03:07bluetail: Hello. I noticed that the Link-Training behavior is differing from w7500 amdgpu to 9070xt. I can no more have a KVM in between without my displays turning off and on again, no matter what mesa version. Direct cabling without adapters works. The issue is when HDMI => KVM => HDMI => DP
05:47dj-death: valentine: looks like the culprit commit is touching NIR stuff
05:47dj-death: valentine: but I can't understand how this is not affecting TGL jobs in the same way...
09:11CounterPillow: yay, more AI spam incoming https://lore.kernel.org/dri-devel/CAOEgbEpEahQVtQxZT-WgT+Wzc8c1qtK1Oh+ZQVhy60Muo2FYXg@mail.gmail.com/T/ 11:18DragoonAethis: FYI: The Intel kernel CI is going down due to lab building maintenance from 17.04 8:00 CEST to 20.04 10:00 CEST - please submit your patches accordingly before the deadline.
12:28linkmauve: My hardware can only scanout from a YUYV buffer in memory, but can composite from RGB surfaces which reside in VRAM, how should I best expose that feature set in a KMS driver?
12:29linkmauve: in main* memory
14:15alarumbe: Hi tzimmermann, I wanted to commit some patches to drm-misc, but was unsure about the right branch to go about it. Since we're in the middle of a merge window, and the bugs these patches fix live in both drm-next and drm-misc-next, which branch do you think would be most suitable for enqueuing them?
14:45tzimmermann: alarumbe, hi. is the bug only in drm-next? then drm-misc-next-fixes. if the bug is already in upstream linux, then drm-misc-fixes
14:49bbrezillon: tzimmermann: OOC, does that apply during the MW too?