ChanServ changed the topic of #asahi-alt to: Asahi Linux: porting Linux to Apple Silicon macs | User-contributed/unofficial distribution ports | Logs: https://alx.sh/l/asahi-alt
<marcan>
jannau, j`ey: thanks, pushed asahi-5.19-5 with the fixes (building the package now)
<marcan>
that said: building modules as y with producer depencdencies as m is wrong. It won't explode, but you will get kernel warnings if the module isn't loaded for some reason, saying the probe thread is stuck
<marcan>
fine for reproducing this problem, not actually fine to ship like that
<Glanzmann>
marcan: Is CONFIG_GPIO_MACSMC the producer and CONFIG_HID_DOCKCHANNEL the consumer?
<marcan>
Glanzmann: yes
<Glanzmann>
marcan: Is it fine to have both as '=y'?
<marcan>
and you should be building in the whole macsmc thing anyway, because that's a dependency of the framebuffer stuff right now for gpio, and presumably you want your framebuffer to work without the initramfs
<marcan>
(I still don't know why you don't just use our config as a reference for these things...)
<marcan>
it's fine to have everything as =y if you feel like it... module policy is up to the distro, but unless you actually understand the dependencies involved, you'd do well to follow what Asahi does to start with
<Glanzmann>
marcan: Okay, both are '=y'. I want to stick as close as the debian kernel config as possible, but I'll sync the asahi config options with asahi-config.
<Glanzmann>
possible*
<marcan>
that's what I mean, for the Asahi-related options
<Glanzmann>
I'll sync them.
<mps>
distros like to build as much as possible of modules/drivers with =m, but for development this 'likes' to break things
<Glanzmann>
marcan: With the dockchannel I had CONFIG_HID_DOCKCHANNEL, CONFIG_APPLE_DOCKCHANNEL and CONFIG_GPIO_MACSMC as =y most of the time the dockchannel was to fast for gpio but on once boot it was fast enough with j'eys patch it always works. But I'll sync now all asahi options with the reference config.
<marcan>
with the patch it will work, but it is still wrong because if the module doesn't load it will just get stuck in the probe function forever, and warn after one minute
<marcan>
it makes no sense to build in something and not its dependencies
<marcan>
mps: I know, and that is exactly what I do
<marcan>
everything that is =y in the asahi config is =y for a good reason
<mps>
right
<jannau>
macsmc_gpio was built-in too so there was no missing dependency (at least at the point I asked about)
<jannau>
the wrong -EPROBE_DEFER check was the only reason it was broken
<Glanzmann>
marcan: To my understanding both the were built-in.
<Glanzmann>
something and the dependencies.
<Glanzmann>
jannau: Yes, it was always built in. I tend to have more '=y' than '=m', but I now synced my config with the asahi reference config and put more things to =m ...
<bluetail>
had tried it on my amd system and it would get stuck at initramfs... secureboot is off. So I currently install the mainline aur hope it does the 'work' for me, otherwise, I dont even know if its working on AMD yet.
<Glanzmann>
bluetail: I tried it and it worked for me without sound, but I'm back on 5.19 until it will be released.
<leif>
Ok if you type 'bluetoothctl' to go into the bluetoothctl shell, then just type 'scan on' -- hopefully you should now see your device show up here (make sure it's in pairing mode), then just 'connect mac-address'
<isoriano>
leif, Great .. it works from the shell and connecting as you mention it!
<Glanzmann>
Cool.
<leif>
nice! yeah, it's always better to try things at the lower level first. The higher-level stuff will eventually get sorted out.
<isoriano>
Thx .. ok, transmission seems distored in between but at least some sounds.
<j`ey>
isoriano: 2.4 or 5ghz wifi?
<isoriano>
5 GHz .. BandModifier5Ghz=10
<isoriano>
in /etc/iwd/main.conf
<isoriano>
[Rank]
<isoriano>
Btw .. just in case you are interested in Debian aarch64 use cases :-) I do not give it up .. so running illumos-aarch64 in qemu-aarch64 on debian-aarch64 :-) https://ibb.co/vkGZ1P3
leif has quit [Quit: Leaving]
<isoriano>
ok sound works properly .. if you leave the bluetoothctl shell .. seems the scanning was causing the distortions
<isoriano>
exit
<isoriano>
exit the shell
<Glanzmann>
I see. Good to know.
<isoriano>
Wow .. so happy .. was even playing a round RTCW and it is really playable at 25xx resolutions
isoriano has quit [Quit: Leaving]
isoriano has joined #asahi-alt
kode54 has joined #asahi-alt
<mps>
I just wanted to look at BT and see this in dmesg out `2.344168] hci_bcm4377 0000:01:00.1: Unable to load firmware; tried 'brcm/brcmbt4387c2-apple,madagascar-u.bin' and 'brcm/brcmbt4387c2-apple,madagascar.bin'`
<mps>
so I need to unpack new firmware
<mps>
it is /boot/efi/asahi/all_firmware.tar.gz I guess
<mps>
but I'm not sure that simple tar will work
<mps>
oh, looks like untar /boot/efi/vendorfw/firmware.tar will be enough
<Glanzmann>
Yep, if you stub is new enought, than this is enough.
<Glanzmann>
your*
<_jannau_>
depends on when you installed. asahilinux has asahi-fwextract that regenerates the vendorfw firmware.tar from all_firmware.tar.gz
<_jannau_>
it's a packaged version the installer's firmware code
<mps>
_jannau_: yes, I know, and for alpine I used this alarm script and changed it a bit
<mps>
but my experience with BT is nearly zero, didn't used for more than a decade or two
<mps>
installed alpine bluez pkg, run /etc/init.d/bluetooth, run bluetoothctl, and stopped there
<mps>
at least driver is initialized and daemon works, rest will be on some alpine dev/user to test
<_jannau_>
depending on where you are you could scan in bluetoothctl and see if there any devices
<isoriano>
Good evening, so nothing to test tonight? :-)
<isoriano>
But hey one thing .. if I remember correctly on Asahi I could switch screen resolutions .. are these some presets that get shipped with Asahi?
_jannau_ has joined #asahi-alt
<mps>
isoriano: I think only with m1n1
<mps>
i.e. on boot and not on macbooks, but minis
Votes78 has joined #asahi-alt
_Votes78 has quit [Ping timeout: 480 seconds]
isoriano has quit [Quit: Leaving]
isoriano has joined #asahi-alt
<isoriano>
@Glanzmann, sorry forgot to ask .. as I wanted to give my qemu-aarch64 a run and as part I setup the NIC and get the following error with the kernel you send me yesterday
<isoriano>
iptables: Failed to initialize nft: Protocol not supported