#asahi-dev · 2026-10-06 04:07
myriaI saw! thank you, sorry was in a lecture#asahi-dev · 2026-10-06 04:07
MichaelReeves[m]myria: looked through ur PMU pre-review branch, it looks really good I just left a couple comments :)#asahi-dev · 2026-10-05 20:41
noopwafel(although tbh I'm assuming it's easier for me to continue just using a script around tuxvdmtool+initramfs, given it'll only be a few machines)#asahi-dev · 2026-10-05 20:38
noopwafelis the plan still the draft MR using the CS board?#asahi-dev · 2026-10-05 20:37
noopwafel(for drm/mesa in particular)#asahi-dev · 2026-10-05 20:37
noopwafelI was wondering about setting something up for the M1 at home, because otherwise it's easy to break things#asahi-dev · 2026-10-05 20:30
DragoonAethisjannau: okie, thanks!#asahi-dev · 2026-10-05 20:16
jannauno, but we plan to use ci-tron (https://gitlab.freedesktop.org/gfx-ci/ci-tron) and there is some progress implementing support#asahi-dev · 2026-10-05 20:12
DragoonAethisG'day, I'm talking with some CI friends and wanted to ask: For kernel changes, do you have working hardware CI with preemption?#asahi-dev · 2026-10-05 17:35
Hel3ziupd: cat /proc/cpuinfo | grep processor processor : 0 processor : 1 processor : 2 processor : 3 processor : 4 ~ ##asahi-dev · 2026-10-05 17:15
Hel3zicat /proc/cpuinfo | grep processor processor : 0 processor : 1 processor : 2 processor : 3#asahi-dev · 2026-10-05 17:14
Hel3ziyuka: Further confirmation on M4 Pro: -C 4,5,6,7 brings up all 4 P-cores cleanly!#asahi-dev · 2026-10-05 17:04
fl0_idnice#asahi-dev · 2026-10-05 17:03
Hel3ziUptime 12+ min#asahi-dev · 2026-10-05 16:52
yukanice!#asahi-dev · 2026-10-05 16:52
Hel3ziyuka: NVMe is fully working on M4 Pro! Rebuilt with built-in RTKIT/NVME_APPLE/SART, partitions are visible and benchmark passed cleanly: code Text ~ # fdisk -l /dev/nvme0n1 Disk /dev/nvme0n1: 1954210120 sectors, 2785M Number Start (sector) End (sector) Size Name 1 6 128005 500M iBootSystemContainer 2 128006 242965550 926G Container 3 242965551 244276259 5119M RecoveryOSContainer #asahi-dev · 2026-10-05 16:16
yuka(as part of the rtkit crashlog)#asahi-dev · 2026-10-05 16:16
yuka> Message (id=1): assert failed: [20646]:Dispatcher: Error interrupt. status = 0x2, valid = 0x2, specific_status = 0x20000#asahi-dev · 2026-10-05 16:16
yukastill fighting with t6050 nvme#asahi-dev · 2026-10-05 12:24
Hel3ziNothing from the actual NVMe driver in dmesg, only PMGR power state lines: [ 11.502775] apple-pmgr-pwrstate ... sync_state() pending due to 44dcc0000.nvme [ 11.504760] apple-pmgr-pwrstate ... sync_state() pending due to 40dc50000.sart So Yuka is likely right - CONFIG_NVME_APPLE is probably =m in defconfig, and no modules are loaded in this bare initramfs.#asahi-dev · 2026-10-05 12:16
yukaMaybe the kernel module is not builtin?#asahi-dev · 2026-10-05 12:16
chaos_princessanything suspicious nvme-wise in dmesg?#asahi-dev · 2026-10-05 12:16
chaos_princessnvme is not quite pcie here, so it should not matter iirc#asahi-dev · 2026-10-05 12:15
Hel3ziUpd: uptime 00:18:48 up 18 min, 0 users, load average: 0.00, 0.00, 0.00 ~ ##asahi-dev · 2026-10-05 12:15
Hel3ziyuka: No /dev/nvme* (No such file or directory), but uptime is beyond 10 minutes without any crash: 00:10:24 up 10 min, load average: 0.00, 0.00, 0.00 (Keep in mind pcie_init is stubbed out in m1n1).#asahi-dev · 2026-10-05 12:14
Hel3ziHaha, oh no, what have I gotten myself into :D There's no turning back now! Thanks a lot for guiding me through all of this, it was an awesome experience!#asahi-dev · 2026-10-05 12:08
yuka(this is how it all starts... and at some point you can not go back and become a kernel developer)#asahi-dev · 2026-10-05 12:07
yukaalso congrats, you are now cursed with the ability to test things on this platform :D#asahi-dev · 2026-10-05 12:05
yuka*beyond#asahi-dev · 2026-10-05 12:05
yuka*beyone#asahi-dev · 2026-10-05 12:05
yukaTwo more questions: do you have /dev/nvme0n* and does it keep uptime beyone 10min#asahi-dev · 2026-10-05 12:03
Hel3ziyuka:Keyboard backlight is physically ON! echo 255 > /sys/class/leds/kbd_backlight/brightnesssuccessfully lit up the keyboard on the 14" M4 Pro! The internal display (simpledrm) is also working perfectly with the framebuffer console. Both test requests are complete! (Caps Lock indicator isn't working yet as expected without dockchannel-hid).#asahi-dev · 2026-10-05 11:58
myriaGoodnight!#asahi-dev · 2026-10-05 11:55
integralpilotok im going to bed now so ill send the KVM patch + test android stuff tomorrow! goodnight#asahi-dev · 2026-10-05 11:55
svenif it happens more often or something else happens there’s another bug somewhere that I never hit #asahi-dev · 2026-10-05 11:54
svenand when it doesn’t work it will just hang when the secondaries are started #asahi-dev · 2026-10-05 11:54
sveneven without C1 it should work like 95% of the time #asahi-dev · 2026-10-05 11:54
Hel3zi@sven @yuka: Makes sense! With -C 1 I was able to get all the way to an interactive Alpine rootfs shell (/bin/sh). Just recompiled the kernel with DRM_SIMPLEDRM=y and PWM_APPLE=y built-in, testing if the framebuffer console and keyboard backlight come up now.#asahi-dev · 2026-10-05 11:52
svenbut it works most of the time for me now #asahi-dev · 2026-10-05 11:52
chadmedhmmm from the sounds of it p3s requires the loader to do some extra magic?#asahi-dev · 2026-10-05 11:51
sventhere’s probably another rare cache race bug or something similar #asahi-dev · 2026-10-05 11:51
svenI very rarely see it now #asahi-dev · 2026-10-05 11:51
svensometimes m4 pro smp bringup hangs but that should already be much better in the current main branch #asahi-dev · 2026-10-05 11:41
integralpilotunless someone else would like to test?#asahi-dev · 2026-10-05 11:41
integralpiloti dont have one#asahi-dev · 2026-10-05 11:40
jannauwould be easier to test on M1/M2#asahi-dev · 2026-10-05 11:37
integralpilotjust need to cherry-pick m3 dts and other stuff now and will give a go#asahi-dev · 2026-10-05 11:37
myriadefinitely!#asahi-dev · 2026-10-05 11:36
integralpiloti probably stuffed it up but only 1 way to find out#asahi-dev · 2026-10-05 11:36
integralpilotoooh ok so it applies fully cleanly now