#asahi · 2022-01-23 20:49
jannaudmesg | grep dcp#asahi · 2022-01-23 20:47
jannauxorg works well enough for sddm here. can you check if there's anything in /dev/dri/#asahi · 2022-01-23 20:24
jannauGlanzmann: try again, I forgot to push#asahi · 2022-01-23 20:24
jannauChaosPrincess: it is asahi + dcp (+ spi hid patches)#asahi · 2022-01-23 20:15
jannauthe dcp driver will take over control from simpledrm, not sure how that works. tested with simpledrm compiled in and dcp as module#asahi · 2022-01-23 20:13
jannauGlanzmann: use asahi-dcp, it also needs patches for rtkit and nvme (dependency of the rtkit changes)#asahi · 2022-01-23 19:59
jannausetting the resolution should work as well if screen is not full HD#asahi · 2022-01-23 19:57
jannauGlanzmann: rotation works but won't with m1n1 alone#asahi · 2022-01-23 19:24
jannaureplugging is still broken, screen blanking is powering the display off? if so it doesn't work. it dims my display thoguh#asahi · 2022-01-23 19:20
jannauGlanzmann: now that that m1n1 inits the display there is no reason to use the dcp in itss current state#asahi · 2022-01-23 15:48
jannauI suspect that will be try for other samples as well#asahi · 2022-01-23 15:48
jannauJamie[m]1: the problem for h264-conformance-aud_mw_e is that VTB outputs nv12 instead of yuv420#asahi · 2022-01-23 14:57
jannauthe samples with pixel differences are unexpected, either something wrong in how ffmpeg uses vtb or vtb does some addtional processing#asahi · 2022-01-23 14:56
jannauI don't remember what the conformance samples are testing, it's possible that they require features not supported by the HW decoder but the list is longer than what I woud expect#asahi · 2022-01-23 14:54
jannauI would not expect a full bitstream HW decoder to produce bit-exact (or any at all) output with a sample with a broken SPS#asahi · 2022-01-23 14:46
jannauJamie[m]1: are any of the fate-h264-conformance tests failing?#asahi · 2022-01-23 14:44
jannauwhich codecs have non bit-exact frames? just a certain frames or all frames#asahi · 2022-01-23 14:41
jannauJamie[m]1: what do you mean with recover? I wouldn't expect tests with brokens bitstream to behave the same with hw-accel#asahi · 2022-01-22 13:31
jannaubut only for mps's error#asahi · 2022-01-22 13:29
jannauVladimirMateev[m]: https://lore.kernel.org/all/20220114140741.1358263-1-maz@kernel.org/ seems relevant#asahi · 2022-01-20 20:48
jannauI guess it's called always-on because it implments the always-on feature. the device acts as if it is always-on even when it is sleeping#asahi · 2022-01-19 15:15
jannauGlanzmann: I would be surprised if the u-boot dt has the audio nodes from asahi#asahi · 2022-01-17 21:39
jannauit should be possible to bundle asahi-installer on the image (readable by macos recovery) for offline installations#asahi · 2022-01-17 21:32
jannau2. has the advantage that it doesn't require writing an installer image onto an usb stick#asahi · 2022-01-17 21:31
jannau1. has the advantage that asahi-installer and the distro installer are completly separate#asahi · 2022-01-17 21:30
jannauin the simplest case ith will be just a shell script / binary asahi-installer executes after it has setup the efi system partition (maybe only after step2.sh)#asahi · 2022-01-17 21:28
jannau2. integration in asahi-installer. too early for this but I don't think it will require much integration work into asahi-installer#asahi · 2022-01-17 21:26
jannau1. installer iso/usb stick which can be bootet via u-boot/efi#asahi · 2022-01-17 21:25
jannaumps: there are two possible integration paths for custom distro installers#asahi · 2022-01-17 21:19
jannaua little annoying but manageable#asahi · 2022-01-17 21:18
jannauthe installer needs cutom handling to make them available before it loads brcmfmac#asahi · 2022-01-17 21:16
jannauthe installer will create the efi system partition and save a tarball with the extracted firmware there#asahi · 2022-01-17 19:48
jannauit's linked under "Network boot" on https://www.debian.org/distrib/netinst#asahi · 2022-01-17 19:43
jannauGlanzmann: was using the tiny option with just the initrd https://deb.debian.org/debian/dists/bullseye/main/installer-arm64/current/images/netboot/debian-installer/arm64/#asahi · 2022-01-17 19:34
jannauGlanzmann: the net-installer is iirc text-only#asahi · 2022-01-17 12:18
jannaukov: https://github.com/AsahiLinux/docs/wiki/HW:USB-PD#about-reboot-command#asahi · 2022-01-17 08:12
jannauVinDuv: did that monitor work with macos 11 iboot? does it work after a chainload?#asahi · 2022-01-16 21:58
jannauVinDuv: https://github.com/AsahiLinux/m1n1/pull/144 for display detection at boot#asahi · 2022-01-16 21:50
jannauhah, see display_choose_timing_mode() in src/display.c for the mode selection#asahi · 2022-01-16 21:49
jannauVinDuv: which mode does m1n1 report? I see with a FullHD display "display: timing mode: valid=1 1920x1080 60.00 Hz"#asahi · 2022-01-16 21:36
jannauchainloading the same m1n1 should work#asahi · 2022-01-16 21:36
jannauVinDuv: dcp training takes a while. around 400ms with my display. I'm preparing a PR#asahi · 2022-01-16 21:00
jannauwell, 7 times display_configure() + mdelay(10)#asahi · 2022-01-16 20:41
jannautakes about 70ms of delay until dcp reports modes#asahi · 2022-01-16 20:20
jannau0% success rate in ~15 reboots#asahi · 2022-01-16 20:20
jannauat other times connected, timing_cnt and color_cnt are all 0#asahi · 2022-01-16 20:18
jannauand "display: failed to get display status"#asahi · 2022-01-16 20:16
jannauyes, it fails: "display: connected:320633856 timing_cnt:1548293737 color_cnt:0"#asahi · 2022-01-16 19:55
jannauthe dcp driver works on devices with integrated display so most of the potential issues would have been visible there#asahi · 2022-01-16 19:53
jannaudcp works with the display initialized by m1n1, everything else would have been surprising