<Retr0id>
APFS Container Resize error code is 49187
<Retr0id>
Error: -69606: A problem occurred while resizing APFS Container structures
<tpw_rules>
bleh, there was another person with corruption too
AmyLarane[m] has joined #asahi
<tpw_rules>
open disk utility and run the First Aid
<rqou_>
been playing around and noticed that running qemu-system-aarch64 with KVM enabled seems to make the trackpad more "stutter-y" (and with possibly more "Read package crc mismatch" errors in dmesg). is this a known issue?
<Retr0id>
side note, my system has kernel panicked a few times in the past
<Retr0id>
so my fs potentially had latent corruptions
<tpw_rules>
not a good look for apple if this is the first time a lot of users are running fsck
<tpw_rules>
anyway that fixed the last person's problem. make sure your backups are good first :)
<marcan>
yeah... let's just say I'm not impressed with Apple's APFS implementation
<marcan>
I've had to do APFS data recovery for friends twice already
<marcan>
the installer resize just calls diskutil, so if that explodes it's Apple's fault :-)
<tpw_rules>
i still am heartbroken that ZFS didn't make it
<marcan>
Retr0id: the trackpad clicking is indeed handled in software in macOS
<marcan>
not on linux though
<marcan>
it's a different mode or something
<marcan>
it's very obvious when you run macOS under the hypervisor with a lot of tracing :)
<rqou_>
er, i meant that load on the guest seems to make the _host_ trackpad laggy
<marcan>
I wasn't talking about your comment
ichi[sheher][m] has joined #asahi
<rqou_>
ah nvm
<nico_32>
tpw_rules: not sure if it would have been a good fit at the time
<marcan>
and yes, I think there is still some dodginess with SPI
<marcan>
it works well most of the time, and recovers when things glitch
AmyLarane[m] is now known as casually-blue[m]
<marcan>
but there's something going on
<M0x8FF[m]>
this is a very cool project, thank you for not abandoning it or anything :)
sheepgoose has joined #asahi
<_alice>
a lot of people seem to be tripping up on the system freezing while the container resize happens… it might be worth adding a notice in the installer output before it does the command ><
<_alice>
bootcamp assistant has a message like this when it does the same resize for windows installation; "Your Mac will stop responding while resizing the startup volume; this may last for several minutes or hours. Please do not quit Boot Camp Assistant, power off the computer, or restart until the operation is complete."
<Retr0id>
yeah I almost did a hard-reboot when it happened, which would almost certainly have broken my fs even more...
* tpw_rules
was saved being mesmerized by the progress bar
<tpw_rules>
it won't freeze if you don't touch it :)
<nicolas17>
marcan: I have seen you say "this is like the third APFS data recovery situation I find myself in, I have no idea how Apple shipped this..." maybe you're supressing bad memories :P
<Retr0id>
on that note, first aid succeeded, but the shrink still errored out with exactly the same output - will try running first aid again from recovery
<brentr123[m]>
OMG THIS IS FUCKING AWESOME
<Retr0id>
(but first I will make a proper backup lol)
<brentr123[m]>
I deleted the Asahi partition, how do I get my space back?
<tpw_rules>
i saw another issue and thought it was yours
<cobre[m]>
Ah lol
<tpw_rules>
no, that issue is there is some minor apfs corruption
<tpw_rules>
open disk utility and run the first aid repair
<tpw_rules>
preferably make backups beforehand
<M0x8FF[m]>
does asahi have sound support
<tpw_rules>
partially
<cobre[m]>
Ah it says it found some corruption, so i’ll pop into recovery mode real quick to try to fix that
<cobre[m]>
hmm I should probably make sure I have a time machine backup made
<tpw_rules>
yeah, the bug is that the asahi linux installer unceremoniously died instead of telling you to do this
<nico_32>
making a backup is always a good idea
<tpw_rules>
the real bug is that apfs appears to be... fragile
<cobre[m]>
Thanks for the pointer, i'll run make a backup and try again!
<nicolas17>
...I just remembered I probably still have an APFS container for sandcastle on my phone >.>
<nico_32>
i hope linux-apfs have a lot of sanity check
turo has quit []
<M0x8FF[m]>
can i just erase my macOS partition now
<M0x8FF[m]>
from recoveryOS
<tpw_rules>
no
turo_ has joined #asahi
<tpw_rules>
you'll need it for future macOS updates. currently, you have to have installed an OS with macOS so Apple's stuff can update the system firmware before Asahi can use it. afaik fixing that is on the roadmap
<jaime10[m]>
Hey guys, I just finished configuring sway on my M1 air and made a Guide for anyone interested: https://github.com/jaime10a/SwayM1
<jaime10[m]>
I think it should be usefull to configure the touchpad and keyboard.
<jaime10[m]>
If you find any mistakes pls feel free to contribute:)
<M0x8FF[m]>
I see
<M0x8FF[m]>
I have one more question: what if I want to remove the asahi instal
<M0x8FF[m]>
0x8FF: "How do I uninstall it?" is on the alpha release post
<M0x8FF[m]>
0x8FF: "How do I uninstall it?" is on the alpha release post
<nicolas17>
what
<M0x8FF[m]>
Why did that send???
<M0x8FF[m]>
Accident
<M0x8FF[m]>
Sorry fluffychat/matrix is being strange
<tpw_rules>
Retr0id: maybe, try from recoveryOS?
<Retr0id>
That's what I tried second-last
<tpw_rules>
oh
<tpw_rules>
well, that sucks. i guess good that you found out. do you have any snapshots?
<nico_32>
i wonder how many computer have their apfs container corrupted
<tpw_rules>
seems like a lot...
<nico_32>
and how many are unable to apply macOS update
<nicolas17>
damn just freed 4GB on my iPhone by deleting the android-sandcastle volume I forgot I still had
<cobre[m]>
sandcastle was cool, shame it hasn’t been updated in a while
<nicolas17>
asahi's patches for M1 are much cleaner, and now I wonder if support for other SoCs like A10 could be added more easily
<clover[m]>
:[
<Retr0id>
it is simultaneously impressive and concerning how apfs continues to mostly function while in a corrupted state
<brentr123[m]>
macOS protects
<brentr123[m]>
You from diskutil erasing the Mac disk right?
<tpw_rules>
i'm not sure
<clover[m]>
getting segfault when trying to run VS Code arm64 binary hmm
<tpw_rules>
at least with Disk Utility, if you try, it will prompt you go to through a "reset the whole system" flow
<tpw_rules>
clover[m]: expected. chromium is broken with 16K pages right now
<tpw_rules>
(you can install nixos and use its 4K kernel if you like...)
<clover[m]>
oh yeah! i forgot electron is basically chromium
bisko has joined #asahi
nyanpasu64 has joined #asahi
<nico_32>
Retr0id: i know a commercial backup system that have the same issue: silent corruption
<cobre[m]>
to use nixOS, you’d use the Asahi UEFI environment right?
<nico_32>
you discover the issue when you try to restore
<tpw_rules>
cobre[m]: yes. it's all covered step by step in the guide
* tpw_rules
prefers to talk too much than too little
<tpw_rules>
you can even install it with asahi linux already installed
<nyanpasu64>
apparently the apfs resizing hung my m1 mba for over 1 minute, in what started out as a disk io lockup
<nyanpasu64>
it unfroze right as i was in the process of joining this chat to report
<tpw_rules>
nyanpasu64: yeah that's expected, it really should print a scary warning before doing it
<tpw_rules>
macos does if you use the graphical disk utility
<nicolas17>
<_alice> a lot of people seem to be tripping up on the system freezing while the container resize happens… it might be worth adding a notice in the installer output before it does the command ><
<nyanpasu64>
anyway i clicked away from the terminal, so you'd have to print a warning right before/after asking for confirmation (y), not when the freeze occurs
<clover[m]>
yeah apfs resizing freezing everything almost gave me an aneurysm
<tpw_rules>
it doesn't freeze if you just let it do its thing without touching it
<cobre[m]>
yeah it startled me when I first tried it…should definitely add a warning
zx2[m] has joined #asahi
kameks has joined #asahi
<zx2[m]>
agree about the warning, I didn't know it was expected until I saw this chat. Going well now though!
<nyanpasu64>
dual boot on 256 GB SSD is really pushing the limits lol
<tpw_rules>
ftr i also found that if you ctrl-c it while it does the check, it doesn't seem to release whatever disk io lock it has
<zx2[m]>
apart from the lack of warning this is a pretty awesome installer
<nicolas17>
is that "extracting" step also downloading in parallel?
<nyanpasu64>
tpw_rules: you mean the os doesn't freeze if you don't use other apps during resizing (even though that's no better than having a frozen computer)?
<_alice>
the "extracting" step is streaming it from the zip file
<nicolas17>
you mean getting firmware out of the macos ipsw?
<_alice>
it does that for both the getting-firmware-from-apple-cdn step, and the root.img extraction too
<nicolas17>
I thought 'extracting' was for the arch rootfs, not the firmware components
<nicolas17>
ah
<nicolas17>
so root.img is *also* in a zip
<_alice>
yeah, on cdn.asahilinux.org
<nicolas17>
I understand now
<tpw_rules>
nyanpasu64: exactly
<tpw_rules>
i've run it dozens of times, but i have this personal flaw where i get mesmerized and incapacitated by progress bars and scrolling messages
<tpw_rules>
so i've never actually seen it freeze
zamadatix has joined #asahi
<Retr0id>
I think I fixed my fs issues, my fix was: boot recovery, open terminal, diskutil unmountdisk /dev/disk3, fsck_apfs /dev/disk3 (and hit y a bunch of times)
<zamadatix>
Congrats on the alpha everyone
chadmed has joined #asahi
<nicolas17>
tpw_rules: I spent most of today staring at 3 parallel scrolling build logs
<nyanpasu64>
will there eventually be a plan to swap alt and super?
<nyanpasu64>
will speaker dsp be on the level of easyeffects or lower level?
<chadmed>
keyboard reassignment is a user preference thing so we probably wont touch it
<chadmed>
speaker dsp will (hopefully) be inside pipewire itself, exactly as the repo has it with some tweaks pending on advice from the pipewire devs
<jaime10[m]>
Ive been playing around for a while and now I my password doesnt work for sudo anymore.. But when I su $user it works. Any idea how to fix?
<jaime10[m]>
also cant su root cause it has no password..
<chadmed>
the thing is this problem exists for many many devices, and that number is only growing
<chadmed>
so it would be nice to have a universal mechanism for what i do in my asahi-audio repo
jx0 has joined #asahi
<jaime10[m]>
Ok nevermind it worked from tty.. so weird
<sting[m]>
Congratulations on the alpha release everybody
<nyanpasu64>
interesting
<nyanpasu64>
I'm a small time pipewire developer (kinda sorta)
jx0 has quit []
sa1 has joined #asahi
doggkruse has joined #asahi
<chadmed>
the main issue for me is that wim can say "wireplumber already does this" as much as he wants, it doesnt mean that it's the correct, easy, most flexible or user-friendly way to handle it
<chadmed>
in fact its probably the most user-hostile way
bisko has joined #asahi
<nyanpasu64>
personally i dislike speakers that rely on non-integrated dsp, like these speakers i dug out of my garage and have missing treble and ??? (possibly boomy) bass
<nyanpasu64>
and i have to set easyeffects rules on my desktop, and find a different program to play to the speakers from my phone
<chadmed>
yeah but it is part and parcel of these kinds of devices now, and simply refusing to support them properly wont make them go away
<chadmed>
and so what we propose makes it easy and user-friendly to do so
<nyanpasu64>
can you elaborate on wireplumber?
<chadmed>
not really because it has negative documentation and actually doesnt do what wim says it does :P
<chadmed>
i actually dont know what he means by "wireplumber already does this" because it literally doesnt
<nyanpasu64>
and pipewire and wireplumber keep changing around behavior
<nyanpasu64>
like handling requests to listen to monitors, they made it rewrite a flag then reverted then reinstated
<nyanpasu64>
and repeatedly shuffling the id reuse scheme (id reuse used to break pulseaudio apps until they found a way to avoid exposing reused ids to pulse)
<chadmed>
yeah the whole session manager thing is just frought with bugs and showstoppers
<chadmed>
it literally just doesnt work sometimes unless you write a bunch of custom rules
<nyanpasu64>
and last time i checked, id reuse still caused races and ABA problems with wireplumber
<chadmed>
i think wireplumber is too complicated for its own good
<nyanpasu64>
yeah
<nyanpasu64>
it was literally designed for cars
<chadmed>
im currently only interested in supporting pipewire-media-session with these machines because its so dumb that it just obeys what ive told pipewire to do and doesnt try to be a smartarse and apply its own rules
<doggkruse>
Hello, I had a repeating error with the installer script for today's alpha release. Looking for info on how to best provide the feedback
<nyanpasu64>
oh lol
<tpw_rules>
doggkruse: put the log on pastebin.com or similar
bisko has quit [Ping timeout: 480 seconds]
<nicolas17>
chadmed: so speakers went the same route as softmodems now?
<nicolas17>
:D
<nyanpasu64>
(there is no way to tell which speaker is plugged into a headphone jack and apply different effects. ideally you'd put an equalization dongle on each speaker's cable. sadly every extra dongle costs money unlike software DSP.)
<nyanpasu64>
(what about soft modems? i wasn't around then)
<tpw_rules>
looks like it's just having trouble downloading the install data from apple
<chadmed>
software modems would do DSP on the signal in software rather than a real modem which would do it with a dedicated IC
<nyanpasu64>
just for confirmation, pulse/pipewire profiles are a layer of audio/volume/mute routing and volume/mute defaults, on top of the raw alsa controls in alsamixer with dozens of parameters I'm clueless about?
<tpw_rules>
do you have a reliable internet connection? are you in a foreign country?
<doggkruse>
yeah, that's what I gathered but unsure why. I'm in the US with 400/40 cable and its very reliable
<nicolas17>
I think the script was downloading in 1MB chunks right?
<doggkruse>
after the 3rd try it went through
<tpw_rules>
10MB i think
<chadmed>
nyanpasu64: well thats the problem we are trying to solve with this pipewire functionality. it would look at a database of installed profiles and apply the correct one for the plugged in device/s
<tpw_rules>
ok. is it working fine now?
