ChanServ changed the topic of #asahi-dev to: Asahi Linux: porting Linux to Apple Silicon macs | Non-development talk: #asahi | General development | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Logs: https://alx.sh/l/asahi-dev
derzahl has joined #asahi-dev
<derzahl>
i just saw the blog entry about switching to fedora
<derzahl>
curious, are any redhat folks involved?
derzahl has quit [Remote host closed the connection]
gabuscus has quit []
jeisom has quit [Ping timeout: 480 seconds]
gabuscus has joined #asahi-dev
crabbedhaloablut has joined #asahi-dev
<lina>
eiln: Fake initializing ANE doesn't work. I got it to get through the buffer allocs, but then it actually tries to initialize the ML model and times out on ANE not actually replying...
<lina>
On the other hand initializing ANE has a nice trivial arbitrary u32 write... so if we ever need to mess with firmware memory and can't do it directly... that's one way... ^^;;
loki_val has joined #asahi-dev
crabbedhaloablut has quit [Read error: Connection reset by peer]
tristan2_ has joined #asahi-dev
tristan2 has quit [Ping timeout: 480 seconds]
<lina>
eiln: OMG I got it to "work" with a horrible hack writing to firmware memory... except now I have exactly the same problem as profile #0
<lina>
Sooo.... we need to solve that either way...
<lina>
(the meta thing)
<lina>
I found out the logic changed somewhere in 13.5 (vs 13.2 which I was looking at) and it only enables MLVNR for certain platforms... so I tried clobbering the platform ID in firmware data directly, and restoring it after the config set so we don't break too much, and it worked. Just like profile #0, no MLVNR, no buffer allocs... but still fails on the meta thing.
<lina>
eiln: I think it's just a bug in the buffer logic in the driver... it looks like we only submit one buffer on startup.
<lina>
I tried making it submit more and now it looks like we don't get frames back...
<lina>
I think there's another problem in the buffer callback logic
<lina>
aaaaaa the top bits difference!!!!
<lina>
I think I got some frames... ^^;;
<lina>
Now it crahshes probably because the way I hacked up the buffer logic is wrong...
<lina>
I can probably fix this on wednesday ^^
elvishjerricco has quit [Remote host closed the connection]
<janneg>
sven: did your kernel recipes talk get rescheduled into the void? I remember seeing it scheduled for today at 16:00 (CEST) but currently I don't see it at all on the schedule page
le_petit_segfault has joined #asahi-dev
hightower2 has quit [Ping timeout: 480 seconds]
le_petit_segfault has quit [Ping timeout: 480 seconds]
hightower2 has joined #asahi-dev
jjanzic has joined #asahi-dev
c10l9 has quit []
c10l has joined #asahi-dev
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-dev
<sven>
I’m sick in bed with a fever, so yeah :/
<chadmed>
povik, janneg: the ucm2 files are fucked up on many many levels
<chadmed>
ill fix them and submit a PR
<ellyq>
get well soon, sven!
AnuthaDev has joined #asahi-dev
cylm_ has quit [Ping timeout: 480 seconds]
AnuthaDev has quit [Ping timeout: 480 seconds]
pg12_ has joined #asahi-dev
pg12 has quit [Ping timeout: 480 seconds]
pg12_ has quit [Ping timeout: 480 seconds]
pg12 has joined #asahi-dev
jeisom has joined #asahi-dev
<leio>
I reinstalled Fedora now (though the ESP mount just needed a sudo for diskutil), took the ESP uuid for the m1n1 append from the pre-installed stage1 console output, but still all the same - no working keyboard, stuck at finding the rootfs and can't intervene with keyboard. guest kernel time I guess
<leio>
janneg: Sep 18 00:00:05 fedora dracut-pre-udev[500]: Warning: :: Asahi: Timed out waiting for NVMe device
<leio>
So I just don't have a NVMe device to find a rootfs from
<leio>
oh... dracut-pre-udev[508]: modprobe: FATAL: Module nvme-apple not found in directory /lib/modules/6.5.0-asahi+
* leio
goes updates the initramfs to be the same, sorry for the noise
<janneg>
sven: oh. I hope you'll get better soon
<janneg>
chadmed: even more? I already fixed 1 1/5 issues last week breaking speaker detection
<janneg>
chadmed: I don't think you should hardcode the CardId
<chadmed>
yeah im not going to, thats just for debug right now
<chadmed>
pipewire is doing this annoying thing where its not hiding the speakers when the jack is plugged in, but the jack disappears if theres nothing plugged into it
<chadmed>
i dont think that's a particularly bad thing since it lets users switch between the two pcms at will, but wp folks are telling me its not supposed to do that so...
