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
chadmed_ has quit [Remote host closed the connection]
zumi_ has joined #asahi-alt
zumi has quit [Remote host closed the connection]
zumi_ is now known as zumi
chadmed_ has joined #asahi-alt
cow123 has joined #asahi-alt
cow123_ has joined #asahi-alt
cow123 has quit [Ping timeout: 480 seconds]
chadmed_ has quit [Remote host closed the connection]
cow123_ has quit [Ping timeout: 480 seconds]
KxCORP5894 has quit [Quit: Bye!]
KxCORP5894 has joined #asahi-alt
cow123 has joined #asahi-alt
tobhe_ has quit [Quit: leaving]
tobhe has joined #asahi-alt
tobhe has quit []
tobhe has joined #asahi-alt
chadmed_ has joined #asahi-alt
john-cabaj has quit [Quit: john-cabaj]
john-cabaj has joined #asahi-alt
john-cabaj has quit [Ping timeout: 480 seconds]
Halian|Jardin has joined #asahi-alt
<Halian|Jardin>
Hemlo! I got Gentoo running on my M1 MBA (for real this time), but I can't get KDE working. I can start X and get into SDDM, but it doesn't list Plasma as an available DE, and entering my password kicks me back to the login screen. Pls help?
john-cabaj has joined #asahi-alt
cow123 has quit [Remote host closed the connection]
cow123 has joined #asahi-alt
cow123 has quit [Remote host closed the connection]
cow123 has joined #asahi-alt
<chadmed_>
xorg is not supported, please use the wayland session
<chadmed_>
also did you emerge plasma-meta or some random combination of plasma-adjacent packages
<chadmed_>
also did you set your profile to desktop/plasma
<Halian|Jardin>
chadmed_: oh ok. plasma-meta. Yes.
cow123 has quit [Remote host closed the connection]
<chadmed_>
once youve emerged plasma-meta youll have to remove all the packages you manually emerged from @world so that portage doesnt keep them around when there are no dependencies on them
<Halian|Jardin>
Wat
<Halian|Jardin>
Why do I have to uninstall the entire system? </very confused>
cow123 has joined #asahi-alt
cow123 has quit [Remote host closed the connection]
<chadmed_>
no not the entire system, just the plasma packages you manually emerged
<chadmed_>
e.g. plasma-meta depends on plasma-desktop, so if you manually emerged plasma-desktop you will need to run emerge --deselect plasma-desktop
<chadmed_>
sorry that wasnt clear from my initial message
<Halian|Jardin>
OH okay
<Halian|Jardin>
Sorry about that
<chadmed_>
no thats my fault, i said "all packages" which is silly
<chadmed_>
i meant just the plasma-related ones :)
<Halian|Jardin>
Alright, I depcleaned plasma-meta and xorg-server; I don't believe I manually emerged anything else that requires those
* Halian|Jardin
installs Wayland and KDE
<chadmed_>
emerge -a plasma-meta should be all you need to do if your profile is set properly
<chadmed_>
and then the kde application metapackages you want
<Halian|Jardin>
Yeah, that's what I did
zerdox has quit [Ping timeout: 480 seconds]
<Halian|Jardin>
Hrm, there's no `startplasma-wayland` command :(
<chadmed_>
can you paste the output of `eselect profile list` somewhere
<Halian|Jardin>
I distinctly remember running `sudo emerge -av plasma-meta`
<chadmed_>
maybe a package failed to emerge
<Halian|Jardin>
Maybe
* Halian|Jardin
tries again
john-cabaj has quit [Ping timeout: 480 seconds]
<Halian|Jardin>
Hrm, ffmpeg failed to emerge
<chadmed_>
add -soc to your USE
<Halian|Jardin>
OK
<chadmed_>
some stupid hand-tuned bullshit assembly for rpi that got magically enabled for all arm64 builds because rpis are the only arm64 machines ever apparently
<chadmed_>
i should probably raise a bug about that
<Halian|Jardin>
Ugh
<Halian|Jardin>
Third time's the charm, hopefully
<chadmed_>
if you use mpv youll have to add -soc to USE for that as well
<Halian|Jardin>
I just put it globally for now
<Halian|Jardin>
motherFUCKER it still failed to compile
<sam_>
it's a gcc thing with native on Some ARM64 CPUs Where I Haven't Worked Out The Pattern
<Halian|Jardin>
Ugh
<chadmed_>
oh more gcc nonsense no wonder i havent seen it
<chadmed_>
for m1 you can safely set march=armv8.5-a and for m2 you can do armv8.6-a
<Halian|Jardin>
Alrighty
<chadmed_>
or you can try bootstrap a systemwide clang toolchain but here be dragons
<sam_>
what annoys me is every so often I get in the mood to at least try figure it out a bit more, and then I can't reproduce it in the chroot I try
<sam_>
just pisses me off
<Halian|Jardin>
:(
<sam_>
it's a mix of ffmpeg being clever and obviously gcc getting something wrong
* Halian|Jardin
offers virtual hugs
<sam_>
<3
<chadmed_>
its probably related to gcc also exploding and dying at trying to figure out core features on these machines (and perhaps other heterogeneous systems?)
