[Intel-gfx] [PATCH 10/11] drm: Use vblank timestamps to guesstimate how many vblanks were missed

[Michel Dänzer]

On 15.09.2015 04:43, ville.syrjala at linux.intel.com wrote:
> From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> 
> When lacking am accurate hardware frame counter, we can fall back to
> using the vblank timestamps to guesstimagte how many vblanks have
> elapsed since the last time the vblank counter was updated.
> 
> Take the oppostunity to unify the vblank_disable_and_save() and
> drm_handle_vblank_events() to call the same function
> (drm_update_vblank_count()) to perform the vblank updates.

It would be nice to keep the drm_update_vblank_count unification
separate. As it is, it's very hard to keep track of which parts of the
patch are for each logical change.


BTW, I think the fact that I was hitting the problem fixed by 209e4dbc
("drm/vblank: Use u32 consistently for vblank counters") within a few
days indicates that there's another bug which causes the counter to jump
forward with drm_vblank_on/off(). It may not manifest itself with
current Intel hardware because that has a full 32-bit hardware frame
counter, turning the related calculations into no-ops. I haven't had
time to investigate this further yet.


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer