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
pjakobsson has quit [Remote host closed the connection]
lactose has joined #asahi-alt
pjakobsson has joined #asahi-alt
chadmed has quit [Ping timeout: 480 seconds]
chadmed has joined #asahi-alt
lactose has quit [Read error: Connection reset by peer]
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-alt
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-alt
chadmed has quit [Read error: Connection reset by peer]
chadmed has joined #asahi-alt
lactose has joined #asahi-alt
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-alt
lactose has quit [Read error: Connection reset by peer]
lactose has joined #asahi-alt
lactose has quit []
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-alt
chadmed has quit [Remote host closed the connection]
chadmed has joined #asahi-alt
minecrell has quit [Quit: Ping timeout (120 seconds)]
minecrell has joined #asahi-alt
Tofu has joined #asahi-alt
Tofu has quit []
Tofu has joined #asahi-alt
Tofu has quit []
kidplayer666 has joined #asahi-alt
mkurz has quit [Ping timeout: 480 seconds]
mkurz has joined #asahi-alt
RoundDuckKira has quit [Quit: RoundDuckKira]
RoundDuckKira has joined #asahi-alt
john-cabaj has joined #asahi-alt
kidplayer666 has quit [Quit: Connection closed for inactivity]
RoundDuckKira has quit [Quit: RoundDuckKira]
john-cabaj has quit [Quit: john-cabaj]
john-cabaj has joined #asahi-alt
kidplayer666 has joined #asahi-alt
xcpy00 has joined #asahi-alt
minecrell3 has joined #asahi-alt
pounce_ has joined #asahi-alt
snuck has joined #asahi-alt
qyliss_ has joined #asahi-alt
yuka_ has joined #asahi-alt
mini__ has joined #asahi-alt
mx08_ has joined #asahi-alt
mps_ has joined #asahi-alt
minecrell has quit [reticulum.oftc.net helix.oftc.net]
flokli has quit [reticulum.oftc.net helix.oftc.net]
qyliss has quit [reticulum.oftc.net helix.oftc.net]
yuka has quit [reticulum.oftc.net helix.oftc.net]
pounce has quit [reticulum.oftc.net helix.oftc.net]
xcpy0 has quit [reticulum.oftc.net helix.oftc.net]
sneak has quit [reticulum.oftc.net helix.oftc.net]
mps has quit [reticulum.oftc.net helix.oftc.net]
mx08 has quit [reticulum.oftc.net helix.oftc.net]
mini_ has quit [reticulum.oftc.net helix.oftc.net]
lucus has quit [reticulum.oftc.net helix.oftc.net]
ahdeb[m] has quit [reticulum.oftc.net helix.oftc.net]
LeandroCunha[m] has quit [reticulum.oftc.net helix.oftc.net]
matrix_ds has quit [reticulum.oftc.net helix.oftc.net]
mixi has quit [reticulum.oftc.net helix.oftc.net]
mini__ is now known as mini_
xcpy00 is now known as xcpy0
mini_ has quit [Killed (charon.oftc.net (Nick collision (new)))]
mx08 has joined #asahi-alt
mini_ has joined #asahi-alt
yuka has joined #asahi-alt
LeandroCunha[m] has joined #asahi-alt
minecrell has joined #asahi-alt
ahdeb[m] has joined #asahi-alt
pounce has joined #asahi-alt
flokli has joined #asahi-alt
mixi has joined #asahi-alt
qyliss has joined #asahi-alt
sneak has joined #asahi-alt
mps has joined #asahi-alt
lucus has joined #asahi-alt
matrix_ds has joined #asahi-alt
mini__ has joined #asahi-alt
mixi has quit [Ping timeout: 485 seconds]
mixi has joined #asahi-alt
flokli has quit [Ping timeout: 481 seconds]
yuka has quit [Ping timeout: 496 seconds]
qyliss has quit [Ping timeout: 486 seconds]
sneak has quit [Ping timeout: 486 seconds]
mx08 has quit [Ping timeout: 486 seconds]
LeandroCunha[m] has quit [Ping timeout: 496 seconds]
larsjel has joined #asahi-alt
minecrell has quit [Ping timeout: 484 seconds]
lucus has quit [Ping timeout: 484 seconds]
mps has quit [Ping timeout: 484 seconds]
pounce has quit [Ping timeout: 484 seconds]
mini_ has quit [Ping timeout: 484 seconds]
mini__ is now known as mini_
flokli has joined #asahi-alt
qyliss_ is now known as qyliss
LeandroCunha[m] has joined #asahi-alt
possiblemeatball has joined #asahi-alt
eiln has quit [Ping timeout: 480 seconds]
mps_ is now known as mps
opticron has quit [Read error: Connection reset by peer]
opticron has joined #asahi-alt
<mps>
I see mesa asahi-20240123 is tagged. Is it ok to use it witj asahi-6.6-15 kernel
<mps>
with*
<cy8aer>
looks like, I just try it out...
