00:18 DemiMarie: karolherbst: if nouveau was written in Rust these issues would not exist _ducks_
00:23 zmike: DavidHeidelberg[m]: I thought you were going to add those zink-tu flakes? or was I supposed to be doing it
05:30 airlied: has mesamatrix not updated in a little while
05:30 airlied: ?
05:33 jannau: what's a little while? vulkan has nvk with a date of 2021-08-04 for me
05:33 jannau: err, 2023-08-04
05:36 airlied: jannau: I'm getting 4 days ago
05:37 airlied: and we've pushed a few features since then
05:39 jannau: airlied: yes, I see 4 days ago as well. a little while sounded like it might have been more than 4 days
09:26 DavidHeidelberg[m]: zmike: I see... I'll add them, something else took my attention back then :P
10:48 cwabbott: airlied: I just saw https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20389/diffs?commit_id=589d4ff0cb624739101ac878d5af38ab92d4a88c after rebasing my series that does more-or-less the same thing
10:50 cwabbott: in vulkan and gallium, we've been using separate RGB formats for these that don't apply colorspace conversion to sampling, so I don't think just translating to P010 like that is the correct approach
10:51 cwabbott: see e.g. https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6693 and how NV12 formats is mapped currently in vk_format.c
10:52 cwabbott: *are mapped
10:52 cwabbott: nothing is actually using the code added there, right?
10:54 cwabbott: (for context, I was adding support for sampling P010 to turnip)
10:59 cwabbott: ah, I guess it's currently returning NV12 but there's a comment in turnip that this is wrong and we shouldn't be doing this
12:03 mlankhorst: I'm looking at using async flip support for cursor updates, but one thing I did was let it complete before flip_done, so I can wait in the cleanup_done callback for a vblank
12:03 mlankhorst: this breaks an igt test
12:09 mlankhorst: I'm not sure how to solve it, unless I call cleanup_done early, but that seems wrong
12:37 mlankhorst: I think I'll cheat, unset legacy_cursor_update and call drm_atomic_helper_wait_for_vblanks() in the cleanup helper, hope nobody notices
12:40 mlankhorst: I gues that would fail because of each cleanup doing a vblank :/
19:11 zmike: any other nir people want to give feedback on https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24565
19:28 airlied: cwabbott: my memory is vague on that, I'd have to page it back in, I know my enabling some more formats on anv had some fallout, gfxstrand also suggested we should translate the yuv formats to just the separate RGB planes which is probably what we want
23:03 Mangix: rusticl have a channel?
23:19 nchery: is it just me or does the windows build for piglit regularly timeout?
23:19 nchery: https://gitlab.freedesktop.org/mesa/piglit/-/jobs/47042630
23:26 zmike: windows everything is timing out lately
23:27 zmike: mesa hasn't successfully merged anything in like 4 hours
23:30 nchery: ah, good to know I'm not alone. thanks
23:56 abhinav__: pq Hi, I had a question on https://gitlab.freedesktop.org/drm/igt-gpu-tools/-/blob/master/lib/igt_fb.c#L4716. So is this API not intended to work with formats having 2 planes? Like I would like to extend this to validate even NV12