<doggkruse>
3rd try worked fine
<nicolas17>
seems like it could be smarter and download the whole subfile but that would need zip format knowledge
<doggkruse>
just wanted to provide the feedback that it failed
<nicolas17>
wonder if partialzip kinda functionality exists as a python library :P
<tpw_rules>
ok. that's between the script and apple's servers. not sure we can do a lot about that. but thanks for coming to report it
<doggkruse>
np
<tpw_rules>
nicolas17: that's how it works i think? what do you mean by subfile
<doggkruse>
time to reboot into my new install :D
<nicolas17>
tpw_rules: I mean like if a file inside the zip is 25MB (compressed) then make a range request for exactly that 25MB, instead of getting chunks as the zipfile module asks for them
<nicolas17>
but maybe all the stuff being downloaded here is small enough?
<nyanpasu64>
tbh pushing DSP to software, requiring one driver per machine per os, makes me sad... but the company making hardware doesn't care about alternative oses
<tpw_rules>
nicolas17: oh. not sure that would meaningfully improve things? i think it's only downloading like one 1GB file or so
<nyanpasu64>
iirc modern usb audio interfaces can't expose all their capabilities through standard UAC profiles, and rely on Windows-only drivers
<nicolas17>
in a hundred 10MB requests?
<tpw_rules>
yeah. i think each + is another 10MB
<chadmed>
eh, this sort of thing belongs in software these days imo. in fact as i say in the readme of my repo, its a good thing because it means we can replace apple's shit sounding DSP with our own
<chadmed>
it really isnt that much work to "support" devices, and whether or not its actually good or bad is irrelevant
<chadmed>
it is happening, and has been happening for over a decade, and is going to keep happening
doggkruse has quit [Read error: Connection reset by peer]
<nyanpasu64>
i don't exactly like dsp the way it's been done before
<chadmed>
we can either choose to support it properly so that users have a halfway decent experience with their devices, or we can sook and whinge about it and refuse to implement basic support for it and alienate our users
<nyanpasu64>
dell laptops which don't detect headphones plugged in without installing realtek waves bloatware
<nyanpasu64>
and if i disable all effects the speakers sound bad, and otherwise the headphones sound bad with effects i don't want
<chadmed>
well yeah thats a silly way to do it
<chadmed>
luckily pipewire already has the functionality to apply effect chains to only a specific node
<chadmed>
which is what i leverage
<nyanpasu64>
interesting
<chadmed>
what we need from pipewire is the ability to hide the targeted hardware device from users so that they cant select it
<chadmed>
as well as a generalised way to apply these parameters so that we dont have a billion different projects duplicating the functionality randomly
<chadmed>
which is how ALSA UCM and other associated crap functions now
<chadmed>
which as i say, is fine for embedded devices that are usually closed off/limit user interaction to a walled garden anyway
<chadmed>
but is wholly unacceptable for these sorts of machines
<nyanpasu64>
lol macos stereo widener
<nicolas17>
I saw a twitter thread (in spanish) about the state and history of the Linux audio stack
<chadmed>
nicolas17: the sad thing is that the stack doesnt actually suck, what sucks is that you need to be proficient with at least 3 different scripting languages and an audio engineer to actually make it not suck
<nyanpasu64>
one of the neat things about pipewire (and jack) is having signal graphs with a fixed latency no matter how many components you chain
<chadmed>
it just sucks OOB for end users, which is the worst possible thing
<nyanpasu64>
so you can erase the latency of convolutional reverb by piggybacking off the period/quantum size
<nyanpasu64>
I'm half an audio engineer lol
<nyanpasu64>
(convolutional eq as well)
<chadmed>
i was one for the better part of a decade, doing serious live reinforcement stuff for 5-6 of those years
<chadmed>
but like, joe blow isnt
<chadmed>
and joe blow isnt going to bother setting up jack autostart scripts and learning lua + asound syntax + pipewire.conf syntax + carla + REW just so his laptop doesnt sound bad
<nyanpasu64>
you mean laptops can not sound bad?
<chadmed>
apply the FIRs in my repo and find out for yourself ;)
<nicolas17>
my father gifted me a Sony wireless headset that was so bad that he returned it
piroko has joined #asahi
<nicolas17>
trying to make a phone call (which would switch to the other bluetooth profile that supports microphone) would sometimes do nothing (silence), sometimes disconnected from bluetooth, sometimes started making horrible noises on the headphones and firmware was so stuck that the power button didn't respond... and sometimes worked
<nyanpasu64>
some of the worst things i've seen are bluetooth receiver audio with over 1 second of latency
<nyanpasu64>
and the car audio with over 1 second of latency, aptX audio compresion (ok), and entering a phone call would switch to 64khz CVSD compression (dreadful)
<nicolas17>
now I have a refubrished Bose headset that even cost less than the Sony
kov has quit [Quit: Coyote finally caught me]
<chadmed>
Buy Other Sound Equipment
<piroko>
asahi successfully installed \o/
phiologe has joined #asahi
<nicolas17>
chadmed: meh I'm not an audiophile :P (surely if I was I wouldn't be using bluetooth?)
matqua[m] has joined #asahi
doggkruse has joined #asahi
<doggkruse>
Hello from Linux on my M1 MBP!
<doggkruse>
The theme of the night: everything works on the third try.... Wifi didn't load FW on the first boot, loaded but non-operational on the second and the third boot it works
dirbaio[m] has joined #asahi
dirbaio[m] is now known as dirbaio
<opticron>
hmmm...I'm thinking the kernel I've got on debian doesn't have the requisite audio changes to test speakers
<chadmed>
its disabled in the dt
<chadmed>
you need to edit the dt
<opticron>
ah, that's the step I missed
<chadmed>
yeah ill need to update the readme
<chadmed>
or maybe not, i dont want people yelling at me that "i" blew up their speakers
doggkruse has quit [Quit: Konversation terminated!]
<opticron>
yeah...I'm making sure to reduce volumes to 30% or so before trying anything
doggkruse has joined #asahi
<nyanpasu64>
wifi dies when i logout and back in
<nyanpasu64>
and possibly randomly otherwise
<nyanpasu64>
i have to reconnect to a specific network
<chadmed>
you can reduce to 70% and it still works
<chadmed>
any higher and it starts clipping, but nothing should blow up
<chadmed>
yeah thats common with brcmfmac chips, happens on my other machines too
AbhayNayar[m] has joined #asahi
<nicolas17>
fullmac means the wifi protocol is implemented in the firmware and not in the kernel, right?
<chadmed>
not sure
bisko has joined #asahi
marvin24_ has joined #asahi
linearcannon has joined #asahi
silverocean[m] has joined #asahi
<nicolas17>
"A FullMAC device hides the complexity of the 802.11 protocol from the main CPU, instead providing an 802.3 (Ethernet) interface; a SoftMAC design implements only the timing-critical part of the protocol in hardware/firmware and the rest on the host."
<nicolas17>
hrm
marvin24 has quit [Ping timeout: 480 seconds]
bisko has quit [Ping timeout: 480 seconds]
jochum has joined #asahi
jochum_ has quit [Ping timeout: 480 seconds]
<nicolas17>
I'd be interested in accessing AWDL from Linux >.>
<silverocean[m]>
Has anyone ran into "Failed to find boot partition `run distro_bootcmd` to boot from any partition"?
eroux has joined #asahi
<tpw_rules>
silverocean[m]: you are using the official asahi linux distro?
<silverocean[m]>
tpw_rules: Yes, I am. It doesn't seem to get past U-Boot.
<tpw_rules>
ok. this is the first time you've tried?
<silverocean[m]>
tpw_rules: I tried on a MacBook Air and it worked great. Not so much on my 16-inch MacBook Pro (M1 Max).
<tpw_rules>
but i mean on that machine
<silverocean[m]>
tpw_rules: Yes, it is!
<tpw_rules>
ok, that's strange. and this is also the first linux install on it? you just have macos and then this is the first time you've run asahi and you haven't been able to boot into it at all on that machine?
<silverocean[m]>
tpw_rules: yep, I only had macOS before, and it's my first time installing Asahi on it. Doesn't boot into it at all.
<tpw_rules>
ok. does it boot if you do that `run distro_bootcmd` (without the `)?
<silverocean[m]>
tpw_rules: It doesn't boot. Instead I get "Device 0: unknown device".
<tpw_rules>
can you take a picture or so of the output of `part list nvme 0`?
<nicolas17>
everything went fine in the installer and it's failing on first boot?
<silverocean[m]>
nicolas17: yep, everything went fine in the installer
<tpw_rules>
there should be 5 partitions. but it sounds like something went wrong writing the partitions in the installer
<silverocean[m]>
tpw_rules: there's no output when I run that.
<tpw_rules>
you did select one of the asahi linux options and not "UEFI environment only" right?
<silverocean[m]>
looks like the nvme drive isn't recognized?
<silverocean[m]>
tpw_rules: yep, I chose option 1 (the full distro)
<tpw_rules>
that's worrying. try rebooting the machine. i wonder if it's just some fluke
<silverocean[m]>
tried to reboot several times
<nicolas17>
does macOS boot work? ._.
<silverocean[m]>
I wonder if this has anything to do with the fact that the Mac has 4TB storage
<silverocean[m]>
macOS boots fine!
<tpw_rules>
does anything show up under `nvme info`?
<silverocean[m]>
tpw_rules: nothing shows up
<tpw_rules>
what if you do `nvme scan`?
<silverocean[m]>
tpw_rules: let me try that in a moment
<nyanpasu64>
...and the classic hang at shutdown
<nyanpasu64>
personally I'd increase the systemd tty log level by default
<nyanpasu64>
i also need to enable alt sysrq
<nyanpasu64>
is this stuff possible on asahi?
<tpw_rules>
it's 99.8% arch linux so sure
<tpw_rules>
my nixos configs come by default with console.loglevel=7 debug earlycon
Skirmisher has joined #asahi
<roxiun[m]>
probably a stupid question but do all wms work on asani? like i3 or hypr
<tpw_rules>
i repeat my answer :)
<tpw_rules>
it's 99.8% arch linux, so if there's builds of it for aarch64 it will work
<roxiun[m]>
so there isn’t any difference because its arm right?
<roxiun[m]>
ah okay
<roxiun[m]>
its just on a basis if there are aarch64 builds
<tpw_rules>
well i can't tell you that, but the problem wouldn't be specific to m1 or asahi
<Denver>
I want to switch so bad, but I fear it's just not ready for my use case.
brent_ has left #asahi [#asahi]
bisko has joined #asahi
<Dcow[m]1>
for me - it's dual boot now
<brent>
is it normal for irc clients to show ip addresses?
<Denver>
You can dual boot it?
<brent>
when people join/leave
<Dcow[m]1>
also, I found myself too used to macos shortcust, it's killing me
<Dcow[m]1>
shortcuts*
<Denver>
Sorry for my ignorance, but you are currently dual booting MacOS and Ashi?
<Dcow[m]1>
yeah
<Dcow[m]1>
sure
<Dcow[m]1>
it's by default with installer
<Dcow[m]1>
even preferred for now
<nicolas17>
Denver: FAQ says "Can I dual-boot macOS and Linux? Yes! In fact, we expect you to do that, and the installer doesn’t support replacing macOS at this point"
<Denver>
oh interesting... well, now I'm going to go at it tonight
<Denver>
Sorry, thank you.
<nicolas17>
also note that these Macs are un-brickable
<Dcow[m]1>
also, I am mostly working in bluetooth headphones nowadays. will be great when bluetooth come to asahi, but afaik it's not in the top of the todo list now
<nicolas17>
no matter how badly you screw up (think wiping the entire storage), you can recover macOS via USB from another computer
<Dcow[m]1>
but you still CAN wipe the whole disk from linux
<Denver>
I'm a junior dev and dream of the day when I can contribute.
<Dcow[m]1>
it's still a pretty bad thing for many of us
<Denver>
I'm going to give it a shot tonight
<leon_kernan[m]>
I use a bluetooth mouse with this macbook because i don't have one that plugs in. Can live with trackpad for now.
<nicolas17>
oh you *can* wipe the disk, I mean it won't permanently screw up the machine
<nicolas17>
and you have backups of your data right? ;)
<Denver>
What you are saying is there is always a way to at least get a fresh install of MacOS back
<nicolas17>
yes
<Denver>
Awesome, thank you.
<nicolas17>
and the code to bootstrap that process is in actual ROM
<Dcow[m]1>
leon_kernan: I am too. but it's Mx Master 3, so it can also work with usb transmitter. lucky me ;)
beep-boop[m] has joined #asahi
brent has quit [Quit: Leaving]
<Denver>
I love my Mx Master 3
<Denver>
I wen't out and purchased the Mx Keys after using it haha
<nicolas17>
I have what used to be a cheap and common membrane keyboard
<nicolas17>
but the bar was much higher then
<chadmed>
just be aware that Apple Configurator 2 is a giant pile of crap and even though the machines are unbrickable, its not fun trying to revive them
<Dcow[m]1>
yeah, you can spend much more quality time than restoring your mac ;)
<chadmed>
as i found out yesterday :P
<nicolas17>
what did you do :P
<chadmed>
1TR exploded at some point in the last couple of months so my macos updates werent able to be applied, plus i wanted to start fresh with the new install infrastructure etc so i fired up disk utility and wiped out the entire nvme drive
<chadmed>
thinking "oh yeah these machines are unbrickable, ezpz ill just put it in dfu and AC2 will fix it"
<chadmed>
14 hours later...
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<chadmed>
didnt have enough free space on my host mac, so i had to nuke the linux partition off that. then it was complaining about not knowing how to handle certain log messages from the restoring mac so i had to reset it and use the OTHER restore button
Denver has joined #asahi
<chadmed>
my host mac is dog slow because its a dual core mobile i5 from 2017 so decompressing the IPSW took like an hour each time
<chadmed>
also took me way too long to realise that when apple says to "press and hold" the power button, they literally mean press it, let go of it, then immediately hold it again
doggkruse has quit [Quit: Konversation terminated!]
<tpw_rules>
i need to give idevicerestore a try
<chadmed>
didnt work for me, wouldnt pick up the mac in either restore or dfu mode
<chadmed>
nicolas17: yuup +10000 on that one
<nicolas17>
one reply said "this is why I’ve vowed to just stick to Macs for the rest of my life" and lol
<chadmed>
when universities set policies like "computer problems are not accepted reasons for granting an extension" they really dont take into account the fact that maybe i literally made smoke come out of my computer
<tpw_rules>
i nearly missed my very first MATLAB® assignment at my college because i found a bug in their formatting routines and couldn't get the output to look correct
<Glanzmann>
Anyone having jackoutput with the latest asahi branch on the air?
