#
✕ nick:janneg
#asahi-dev · 2023-10-08 18:29
jannegthe issue seesm to be that kwin apparently can't deal with simpledrm + asahi failing to work
#asahi-dev · 2023-10-08 18:22
jannegI don't see any of efifb's messages in the kernel log, blocking efifb_probe makes no difference
#asahi-dev · 2023-10-08 18:21
jannegsomething breaks once the system is booted, on shutdown I see again messages
#asahi-dev · 2023-10-08 18:19
jannegdespite of the error simpledrm still seems to work. error is before the kernel calls init but with "rhgb quiet" removed from cmdline args it still print systemd startup messages
#asahi-dev · 2023-10-08 18:06
jannegI tried to block eififb's initcall but that didn't change anything, also did not look like efi grabbed the framebuffer from sysfs
#asahi-dev · 2023-10-08 18:05
jannegI noticed that fedora builds with efifb
#asahi-dev · 2023-10-08 18:05
jannegj`ey: no, simpledrm intentionally doesn't do that. simpledrm should only grab the aperture if it is free. otherwise it could race with the proper drm driver if both are built-in or modules
#asahi-dev · 2023-10-08 16:29
jannegthe devicetree in /proc looks fine
#asahi-dev · 2023-10-08 16:29
jannegany idea why simpledrm could end up with such a bogus looking resource on fedora: "simple-framebuffer simple-framebuffer.0: [drm] *ERROR* could not acquire memory range [bus ffff5581cc480580-ffff80008006bae0 flags 0xffffc578ca683088]:"?
#asahi-dev · 2023-10-08 15:47
jannegcopying now 12.3 and 12.4 appleh13camerad
#asahi-dev · 2023-10-08 15:43
jannegmarcan: see https://github.com/AsahiLinux/linux/issues/207 (5/6 probably 12.4) and https://tpaste.us/PEvJ (maybe 12.3)
#asahi-dev · 2023-10-08 14:11
janneglooks like we need to handle missing calicration files in older versions of appleh13camerad gracefully
#asahi-dev · 2023-10-08 11:32
jannegasahi-6.5-11 / m1n1-1.4.x works on j375d/j493
#asahi-dev · 2023-10-08 08:16
jannegchadmed was planning to remove the speaker detection so I would expect that that breaks again without speakers
#asahi-dev · 2023-10-08 08:15
jannegthe ucm files apply the necessary configuration
#asahi-dev · 2023-10-08 08:14
jannegheadphones with alsa should work without ucm files on j475d if the alsa device is configured correctly, not sure if pipewire will interfere with the config
#asahi-dev · 2023-10-08 08:02
jannegheadphone jack should work just with the UCM files which haven't been updated for the 2023 machines
#asahi-dev · 2023-10-08 08:00
jannegI guess for the m2 someone needs to run the audio tracer to see how macos configures the amplifier
#asahi-dev · 2023-10-07 17:19
jannegso it was a while since you ran `make apple_m1_defconfig`? distro packages have advantages sometimes
#asahi-dev · 2023-10-07 17:17
jannegI see it both in asahi and asahi-releng
#asahi-dev · 2023-10-07 16:53
janneghow does it look with the alias removed?
#asahi-dev · 2023-10-07 16:32
jannegthat should be the DT
#asahi-dev · 2023-10-07 16:29
jannegthe first entry in rsvmem has a non page-aligned size
#asahi-dev · 2023-10-07 16:22
jannegthe nvram partition size differs between 12.x and 13.x os firmware so it is now calculated and filled by m1n1
#asahi-dev · 2023-10-07 16:19
jannegChaosPrincess: nvram partition size might be worth checking
#asahi-dev · 2023-10-07 16:17
jannegChaosPrincess: can you try to disable the isp node and remove the alias
#asahi-dev · 2023-10-07 16:16
jannegnot sure why u-boot would care about the regs of a device it has no drivers for
#asahi-dev · 2023-10-07 16:16
jannegwas easier that way to support the different register layout that way but should be fixed
#asahi-dev · 2023-10-07 16:14
jannegthe only thing I can think of which was added recently and hasn't a page aligned size are resources in the isp node
#asahi-dev · 2023-10-07 16:12
jannegseems to be t6000 specific, I have no issues on t6001
#asahi-dev · 2023-10-07 16:11
jannegI remember something about u-boot cosuming the whole memory due to something not being page aligned
#asahi-dev · 2023-10-07 16:09
jannegChaosPrincess is seeing those with m1n1-1.4.0 and the dts from the current asahi-wip branch
#asahi-dev · 2023-10-07 16:09
jannegkettenis: any idea what could cause this allocation errors in u-boot https://paste.bingner.com/paste/rbtdx
#asahi-dev · 2023-10-07 15:53
jannegkettenis: quoting from somewhere else: "there's a join_pref error I haven't looked at yet (it looks like a variable size/tagged struct, not sure why it fails) but it's nonfatal"
#asahi-dev · 2023-10-07 13:15
jannegjust additional pointers or changes in the decoder headers as well? I would expect 2 reference frame lists. might be hard to detect if you just decode 2 frames
#asahi-dev · 2023-10-07 11:53
jannegH.264 (03/2005) is the best spec document to look at
#asahi-dev · 2023-10-07 11:52
jannegI guess the dart addresses could point to reference frames or scaling_lists
#asahi-dev · 2023-10-07 11:49
jannegis the 10 constant? does your h264 bitstream has a single slice or multiple?
#asahi-dev · 2023-10-07 11:43
jannegi.e. SPS.pic_width_in_mbs_minus1
#asahi-dev · 2023-10-07 11:42
jannegI would expect ((width + 15) >> 4) - 1
#asahi-dev · 2023-10-07 11:42
jannegany idea why the dimensions are effectively just shifted right by 3?
#asahi-dev · 2023-10-07 11:40
janneg64 and 4096 are reasonable minimal/maximal dimensions
#asahi-dev · 2023-10-07 11:38
jannegthe only appeal of v4l2 m2m stateless is that it is a standardized API
#asahi-dev · 2023-10-07 11:36
jannegnot sure how happpy v4l2 is about a custom which pipes basically binary data through
#asahi-dev · 2023-10-07 11:34
jannegvulkan video would probably mean a drm mode
#asahi-dev · 2023-10-07 11:33
jannegI haven't looked at libva internals but I suppose we could / have to whatever we want as backend and driver
#asahi-dev · 2023-10-07 11:31
jannegif we need an userspace part we should reconsider va-api or vulkan video as those have already userspace
#asahi-dev · 2023-10-07 11:30
jannegif svc and mvc were just annexes they could be easily ignored but unfortunately they seeep through the whoole standard and add complexity
#asahi-dev · 2023-10-07 11:28
jannegeiln: they do what (RE: neon)! neither scaler or gpu can handle the AVD tiling?
#asahi-dev · 2023-10-07 00:07
janneginitial standard is even less but I think the fidelity range extension doesn't make the standaard unreadable

plus ›