<Halian>
sam_: <3
<sam_>
I did try to get one of the gcc arm people's attention and I sort of succeeded but then they got distracted with sometihng else
<sam_>
thing*
<Halian>
With -march=armv8.5-a -mtune=armv8.5-a it fails in the EXACT SAME WAY. With just -march=armv8.5-a it compiles. Grumble grumble.
<chadmed_>
forget mtune gcc has no core model for these machines
<Halian>
Ahh
<Halian>
As a hackintosher, I recognize Works On My Machine™ when I see it lol
<chadmed_>
clang is slightly better but its core model is for cyclone or something which was one of apples first arm64 cores lmfao
<Halian|Jardin>
LMAO
<chadmed_>
i still just recommend to everyone to switch to clang systemwide on these machines because youll get random crap like this with gcc because it can't into HMP
<chadmed_>
sam_: do you actually have an AS mac?
<sam_>
nope
<chadmed_>
ah bugger, was gonna ask if you could quickly flash the livecd and try boot it
<sam_>
it just happens for me on one of our dev machines
<chadmed_>
i want at least one other successful boot before i go and spruik its existence
<sam_>
sorry!
<sam_>
leio should be up soon at least
<chadmed_>
yeah
<sam_>
honestly I would've bought one of them but I got screwed hard with a butterfly kb and I'm sour about it still
<chadmed_>
if we are going to go with this for an official install mechanism now that it works i cant host it on my personal infra because its too slow and im on the arse end of the world
<Halian|Jardin>
D:
<Halian|Jardin>
I got the M1 MBA because Walmart had them for $700 and I was in the market for a new laptop
<sam_>
we can host it but we'd like to do the build in the official machinery if so
<chadmed_>
yeah absolutely, im just cleaning up the spec files to push to GH now :)
<chadmed_>
i dont expect official infra to host my scuffed unsigned builds
<Halian|Jardin>
I would offer but I don't think off the top of my head that my host would appreciate getting hit that hard ^^;
<Halian|Jardin>
Otherwise I'm not limited by space, just file count
<chadmed_>
i dont care about space, more just speed and my ISP asking questions :p
<Halian|Jardin>
Lol
chadmed_ has quit [Quit: Page closed]
<Halian|Jardin>
I got KDE working \o/
zerdox has joined #asahi-alt
zerdox has quit [Remote host closed the connection]
zerdox has joined #asahi-alt
<chadmed>
lfggg
zerdox has quit [Remote host closed the connection]
zerdox has joined #asahi-alt
espo has joined #asahi-alt
<chaos_princess>
chadmed: i can test out your iso, what's the process, just install it with genstrap.sh?
zerdox has quit [Ping timeout: 480 seconds]
zerdox has joined #asahi-alt
<chadmed>
chaos_princess: nope! flash to usb, boot from usb
<chadmed>
you can use an existing install's u-boot, i just need to know it boots. its essentially just a standard gentoo arm64 minimal image apart from the kernel
<chadmed>
if this works i can retire genstrap and the whole install-via-fedora thing entirely which is a massive PITA
<chadmed>
(we would also be the first install-like-you-would-on-a-pc distro which is neat)
<chaos_princess>
install-like-on-pc is more effort, you need to find a usb stick to do it :P
* chaos_princess
might be spoiled from only using x86 machines that have a bmc for quite a while
<chadmed>
i had trouble finding one lying around to test it too heh
<chaos_princess>
chadmed: `run bootcmd_usb0` from u-boot shell? Cause that ended up booting systemd-boot from my existing install on internal
<chadmed>
chaos_princess: i actually dont think that works anymore, have a look at the logs for this channel someone found a set of commands that work last night some time i think
<chaos_princess>
Ended up "deleting" the esp. it boots, but then fails to find the modules for apple-mailbox and nvme, and then proceeds to reboot.
<chadmed>
ergh
<chadmed>
i mustve busted something then
zerdox has quit [Ping timeout: 480 seconds]
zerdox has joined #asahi-alt
u154ss has joined #asahi-alt
<u154ss>
@chadmed "if this works i can retire genstrap and the whole install-via-fedora thing entirely which is a massive PITA ". Would heartily agree... I will test your new .iso soon. Busy at the moment.
<leio>
I'm around, but I'm not sure I can really help beyond chatting atm until next month
u154ss has quit [Remote host closed the connection]
<chadmed>
leio: no worries at all
maxmatteo has joined #asahi-alt
<j`ey>
chadmed: so this mean you can just asahi-installer -> UEFI -> USB into gentoo? no intermediate fedora step
<chadmed>
yep! once it works of course...
<cy8aer>
That is the way we install e. g. debian dists.