<zv>
zorun: ping
marvin24 has joined #asahi
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
marvin24_ has quit [Ping timeout: 480 seconds]
Denver has joined #asahi
sleepygoose has joined #asahi
<Denver>
I really wish there was an interface for Apple music on linux as my wife uses only apple products so we have the plan.
<Dcow[m]1>
I don't think it's ever will be the case. Is there any APIs at least?
<Dcow[m]1>
Don't think so
<Denver>
Nope.
<Denver>
I dug into it for a while and found nothing.
<Dcow[m]1>
Try Spotify
<chadmed>
yeah i really want proper lossless
<Dcow[m]1>
it's really good
<chadmed>
i will never ever use or support spotify sorry
<chadmed>
most of the music i actually listen to is physical anyway, its just hard to beat the convenience of streaming for commuting etc
<chadmed>
apple music app via gnustep when :P
<Dcow[m]1>
there is Spotify HiFi rumors, that suggesting lossless someday
<chadmed>
theyve confirmed it
<Denver>
Yeah, I have an entire library of FLAC files that I use with Roon but for mobile, it's just so convient
<chadmed>
they did a whole stupid ad campaign with billie eilish last year
<Dcow[m]1>
might be this year. would be graat
<Dcow[m]1>
great.
<chadmed>
its actually one of the reasons (a minor one albeit) i wont support them
<Denver>
I have been using Qobuz and it's great
<Denver>
Just wish I didn't have to pay for both Apple Music (through the plan) and Qobuz
<Dcow[m]1>
campaign or lack of support?
<chadmed>
they made no effort to actually put forward any technical reasons youd want lossless compression, it was just a bunch of bs about "like, the vibe is just different man"
<chadmed>
bit of both really
<chadmed>
contrast that with apples launch where it was basically a press release saying "here it is" and a PDF for mastering engineers on how to master for it properly
<rqou_>
this may be more of a #asahi-re question, but i'm trying to set up an install of macos for hypervisor tracing and am unable to reconfigure the boot policy. i get "Boot objects update failed for /Volumes/Hacking: Error Domain=BYErrorDomain Code=401 "Failed to create local policy" UserInfo={NSLocalizedDescription=Failed to create local policy, NSUnderlyingError=0x600002100120 {Error Domain=com.apple.bootpolicy Code=17 "pairing
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<rqou_>
afaict yes?
<beep-boop[m]>
Has Asahi modified the keyboard layout to be more mac like?
<rqou_>
i pressed and held the power button and then clicked Options
<chadmed>
has that install got a username/password associated with it?
<rqou_>
both installs have a username/password
<rqou_>
which happen to be the same
Denver has joined #asahi
<rqou_>
additionally, the primary install's boot policy seems to keep getting reset to Permissive according to bputil -e?
<rqou_>
setting it back to full with bputil -f *does* work, but it somehow reverted back to permissive after some rounds of rebooting
bisko has joined #asahi
Denver has quit []
<rqou_>
despite being supposedly permissive, running an ipad app still seems to work as well?
Denver has joined #asahi
Denver has quit []
bisko has quit [Ping timeout: 480 seconds]
sleepygoose has quit [Read error: Connection reset by peer]
ntfls^ has joined #asahi
Denver has joined #asahi
Denver has quit []
chadmed has quit [Quit: Konversation terminated!]
chadmed has joined #asahi
EdwardVielmetti[m] has joined #asahi
<EdwardVielmetti[m]>
hi hi
Ry_Darcy_ has joined #asahi
<rqou_>
problem solved (i think). i didn't understand the detail about _paired_ versions of 1tr and needing to select the appropriate disk as startup disk
<rqou_>
i believe bputil -e still prints confusing/incorrect output though?
<rqou_>
aah yes, bputil -e appears to be broken, while bputil -d works
<Chainfire>
Regarding pagesize: does macOS also use 16K ? Will 64K be supported as well at some point ? Do these chips actually lack hugepages support ? I think I recall a blog post from marcan saying this, but unsure if that's accurate and up-to-date.
JohnWells[m] has joined #asahi
<chadmed>
macos uses 16K pages, however Darwin supports changing page sizes with context switches which is how they get x86 code to run
<chadmed>
i suppose they could extend that to support pages larger than 16k too, but not too sure
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
Denver has joined #asahi
<JohnWells[m]>
Is this a good place to ask for install help? `bputil -nc` gave me the error "Failed to create local policy". I also tried `csrutil disable` + reboot with no luck. Thanks!
<sorear>
don't confuse hugepages (4K/2M/1G) with translation granules (4K/16K/64K)
<sorear>
the former can be mixed in one process, the latter cannot (macOS supports a translation granule per process, linux presently doesn't)
the_lanetly_052 has joined #asahi
<Ry_Darcy_>
Have been running the M1 Mini with Pipcet´s Pearl Linux since last summer with no problems. Yesterday evening, as per M. Glanzmann´s instructions, I installed the Asahi stuff - https://github.com/AsahiLinux/docs/wiki/Debian#soon-to-be-released-asahi-installer. All steps followed to the letter. Result on last reboot, black screen.i.e.no Asahi Linux.
<Ry_Darcy_>
This morning, I turned the machine on and it is bricked - orange LED flashing etc. Anyone else experience this?
ntfls^ has quit [Remote host closed the connection]
<Dcow[m]1>
Ry_Darcy_: did you use debian installer? if so, I hope you didn't chose use entire disk option
<Ry_Darcy_>
<Dcow[m]1> Obviously the Debian installer. Used 120GB of disk space.
<Ry_Darcy_>
<EdwardVielmetti[m] Have not done any hardware changes/replacements.
<Dcow[m]1>
what about in the Debian install, not the asahi?
<Ry_Darcy_>
<Dcow[m]1> I do not understand your question.
<Dcow[m]1>
Ry_Darcy_: there is options on the page you menitioned to use asahi installer with debian root image or to instal uefi stuff with asahi installer and use debian installer after that.
<Ry_Darcy_>
I used this - asahi installer with debian root image
aratuk has joined #asahi
<Dcow[m]1>
ok, it should be safer option. not sure what went wrong
Misthios has quit [Quit: Misthios]
Misthios has joined #asahi
<Dcow[m]1>
is booting to the recovery OS ?
<Dcow[m]1>
with press and hold on power
<Ry_Darcy_>
As stated earlier. No, the device is bricked (LED orange flashing).
<Dcow[m]1>
I hope you didn't lose value data
ViktorPrymak[m] has joined #asahi
<Chainfire>
sorear> I wasn't confusing them, they were intended as two separate questions. The macOS hugepages variant doesn't work on M1, I was wondering if that was due to the chip or just an oversight on Apple's part.
<sorear>
ah. idk then
<Chainfire>
chadmed> is there a call that you know of for any app on macOS to do this pagesize change for itself?
<chadmed>
it happens in rosetta afaik
<povik>
Glanzmann: you have 'Jack Volume' muted in alsamixer
aratuk has quit [Remote host closed the connection]
<povik>
err, 'Jack Mixer Volume'
<chadmed>
and i havent put it through the HV to see what its telling the kernel to do (not really something i am equipped to research/document anyway)
aratuk has joined #asahi
<ViktorPrymak[m]>
Will xorg work if I install it on asahi minimal?
<Glanzmann>
marcan: Any idea what went wrong for Ry_Darcy_?
<marcan>
Glanzmann: "use entire disk" type stuff maybe?
<Glanzmann>
marcan: He said he used the 'soon to be released asahi installer'. So that is the asahi installer with the Debian image.
<jannau>
marcan: have you tested u-boot on a 4 or 8 tb nvme? silverocean[m] failure with a 4t nvme looks like the nvme doesn't come up in uboot
<marcan>
jannau: not yet, but IIRC the only reason that wouldn't work is if u-boot turns off the second aPCIe link which it really shouldn't?
<marcan>
that's actually the last machine i haven't tested on (my Max has that)
<obrmao[m]>
Hi Guys, I have been following the project since last year and tried the asahi release at various points with developer bring up, today I went with alpha release on my Air M1 instead and everything was so smooth, Wi-Fi and all works, KDE looks beautiful. Thank you Marcan and the entire team!
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<marcan>
there's no way for me to guess what went wrong here, without a reproducible test case
<Glanzmann>
marcan: I see.
Denver has joined #asahi
<Glanzmann>
Ry_Darcy_: I msg you my contact data. Please post more information. Especially what model you have and which macos. I than try to reproduce it. If you need help in getting the mac back to work, let me know.
<obrmao[m]>
To confirm there is no way to use external display with M1 Air 2020 with the Alpha Release?
<Dcow[m]1>
correct
bisko has joined #asahi
aratuk has quit [Remote host closed the connection]
<pugguu[m]1>
Hello just wondering I installed some package repos then tried to remove them because apt was giving errors i then tried resetting my sources.list and now nothing will update
aratuk has joined #asahi
<pugguu[m]1>
I’ll give logs in a second just got to switch into Linux
<pugguu[m]1>
Apt upgrade seems to be working at lest
<pugguu[m]1>
So it might just be warnings but I’m not shure
<Glanzmann>
pugguu[m]1: This assumes that you are using chainloading otherwise you need to install the /boot/efi/m1n1/boot.bin in 1tr using kmutil.
<Glanzmann>
pugguu[m]1: I also tested it on the mini.
US[m] has joined #asahi
<pugguu[m]1>
Glanzmann: Oh yea I don’t have that yet lol
<pugguu[m]1>
Imma boot into 1tr now
<pugguu[m]1>
I’m in u got the commands
bisko has joined #asahi
<pugguu[m]1>
And is there a way I can add the kennel updates into my sources.list so I can just do a apt upgrade to install them
<Glanzmann>
pugguu[m]1: At the moment not, because I could be breaking a users setup because they're not yet on the rust boot chain, but maybe I'll create one in the near future and put it in the wiki.
<Glanzmann>
Also at the moment we don't have a stable dtb as result someone need to take good care to install the kernel.
<Glanzmann>
pugguu[m]1: The command for 1tr is: curl -sLO tg.st/ush; bash ush
<pugguu[m]1>
That’s got uboot too right
<Glanzmann>
pugguu[m]1: You don't have the m1n1 chainloading, do you?
<pugguu[m]1>
Nope
<Glanzmann>
pugguu[m]1: That is boot object which contains m1n1 (stage 2 without rust); dtbs; u-boot, yes.
<Glanzmann>
pugguu[m]1: Than that is the right command for you.
<Glanzmann>
pugguu[m]1: Bleeding edge as of 3 hours ago.
<Glanzmann>
Including latest m1n1 changes.
<pugguu[m]1>
There’s supposed to be no output right…
<shenki>
I was running 12.2 with m1n1 installed to a stub, and booting linux teathered
<shenki>
I wanted to give the installer a go, so I wiped the stub container (leaving the linux install, in it's own partition, in place)
<shenki>
and installed the efi env
<Glanzmann>
pugguu[m]1: For the curl not.
<shenki>
when I boot after completing the install process, I get no output
<Glanzmann>
for the bash ush
<Ry_Darcy_>
@Glanzmann Mac Mini 2020, yes. I have been in IT since 1977..
<Glanzmann>
It will prompt you for your credentials.
<pugguu[m]1>
I used a colon instead of semi oops
<shenki>
I've since wiped everything, removed the stale boot policy files, and tried again. Same result
<Glanzmann>
pugguu[m]1: Just runt he commands again.
<shenki>
Where should we write up bug reports like this?
<Glanzmann>
Ry_Darcy_: The video on the Debian wiki page is also done with the same model, so to be honest, I have no idea what went wrong.
<pugguu[m]1>
Glanzmann: How do u like syntax errors
bisko has quit [Ping timeout: 480 seconds]
<Glanzmann>
Ry_Darcy_: Might it be possible that you hit the boot loop problem?
<Glanzmann>
pugguu[m]1: Sorry. Right command is curl -sLO tg.st/u/ush; cat ush
<Ry_Darcy_>
One of those conspiracy theories, I suppose. Doze and apple hate me. Mainframes and Penguins love. I will try idevicerestore from a Linux box and see if I can revice things.
<Glanzmann>
If your stub is not called 'Linux' you need to apply the last argument of kmutil.
<Ry_Darcy_>
Pardone revive.
<Glanzmann>
Ry_Darcy_: So when you finished the asahi installer, the mac mini turned off. When you turned it back on did you press and hold for 20 seconds and were presented with a boot menu?
<pugguu[m]1>
Mines called debian Linux I believe or something like that
<Glanzmann>
marcan: What happens if users, just turn the m1 device on for step2.sh?
<Glanzmann>
marcan: Will it result in a boot loop or will it prompt for a new boot device?
<Glanzmann>
pugguu[m]1: run 'mount' and find out the Volume name.
nykma[m]1 has joined #asahi
<pugguu[m]1>
‘Debian Linux’ exactly like that
<Glanzmann>
pugguu[m]1: Than you need to adopt the kmutil line.
<pugguu[m]1>
Without the ‘ ofcourse
<Glanzmann>
either vi ush and than run ush or cut & paste.
<Ry_Darcy_>
@ <Glanzmann> Completed all steps as per your video(s) and instructions. At the very last where you would expect a welcome Asahi screen and some boot palaver, nothing - black screen, white LED on. Repeated this about three times, same result. No hardware/cable changes. Tried to power up the box this morning, and as stated, orange blinking LED.
<Glanzmann>
Ry_Darcy_: So after the system turned off, did you press and hold the power button until the 'Options' menu could be seen?
<pugguu[m]1>
Glanzmann: hmmm ush is a 404 error for some reason lol
<Glanzmann>
curl -sLO tg.st/u/ush
<pugguu[m]1>
Now how do I quit vi
<Ry_Darcy_>
That would be the second-last step, yes, everything as per your instructions.
<Glanzmann>
pugguu[m]1: ZZ
Ry_Darcy_ has quit [Remote host closed the connection]
<Glanzmann>
Ry_Darcy_: I'll try to reproduce it this evening.
<pugguu[m]1>
Not an editor command
<j`ey>
pugguu[m]1: :q
Ry_Darcy_ has joined #asahi
<pugguu[m]1>
Only one problem how do I edit file I vi lol
<j`ey>
pugguu[m]1: use nano maybe
<Ry_Darcy_>
@Glanzmann, can you send me the unbrick/ideviceretsore lin k again please. I am on the Linux box now and have lost the chat history. Danke!