<janneg>
I suppose that would be a use case covered by "Pro Audio"
<chadmed>
i got blasted for using Pro Audio outside of its intended use case so we have to make this work
<chadmed>
wim and george will not upstream anything that relies on abusing pro audio
<janneg>
no, I mean using both speakers and headphones at the same time is a "Pro Audio" feature, or I'm misunderstanding what ConflictingDevice does
Dementor has quit [Remote host closed the connection]
<chadmed>
its supposed to make the devices mutually exclusive
<chadmed>
so if the headphones are plugged in, speakers should disappear as an option and vice versa
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-dev
<leio>
as a personal note - I hate this behaviour, I want to be able to switch between my headset and speakers without having to pull the cable out and add it back in all the time, just choosing the device in GNOME UI
<chadmed>
yeah thats why i dont really care if it stays like this, its a feature
<chadmed>
but im being told its not _supposed_ to so idfk
<leio>
my Y splitter is already broken after 8 months from all this cable pulling 6 times a day
<leio>
though from the Mac Studio I'm hoping I can just switch between HDMI and headset in the plug, not internal speakers, but that's that and just me
<leio>
(as I don't have separate line out and headset on there like on a PC)
<leio>
(hence got to use the sound output from the monitor for the external speakers)
<povik>
chadmed: i did my best *shrug*
<chadmed>
povik: yeah thats fine, no one handles ucm sanely anyway which is why i was hoping to just use pro audio and not care about any of this
<povik>
there are pitfalls every step of the way, was my experience writing the ucm we have
<povik>
and some i didn't know i fell into until jannau came around and fixed them
<chadmed>
yeah its awful, and like with everything alsa theres like... negative documentation
<janneg>
I'm not seeing anything which would explain why the headphones fail on some system but work on others
<janneg>
I detected the pitfall only via a typo. I deleted random things and it started working aafter I typoed "spk" as "skp"
<povik>
ha! :D
ella-0 has quit [Remote host closed the connection]
dubiousness has quit [Remote host closed the connection]
pldtf has quit [Remote host closed the connection]
dk_ has quit [Remote host closed the connection]
alethkit has quit [Remote host closed the connection]
cds has quit [Remote host closed the connection]
akspecs has quit [Remote host closed the connection]
d4ve has quit [Write error: connection closed]
signaryk has quit [Remote host closed the connection]
probie has quit [Remote host closed the connection]
tsujp has quit [Remote host closed the connection]
coder_kalyan has quit [Remote host closed the connection]
amada95 has quit [Remote host closed the connection]
handlerug has quit [Remote host closed the connection]
rosefromthedead has quit [Remote host closed the connection]
<leio>
janneg: I am skipping the initramfs, made all relevant stuff I could find for NVMe builtin instead of module and it just doesn't see a nvme at all still; any chance it could be something with dtb still?
le_petit_segfault has quit [Ping timeout: 480 seconds]
<leio>
should I see anything about nvm in /sys/firmware/devicetree/base ?
<janneg>
leio: there should be a soc@22000.../nvme@3... directory
<leio>
ls -R /sys/firmware/devicetree/base |grep nvm was silent
<janneg>
that's strange, does it contain anything at all? it should also match on "nvmem"
<leio>
checking; waiting for the timeout to hit to get emergency console
<leio>
dracut does complain about not finding nvme modules and such, but I have them builtin in the kernel and it doesn't find anything with rootwait and no initramfs either
<leio>
(the modules path doesn't match with what's in the initramfs, just using it now to get a shell)
<janneg>
is there anything in /sys/kernel/debug/devices_deferred ?
<leio>
I do have it in the soc stuff, I'm confusing something from before with other checks
<leio>
the debug folder is empty
<leio>
there is a /sys/firmware/devicetree/base/soc@2200000000/nvme@34bcc0000 folder
<janneg>
you have to mount debugfs yourself in the emergency shell
<leio>
devices_deferred is empty
<janneg>
is there anything matching nvme-apple or nvme0 inthe kernel log?
