#asahi-dev · 2026-09-01 15:01
jannauthe dev installer has no M3, M3 Pro/Max compatible Fedora-Asahi-Remix images#asahi-dev · 2026-09-01 14:04
jannaunot yet pushed. were 10 to 15 indent by space issues and 2 more multi die dtsi files not in upstream#asahi-dev · 2026-09-01 14:03
jannauyes, I did that. I also fixed relaxed style issue in the downstream kernel#asahi-dev · 2026-09-01 14:01
jannauI was thinking about the churn of cleaned upstream changes#asahi-dev · 2026-09-01 13:20
jannaumoving rules from strict to relaxed might be an easier way to upgrade strictness gradually#asahi-dev · 2026-09-01 13:19
jannauwould be nice to test for strict in the parts a commit changes but still annoying with all our downstream only dt changes#asahi-dev · 2026-09-01 12:53
jannauwrong way#asahi-dev · 2026-09-01 12:19
jannauor do we want to fix strict mode complaints#asahi-dev · 2026-09-01 12:19
jannausven: what's the goal for dt-check-style? `dt-check-style arch/arm64/boot/dts/apple/*.{dts,dtsi}` passing without complaints?#asahi-dev · 2026-08-31 16:28
jannauis it possible to change the target branch of a PR?#asahi-dev · 2026-08-30 15:49
jannausee the fixes commits for aop and pmp on how to adapt to 7.2 rust#asahi-dev · 2026-08-30 15:48
jannautryhard[m]: or onto asahi-wip-7.2 (to become asahi-wip as soon as the GPU driver works)#asahi-dev · 2026-08-30 09:58
jannauI've tagged asahi-7.1.12-1#asahi-dev · 2026-08-30 09:49
jannausofus: that fixes 1 sample (net, reports now 77/135 ass well)#asahi-dev · 2026-08-30 09:43
jannausofus: t8103 h.264 reports 76/135#asahi-dev · 2026-08-30 09:38
jannausofus: my fault, that system had still incomplete h.264 test samples#asahi-dev · 2026-08-30 09:07
jannausofus: was GStreamer-H.264-V4L2SL always at 28/135? Seeing that after merging your PR on t6020#asahi-dev · 2026-08-29 13:24
jannaunot sure how mesa's common yuv code handles hw specfic modifiers, it might just work as long as the drivers accept the the modifier#asahi-dev · 2026-08-29 13:22
jannauuser space software needs to explicitly handle the v4l2 format to drm modifier mapping#asahi-dev · 2026-08-29 13:19
jannauyuka: the kernel always uses /chosen/bootargs unless the kernel config has a CMDLINE and CONFIG_CMDLINE_FORCE is defined#asahi-dev · 2026-08-29 12:46
jannauyuka: looks like it'll easier to handle in kboot.c:dt_set_chosen() directly#asahi-dev · 2026-08-29 12:41
jannauand adding m4 support to u-boot#asahi-dev · 2026-08-29 12:40
jannauafter implementing kboot_append_chosen()#asahi-dev · 2026-08-29 12:40
jannausome thing like `if (m4) kboot_append_chosen("bootargs", "idle=none");` should work#asahi-dev · 2026-08-29 12:39
jannauu-boot handles bootargs via /chosen/bootargs, see linux.py / kboot_set_chosen#asahi-dev · 2026-08-28 13:27
jannaudifference just due to the different vo are unexpected#asahi-dev · 2026-08-28 13:26
jannautraces are (word-)diffable, especially when decoding the same video stream with the same program#asahi-dev · 2026-08-28 13:21
jannaua little annoying to read since there will be multiple files (one per thread) but the main decoding should run from a single thread#asahi-dev · 2026-08-28 13:20
jannaulibva produces text trace files with LIBVA_DEBUG={wayland,gpu} using the value as prefix#asahi-dev · 2026-08-28 13:18
jannauthis has to happen after the decoder because no other code would be able to handle 422 data correctly when it expects 420#asahi-dev · 2026-08-28 13:17
jannauunsure where the error could be though. yuv handling for asahi/hk in mesa is afaik mostly generic shader code#asahi-dev · 2026-08-28 13:12
jannauyes, looks like interpreting 4:2:2 data as if it is 4:2:0#asahi-dev · 2026-08-28 13:08
jannauif the first chroma channel height lines (instead of the expected height/2) then there is just 1 chroma channel#asahi-dev · 2026-08-28 13:05
jannauit's still weird though as I doubt that image as just approximately 2 colours#asahi-dev · 2026-08-28 12:49
jannauI don't understand this error. looks more like the driver produces 4:2:2 but it's interpreted as 4:2:0#asahi-dev · 2026-08-28 12:45
jannaunot sure actually#asahi-dev · 2026-08-28 12:45
jannauno, that should look different#asahi-dev · 2026-08-28 12:45
jannauthe color information is scaled by factor 2 vertically#asahi-dev · 2026-08-28 12:43
jannauJames image looks like a interpreting nv12 as fully planar yuv400#asahi-dev · 2026-08-28 12:41
jannaudmabuf-wayland is either direct-scanout or nv12 to rgb conversion in the compositor, gpu/gpu-next is yuv2rgb in mpv#asahi-dev · 2026-08-27 11:54
jannaubut my device uses "USI"/-u#asahi-dev · 2026-08-27 11:54
jannauhas anyone had any issues with bluetooth firmware on M3? There are additional firmware files with vendor "AMKOR" (https://github.com/AsahiLinux/asahi-installer/blob/f0469cea0899f3efed8efead604174c7a53c4451/asahi_firmware/bluetooth.py#L20)#asahi-dev · 2026-08-26 09:11
jannauI've tagged asahi-7.1.10-1#asahi-dev · 2026-08-26 07:34
jannaufor example "event_name-bit42 = USB2_wake" in smc-pmu#asahi-dev · 2026-08-26 07:33
jannaunot sure how to integrate that with smc's wake notifications. we're just using lid and power button but there are more#asahi-dev · 2026-08-26 07:31
jannauwe might need to sprinkle "wakeup-source;" over the device tree and use device_set_wakeup_capable() in some drivers#asahi-dev · 2026-08-26 07:22
jannaubut we don't even wake on laptop keyboard presses which we easily could#asahi-dev · 2026-08-26 07:21
jannauI'd expect it to work under macos but it doesn't work under linux, both on the native and PCIe xhci ports#asahi-dev · 2026-08-25 18:46
jannauack#asahi-dev · 2026-08-25 18:08
jannauno pm ops either so likely the explonation why usb devices need to be replugged after suspend