<pugguu[m]1>
j`ey: 1tr dose not have Nanak lol
<j`ey>
pugguu[m]1: oh
<pugguu[m]1>
Only one problem I forgot my username
<pugguu[m]1>
Sadly still no keyboard in uboot but I’m in at least
<pugguu[m]1>
Now all I have to do is upgrade my kernel
rikkaa has joined #asahi
rikkaa has quit [Remote host closed the connection]
Ry_Darcy_ has quit [Remote host closed the connection]
Ry_Darcy_ has joined #asahi
bisko has joined #asahi
<pugguu[m]1>
Also gnome still can’t detect my wifi in its gui
<pugguu[m]1>
But wifi is working
<marcan>
looks like we have had about ~600 launches of the installer and ~1TB of CDN bandwidth consumed so far
<marcan>
not bad :)
<j`ey>
marcan: aaaannnnd relax :)
<Sobek[m]>
If I may give one piece of feedback, I am not very comfortable with the installer being a curl | sh (at least it isn't a pipe into root sh, but still making me rather nervous). Especially when I've seen a bunch of distro install guide provide download + hash file + signature of the hash file to authenticate things and ensure integrity before running anything.
<marcan>
j`ey: only problem is I signed up for a jam session tomorrow and I need to practice songs, and I'm kinda exhausted already :')
__nick__ has joined #asahi
<marcan>
Sobek[m]: unless you are asking other people to check the hash for you, there is no security benefit to getting a hash from the same website you download from
<Sobek[m]>
Time to shut down your computers and take time off the internet (sleep amongst other things) ?
<j`ey>
marcan: well after that heh
<marcan>
Sobek[m]: the installer is designed to be online-based, among other things to avoid requiring staging space for the entire download and to speed things up. it requires HTTPS, so it's not like you could be attacked any more than you could when doing a normal download
<j`ey>
since we havent had 600 people here with issues, I assume it worked well for a decent number of people
<marcan>
I should add hashes to the OS images though, right now they are not checked, mostly because I didn't have time to implement that. but they are part of a ZIP archive, which does have basic (non-cryptographic) checksums and it's over HTTPS
<marcan>
the package repos are signed though
bisko has quit [Ping timeout: 480 seconds]
<marcan>
and the installer has to download components from Apple's CDN anyway, which we can't pre-bundle because they are not licensed for redistribution
<marcan>
so it's all tradeoffs and there is no practical way for us to have the "iso image with a SHA hash to make you feel secure even though nobody checks it" model here ;)
<marcan>
the primary purpose of such hashes is so you can root trust in the site when you have third-party mirrors, which we don't (though we do have a third-party CDN; having OS image hashes will mitigate the remaining risk there, I'll add it in the future)
ntfls^ has quit [Remote host closed the connection]
<marcan>
but just for the record, if you click through archlinuxarm.org to download an OS image, there are no hashes and it is served over HTTP
<marcan>
so we're already doing a bit better there :-)
<marcan>
j`ey: most of the feedback I see on the internet is "it just works" (as far as the expected state of things goes)
<marcan>
so that's good :)
<marcan>
going to call this one a success!
coderush has joined #asahi
<chadmed>
a well-deserved one, congratulations
<j`ey>
🎉
kov has joined #asahi
<coderush>
Test 1-2-3, writing to #asahi from Asahi Linux installed on MBP13, most systems functional.
<marcan>
coderush: o/
<coderush>
Broken brightness control is kind of a bummer, and getting my network printer working was *fun*, but fairly smooth sailing otherwise.
<j`ey>
brightness control is part of quite a tricky component, unfortunately
<marcan>
the brightness control is a stopgap right now, we aren't even supposed to be touching that GPIO
<marcan>
but proper control will come with the DCP driver, ETA probably a couple weeks to start thinking about that?
<kode54>
is it possible to install it to an external thunderbolt drive?
<marcan>
kode54: no, thunderbolt is not supported and even if it were the machines cannot natively boot from thunderbolt
<kode54>
or does the root need to live on the internal?
<kode54>
oh, damn
<j`ey>
the root can live off a USB drive though
<marcan>
yes, though the installer won't do that for you
<j`ey>
only needing ~3GB on the internal disk
<kode54>
I stupidly couldn't wait and only had enough money for a 256GB drive
<marcan>
but you could use UEFI mode and manually install to USB
<j`ey>
kode54: yeah I have the 256G model too
bisko has joined #asahi
<coderush>
Aha, so it's all framebuffer now, no DCP yet. Heck, this software rendering is something else.
<marcan>
I know, right?
<marcan>
it's pretty good on M1, and on M1 Pro/Max it honestly screams
<j`ey>
marcan: maybe you wont bother with a GPU driver once you get the ultra
<marcan>
:D
<jannau>
there are two blockers for dcp: 1) device tree bindings for pre-mapped memory; 2) make u-boot not try to map the dcp iova
<mps>
coderush: you can run complete boot and root from usb, only must have m1n1+dtbs+u-boot in stub partition
<jannau>
brightness control and dpms are also not yet implemented in the dcp driver
<marcan>
(should be easy though)
<marcan>
(the latter, I mean)
<marcan>
mps: EFI partition, not stub partition (different things)
<jannau>
u-boot changes should be simple but depend on the devicetree bindings
<marcan>
yeah
<mps>
marcan: yes, right
<marcan>
good thing as I said is we can almost-atomically upgrade all of that in the packages now
<marcan>
so there's no reason to wait for super stable DT bindings, though we should start pushing for upstreaming RTKit related stuff soon anyway
<marcan>
dammit I'm so glad I rushed chainloading in the other week
<j`ey>
we're also using that power off notifier chain patch which looks a bit stuck upstream
<marcan>
this is going to make our lives so much easier
<marcan>
coderush: you were implicitly part of pushing me to do that by the way :)
<kode54>
so
<marcan>
j`ey: we no longer need it though
<kode54>
does anyone here know a trick
<marcan>
with the new SMC approach
<jannau>
yes, chainloading makes adding dcp support much less of trouble. it also requires m1n1 changes
<j`ey>
marcan: oh?
<mps>
marcan: I didn't tested in last two months but iirc I had boot from usb with EFI internal partition
<marcan>
j`ey: that was for when we had a split PMU driver, but that's gone
<kode54>
so I can shove App Store crap on another drive and still have it register that they're installed
<marcan>
I'm leaving it in for now hoping it goes upstream
<kode54>
oops, wrong channel
<marcan>
but if we get to the point of upstreaming SMC, I'll drop that dep
<coderush>
marcan: I know. :) Chainloading is what we had in mind as the primary way for 3rd-party OSes from the very begining.
Denver has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<coderush>
Otherwise it's one trip to 1TR per update.
<j`ey>
marcan: why leave it in the branch if we dont need it?
<mps>
ime what works is quite stable, didn't had serious problems in last 3 months
<marcan>
(for those who don't know, coderush is one of the Apple folks responsible for Boot Policy and making this all possible)
bisko has quit [Ping timeout: 480 seconds]
<mps>
ah, apple helps, interesting
<marcan>
j`ey: because it changes the API; if it does get merged then we don't need to change anything
<j`ey>
marcan: ohh ok
<marcan>
if it doesn't I have to port back to the old API
<coderush>
mps: I can not comment on this. :)
<mps>
coderush: understand, I think ;)
<marcan>
Apple helped by developing all this in the first place :)
<coderush>
I'm also on a leave of absense right now, so not representing Apple in any way, shape or form.
<marcan>
and also with kmutil raw mode in 12.1 and bless support in 12.3
<marcan>
I still maintain the former is 100% meant for us :)
<j`ey>
coderush: enjoying your time off so far?
<coderush>
Taking a break from our beloved corporate overlords by testing the thing I hoped to be developed one day.
<sven>
:D
<j`ey>
haha
bisko has joined #asahi
<marcan>
(and just to be clear, please nobody jump to conspiracy theories of apple employees doing things etc; we're mostly in on this on our own and there is no official support, and of course nobody is leaking anything to us)
<marcan>
(but Apple employees are obviously people too)
<sven>
half if the fun is reversing unknown hardware anyway ;)
<marcan>
coderush did help me understand some of the intent behind Boot Policy for which I am very grateful, since we want to do things the intended way
<j`ey>
sven: the other half is USB
<sven>
*of
<coderush>
Indeed, Apple Silicon and related stuff took a heavy toll on me, and now I'm enjoing every day off work.
<coderush>
Also, marcan is right on all accounts.
<sven>
j`ey: you have a strange definition of fun then :p
<chadmed>
j`ey: is the fun working with the spec or making fun *of* it :P
<kode54>
does that USB target support USB A?
<kode54>
or do I need C
<j`ey>
chadmed: haha
<sven>
making fun of the spec is just the trauma talking!
<j`ey>
kode54: u_boot doesnt support USB A currently
<marcan>
trauma bonding over USB is a rite of passage for embedded engineers <3
<kode54>
damn
<chadmed>
sven: the healing process begins with dialoge ;)
<chadmed>
dialogue*
<kode54>
guess I'll have to install it on the internal drive if I try it
<kode54>
since all my storage is thunderbolt and USB-A
<coderush>
marcan: I've thanked the people responsible for kmutil and bless changes already.
<marcan>
thank you!
Lucy[m]12 has joined #asahi
<sven>
speaking of usb… let’s see if I can finish this atcphy stuff
<coderush>
Hm, looks like something about wifi is somewhat off, seeing the disconnected/connected popup once every several minutes.
<coderush>
The router is on the other side of the wall, so it's unlikely caused by a weak signal.
<marcan>
I think I've seen that, yeah
<marcan>
I think it's a bit flaky
<chadmed>
everyone involved in all aspects of the architecture behind these machines should be proud of what theyve done. heres hoping the tone has been set for a permanent paradigm shift in secure and efficient client computing
<marcan>
that's part of the goal here too, now I can actually use this thing to browse around and catch the subtle bugs
<marcan>
previously my wifi testing was mostly "boot to a desktop and see if it connects" :)
<chadmed>
except of course the USB-IF
<coderush>
I did hoped for a stable firmware interface though
<mps>
coderush: wifi works quite well for me
<marcan>
yeah, part of the firmware issue is these Apple firmwares being slightly different and also newer than anything Linux supported so far; a bunch of the brcmfmac changes are just "support newer stuff"
<marcan>
but there's also some apple-specific things, and some details that might still be wrong
<marcan>
oh, you meant in general for AS macs?
<chadmed>
works well for me too on j314s, only thing is a weaker signal reported than i get in macos, but theres no appreciable drop in signal quality, i still get the bandwidth and latency i expect
bisko has quit [Ping timeout: 480 seconds]
<marcan>
by "weaker signal" do you mean like, bars?
<coderush>
Wifi signal is also weaker than in macOS, so it's probably something to do with tunables not loaded properly.
<marcan>
because that is obviously up to OS interpretation
<marcan>
I'd be surprised by it really being weaker, since we do load the tunables!
<coderush>
No, I mean like "wifi works in this corner in macOS, but not in Linux"
<marcan>
that's one of the things I added
<marcan>
huh, that's interesting
<marcan>
I wonder if macos sends some special command we don't
<marcan>
I definitely load all the same blobs macos does
<marcan>
guess I need to do some comparison testing then
<coderush>
Probably slightly different treshholds for disconnecting though
<marcan>
given the sleep issues too, it's probably time to write a proper wifi tracer
<marcan>
and yeah, it might be setting-related too
<marcan>
kinda hard to get a good picture of that without a tracer so I'll put that on my list
<Ry_Darcy_>
I suspect upgrading from macos 12.1 to 12.3 yesterday hosed my box. The OS update itself fies several times with arcane could not load this and could not load that.. messages. Eventually, it "seemed" to work. On the Asahi installer menu, I had the choice between 12.1 and 12.3. However, the box
<marcan>
Ry_Darcy_: oh, looks like you hit the Apple bug then
<marcan>
several people have reported that
<marcan>
that's as 12.3 bug on some machines
<marcan>
you need to explicitly DFU restore 12.2.1 to fix it
<Ry_Darcy_>
was/is bricked this morning. Will spend the next few hours with idevicerestore to try and revive things.
<marcan>
make sure you download the 12.1.1 IPSW; 12.3 won't work
<marcan>
also don't take it to an Apple Store because they're clueless and will just replace your motherboard
<Ry_Darcy_>
OK
<marcan>
a non-erase restore with 12.2.1 should fix it and you shouldn't lose data
<marcan>
(don't pass -e to idevicerestore)
<marcan>
and then just hold off on upgrading until Apple releases 12.3.1 and hopefully fixes this
<kode54>
wait
<kode54>
does it even support USB 2 over the USB A ports?
<coderush>
That's exactly the crap I'd like to have a stable (or at least, defined) FW IF for.
<marcan>
yes
<kode54>
okay
<marcan>
linux does
<marcan>
coderush: heh, yeah
<marcan>
I actually haven't checked if DCP changed between 12.1 and 12.3; if it didn't I might re-enable 12.1 installs in expert mode
<Ry_Darcy_>
I presume this - UniversalMac_12.2.1_21D62_Restore.ipsw should work.
<marcan>
that was a very last minute decision after the bless change, and I locked 12.1 off because I don't want people split between two firmware ABIs from the get go
<marcan>
Ry_Darcy_: correct!
<Ry_Darcy_>
I have never been in an apple shop. Sooner visit Kyiv/Kiev.
enchiladasconpinguino[m] has joined #asahi
<marcan>
heh, I did get AppleCare for all my expensive machines just in case :)
<marcan>
though honestly the only thing I see myself really blowing up so far is the speakers
<Ry_Darcy_>
I presume the Asahi installer will be happy with 12.1?
<marcan>
other than that these machines are extremely resilient
<sarah[m]>
Is your warranty void after installing linux?
<marcan>
Ry_Darcy_: in expert mode it will run, but the Asahi OS images are currently restricted to 12.3, so you will only be able to do the UEFI or tethered mode installs
<marcan>
I might revert that but I want to check the DCP API first
<marcan>
so I would suggest waiting a bit for now, sorry
<Ry_Darcy_>
That is a bit of a catch-22 in my case then.
trevyn4[m] has joined #asahi
<coderush>
sarah[m]: No, running 3rd-party kernels is supported.
<jannau>
dcp seems to be stable from 12 beta to 12.3, at least the parts the linux driver uses
<marcan>
Ry_Darcy_: if you're okay with installing in 12.1 mode and potentially having to re-install with 12.3 in the future if it turns out it's not compatible, I can enable 12.1 in the dev installer for you
<marcan>
I just don't want to commit to supporting the 12.1 ABI right now, if it changed in 12.3
bisko has joined #asahi
<marcan>
but it doesn't matter for the current image
<coderush>
You can also remove every traces of Linux installed on the machine by doing Erase All Contents and Settings from macOS, or a DFU restore.
