#
#asahi-dev · 2026-01-25 00:32
fl0_id__baremetal booting works too, just no serial :/
#asahi-dev · 2026-01-25 00:29
fl0_id__(and i had specified init=/init option)
#asahi-dev · 2026-01-25 00:27
fl0_id__now ideally I just need glasses or a bigger font lol
#asahi-dev · 2026-01-25 00:27
fl0_id__yay i have a shell :) turns out it somehow didn't like the /init, but just specifyig alpine /sbin/init worked
#asahi-dev · 2026-01-25 00:24
fl0_id__(and baremetal)
#asahi-dev · 2026-01-25 00:24
fl0_id__maybe once serial is working, this can be submitted as a minimal device tree
#asahi-dev · 2026-01-25 00:07
fl0_id__smp: Brought up 1 node, 16 CPUs [ 0.044822] SMP: Total of 16 processors activated.
#asahi-dev · 2026-01-25 00:05
fl0_id__https://www.pastery.net/kjqmsd/
#asahi-dev · 2026-01-25 00:04
fl0_id__well at least smp seems to come up, even if init is not working...
#asahi-dev · 2026-01-24 22:25
JamesCalligeros[m]Yeah there we go
#asahi-dev · 2026-01-24 22:25
jannauit uses 8-bit and it uses software cursors with dual displays
#asahi-dev · 2026-01-24 22:22
JamesCalligeros[m]Well, the room
#asahi-dev · 2026-01-24 22:22
JamesCalligeros[m]Ill go back to the hotel and check
#asahi-dev · 2026-01-24 22:22
JamesCalligeros[m]Oh hang on i might be misremembering
#asahi-dev · 2026-01-24 22:21
JamesCalligeros[m]Nope still argb8888
#asahi-dev · 2026-01-24 22:21
jannaukwin does not use ARGB2101010 for cursors now that it is available?
#asahi-dev · 2026-01-24 22:18
JamesCalligeros[m]And make the sw fallback very obvious
#asahi-dev · 2026-01-24 22:18
JamesCalligeros[m]Yeah but will also make cursors look bad
#asahi-dev · 2026-01-24 22:16
jannauI wanted to force sRGB for 8-bit RGB formats but I forgot. I think that would solve the oversaturated colors on TTYs
#asahi-dev · 2026-01-24 22:14
jannauthe solution for the problem for SDR is using a color profile. The only reason kwin dowsn't use the internal one by default is that displays with broken color space information are not uncommon
#asahi-dev · 2026-01-24 22:12
jannauto my understanding oversaturated colors are expected for the Default colorspace with wide gamut displays and without active color profile (at least not uncommon)
#asahi-dev · 2026-01-24 22:10
JamesCalligeros[m]Yeah thats how i was getting around it before
#asahi-dev · 2026-01-24 22:10
jannauI did not look into external displays yet but for the internal mini led display selecting Display P3 as icc profile has the same result as using dcp's sRGB colorspace
#asahi-dev · 2026-01-24 22:09
JamesCalligeros[m]It would be very nice if the drm colourspace property wasnt just used for yuv surfaces but alas
#asahi-dev · 2026-01-24 22:09
JamesCalligeros[m]I dont think theres a sane way to deal with this until the colour pipeline stuff is bedded down
#asahi-dev · 2026-01-24 22:08
JamesCalligeros[m]Your comment in the new code is accurate iff kwin is using the monitors builtin colour profile, but it doesnt do this by default
#asahi-dev · 2026-01-24 22:07
JamesCalligeros[m]Nah its every non-srgb display unless you explicitly enable a colour profile in kwin
#asahi-dev · 2026-01-24 22:06
jannauI thought you wanted sRGB primarily for the internal displays as those are oversatuted
#asahi-dev · 2026-01-24 22:05
JamesCalligeros[m]Hopefully we can do something more sane with the new colour pipeline stuff
#asahi-dev · 2026-01-24 22:04
JamesCalligeros[m]I wont rule out the possibility of the builtin colour profile being wonky, but i dont expect it to be for these monitors
#asahi-dev · 2026-01-24 22:02
JamesCalligeros[m](the tty especially is massively oversaturated but its the tty so not really important)
#asahi-dev · 2026-01-24 22:02
JamesCalligeros[m]Everything still looks awful with native on my bt.2020 monitors
#asahi-dev · 2026-01-24 22:01
JamesCalligeros[m]jannau: which monitors/compositor did you observe doing proper colours with dcp native?
#asahi-dev · 2026-01-24 20:11
fl0_id__jannau, thx. still fails but a bit later maybe. does the compression matter? I used gz
#asahi-dev · 2026-01-24 19:44
jannaufl0_id__: I use https://paste.centos.org/view/faaef47f
#asahi-dev · 2026-01-24 19:37
jannauI've pushed asahi-6.18.7-1 to github with dcp fixes, a rust compile fix and small preparations for m3/m4
#asahi-dev · 2026-01-24 19:35
fl0_id__integralpilot - can you maybe share your minimal init script? I packaged the alpine minimal archive + init script, but it panics, "attempted to kill init" I just mounted proc, dev, sysfs and pts and then did exec /bin/sh
#asahi-dev · 2026-01-24 13:56
sventrue, and iirc there was some fixup of the power-domains we had to do for some chip as well where some parent-child relation was missing
#asahi-dev · 2026-01-24 13:55
jannau I haven't looked at pcie. maybe also additional power states as nmve and pcie are closely related. pcie physical layer is used to connect the actual NAND chips (but not using PCIe as protocol)
#asahi-dev · 2026-01-24 13:51
svencould be that something is missing for M3 there
#asahi-dev · 2026-01-24 13:51
svenm1n1 does quite a bit of work to prepare pcie iirc, see https://github.com/AsahiLinux/m1n1/blob/main/src/pcie.c
#asahi-dev · 2026-01-24 13:46
noopwafelso likely I didn't enable something
#asahi-dev · 2026-01-24 13:46
noopwafelthat would make sense.. I was poking at pcie on my M3 (not sure if anyone looked already?)
#asahi-dev · 2026-01-24 13:43
jannaunot sure about the exact error but it can be also caused by the HW block being powered down
#asahi-dev · 2026-01-24 13:42
noopwafeland thanks, I guess I still need to enable something..
#asahi-dev · 2026-01-24 13:41
noopwafelhmm thank you I was just typing to ask :D
#asahi-dev · 2026-01-24 13:40
pitust(there is a nice ESR decoder at https://esr.arm64.dev, btw)
#asahi-dev · 2026-01-24 13:40
pitustprobably
#asahi-dev · 2026-01-24 13:40
pitustso i guess there isn't mmio behind that address
#asahi-dev · 2026-01-24 13:40
pitustso that's a "Synchronous External abort, not on translation table walk or hardware update of translation table."

plus ›