<mps>
cy8aer: yes, also I rebooted to this combo
<cy8aer>
... still compiling ...
<mps>
for now all works
<mps>
except soffice, but this is another story, probably I messed something, it didn't worked earlier so no problem
<cy8aer>
Ok, it is now in my unstable repo and it looks like it is working. How can I check if pipewire uses CONFIG_UCLAMP_TASKS?
<cy8aer>
Yep, I had a typo in my config. Need to compile the kernel with CONFIG_UCLAMP_TASK again!
<j`ey>
you manually edit the config?
<cy8aer>
No the Glanzman construct has a config.txt which is then put into the kernel's .config
<cy8aer>
But it is a text file. And the kernel compilation does changes in the config then.
<cy8aer>
What about CONFIG_UCLAMP_TASK_GROUP - it is not set by now?
<j`ey>
fedora does also set tha
<cy8aer>
sound logical somehow if you want to set a cgroup...
<mps>
I didn't set UCLAMP_TASK_GROUP
<cy8aer>
(sounds)
<mps>
enabled*
<cy8aer>
"UCLAMP_TASK_GROUP" (and the other UCLAMP stuff) enables the user space setting which enables a process running on low power cores with low frequencies. Saves power when running pipewire sound (I think it is running on the low power cores with 200MHz then?)
<j`ey>
I dont think they go that low, I think it's 600mhz, but it just places them on the E cores, they might still be at 1-2ghz
<cy8aer>
Oh, then I misinterpreted some old mastodon post (?) of marcan.
<mps>
I never seen lower than 600MHz
<cy8aer>
but audio is then stuck on the E-cores.
eiln has joined #asahi-alt
<opticron>
For some reason, I didn't think it *forced* it to be on the e-cores, just told the kernel that its time-averaged processor demand was small enough to run on an e-core at its lowest freq
<j`ey>
yes that sounds correct
<cy8aer>
Is there a way to see if this works?
<opticron>
glad I have some amount of information retention :D
<j`ey>
forcing it would be with taskset
<chadmed>
cy8aer: uclampset -p [pid of pipewire]
<cy8aer>
chadmed: thanks for the command - and it works.
nela has quit [Ping timeout: 480 seconds]
eiln has quit [Ping timeout: 480 seconds]
ashi has joined #asahi-alt
<ashi>
cy8aer: Thank you for the new packages, everything works as expected in linux, however for some strange reason in m1n1 I cannot use the keyboard on my M2 Air 13" anymore. Good that grub auto-boots ;)
<j`ey>
did you update u-boot?
<cy8aer>
Oups, let me see...
<cy8aer>
I am still on asahi-v2023.07.02-4 - what should I use?
<j`ey>
well I was trying to figure out what changed for ashi
<ashi>
j`ey: yes I did. I can call update-m1n1 again.
<ashi>
u-boot-asahi:arm64 2024.01+dfsg-1 arm64
mkurz has quit [Ping timeout: 480 seconds]
<j`ey>
is that from cy8aer or the team banaanas stuff?
<cy8aer>
my stuff is Glanzmann and does not use update-m1n1!
<ashi>
j`ey: debian upstream
<j`ey>
debian.org is slooooow
<j`ey>
it might be that debian's uboot is using upstream u-boot
<j`ey>
ashi: so did that upstream debian install automatically, or did you manually do it?