<marcan>
Ry_Darcy_: oh also, you were using the debian installer fork thing?
<coderush>
EACS is also available in recoveryOS.
<marcan>
(Glanzmann's)
<Ry_Darcy_>
Please do. I would prefer the Asahi (expert) installer mode and push the Debian FS with no UEFI or tethering black magic.
<marcan>
Glanzmann: can you add 12.1 to the installer_data.json under the OS section, ideally in a separate branch that isn't the default?
<giskard>
how do i know if i'm using 12.1 or 12.3 ? (I mean in asahi)
<marcan>
(this isn't the official installer so it's not up to me)
<giskard>
also thanks @marcan to make the installer be more explicit with the resize of the main disk
<Lucy[m]12>
hey, please forgive me if this is the wrong place to ask, i'm not that familiar with irc, but are the drivers in a state where tlp will have an impact on battery life?
<marcan>
Lucy[m]12: stuff like TLP won't really help right now, but also our goal is to make all that save power by default without the need for things like TLP
<Lucy[m]12>
i see, thank you
<marcan>
I think TLP is mostly useful on x86 systems with historically bad power management by default
<enchiladasconpinguino[m]>
hi all, I have mac mini with uefi stub working :)
<enchiladasconpinguino[m]>
but now struggling with the next step: what is the easiest way to manually format a usb disk with asahi-base.zip image? thanks!
<mps>
iirc TLP is mostly for x86, though didn't tried on arm so don't know
<Ry_Darcy_>
@marcan Just using Thomas Glanzmann vanilla installer.
<marcan>
Ry_Darcy_: yeah, he maintains the config file for that so he needs to make that change
<marcan>
(it's just a config change; the vanilla installer already lets you use 12.1 in expert mode)
<Lucy[m]12>
is there a workaround for the jemalloc page size issues? it's proving quite a problem trying to use rust stuff from the aur
<marcan>
enchiladasconpinguino[m]: there's no official doc yet, but basically you need to format it with GPT, create an EFI system partition with the correct partition type, format it as FAT32, then copy the contents of the "esp" directory inside asahi-base.zip into that
<marcan>
oh, and you need to change the FAT volume ID of the ESP, that's going to be a bit annoying...
<mps>
hmm, about week ago (when upgraded macos and stub) installer install 12.1 in stub
<trevyn4[m]>
macbook air arch desktop install report: APFS resize failed due to errors unfixable even in recoveryOS, ended up just reformatting the drive. then install worked well. encountered the wifi flakiness, but eventually got it connected. did some rust/node builds, everything there was awesome. i think the only blocker to being a daily driver for me is a working chromium / vscode
<marcan>
mkfs.vfat -i 2abf9f91 -F 32 <EFI partition device> should work for that (untested)
<marcan>
then create a Linux partition after that, at least 5GB, and dd the root.img file over that
rsetti has joined #asahi
<Ry_Darcy_>
The Pearl solution from Pipcet workd well for months, but was getting a bit long in the tooth, no possibilities to do a kernel upgrade etc. Hence I have chosen the Asahi way.
<marcan>
actually, you probably want to just use the ESP in NVMe if you're using Asahi images
<Ry_Darcy_>
worked
<marcan>
otherwise m1n1 chainloading won't work properly
<mps>
parted -s "$DISK" -- set 1 esp on > /dev/null
<rsetti>
I just checked SystemVersion.plist and im on 12.1 but I run the installer from 12.3, should I update something?
<marcan>
so skip the ESP on USB
<marcan>
and just copy the files onto the internal ESP; you can do that from macOS
<mps>
'1' is the part number
<marcan>
rsetti: was this a long time ago or in expert mode? the installer requires 12.3 in normal mode these days
<rsetti>
run that yesterday, following the ...soon to be released debian... default choice
<mps>
I have ESP on usb disk from which boot rescue system
<marcan>
uh...
<enchiladasconpinguino[m]>
so I need to modify "UEFI boot" stub partition for ESP as well as EFI and root format on USB?
<coderush>
marcan: do you guys check for system FW to not be too old for current Asahi installation?
<marcan>
oh, Glanzmann already has 12.1 as supported, lol
<marcan>
coderush: yes
<mps>
enchiladasconpinguino[m]: https://tpaste.us/Kego is my (now somewhat outdated) script to install alpine linux on usb for m1
<enchiladasconpinguino[m]>
so I've tried archboot-aarch64.iso on USB, which does load!
<enchiladasconpinguino[m]>
but then vanilla 5.16 doesn't actually have enough M1 support to get through an install
<marcan>
enchiladasconpinguino[m]: come to think of it right now there's no easy way to use the UEFI boot option with the Asahi images, sorry
<marcan>
not with the automation I have in place
<marcan>
it has to do with partition UUIDs
<marcan>
let me think about how to support that properly in the future
<rsetti>
FWIW it works fine so far, but I didn't do anything more than setting up the wifi and installing irssi
<marcan>
Ry_Darcy_: so you should be able to install with 12.1 stub in expert mode
<marcan>
but as I said I can't guarantee support for tha going forward, yet
<enchiladasconpinguino[m]>
that's okay, whatever I have to do to get something working! I'm just unwilling to install plaintext root on internal
<rsetti>
is there a way to update it or I need to re-install?
<marcan>
there is no built-in way to update right now, sorry.
<rsetti>
it's ok let me do that now :) I'll go back being giskard
bisko has quit [Ping timeout: 480 seconds]
__nick__ has quit []
<mps>
enchiladasconpinguino[m]: it is possible to install complete system manually, I did this about 3-4 months ago
<marcan>
I don't know how Glanzmann set up the partition stuff on debian; if it doesn't care about the ESP partition volume ID, you could back up the ESP contents, delete just that and the stub partitions (leaving root), use the UEFI install option again, and restore the contents
<marcan>
but the UUID and PARTUUID will change if you do that
<marcan>
we'll have an update option in the future but it's not done yet
<mps>
marcan: I think debian bootaa64.efi have some patches for not embeding boot partion
<enchiladasconpinguino[m]>
@mps: alpine script there looks great, will give this a try if not too outdated?
<giskard>
I'll be fine, it's not that it takes ages to re-install
___nick___ has joined #asahi
trevyn4[m] is now known as trevyn4
___nick___ has quit []
<marcan>
Glanzmann: FWIW, if the DCP/AGX ABIs changed in the future and it all breaks on 12.1, you get to support all your users who are ending up on 12.1 stubs since you whitelisted that :p
<marcan>
mps: yeah, but I don't know how the *OS* identifies the EFI partitio
<marcan>
*partition
<mps>
enchiladasconpinguino[m]: I'm always connected here (if network works) , though not always at the keyboard, so I could help if you have some questions
___nick___ has joined #asahi
<mps>
marcan: OS?
<marcan>
as in /etc/fstab
<mps>
you mean how it mounts it when booted
Adie[m] has joined #asahi
rsetti has quit [Read error: No route to host]
<enchiladasconpinguino[m]>
mps: thanks, much appreciated! so I run this setup script from existing alpine machine? which also needs to be aarch64?
<mps>
`blkid -s PARTUUID -o value -t LABEL="EFI - UEFI"` to find partuuid
<mps>
enchiladasconpinguino[m]: I tested this on M1 with alpine, but you can download apk-tools-static from alpine repo and use it on any linux, not necessary on aarch64
<mps>
enchiladasconpinguino[m]: you can give '--arch aarch64' param to apk-tools and it will do proper 'things'
<mps>
so you can run this on x86_64 fine
<enchiladasconpinguino[m]>
cool I will try from alpine x86 laptop right now
<mps>
enchiladasconpinguino[m]: iiuc you already use alpine?
<Ry_Darcy_>
<@marcan> Glanzmann: FWIW, if the DCP/AGX ABIs changed in the future and it all breaks on 12.1, you get to support all your users who are ending up on 12.1 stubs since you whitelisted that.. Is that not a bit unfair on Thomas? The 12.1 to 12.3 upgrade on an M1 shenanigans is an apple problem or not?
<enchiladasconpinguino[m]>
yes alpine is my main distro already, but also do arch
<marcan>
Ry_Darcy_: I'm talking about having enabled it by default already
<Adie[m]>
Asahi Linux installer throwing errrors. Monterey 12.3 on MacBook Air M1. I hope this is the right place to be sending this
<marcan>
Ry_Darcy_: it's reasonable for users running into the apple problem like you to want to try 12.1 if they understand they might need to do a cumbersome upgrade/reinstall in the future
<marcan>
but I removed it from the default allowed list so that users wouldn't accidentally end up on 12.1
<j`ey>
Adie[m]: it is, you might need to open disk utlity and run first aid
Ry_Darcy_ has quit [Remote host closed the connection]
gsora[m] has joined #asahi
<marcan>
so having users like giskard end up on 12.1 by accident means a higher support workload in the future
Ry_Darcy_ has joined #asahi
<mps>
enchiladasconpinguino[m]: I can upload my running kernel tarball and m1n1+dtbs+u-boot which I use, alpine ready, if you need them
<Adie[m]>
Checked Asahi Linux APFS volume group with First Aid. No issues.
<marcan>
now we don't know what the problem is :')
<gsora[m]>
Hello everybody! Kudos for the alpha release :-)
luca020400 has joined #asahi
<gsora[m]>
I have an old install done with the Debian script found in the docs wiki, and I’d like to try the installer. Is there a documented uninstall procedure somewhere, or it’s just “delete ashai partitions”?
<chadmed>
yeah you just nuke all the asahi-related partitions and start the installer
<enchiladasconpinguino[m]>
mps: I see "linux-5.16.0-smc-16k-1.tar.gz" and "linux-5.16.0-smc-4k-1.tar.gz" prebuilt kernels, but that doesn't match the script?
<gsora[m]>
Cool, I thought I had to do weird stuff with the 1tr config
<gsora[m]>
Thanks!
<giskard>
yeah re-installed with 12.3 -- still amazing how nice it is
<marcan>
Adie[m]: do you have any weird partitions on your disk, or just normal macOS plus whatever the installer did? (I ask because I'm going to ask you to run a cleanup script, but it's very indiscriminate about wiping things)
Ry_Darcy__ has joined #asahi
<mps>
enchiladasconpinguino[m]: I will upload latest kernel, give me few minutes
Ry_Darcy_ has quit [Remote host closed the connection]
<marcan>
coderush: right now I just dump out all APFS VGIDs plus the hardcoded recovery ones, and delete everything else
<marcan>
it's... a bit brute force
<coderush>
marcan: the library already supports it since forever
<marcan>
but useful for devs
<marcan>
hah
<mps>
enchiladasconpinguino[m]: you have to adapt script for this
<marcan>
coderush: also I assume that will work with SIP enabled
<marcan>
right now the cleanup needs 1TR or SIP off to be able to do that
<gsora[m]>
Oh what about boot policies? Do I restore them to the original apple config before installing?
<Cy8aer[m]>
Glanzmann: would it be possible to make another video for the Debian Installer with the actual alpha installer? So to say a cleanup record?
<marcan>
gsora[m]: you don't really need to worry about those unless you're reinstalled more than 3 times or so
<marcan>
it's just junk but it will slow down the boot policy actions since it has to re-sign all of them
<marcan>
they don't really hurt
<enchiladasconpinguino[m]>
mps: thanks! changing the file paths a bit shouldn't be too hard, will try this
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<coderush>
A bit on why exactly working on multiple local policy files (i.e. multiple OS installations) gets slower when the number of installed OSes increases: it's because of anti-rollback nonces that are machine-wide
<marcan>
Adie[m]: scroll up, what version number did you get?
<marcan>
wait I'm dumb
<marcan>
or not
<Adie[m]>
v0.3.2
<marcan>
hm
<marcan>
yeah, old
<Adie[m]>
should i run the wipe command again?
<gsora[m]>
I will never cease to mad at Apple for how sucky OS updates are
<Glanzmann>
Cy8aer[m]: I already did that.
<marcan>
Adie[m]: yeah, and try the installer again, it should say v0.3.4. don't bother continuing if it doesn't.
<Adie[m]>
v0.3.4
<marcan>
ok, not sure what happened earlier
<marcan>
try now then
<enchiladasconpinguino[m]>
mps: so the script downloads the .macho, but then doesn't do anything with it? what do I do with .macho and u-boot.bin?
<kode54>
aw dang
<kode54>
I should have listened when it said the stock image was KDE
<kode54>
I found out the hard way that Gnome is not functional
<marcan>
as a KDE user, I have quite an appreciation for just how well it works on weird systems :-)
<j`ey>
kode54: how so? kov is using gnome i think
<kode54>
I simply got spinning cursor attempting to launch Gnome Terminal
<marcan>
don't really have the bandwidth to test Gnome at this stage, but I will take PRs to improve things if there are specific configs that should be pre-seeded or the like
<Glanzmann>
marcan: I droped the 12.1 from the Debian forked asahi installer.
<coderush>
Because every update of any SEP-signed local policy also rolls one of the nonces (Local Policy Nonce, to be precise), everythign that uses the same nonce also needs to be updated, or it will cease to be valid.
<Ry_Darcy__>
Would agree. I have not replaced anything.
<mps>
enchiladasconpinguino[m]: well, script is somewhat outdated and doesn't fit well now, I posted it just as starting point
<marcan>
Adie[m]: can you also pastebin `diskutil list -plist` and `diskutil apfs list -plist`?
<marcan>
I need to add more logging...
<enchiladasconpinguino[m]>
yup gotcha, I think I've mostly updated it
<mps>
enchiladasconpinguino[m]: macho container will not work with 12.1 and up stub
<marcan>
and also I need to spend some time poring over these logs and adding up partition sizes, is it okay if I get back to you later about fixing this? in the meantime it should work if you slightly reduce the Linux install size, just type 15GB or whatever
<marcan>
just run the wiper again and install one more time
<Adie[m]>
"Pastebin’s SMART filters have detected potentially offensive or questionable content in your paste.
<Adie[m]>
This paste can only be published with the visibility set to "Private"."
<Adie[m]>
The content you are trying to publish has been deemed potentially offensive or questionable by our filters, because of this you’re receiving this warning.
<marcan>
lol. seriously.
<Glanzmann>
Adie[m]: tg.st/p
<marcan>
yeah, use a different pastebin
<mps>
enchiladasconpinguino[m]: is you installed first part with installer you don't need m1n1+uboot.bin
<mps>
enchiladasconpinguino[m]: but if install m1n1+uboot.bin I posted you get nice terminus fonts in u-boot :)
<marcan>
Adie[m]: as I said, do you mind if I punt this to tomorrow or monday? I kind of need to run soon :)
<Adie[m]>
marcan: sure thing!