<chadmed>
i booted successfully and even got wifi up last night so idk why this scuffed build chaos_princess just tried doesnt work... im spinning another one out now and will hopefully have it available before i go to the gym
<chadmed>
cy8aer: okay so we will be second, not first :p
<cy8aer>
😉
<cy8aer>
But IMHO @Glanzmann already built some modified install.sh
<cy8aer>
So we did it the other way round.
maxmatteo has quit [Quit: Leaving...]
zerdox has quit [Ping timeout: 480 seconds]
<mps>
chaos_princess: this I found yesterday `boot_targets "usb" ; setenv bootmeths "efi" ; boot`
<mps>
u-boot 2024.04
<mps>
hm, I need more coffee, it is `setenv boot_targets "usb" ; setenv bootmeths "efi" ; boot`
<mps>
also, yesterday I updated installer for alpine and tested it, but noticed that I've set 'root=$partition' wrongly in grub.cfg
<chadmed>
yes you should be able to flash it to a usb stick and boot from it using the commands provided by m p s
<chadmed>
leio sam_: assuming this works, how do we go about getting official builds done and advertised? is that an -arm thing or a -releng thing?
<chaos_princess>
chadmed: same error except this time it also fails to find a disk with a "Gentoo" Label
<chadmed>
aaaaaaaaa
<chadmed>
i legit did not change anything between the build i tested that worked
<chadmed>
goofy ahh releng
u154ss_ has joined #asahi-alt
<u154ss_>
@chadmed et all should I hold off?
<chadmed>
u154ss_: yeah i guess if it aint workin
<u154ss_>
@chadmed will wait then..
<leio>
chadmed: both, I suppose
qdot has joined #asahi-alt
<chadmed>
hmm dracut must be being sussy because both modules it says are missing are built and in /lib/modules
<chaos_princess>
Do you have them in initrd or on root fs?
<chadmed>
they should be in the initramfs since theyre installed by our asahi dracut module
Guest10498 has quit [Quit: Bridge terminating on SIGTERM]
rhysmdnz has quit [Quit: Bridge terminating on SIGTERM]
Jamie has joined #asahi-alt
Jamie is now known as Guest10646
rhysmdnz has joined #asahi-alt
zerdox has quit [Ping timeout: 480 seconds]
u154ss has quit [Remote host closed the connection]
<chadmed>
the kernel has apple-mailbox builtin so it should not be eating shit looking for that at all
<chadmed>
oh i think i know why
<chadmed>
jannau: we need to change our localversion from 6.8.9-p7-asahi to 6.8.9_p7-asahi
<chadmed>
ill do that tomorrow im eepy
<jannau>
for asahi-sources? why? at least the kernel eclass has code to deal with '-' vs. '_', I don't remember if I checked other kernel ebuilds doing the same. dracut however should only care about 6.8.9-p7-asahi
<chadmed>
no this is asahi-kernel
<chadmed>
catalyst only supports dracut with dist-kernels
<chadmed>
the issue is that it uses equery to get the kernel version to pass into dracut --kver, and equery uses ${PV}
<chadmed>
so the first build i did, i just hard-coded our actual kver into the call to dracut
<chadmed>
the second time, i wrote a thing to mv the module dir to where the upstream catalyst expects it to be
<jannau>
ok. I was looking at that for asahi-kernel
<chadmed>
obviously once it goes into the initramfs in the wrong dir the kernel cant find it
<chadmed>
so the issue is one of skill on my part, should have realised that little trick wasnt going to work :p
<jannau>
if it uses ${PV} isn't '-asahi-dist' a problem as well?
<chadmed>
no its okay with that, dunno where it gets it from havent looked at the code
<jannau>
or does it append that automatically from $name-kernel?
<chadmed>
maybe but how could it know that
<jannau>
I vaguely remmeber code in the kerne eclass which suggests that this might be expected
<chadmed>
yeah we overrode localversion in the first place because of that
<chadmed>
but we just need to change - to _ in that string and everything will work out fine
<chadmed>
ill do it tomorrow without a revbump since we're not breaking existing installs
john-cabaj has joined #asahi-alt
zerdox has joined #asahi-alt
u154ss_ has quit [Remote host closed the connection]
<sam_>
chadmed: -releng but say leio and I endorse it
espo has quit [Remote host closed the connection]
espo has joined #asahi-alt
cylm_ has joined #asahi-alt
cylm has quit [Ping timeout: 480 seconds]
hightower3 has joined #asahi-alt
hightower2 has quit [Ping timeout: 480 seconds]
jacksonchen666 is now known as Guest10670
jacksonchen666 has joined #asahi-alt
Guest10670 has quit [Ping timeout: 480 seconds]
zerdox has quit [Remote host closed the connection]
zerdox has joined #asahi-alt
cow123 has joined #asahi-alt
cyrinux30 has quit []
cyrinux30 has joined #asahi-alt
<Halian|Jardin>
brb
Halian|Jardin has quit [Quit: WeeChat 4.2.1]
halian2 has joined #asahi-alt
<halian2>
Back
halian2 is now known as Halian|Jardin
<Halian|Jardin>
Now there is the problem of “I get no sound”
cylm has joined #asahi-alt
cylm_ has quit [Ping timeout: 480 seconds]
zerdox has quit [Read error: Connection reset by peer]