<ashi>
linux-image-asahi depends on u-boot-asahi so it gets pulled in
<j`ey>
ah, and it's a later version so takes over from the ones built by that script i guess
<ashi>
anyway there is more confusion, because looking at /etc/kernel/postrm.d/zz-update-m1n1 it will call update-m1n1
<ashi>
there is a check for glanzmanns m1n1 package but the m1n1 package is from debian also
<cy8aer>
yes and no: The postrm and preinstall script look for boot.bin and m1n1.bin in /usr/lib/m1n1 and exits 0 because this situation is on glanzmann installs
<cy8aer>
(scripts)
<ashi>
but then are in /usr/lib/m1n1
<ashi>
it does not exit
<ashi>
if [ ! -e /usr/lib/m1n1/m1n1.bin ] && [ -e /usr/lib/m1n1/boot.bin ]; then
<ashi>
but is does exist so it does not exit
<cy8aer>
so what m1n1 version do you have installed?
<ashi>
ii m1n1 1.4.11-6 arm64 Apple silicon boot loader
<cy8aer>
yes and this has: /usr/lib/m1n1/boot.bin and /usr/lib/m1n1/m1n1.bin and let zz-update-m1n1 ignore the update-m1n1
<ashi>
if [ ! -e /usr/lib/m1n1/m1n1.bin ]
<ashi>
doesnt this mean "if this file does not exist"
<ashi>
there is a "!"
<ashi>
anyway, it is getting late, and I did not want to take your time. I can figure this out myself, and if everything is fine at your side, I probably did something weird.
<cy8aer>
hm, is is possible that I edited some stuff there. Please look at /etc/kernel/postinst.d/zz-update-m1n1. Is there the '!' too?
<cy8aer>
(it is possible)
<ashi>
yes, it is there also, I will reinstall asahi-scripts just to make sure
<cy8aer>
crazy: wth my u-boot-asahi (the new one I just built) update-m1n1 works, with the debian upstream one it fails (because it cannot find /usr/lib/u-boot-asahi/u-boot-nodtb.bin - just a moment
<ashi>
cy8aer: yes paths are different - I put a symlink there :P
<cy8aer>
I am working with the upstream asahi-scripts and there is the wrong path?
<ashi>
but update-m1n1 from asahi-scripts wants /usr/lib/u-boot-asahi/u-boot-nodtb.bin
<cy8aer>
I put this information into #debian-bananas right now. Let's see what the devs do with this.
<ashi>
I symlinked u-boot-asahi -> u-boot/apple_m1 in /usr/lib
<ashi>
then everything "worked", it got triggered and no keyboard before kernel boot
<ashi>
I dont know what would happen if I put an exit 0 in the postrm/postinst scripts in the first line - if I would have ended up with a working u-boot/m1n1
<j`ey>
(no keyboard is expected)
<cy8aer>
because update-m1n1 works now I probably would be able to forget the pre built boot.bin now. But my first tries ended up with killing my stub without an efi backup - some people may remember this in august 😉
<ashi>
j`ey: might be but it is a regression for me after updaten the packages (and creating the symlink to make postinst work)
<j`ey>
ashi: for sure, just saying that it's an expected one if you use their packages
<ashi>
For debian there are only "their packages" for that part because cy8aer removes pacakges as soon as they got upstreamed (if I understood that correctly)
<cy8aer>
ashi: my m1n1 builts a boot.bin at compile time and copies it to /boot/efi/m1n1 at install time (with backup boot.bin-2024...). update-m1n1 creates the boot.bin at kernel install time from the kernel dtbs and the m1n1.bin and the u-boot.bin and a backup boot.bin.old. Actually both work (if the paths are set right)
<cy8aer>
more or less as soon as they got upstreamed, right
<j`ey>
well cy8aer should keep uboot :P
<cy8aer>
I do not build the audio stack anymore.
<cy8aer>
Yes, I'll do until there is a agreement what to do at future time.
<ashi>
cy8aer: so if i put an exit 0 in the hooks and then reinstall m1n1 from your repo, then it is expected to have keyboard back?
<cy8aer>
My version of m1n1/u-boot works.
<ashi>
I mean I will end up with a different boot.bin then?
<cy8aer>
Yes. And - just wait. I'll upload a new m1n1/u-boot with faster keyboard response.
<cy8aer>
By default my m1n1/u-boot is not priorised but m1n1 has a subversion greater than upstream. And this trick priorises my m1n1. I did not document a pinning of my packages
<ashi>
cy8aer: but you do not have a u-boot-asahi package right?