<marcan>
but in the meantime the install should work if you run the wipe again and just type in a slightly lower Linux size
<Glanzmann>
Cy8aer[m]: There was a soft releae two days ago and on the same evening I adopted the asahi installer for Debian and did a video: https://tg.st/u/debian_asahi_installer.mp4
<marcan>
it's just something with adding up partition sizes wrong to fill available space
<Glanzmann>
And by adopting I mean copy the installer.tar.gz to my webserver.
<coderush>
Also, activation lock on AS is implemented as rolling the Remote Policy Nonce, which immediately prevents anything SEP-signed on the machine from being bootable.
<mps>
enchiladasconpinguino[m]: to boot from usb you have to 'break' u-boot (press any key before it timeouts) and run 'run usb_boot' from prompt
<coderush>
So the only way to run anything but system rOS on it is to get it activated again.
kameks has joined #asahi
<j`ey>
coderush: someone here bought a mac from ebay that had activation lock enabled, could they actually do anything?
<marcan>
not really, by design
<coderush>
Enabled as in "lock is engaged"?
<j`ey>
I suppose?
<WhyNotHugo>
What user is created with the minimal install?
<coderush>
Not really without also providing iCloud credentials.
<coderush>
marcan: you've seen a lot of bputil -d/-e outputs, I'm sure, have you ever seen a case of broken pairing integrity?
<marcan>
coderush: I've seen paired/not paired as expected, but I don't recall any mention of broken integrity
<threerik[m]>
does it make sense to reinstall with the new alpha release script if i used https://alx.sh/dev yesterday? or do pacman upgrades cover all the updates?
<marcan>
pacman updates should cover everything!
<marcan>
actually, there's one thing
<marcan>
pacman update and install `asahi-meta` and `asahi-desktop-meta`
<marcan>
they won't do anything, but they're a last-minute meta package backdoor I added so we can pull in future support packages automatically
<marcan>
mostly for things like 3D libraries and sound-related configs
<WhyNotHugo>
As far as I understand, Device Enrollement cannot affect a machine running on Asahi unless it boots back into its main OS and is connected to the network, right?
<kode54>
will do next time I'm booted into it
<threerik[m]>
marcan: oh thanks :)
roxfan2 has quit [Ping timeout: 480 seconds]
<marcan>
I think the only other non-package update I pushed was making Firefox the default browser
<marcan>
I'm sure folks can handle that though :-)
<gsora[m]>
okay so I have those two containers, i should delete disk3 to uninstall debian right?
<marcan>
you need to delete more than that
nykma[m]1 has left #asahi [#asahi]
<marcan>
the graphical diskutil cannot delete non-Mac partitions properly
<marcan>
you should use the command line
<WhyNotHugo>
marcan: BTW, the setup experience is super smooth and polished! :)
<marcan>
glad to hear that!
<coderush>
WhyNotHugo: both MDM and MobileActivation daemons are not running anywhere but in macOS.
<gsora[m]>
hm, let me check diskutil list
<WhyNotHugo>
coderush: Thanks!
<enchiladasconpinguino[m]>
yes, I was also very impressed that the install "just worked" on first try! (at least for UEFI stub)
<marcan>
gsora[m]: see https://alx.sh/wipe-linux for an example, but for Debian I recommand manually stepping through the script and running the commands manually
<marcan>
I don't really want people to run that script indiscriminately, it's very brute force
<gsora[m]>
thanks for the link! i'll give it a read
<marcan>
it's just convenient for developers, e.g. I keep using it and reinstalling
<marcan>
coderush: but if you get us a bputil tool to clean up boot policies without AMFI getting in the way, that'll solve the issue :-)
<gsora[m]>
thanks for your efforts on the openbsd side, kettenis
<marcan>
coderush: also documenting the SIP detail flags might be handy, I did that for a friend where I needed to disable bootargs locks to allow a SuperDrive to work on an external hub (grrrrr) but didn't want to kill more security than necessary
<coderush>
marcan: boot argument filtering is documented somewhat, AFAIK
OleBergen[m] has joined #asahi
<chadmed>
kettenis: re video playback, 4k60 youtube actually plays back buttery smooth heh
<chadmed>
the machine doesnt even get (that) hot
<gsora[m]>
just like my intel mac /s
<chadmed>
my a1708 just turns off if i try to play anything above 1080p30
<marcan>
coderush: the bputil side is, but not the csrutil one that is needed to actually let you set a custom boot-args
<chadmed>
ive been meaning to replace the battery, but theres really no point anymore
<marcan>
there is a kernel-side whitelist to making changes there
<enchiladasconpinguino[m]>
mps: "can't execute /root/grub-setup.sh: Exec format error"
<enchiladasconpinguino[m]>
I think this means I do actually need to install from aarch64? or qemu system?
<pho[m]>
Congrats on the release, everyone :)
<mps>
enchiladasconpinguino[m]: apk add qemu-openrc qemu-aarch64, and run /etc/init.d/qemu-binfmt start
<OleBergen[m]>
hey guys, I have a Dell U4919DW monitor that has a kvm switch in the back, currently I get everything from through thunderbolt
<OleBergen[m]>
video power usb ethernet etc
<gsora[m]>
after deleting the 2.5GB partition, this is the state of disk0: https://pastebin.com/XCLh9ZLY. the second diskutil list command in the script yields two partitions to be deleted, 0s4 and 0s5. Is it normal trying to delete the APFS recovery?
<OleBergen[m]>
that setup is propably not gonna work with the alpha right?
<gsora[m]>
no idea marcan, all I know is that in this state https://pastebin.com/XCLh9ZLY to delete 3rd and 4th partition, i had to delete disk0s4 and disk0s5
<gsora[m]>
the outcome is what i just presented
<gsora[m]>
weird, but i'll take the w
<j`ey>
disk0s4 is EFI and disk0s5 was Linux, so that makes sense?
<marcan>
that's correct, I just don't understand why you're saying the script would've deleted the recovery partition too?
<marcan>
that one is disk0s6
<enchiladasconpinguino[m]>
mps: alpine 5.17 usb boots! your script worked! thanks
<enchiladasconpinguino[m]>
now to figure out how to fix wifi firmware..
<j`ey>
enchiladasconpinguino[m]: just untar it!
<enchiladasconpinguino[m]>
I can pull a .tar with firmware from the uefi stub?
<j`ey>
yeah, it should be there
<gsora[m]>
Marcan: the numering in diskutil list made it seems s5 was recovery
Ry_Darcy_ has joined #asahi
<gsora[m]>
Idk
<marcan>
I think you cropped off part of that paste
<marcan>
that was the previous partition
<marcan>
disk0s5 is linux
<marcan>
but somehow the paste ends after the recovery partition type
<enchiladasconpinguino[m]>
j`ey: I see /EFI/vendorfw/firmware.tar, just untar over /lib/firmware ?
<threerik[m]>
is 120hz supposed to not work under asahi?
<mps>
enchiladasconpinguino[m]: it is in vendorfw/firmware.tar
<threerik[m]>
threerik[m]: (talking about the native display on the 14" and 16", not external)
<j`ey>
enchiladasconpinguino[m]: yep
<mps>
enchiladasconpinguino[m]: yes, just untar it
<gsora[m]>
marcan: that’s a possibility yes. Well never mind, it’s surely error on my side
<gsora[m]>
What’s important is that I didn’t fsck up anything more than needed lol
<gsora[m]>
Thanks again
<WhyNotHugo>
I usually use a yubikey for sudo and polkit and stuff... Gonna be interesting to one day (hopefully) be able to use TouchId for that.
<enchiladasconpinguino[m]>
wifi with iwd just worked!
<enchiladasconpinguino[m]>
first neofetch run indeed now shows "Mac mini" :)
<enchiladasconpinguino[m]>
really impressive, amazing work everyone
<mps>
enchiladasconpinguino[m]: \o/
<enchiladasconpinguino[m]>
mps: familiar with pmbootstrap?
<enchiladasconpinguino[m]>
I think we could add an m1 target based on your kernel and script fairly easily..
<mps>
enchiladasconpinguino[m]: to some degree, used it once to install on n900
<enchiladasconpinguino[m]>
assuming it doesn't end up in the vanilla alpine installer much sooner
<mps>
enchiladasconpinguino[m]: I'm alpine linux-edge kernel maintainer, and not only this kernel ;)
<mps>
and u-boot, m1n1 etc maintainer
<enchiladasconpinguino[m]>
oh awesome
<mps>
enchiladasconpinguino[m]: I already have APKBUILD for linux-asahi in my local repo, but didn't merged to aports yet, waiting for more stable release and upstreamed most things
<enchiladasconpinguino[m]>
well my dream is now that alpine-standard-3.16-aarch64.iso will boot ootb on m1s !
<mps>
'13:25 ... enchiladasconpinguino[m]| oh awesome', yes I use awesome wm ;)
<enchiladasconpinguino[m]>
maybe that is too soon idk, might stay edge?
<mps>
it is to early to have asahi in alpine stable I think
<enchiladasconpinguino[m]>
yeah, that's what I thought
<enchiladasconpinguino[m]>
realistically quite a while
<enchiladasconpinguino[m]>
going to probably try awesome wm and xfce here on the mini next ;)
<mps>
but you are free to prove me wrong
<enchiladasconpinguino[m]>
that's why I was thinking reuse the "embedded device" stuff with pmbootstrap
<enchiladasconpinguino[m]>
where weird kernel and other customizations are well handled, and does installs with edge
<enchiladasconpinguino[m]>
^ at least for alpine
<mps>
yes, but I'm not much interested in pmbootstrap, I waste my free time mostly on alpine base, and because this I wrote 'you are free to prove me wrong' ;)
<mps>
and you reminded me, have to upgrade kernel for alpine now
<enchiladasconpinguino[m]>
situation with arch is much better in the short term I think, existing standard uefi archboot-aarch64.iso could add asahi packages on install without too much work
<mps>
enchiladasconpinguino[m]: yes, but arch is not musl based ;p
<mps>
well, I would rephrase ncopa as 'musgnu linux' because we use gcc as primary compiler
<mps>
enchiladasconpinguino[m]: if you have improvements/fixes/ideas for install script I posted it will be appreciated if you post it to me or somewhere else
coderush has quit []
m6wiq1 has joined #asahi
ciggi has joined #asahi
ciggi_ has quit [Read error: Connection reset by peer]
m6wiq has quit [Ping timeout: 480 seconds]
<Chainfire>
marcan> just out of 'academic' interest, which memory controller optimization isn't enabled yet?
<gsora[m]>
live from asahi! yay!
<marcan>
Chainfire: there is a power management setting that engages when the CPUs increase in frequency
<marcan>
that was in the old cpufreq patchset but I decided to punt on it for now since its impact is negligible and things changed on M1 Pro/Max/Ultra, and it might make sense to wait until more of the memory controller is understood
riker77 has joined #asahi
<enchiladasconpinguino[m]>
question, should I expect virtualization/VMs to work with latest kernel?
<enchiladasconpinguino[m]>
I see "kvm: virtual gic" in the dmesg, which is encouraging ..
Ry_Darcy__ has joined #asahi
<Chainfire>
well that code I mentioned is absolutely pointless, I just like to play with optimizing it, so no biggy there, just interest.
<j`ey>
enchiladasconpinguino[m]: yes
<j`ey>
enchiladasconpinguino[m]: assuming a new enough qemu
<rikkaa>
qemu-system-aarch64: -accel kvm: -m and ,maxmem option values require an IPA range (40 bits) larger than the one supported by the host (36 bits)
<rikkaa>
qemu 6.2 not work, I will build qemu form source.
MajorBiscuit has joined #asahi
<mps>
rikkaa: qemu 6.2.0 with patches
<mps>
6.2*
<rikkaa>
mps: where is patches
<mps>
tpw_rules: I see, thanks. but I don't think alpine is ready to be added there, yet
<kode54>
I have displayport working .... using an HDMI to DisplayPort active cable lol
<rikkaa>
mps: Thank you!
<tpw_rules>
mps: that's fair. just didn't want you to be left out
<mps>
tpw_rules: thank you for care
<kode54>
hey
<kode54>
maybe you can get this right where macos doesn't
<kode54>
when you get DP working
<kode54>
support Variable Refresh Rate
<kode54>
and actually make the setting stick across reboots
<kode54>
macOS Monterey added the setting, but it either doesn't work, or it just forgets to reset it on reboot
<mps>
tpw_rules: though alpine is easy as enchiladasconpinguino[m] demonstrated
<ah-[m]>
is there a good way of upgrading an old manual asahi installation to the alpha setup?
<mps>
ah-[m]: ?
<mps>
you mean from Arch to alpine?
<psykose>
to the installer version
<mps>
ah-[m]: sorry, I misread
<ah-[m]>
I meant from an asahi installation from about 1-2 months ago to the alpha version. i currently have m1n1 + u-boot installed manually, and just a default alarm setup
<mps>
I read 'alpha' as 'alpine'
<rikkaa>
tpw_rules: what desktop environment do you use at nixos? i try xfce, but webkitgtk build fail.
MajorBiscuit has quit [Quit: WeeChat 3.4]
<enchiladasconpinguino[m]>
bug: machine cannot reboot properly yet? poweroff/halt works
<j`ey>
reboot should work
<enchiladasconpinguino[m]>
reboot just hangs, never comes back
Major_Biscuit has joined #asahi
Major_Biscuit has quit []
Major_Biscuit has joined #asahi
<enchiladasconpinguino[m]>
maybe its a pmu fix in recent patch
<j`ey>
it's worked for a while
<enchiladasconpinguino[m]>
hrmm then probably alpine specific issue..
<mps>
enchiladasconpinguino[m]: reboot works fine
<mps>
in alpine I mean
<mps>
and poweroff ofc
<mps>
enchiladasconpinguino[m]: I'm using asahi with alpine as my primary workstation for nearly 4 months
<zorun>
nice job about the release!
<mps>
enchiladasconpinguino[m]: what is your machine?
