
>> 2. Can reduce the framerate due to throttling the client, especially
>> on systems that are already challenged wrt. to their irq timing.
>>
>> Is the vblank period very short on these kind of devices? From Felix
>> description is sounds as if it is only 2 scanlines?
> It looks like that.

Turns out that that's not correct. Smaller negative values of vpos never
showed up in my log output because I didn't print it in case
update_pending was 0. The actual vblank period is 8 scan lines on this
device. Still not much compared to the ~40 I was seeing with an external
monitor. Anything > -4 would result in flickering in my experiments, so
only 5 scan lines worth of time are available for submitting the page
flip in time for the next frame. If I miss that time window, the flip is
deferred by an extra frame. In practice that seems to occur in about 25%
of cases on this particular device.

Regards,
  Felix

> Thanks for the feedback,
>   Felix
>
>> thanks,
>> -mario
>>

-- 
 _____    Felix Kuehling
 \ _  |   MTS Software Development Eng.
 /|_| |   SW-Linux Base Gfx | AMD
|__/ \|   T 905.882.2600 x8928