On 12/12/2020 13:17, H. Nikolaus Schaller wrote:
Am 12.12.2020 um 12:12 schrieb David Shah dave@ds0.me:
I believe the atomic timeouts when doing DPMS started happening sometime around 5.7/5.8, they aren't even a 5.9 issue.
It seems like what might be happening is that vblanks are stopping because the display controller is stopped (as you'd expect doing DPMS), but something in the atomic code is expecting them still.
Unfortunately, ime the atomic DRM stuff tends to be quite hard to debug when it goes wrong.
Indeed...
I think we should wait until the new omapdrm drivers are upstream (v5.12-rc1?) and we can use the new drm/panel driver and look if it still happens.
I have pushed all the patches to drm-misc-next.
Tomi