<mps>
j`ey: you use alpine also? or just tested it
<j`ey>
im running alpine, but not as a daily driver
<enchiladasconpinguino[m]>
mps: m1 mac mini
creechy has joined #asahi
<mps>
enchiladasconpinguino[m]: maybe I didn't enabled something in kernel for mini, I don't have this machine so couldn't test
<bligma[m]>
wait is my laptop supposed to freeze during partition phase of install? sorry if wrong channel
<j`ey>
yes
tt022[m] has joined #asahi
<zorun>
marcan: is there a contact email address for the project? I'd like to contact lwn.net about the M1 in the gcc farm, and keep you in copy
<marcan>
there's no project email per se
<marcan>
I'm marcan@marcan.st
joshtau has joined #asahi
<enchiladasconpinguino[m]>
question, how do I load the aquantia kernel module? I saw ethernet was marked as working yes? only getting wifi interface
<joshtau>
it. The screen goes black, I can access it via ssh but can't change to a virtual console. Is there a better place to log this issue? (sorry)
rikkaa` has joined #asahi
rikkaa has quit [Remote host closed the connection]
MZG[m] has joined #asahi
<jannau>
joshtau: does the screen comes back after pressing "brighter"/F2?
<joshtau>
yes! it does. no idea why i didn't think of that first. thanks so much!
Ry_Darcy_ has joined #asahi
<Chainfire>
marcan> you note Chromium has the page-size issue, so I assume someone tried to build it? Cross from x86 or on the m1?
<jannau>
Chainfire: it's known what needs to be changed and someone was working on it and in contact with the chromium team
rikkaa has joined #asahi
<mps>
enchiladasconpinguino[m]: you use kernel I posted?
<Chainfire>
jannau> the guy from the bugreport dropped it due to lack of interest and time
<jannau>
joshtau: brightness is currently just off and on, strange that something turns it to 0% automatically
chir4gm has joined #asahi
rikkaa` has quit [Remote host closed the connection]
<Ry_Darcy_>
For anyone who might be affected by the macos 12.3 update blight (brick): idevicerestore *.ipsw will not work as so, needed to supply the -e (erase) flag to get the process to move as idevicerestore was stuck on the "waiting for device to enter restore mode" even though the device was in this mode already.
<mps>
enchiladasconpinguino[m]: I see I didn't enabled CONFIG_NET_VENDOR_AQUANTIA, is this ethernet in mini? if yes I can do build it now and upload
<jannau>
AQUANTIA is the 10GB ethernet in the mac mini and studio
<mps>
jannau: thanks, will add it now and rebuild kernel
<mps>
but I heard there are two ethernets in mini, is that right?
Ry_Darcy_ has quit [Remote host closed the connection]
<mps>
s/right/true/
<jannau>
it has 1GB thernet with and a 10GB ethernet option
<mps>
same driver?
<enchiladasconpinguino[m]>
wiki just says "linux-asahi (5.17)" for ethernet support
<enchiladasconpinguino[m]>
so I think there is an Asahi patch on top of the aquantia driver (?)
<jannau>
mps: no, 1GB is CONFIG_TIGON3, 10GB is CONFIG_AQTION
<mps>
jannau: ah, I remember now, and because that I have CONFIG_TIGON3=m
<mps>
didn't know that aquantia also needed
<jannau>
the only change necessary for the aquantia nic was a change to read the mac address from the device tree but that should be already upstream
<mps>
enchiladasconpinguino[m]: uploading, will be soon on the url with same file name
<tpw_rules>
rikkaa: xfce works on the 4K kernel
<tpw_rules>
rikkaa: webkitgtk should be able to build for 16K if you have 16GB ram and a few gigs of swap. i've done it before
<tpw_rules>
it just takes an hour at least
<enchiladasconpinguino[m]>
mps: other missing module is btrfs btw
<enchiladasconpinguino[m]>
not a lot of software supports 16k pages, gonna be awhile on that one tbh
<enchiladasconpinguino[m]>
so .. why not 4k by default for now ?
<tpw_rules>
almost all software does. just some big important names don't
<j`ey>
enchiladasconpinguino[m]: most software supports 16k
<enchiladasconpinguino[m]>
no chrome, no electron, no games ..
<j`ey>
that's why the option for a 4k kernel is there too ;)
<tpw_rules>
that's just chrome twice, and they did it to themselves
<enchiladasconpinguino[m]>
lol correct
<mps>
enchiladasconpinguino[m]: you use btrfs? I don't know anyone who does ;)
<sven>
:D
<enchiladasconpinguino[m]>
for root yes I do btrfs on some machines, but all data is zfs
<enchiladasconpinguino[m]>
btrfs is quite reliable now for single disks
<piroko>
yeah just don't RAID it
<piroko>
lesson learned thanks to my synology NAS -_-
<rikkaa>
tpw_rules: oh! thank you! I will try 4k
<mps>
I mostly use f2fs on everything, but most of my machines are with e/mmc
<rikkaa>
tpw_rules: Why do many things need to be compiled, is it because there are not enough nix compilation servers or unstable is too new
<tpw_rules>
rikkaa: because you chose to use the 16k kernel. libunwind is a base package of a lot of things in nixos, and it doesn't work on 16k
<tpw_rules>
you shouldn't need to compile anything really if you set the 4k kernel option and the cross compiled kernel option appropriately
<mps>
enchiladasconpinguino[m]: kernel is uploaded
vmeson has quit [Quit: Konversation terminated!]
<tpw_rules>
(and you used the version of nixpkgs in the iso)
<enchiladasconpinguino[m]>
mps: I can safely just re-extract tarball over root?
<mps>
enchiladasconpinguino[m]: yes
<enchiladasconpinguino[m]>
kk thanks
<mps>
I did every day few times
<rikkaa>
tpw_rules: I updated nix-channel
vmeson has joined #asahi
<tpw_rules>
rikkaa: ah, well that i can't say. usually unstable won't advance until stuff is compiled. but try 4k first
<tpw_rules>
it will probably have to compile a new kernel and u-boot, but that should only take 5 or so minutes
<mps>
enchiladasconpinguino[m]: I can add btrfs, do you know which options are needed
<rikkaa>
tpw_rules: I switch to 4k, thank you
diego[m]1 has joined #asahi
diego[m]1 is now known as xyaman
<kov>
j`ey, kode54 indeed, been using gnome for a while now, with debian and fedora, both working fine
<kov>
I am actually going to reinstall the air now and see if I can write up a guide for manually installing fedora
<kov>
I need to do the reinstall anyway to get the chainloading stuff and be on the new firmware ABI
m6wiq has joined #asahi
<enchiladasconpinguino[m]>
mps: should only be `CONFIG_BTRFS_FS=m`
<mps>
CONFIG_BTRFS_FS_POSIX_ACL?
<enchiladasconpinguino[m]>
that one is optional, but doesn't hurt afaik `CONFIG_BTRFS_FS_POSIX_ACL=y`
<mps>
enchiladasconpinguino[m]: uploaded, though with CONFIG_BTRFS_FS=y
Ry_Darcy_ has joined #asahi
bisko has joined #asahi
<Ry_Darcy_>
Managed to restore to 12.1 and repeated Thomas G ´s steps from curl -sL to the last messages "wrapping up", "press enter to continure" Still getting a black screen i.e. no screen output. Not a tty nada. No hardware/cable change in the meantime. Ideas?
<j`ey>
Ry_Darcy_: can you get into recovery mode?
<Ry_Darcy_>
Yes, and then?
<j`ey>
well I was just wondering if it was bricked again
<j`ey>
why not wipe the new partitions and try the official installer?
<Ry_Darcy_>
No, I have recovered to 12.1...
<Ry_Darcy_>
In recovery mode, I have three menu items - I) Macintosh HD ii) Debian iii) Options.
marvin24_ has joined #asahi
<Ry_Darcy_>
I might be forced to... Why does Thomas G 's installer not work - at least for me. I follow all of the instructions to the letter. Scratches head...
<Ry_Darcy_>
A prereq for the Asahi installer is 12.3 which for obvious reasons is a no-go for me.
bisko has quit [Ping timeout: 480 seconds]
<j`ey>
if you use expert mode, you can use 12.
<j`ey>
1
bisko has joined #asahi
<Ry_Darcy_>
Will give it a whirl and report back - soon hopefully...
marvin24 has quit [Ping timeout: 480 seconds]
Nspace has quit []
bdju_ has quit [Quit: Reconnecting]
bdju has joined #asahi
<Ry_Darcy_>
I prefer Debian anyway... but..
Nspace has joined #asahi
<Glanzmann>
Ry_Darcy_: If you want you can msg me your email address. I'll be in one hour back home and test the installation on the mini and also record it and let you know how it went.
<Glanzmann>
Otherwise I'll report back in the irc channel. you can check the irc logs here: alx.sh > community
<Ry_Darcy_>
How to message?
<Ry_Darcy_>
Here
<Glanzmann>
you can also send me an email thomas at glanzmann dot de
<Ry_Darcy_>
sean at jdcassidy dot eu
<rikkaa>
tpw_rules: xfce scale not work (black screen) do you do the same?
<Ry_Darcy_>
The Asahi installer installer in expert mode will not accept a base macos of 12.1. Message "Your system firmware is too old"... shite.
<tpw_rules>
rikkaa: sorry, i'm on a mac mini, so i don't use the scale settings
<tpw_rules>
Ry_Darcy_: why can't you upgrade to 12.3 again? do you have one of the mbps with a replacement motherboard?
Lazurus has joined #asahi
<Ry_Darcy_>
Nothing has been replaced on my M1 (2020). The 12.3 bricked the machine nevertheless.
<rikkaa>
tpw_rules: I just for scale to use xfce
<Ry_Darcy_>
update
<tpw_rules>
you might have to be patient then. you need to have installed 12.3 on the machine at least once before so the system firmware is updated before the official installer will work
<Sobek[m]>
Except he has one of those machine that currently cannot update to 12.3 until apple fix their mess
QuixoteDreamPool[m] has joined #asahi
<tpw_rules>
that's why i said he has to be patient
<tpw_rules>
(also he just said he didn't)
<QuixoteDreamPool[m]>
hi everybody!! Congrats for this amazing work!
QuixoteDreamPool[m] is now known as Quixote[m]
<Quixote[m]>
I'm trying to set up ufw but doesn't work, i tried to put iptables (legacy) but still is not working, any idea to fix that?
<Ry_Darcy_>
Patience is not one of my virtues.
<Sobek[m]>
tpw_rules: He said that his machine hadn’t had any replacement but was still bricked by 12.3 (and he has had quite a bit of trouble reviving it as the backlog shows).
<tpw_rules>
maybe it was just a random fluke then?
<Ry_Darcy_>
@Thomas G. E-Mail unterwegs.
<tpw_rules>
Glanzmann is Thomas G's nick here
<Ry_Darcy_>
Not sure, doze and apple hate me. Penguins and Mainframes love me.
<Ry_Darcy_>
Fair enough. Thanks.
<Lazurus>
just wanted to chime in and congratulate you all on the project, an amazing thing. Probably gonna run it at some point in the future
<Ry_Darcy_>
No. I spent most of the afternoon reverting to 12.1. This is or was the weekend... :)
creechy has quit [Ping timeout: 480 seconds]
<mattebullet[m]>
I had a small problem with the first run of the install script that I wanted to report. It failed while preparing target volumes on a diskutil command. I run it a second time and the second time it worked. https://pastebin.com/8Xy3wbWR
<Glanzmann>
Ry_Darcy_: Got your email. Can I give you a call?
<mps>
enchiladasconpinguino[m]: fine, thanks for info
<Glanzmann>
Ry_Darcy_: My current working assumption is the following: Macos 12.3 works on your mini because - to my knowledge - you need to have it to be able to choose to install the stub 12.3.
<Glanzmann>
So I think what went wrong is, that after the installer turns of the machine, you did not press the power button long enough. As a result it tries to boot from a non existing partition.
<sven>
uhhhh…
<Glanzmann>
You don't see anything because the mini has problems with the hdmi.
<Glanzmann>
sven: Do you agree or is something wrong with my reasoning?
<sven>
didn’t he have to DFU restore back to 12.1?
<mps>
do we need SND_SOC_TAS571X in kernel?
<sven>
he couldn’t even get into recovery when 12.3 was installed iirc
<Glanzmann>
mps: To my knowledge only CONFIG_SND_SOC_TAS2770.
<sven>
that very much sounds like that 12.3 bug
kameks has quit [Ping timeout: 480 seconds]
<Glanzmann>
sven: But now he reproduced his problem with 12.1
<Glanzmann>
Can you install a 12.3 stub with a 12.1 macos?
<sven>
but he can still access recovery now
xeq has joined #asahi
<Glanzmann>
sven: Okay, let me try to repeat his symptoms.
<sven>
no idea
<sven>
the reason the installer requires 12.3 mainly is because bloss started to work there
<Glanzmann>
sven: What is bloss?
<sven>
bless
<mps>
Glanzmann: I asked because you have it in your config
<Glanzmann>
Oh, I see.
<Glanzmann>
mps: I'm running a debian distro config now for my kernel so I have all sorts of crap in my config. But I try to stay as closs as possible with the debian kernel config as I can.
<Glanzmann>
close*
creechy has joined #asahi
<mps>
Glanzmann: thought so
<Glanzmann>
sven: And about the idevicerestore it is tricky. When I had to do it, I screwed it up multiple times until I suceeded.
<Ry_Darcy_>
It would seem, so. With the (faulty) 12.3 for my M1, the result of Thomas G 's script was the same (black screen). After a power off for the night and subsequent power on this morning I had the blinking orange LED. Having reverted to 12.1 this afternoon, The result after running Thomas G's script - still a black screen.
<Glanzmann>
Ry_Darcy_: But, when you turned the mini on did you press and **hold** the power button for 20 seconds?
<Ry_Darcy_>
Yes, until the options screen appeared and a good five seconds after that.
<Glanzmann>
Did you then select debian?
herbas has quit [Remote host closed the connection]
<Glanzmann>
j`ey / marcan / kettenis / jannau: Is there also a fix for the 4tb issue?
<tpw_rules>
Glanzmann: yes, it's been fixed as of a couple hours ago
<Glanzmann>
Than I'll rebuild debian.
<tpw_rules>
there is a patch to u-boot
wxllow[m] has joined #asahi
<arnidg[m]>
I want to switch from the Asahi Minimal Install to the Asahi Desktop Install. To uninstall my current Asahi Linux install, which partitions should I delete?
<Glanzmann>
arnidg[m]: the ext4, esp and 2.5GB one.
<Glanzmann>
arnidg[m]: Do you delete from Linux or Macos?