<leio>
no, there's nothing, dmesg |grep -i nvm only gives my kernel cmdline
<janneg>
Are all of NVME_APPLE, APPLE_MBOX, APPLE_RTKIT, APPLE_SART built-in?
hightower2 has quit [Ping timeout: 480 seconds]
dsharshakov has joined #asahi-dev
<dsharshakov>
chadmed: UCM proper should probably allow switching between those two at will when headset is in. Look at Audio in KDE System Settings and there's "port" config. Works for me on HDA
<dsharshakov>
maybe that could be implemented in Asahi UCM config
jeisom has quit [Ping timeout: 480 seconds]
<leio>
sorry, was grabbing dinner
<leio>
janneg: yes, all those are built-in in my config (i just went and made all *APPLE* builtin I could, there's CONFIG_APPLE_RTKIT_HELPER=m though)
<janneg>
other nvme configs? NVME_CORE and BLK_DEV_NVME? the first is probably a dependency
jjanzic has quit [Remote host closed the connection]
jjanzic has joined #asahi-dev
<leio>
janneg: yes, they are builtin; I also have /sys/modules/nvme{,_core} for example
<leio>
the mtp_fabric pm_genpd_init failure is ok?
hightower2 has joined #asahi-dev
dsharshakov has quit [Quit: Page closed]
fredmb has quit [Remote host closed the connection]
AnuthaDev has joined #asahi-dev
lsep has quit [Ping timeout: 480 seconds]
eiln has joined #asahi-dev
<eiln>
am slightly sick too. was whining about the heat a week ago and now the cold has gotten me. woke up at midnight rn
<eiln>
lina: the platform_id clobbering is a disgusting hack, I love it. don't we install 13.5 as default though?
<eiln>
will look at the buffer logic in a bit. that bit has always been hacky, but got even hackier with the single planar hack
cylm has joined #asahi-dev
<eiln>
as a last resort I think I can kick up the ane for this sake. we're not calling face detection, we just need the client up. not sure how the hell that's gonna work in v4l2 though
jeisom has joined #asahi-dev
cylm has quit [Ping timeout: 480 seconds]
le_petit_segfault has joined #asahi-dev
<janneg>
leio: pushed changes #180 to fix hdmi turning off without pd_ignore_unused
<janneg>
do you have APPLE_DART built-in? missing log messages for that, not related to nvme though
<leio>
janneg: yup, CONFIG_APPLE_DART=y
<janneg>
are you 100% sure you're booting the correct kernel? do you have /sys/module/nvme_apple/ (should be even there if built-in)
<leio>
janneg: don't think I'm running the right kernel, it doesn't recreate a Image.gz anymore after changes, just Image...
<leio>
I think that must have been a mishap with the kernel config change and then run_guest_kernel preferred the old arch config Image.gz or something :(
<leio>
not preferred, only supports that
le_petit_segfault has quit [Remote host closed the connection]
loki_val has quit []
AnuthaDev has quit [Quit: AnuthaDev]
<leio>
so in that case I can't be sure #192 alone couldn't have helped with the disappearing picture without pd_ignore_unused, but looks like you found improvements to dtb already that helped too
jeisom has quit [Quit: Leaving]
Dementor has joined #asahi-dev
<janneg>
the device tree was broken, #192 was due to other oddities in the devicetree not necessary but is now. I can simulate the setup on a m2 pro mini
<janneg>
regarding the black screen in the beginning. is it possible that your display is slow to show I picture after initializing/configuring it? I have one as well and booting is so fast that the display only shows a picture when linux is already booting
<leio>
I'm not sure about a black screen in the beginning; I was saying that the monitor didn't get log entries shown for a while that did end up in the m1n1-sec output, unlike before
<leio>
basically systemd messages aren't reaching vt1 anymore after the latest dtb updates, just initial kernel stuff and then eventually the emergency shell (I need to correct my root partition)
<leio>
while before I saw all on VT1, including the dracut-initqueue pages of spam before it drops to emergency shell
<leio>
"[ OK ] Reached target basic.target - Basic System." and such, that isn't visible anymore and I believe all I changed was update #180 patchset (and use old kernel image still) - same after re-doing Image.gz too.
<leio>
maybe that's more correct even, dunno, just observation
<janneg>
that sounds more like it's things controlled by your initrd and boot kernel cmdline parameters
<leio>
yeah, very weird, it's like it started behaving 1 boot before it should have like that (still with old kernel afaik)
<leio>
but yeah, it looks like now I'm just fighting with further missing modules and the wrong initramfs for my kernel compile; I shall overcome that with a bunch more builtins, get into Fedora proper and build a clean patched kernel there
<leio>
and now when I re-enabled plymouth, I started seeing systemd messages on monitor again; just too much magic to understand that :)
<leio>
(but no plymouth screen, just took away the plymouth disabling kernel cmdlines)
<leio>
and I have a "Welcome to Fedora Linux Asahi Remix 38!" screen \o/
<leio>
(without keyboard or mouse to do anything in it :( )
<eiln>
update we can decode with the firmware. trigger "mailbox" by literally writing to nvic with the physaddr it jumps to. I just really, really don't want to do it this way. I'm gonna fuzz the params to try to get the intermediate passes
chadmed has quit [Remote host closed the connection]