#asahi-dev · 2026-10-19 08:27
JamesCalligeros[m]just the decoded_fmts bit or that whole commit? #asahi-dev · 2026-10-19 08:11
sofusremove this https://github.com/sofus13/linux/commit/79c4964f53b0678b4e3f37f4ca167546234fbad0#diff-a86fccc634687965763a3bf4ae1daddf5c036ce1fb6c7f7722c73eab15ddf33aR133#asahi-dev · 2026-10-19 08:11
sofusoh sorry. Thats a kernel change i forgot to handle in the vaapi driver#asahi-dev · 2026-10-19 08:01
JamesCalligeros[m]It does try to use IX12 though so its doing… something#asahi-dev · 2026-10-19 08:01
JamesCalligeros[m]I think something broke. mpv complains about the stride being invalid before it gets to touching the gpu#asahi-dev · 2026-10-19 07:51
chadmed:D#asahi-dev · 2026-10-19 07:49
svenwhoa, have we started working at the same place? :P#asahi-dev · 2026-10-19 07:44
chadmedin which case probably not something i can fix in between shifts at the bullshit factory#asahi-dev · 2026-10-19 07:43
chadmedi think the gpu driver may actually just be fundamentally broken with yuv images#asahi-dev · 2026-10-19 07:43
sofusi think its the setup, some alignment thing and a third thing#asahi-dev · 2026-10-19 07:42
chadmedsounds like memory alignment or something like that#asahi-dev · 2026-10-19 07:42
chadmedbuilding the kernel now, ill have a play#asahi-dev · 2026-10-19 07:42
sofussometimes every second tile on the first row look correct#asahi-dev · 2026-10-19 07:41
sofusthere seems to a few different kinds of errors. Sometimes the 'shapes' are kinda correct and its just blocky. Other times it looks mostly uninitialized (the cyan color + a few blocky things)#asahi-dev · 2026-10-19 07:39
chadmedthat layout only refers to the image in memory, and it's correct for that#asahi-dev · 2026-10-19 07:39
chadmedit might not actually be the interchange setup then#asahi-dev · 2026-10-19 07:39
chadmedhmm#asahi-dev · 2026-10-19 07:39
sofusrender incorrectly#asahi-dev · 2026-10-19 07:38
sofusoh you mean asahi-wip...#asahi-dev · 2026-10-19 07:38
sofusi have used direct scanout for the last few weeks#asahi-dev · 2026-10-19 07:38
chadmeddoes anything error or does it just render incorrectly#asahi-dev · 2026-10-19 07:38
sofusi did try to fix the interchange setup, but saw no difference. Also note the metadata setup#asahi-dev · 2026-10-19 07:38
chadmeddirect scanout is completely broken in any case#asahi-dev · 2026-10-19 07:37
sofusthanks, AX makes a lot more sense. I pushed kernel changes to avd-interchange and vaapi to interchange. Please test that direct scanout works first#asahi-dev · 2026-10-19 07:31
chadmedi will try to clean this all up, but no guarantee itll magically just work #asahi-dev · 2026-10-19 07:30
chadmed16x16 tiles, 1x1 blocks, etc#asahi-dev · 2026-10-19 07:30
chadmedoh yeah the ail function to initialise an interchange buffer makes a lot of rgb assumptions#asahi-dev · 2026-10-19 07:27
chadmedin the context of the fourccs i would even go with something like AX instead of IX to make it clear that it is apple interchange#asahi-dev · 2026-10-19 07:27
chadmedohhh i see where you used IC #asahi-dev · 2026-10-19 07:12
chadmedthanks :) #asahi-dev · 2026-10-19 07:12
sofuswill do#asahi-dev · 2026-10-19 07:11
chadmedif thats all thats needed #asahi-dev · 2026-10-19 07:11
chadmeda patch atop asahi-wip would be appreciated#asahi-dev · 2026-10-19 07:09
sofusi can prepare a little patch for you, otherwise it would be the ml version together with the vaapi vrapper branch#asahi-dev · 2026-10-19 07:08
chadmedslash a kernel branch with the avd bits wired up#asahi-dev · 2026-10-19 07:08
chadmeddo you have any test cases i can run easily#asahi-dev · 2026-10-19 07:08
chadmedif thats the only problem then the fix is trivial in the AIL code#asahi-dev · 2026-10-19 07:07
sofusi also think it suffers from the same BOMDAS error for the packed formats, but im not sure if they can even be imported#asahi-dev · 2026-10-19 07:07
sofusJamesCalligeros[m]: no, i tried a few things but could not get fix it / change anything. Mesa ignores the offset and the size calculations seem wrong (including meta)#asahi-dev · 2026-10-18 21:17
JamesCalligeros[m]Nothing has regressed but if yuv never worked in the first place i can look at that#asahi-dev · 2026-10-18 21:16
JamesCalligeros[m]sofus: did you ever figure out what was up with avd interchange looking wrong? Im going to take another look at mesa ~soon to get a merge request ready#asahi-dev · 2026-10-18 19:11
svenhttps://github.com/AsahiLinux/docs/pull/284#asahi-dev · 2026-10-18 18:01
sonoyumiiiSpeaking of updates, I already have version 1.6.1 of Stage 1, so I'll just wait for future updates or try to figure it out on my own. Thanks for the advice.#asahi-dev · 2026-10-18 17:58
sonoyumiiiI admit, I wanted to make the text clear and grammatically correct so that someone could understand it quickly. I'm from Europe and don't know English very well; I wrote this text using a translation tool and won't repeat the two grammatical mistakes I made. :\#asahi-dev · 2026-10-18 17:55
chaos_princessanyway, try re-running the installer and updating stage 1#asahi-dev · 2026-10-18 17:54
chaos_princesswell, there is one more thing that your issue does not meet :P#asahi-dev · 2026-10-18 17:53
sonoyumiiiSorry, my mistake: I filed it with the gh CLI, which skipped the form. I'm on Arch Linux ARM, not Fedora Asahi Remix, so it doesn't meet the template. I'll close it.#asahi-dev · 2026-10-18 17:45
chaos_princessi happen to notice that you decided to ignore the issue template, why is that?#asahi-dev · 2026-10-18 17:40
sonoyumiiiIs its payload format known, or is there any other way to get the backlight working without the visible panel blank? Details and logs: https://github.com/AsahiLinux/m1n1/issues/733 . Happy to test patches on this machine.#asahi-dev · 2026-10-18 17:40
sonoyumiiiHi! j293 (MBP 13" M1), SFW 26.6.2 (mBoot-18000.161.10), stub 13.5, m1n1 1.6.1. The Sequoia power-cycle workaround in display_init() is required here: without it the backlight stays stuck at max in Linux (DCP still reports correct mNits/iDAC, so the change seems to never reach the backlight IC). Same-mode reconfigure without power-off and DPMS off/on from Linux don't help. trace_dcp.py lists iBoot EP 0x23 op 7 as setBrightness (jannau also mentioned