#asahi-dev · 2025-12-28 08:54
nickchanm3 adt iop compatible is iop,ascwrap-v6 i think#asahi-dev · 2025-12-28 08:49
integralpiloti guess the M3 RTKit running on the ANS uses a different message format for shutdown, will have to RE it#asahi-dev · 2025-12-28 08:49
integralpilotand now we get the fun of an RTKit crash when we try to shut it down for the soft-reset :/#asahi-dev · 2025-12-28 08:46
integralpilotoh makes sense, thank you!#asahi-dev · 2025-12-28 08:28
svenThen it’s gonna be complicated :/#asahi-dev · 2025-12-28 08:28
svenwe also have to check if this happens with newer firmware on m1/2 #asahi-dev · 2025-12-28 08:28
svenideally base your nvme work on Torvalds tree and use tXXXX-nvme-ans2 where XXXX is whatever m3 machine with that issue was released first #asahi-dev · 2025-12-28 08:16
integralpilothow do we determine if Apple made a new major version, or if it's just a hardware revision that patched this?#asahi-dev · 2025-12-28 08:13
integralpilothave to make a compadible tag to avoid applying that chicken on M3+, should i call it ANS v2 rev 1 or ANS v3?#asahi-dev · 2025-12-28 08:11
integralpilotlooks like apple patched the chicken at https://github.com/AsahiLinux/linux/blob/asahi/drivers/nvme/host/apple.c line 1134 and the register we use there isn't in the memory map anymore#asahi-dev · 2025-12-28 08:09
integralpilotthanks, got it sorted now#asahi-dev · 2025-12-28 07:51
svenstill weird though#asahi-dev · 2025-12-28 07:51
svenintegralpilot: maybe the timeout isn't high enough in the power domain driver#asahi-dev · 2025-12-28 07:49
oliverfor sdr#asahi-dev · 2025-12-28 07:49
oliverjust double checked, it is XRGB2101010.#asahi-dev · 2025-12-28 07:23
oliverchadmed: mutter uses rbg2101010 for both sdr/hdr iirc.#asahi-dev · 2025-12-28 06:52
integralpilotif it’s reached the desired power state why does it continue saying it’s a failure (every now and then says that it dmesg) and hold up the initialisation chain#asahi-dev · 2025-12-28 06:51
integralpiloton M3 Linux getting a cryptic error from apple-pmgr-pwrstate for the ANS power domain it says “failed to reach power state 0xf (now: 0xf)” and this holds up the SART, ANS and NVMe probing and loading#asahi-dev · 2025-12-28 00:38
integralpilotanyway, I’m going to be doing some M3 Linux bring up work now as that’s a bit more fun but if anyone has any other ideas re. hypervised macOS I’m happy to test things out#asahi-dev · 2025-12-28 00:32
chadmedive commented them all out and the only one that seems to do anything important is the one where we fake a swap with no surfaces#asahi-dev · 2025-12-28 00:31
chadmedshould we be faking vblanks if crtc_state->no_vblank is armed? i kinda feel like the fact that we "have" to when kms already knows how to deal with vblankless hardware is an us problem #asahi-dev · 2025-12-28 00:21
integralpilotjust checked and all corruption issues all happen during the secondary core’s startup, so this looks related, but I’m chainloading a gcc version now and it doesn’t seem to fix the issue#asahi-dev · 2025-12-27 23:56
chadmedit uses the format fourccs for more than just the framebuffer format, e.g. r03w is argb2101010 but so is r01l. only the latter works properly for sdr argb2101010 framebuffers#asahi-dev · 2025-12-27 23:55
chadmedif its xrgb2101010/argb2101010 we will need to rework how we map drm formats to dcp formats#asahi-dev · 2025-12-27 23:51
chadmedoliver: what framebuffer format does gnome use for hdr#asahi-dev · 2025-12-27 23:41
integralpilotsven: the most common corruption symptom (SError) always happens while XNU starts up the secondary core, so might be a variant. I’ll look into it, thanks!#asahi-dev · 2025-12-27 23:21
chadmednot sure but i doubt the light sensor can spit out a ready to use CTM#asahi-dev · 2025-12-27 23:19
lminteresting. is that fed straight from the light sensor?#asahi-dev · 2025-12-27 23:18
chadmedit happens in iomfb. macos feeds it a colour transform matrix and it applies that to screen space after its composited all the swap surfaces#asahi-dev · 2025-12-27 23:16
lmDoes true tone work via DCP or is that managed in software? Just curious if those dynamic colour adjustments are happening at the hardware level#asahi-dev · 2025-12-27 23:14
lmYeah true, just wondering how it all comes to together#asahi-dev · 2025-12-27 23:12
jannauthat's afaik already loaded by iboot. but the driver can provide temperature measurements for temp compensation#asahi-dev · 2025-12-27 23:12
chadmedthere might be a cal blob somewhere, but this is less important than just getting dcp to output in the display's proper colour space#asahi-dev · 2025-12-27 23:10
lmIs there per-device colour calibration data used by DCP?#asahi-dev · 2025-12-27 23:07
chadmedthats okay we can fake it, as long as we can force a modeset into the correct colorimetry and eotf#asahi-dev · 2025-12-27 23:06
jannauColor and Timing Elements seem to derived from te EDID in the dcp firmware and since the internal displays don't have one we get bogus information there#asahi-dev · 2025-12-27 23:03
chadmedagain making sure its idea of what the display can do is important when multiple surfaces have content of different colour/dynamic range. we cant rely on DCP_COLORSPACE_NATIVE doing the right thing.#asahi-dev · 2025-12-27 23:03
chadmed"default" RGB for the builtin screens. we know the 14/16" machines are DisplayP3 and can do both HLG and PQ, and DCP supports all this #asahi-dev · 2025-12-27 23:02
chadmedcurrently ColorElements and TimingElements only show SDR and #asahi-dev · 2025-12-27 23:01
chadmedfaking edid for the builtin screens seems fine though, but we will need to force dcp to put the panel colorimetry and eotf into the correct mode or the resutls will be "wrong"#asahi-dev · 2025-12-27 23:00
chadmeddoing that will also not play nice once we wire up ddc for external monitors#asahi-dev · 2025-12-27 23:00
chadmedthe driver should not force a backlight value. let userspace do what it wants with it#asahi-dev · 2025-12-27 22:02
oliverMy patch does though. I have to check tomorrow what kwin does if I disable that again. #asahi-dev · 2025-12-27 22:01
oliverNo, it doesn't right now. I think I've read somewhere that this might be open for discussion. #asahi-dev · 2025-12-27 21:59
jannaudoes mutter at least initially sets the backlight to 100%?#asahi-dev · 2025-12-27 21:59
jannauyes, maybe based on the values in the metadata#asahi-dev · 2025-12-27 21:58
oliverCurrently in mutter that wouldnt be helpful, as you can not change the real brightness in hdr mode due to mutter only using the virtual backlight control.#asahi-dev · 2025-12-27 21:56
oliverThat is set as soon as you switch into hdr mode, even if only sdr content is on the screen. Ah, but you mean that would be fine, but still unlock the full brightness range.#asahi-dev · 2025-12-27 21:55
jannaubased on HDR_OUTPUT_METADATA set by the compositor#asahi-dev · 2025-12-27 21:50
oliverThat would need at least support by the compositor, in kernel space you don't know if you have pixels above sdr range