<arnidg[m]>
MacOS
xeq has quit [Quit: This computer has gone to sleep]
wxllow[m] is now known as willow[m]
Ry_Darcy_ has quit [Quit: Page closed]
<Glanzmann>
arnidg[m]: Do this: 11:04 <marcan> ok, `sudo -i` to root and run `curl -L https://alx.sh/wipe-linux | sh`
<Glanzmann>
arnidg[m]: This remove all partitions. Than re-run the installer.
<Glanzmann>
arnidg[m]: Yes, but you should have backups of any sensitive data. Because people in here more than once accidently deleted all their data due to not paying attention what they we're doing.
<Glanzmann>
arnidg[m]: I never manged to delete my data, however I once killed my 1tr so I was unable to install updates.
<nyanpasu64>
is asahi-audio unsupported on MacBook Air?
<Glanzmann>
nyanpasu64: The audio jack works.
<marcan>
lol, I just noticed the OS sizing bug, it wasn't a rounding issue
<marcan>
it's just broken
<marcan>
fixing...
<nyanpasu64>
i meant the speaker dsp setup
<nyanpasu64>
also is there no printscr key for alt sysrq?
<Glanzmann>
nyanpasu64: It is possible to enable it in the dtb, but you could accidently kill your speakers if the volume is too high, so better wait until you receive a kernel update from marcan.
<Glanzmann>
nyanpasu64: Yes, this is annoying. You can use an external keyboard. Or press the power button to shut down (at least in Debian). If you ever find out how to sysrq on this machines let me know.
<sandornagy[m]>
I'm trying to install fedora on my m1 air. I set up asahi installer with UEFI environment only (m1n1 + U-Boot + ESP) but I can't seem to be able to get fedora installer to boot. I tried putting both the iso and raw image on a flash drive, am I missing something?
<Glanzmann>
sandornagy[m]: It will not boot.
<tpw_rules>
sandornagy[m]: fedora's installer wouldn't have a kernel that is compatible with these machines yet
<sandornagy[m]>
ah I see, thank you :)
<Glanzmann>
sandornagy[m]: You could try the debian live image and than use that to create the aprtitions, extract the fedora arm rootfs, make it bootable.
joshtau has quit [Quit: Connection closed for inactivity]
<jannau>
nyanpasu64: speaker DSP from asahi-audio are for now macbook pro 14/16" only
<Glanzmann>
sandornagy[m]: Someone earlier said he write a guid for gedore.
<Glanzmann>
I thought about packaging it, because then we have a slight chance that torvalds will use the platform. As a result the patches would go faster to upstream. :-)
<tpw_rules>
rikkaa: i just pushed an update to the m1-support that lets you use 16K pages without recompiling so much
<Glanzmann>
jannau: Air works as well if you modify the dtb, but you probably know that.
<tpw_rules>
try it out, i am interested to see if any other software is broken. firefox and xorg seem to work
<gsora[m]>
The amount of stuff that’s using jemalloc is… bigger than I expected
<tpw_rules>
either not too much seems to use it on nixos, or nixos's build servers run on 64K pages
<tpw_rules>
i'd expect more to be broken too, but i haven't figured out why it's not
<sandornagy[m]>
Glanzmann: I think I will just wait for that. I tried the normal desktop setup and it worked perfectly but I prefer fedora over arch. I'd happy to help out with that guide even if just testing
<Glanzmann>
jannau: I see, but the speakers in general work, I heard it with my own ears.
<jannau>
sure but that was not asked
<Glanzmann>
sandornagy[m]: Okay, maybe I'll be able to put something together this evening. you can send me an email or check the wiki. I'll put it there if I succeed. thomas@glanzmann.de
<Glanzmann>
jannau: Oh, I see.
<marcan>
Adie[m]: that dumb partitioning bug should be fixed
<marcan>
I thought it'd something more subtle, but it was off by 2.5GB... :-)
<nyanpasu64>
is it safe to use speakers without dsp?
<Glanzmann>
nyanpasu64: I don't think so, but povik knows for sure.
<jannau>
not related to the DSP, it's most likely possible to destroy the speaker with too loud audio
<Glanzmann>
nyanpasu64: In short don't do it.
<mps>
marcan: why is jack removed in kernel commit 895419184a2edef280ec9f8b8fadfd68c70e8790
raizo[m] has joined #asahi
<mps>
- {"Jack Mixer Volume", "63"},
<marcan>
that's not removed, it's added
<marcan>
that line was hiding the jack control
<marcan>
povik asked me to change that because people were stuck at 100% volume
<mps>
ok, why hiding it
<Glanzmann>
marcan: Ah.
<marcan>
the problem now seems to be that pipewire doesn't consider that the master volume and it starts off at 0 so you need to turn it up manually
<Glanzmann>
marcan: Yep, I had that with pulse/aslsa as well.
<mps>
marcan: ok, thanks. I will try to revert and see
<marcan>
why woul you revert?
<marcan>
just turn it up
<mps>
marcan: I don't see it in alsamix or amixer
<rikkaa>
tpw_rules: wow! I will update later
<mps>
alsamixer*
<mps>
is there UCM file for macadio somewhere?
Denver has joined #asahi
Denver has quit []
___nick___ has quit []
___nick___ has joined #asahi
<doggkruse>
So I ran the installer 3 times (2 failures) and I now have 3 "Asahi Linux" volumes that show up under mac OS that appear to be 2.5G each
<doggkruse>
I'm assuming 2 of these are macOS stubs I don't need?
___nick___ has quit []
<tpw_rules>
yes
<doggkruse>
tpw_rules: any ide which two? :D
<doggkruse>
they all look identicle
<tpw_rules>
probably the one with the lowest partition number is the real one
<Glanzmann>
mps: For me, I need to start alsamixer, than press F6, select 0:default and than there is the knob.
<tpw_rules>
frankly i would blow all three away and then retry a fourth time
<doggkruse>
tpw_rules: One also shows having macOS 12.3 and the other two do not... the failed installs were at the apple download
<Glanzmann>
doggkruse: Do this: 11:04 <marcan> ok, `sudo -i` to root and run `curl -L https://alx.sh/wipe-linux | sh
<doggkruse>
Glanzmann: that will wipe all including the valid install correct?
Quinten has joined #asahi
<Glanzmann>
doggkruse: Yes.
<doggkruse>
thanks
<Glanzmann>
doggkruse: But if you just want to kill the invalid once.
<Glanzmann>
doggkruse: Than paste me the output of 'diskutil list' to tg.st/p
<Glanzmann>
And than tell me which is the invalid one and I give you the commands to wipe the others.
<Glanzmann>
doggkruse: Or when you're under LInux, do 'parted /dev/nvme0n1 print' and paste that.
<doggkruse>
yeah, I don't know which one is invalid :D
PaterTemporalis has quit [Ping timeout: 480 seconds]
<doggkruse>
the installer was also telling me the minimum partition was ~540G is that expected? I read the blog about the need for some space but didn't see a number called out
<tpw_rules>
probably
<tpw_rules>
due to snapshots
<doggkruse>
interesting, I didn't make the connection between snapshots and the linux partition, I was thinking that applied to the mac partition
<tpw_rules>
maybe we are talking about different things then
<doggkruse>
I'll wipe my install and go through it again... with some screenshots to confirm
<doggkruse>
since I need to clean up anyway
<tpw_rules>
i say wipe them all because if you don't you'll just have this big block of free space you can't give bakc to macos
<doggkruse>
but may be moot since I'm going to wipe it
<tpw_rules>
oh hm, it didn't get far enough to create the non-stub partitions
<tpw_rules>
then do diskutil apfs deleteContainer disk0s4 and disk0s5
<tpw_rules>
disk0s3 should be the real one
<Glanzmann>
doggkruse: I'm pretty sure that the last one is the active.
<Glanzmann>
tpw_rules: I think disk0s5.
<doggkruse>
tpw_rules: yeah, didn't get past the stub since it failed to download from apple
<Glanzmann>
tpw_rules: Because the installer adds the stub first.
<tpw_rules>
oh, that is a good point
<mps>
Glanzmann: what is name of the knob
<Glanzmann>
doggkruse: I would run diskutil apfs deleteContainer disk0s3; diskutil apfs deleteContainer disk0s4 and than use the graphical thing to grow macos an extra 5gb by 'deleting the free space'.
nicolas17 has joined #asahi
<Quinten>
Hi there, just wanted to ask about video playback on Asahi. Can´t get it to work and FF says the openh264 codec will be "installed shortly", but never actually installs :/
<tpw_rules>
can you name a site?
Lazurus has quit [Quit: Leaving]
<tpw_rules>
youtube works fine on my nixos install in firefox, but i don't think that uses h264
<Quinten>
any website, even youtube
<Quinten>
buffer is loaded, but it just shows the spinning wheel
<Quinten>
when I click on the timeline it does show that specific frame
<mps>
Glanzmann: I have `default:0 MacBook Pro J293 integrated audio`
Quinten has joined #asahi
<Glanzmann>
mps That is the right one and there is no jack mixer when you scroll to the right?
<Glanzmann>
Can you send a screenshot of a full screen xterm?
<Glanzmann>
of alsamixer
<mps>
no, left one is just 'Mixer'
<Quinten>
Anyone have an idea why the entirety of Asahi freezes up when I plug in usb-c earbuds? Nothing responds anymore, although if I e.g. clicked on FF during the freeze, it will open after I unplug the earbuds.
<nicolas17>
oh video playback due to audio issues is a thing yeah... on my Linux desktop I once had video in all apps stop working (Firefox, Discord, etc) and I was about to reboot to fix it, it turned out to be something stuck in bluetooth audio, I turned my headset off and everything got instantly unstuck
<jannau>
most people like their audio and video in sync. if audio isn't broken completely (missing audio device) it will affect video playback
zamadatix has joined #asahi
doggkruse has quit [Read error: Connection reset by peer]
the_lanetly_052 has quit [Ping timeout: 480 seconds]
doggkruse has joined #asahi
mefff has joined #asahi
<mefff>
hey guys, got the news about the release, don't have any questions, just wanted to say thanks for the work you done!
zamadatix has quit [Quit: Leaving]
Simonx22 has joined #asahi
testingtestest has joined #asahi
beacon[m] has joined #asahi
<beacon[m]>
hey guys, can I ask a question here
<j`ey>
yes
<doggkruse>
Is there an easy way to transfer files between mac and linux in this setup using internal storage or is a thumb drive / SD card the best bet?
<doggkruse>
can linux mount the mac partition or vice versa?
<Glanzmann>
Fedora boots ..
<beacon[m]>
I keep getting the "preboot folder path [path] doesn't exist couldn't set bless data..." error
<nicolas17>
doggkruse: I suppose you could make another FAT32 internal partition :P
<j`ey>
doggkruse: macos can read vfat, so you stash stuff on the EFI partition
<nicolas17>
also, you can read the mac partition from linux, but I don't think you can write yet
<nicolas17>
oh wait
<doggkruse>
nicolas17: Read is still very helpful
<nicolas17>
you can read apfs filesystems from linux, but actually the encryption in the default user-data mac volume may get in the way...
kreechy has joined #asahi
<mps>
nicolas17: this is interesting, how is possible to read APFS from linux?
<nicolas17>
apfs-fuse has been around for a while, I have used it to get binaries out of iOS ipsw files
<nicolas17>
oh the one you linked seems to be *another* implementation
<mps>
nicolas17: ohm, linux-apfs-rw-src-0_git20210823 is in alpine, and I'm building separate locally, have to follow new pkgs more carefully :_
<beacon[m]>
Hey guys, could someone help me, I keep getting the "preboot folder path [path] doesn't exist couldn't set bless data in reboot volume for device" error
<nicolas17>
interesting, I should try it... apfs-fuse actually has problems with some small compressed files in iOS images
<silverocean[m]>
Hi folks! Is there a way to turn off the 2 second delay in U-Boot? I tried `setenv bootdelay 0`, then rebooted, but if I runt `printenv`, it's automatically reset back to 2.
<mps>
I'm busy to build proper alpine dev pkg but I have this in mind
<mps>
silverocean[m]: you have to fix this in u-boot .config
creechy has quit [Ping timeout: 480 seconds]
* nicolas17
has no M1, is lurking here out of curiosity
<Glanzmann>
beacon[m]: What did you do to get this error?
<mps>
and build your local u-boot
<Glanzmann>
beacon[m]: Which macos version do you have?
<silverocean[m]>
mps: ooh, I see! let me give this a try
<Glanzmann>
beacon[m]: What m1 device do you have?
<mps>
silverocean[m]: this delay is sometimes useful, imo
<doggkruse>
nicolas17: Yeah, I'm assuming none of these support the apfs encryption?
<silverocean[m]>
silverocean[m]: is it expected for `saveenv` to be unavailable?
<beacon[m]>
Glanzmann: did the curl command, partitioned and chose the asahi linux desktop.
<beacon[m]>
Glanzmann: m1 air
<nicolas17>
doggkruse: right, dealing with encryption needs dealing with the SEP too
<mps>
silverocean[m]: no, it don't have persistent storage to save envs
<silverocean[m]>
mps: presumably if i need the delay, I can always boot into macOS and turn it back on
<doggkruse>
ah, that makes sense!
<beacon[m]>
macos version 12.3
<silverocean[m]>
ah, that explains it! thanks
<Glanzmann>
beacon[m]: I see, so when do you get that error?
<silverocean[m]>
mps: thank you!
<beacon[m]>
Glanzmann: after extracting root.img into partition
<Glanzmann>
marcan: 20:37 < beacon[m]> Hey guys, could someone help me, I keep getting the "preboot folder path [path] doesn't exist couldn't set bless data in reboot volume for device" error
<nicolas17>
doggkruse: I suppose you *could* make a new non-encrypted apfs volume/partition, either in internal or external storage, but then you might as well use VFAT and get reliable write support :P
<Glanzmann>
beacon[m]: Can you paste the complete installer output to tg.st/p ?
<mps>
silverocean[m]: it is possible ofc, but I don't think it is worth extra effort and reinstalling u-boot whenever it is needed. two seconds is not much
<Glanzmann>
beacon[m]: So bless selects the boot volume.
<Glanzmann>
beacon[m]: Can you also paste the output sw_ver ?
<nicolas17>
how fast is the full boot atm? from power off to Plasma login?
hellosway[m] has joined #asahi
<hellosway[m]>
Is there a way to manually install without the installer.
<hellosway[m]>
I want to atempt to use systemd-boot instead of grub
<hellosway[m]>
j`ey: I know, but is there a way to install it from scratch with a tarball
<Glanzmann>
beacon[m]: Have you already tried to wipe everything and run the installer again?
<hellosway[m]>
I can create a tarball with systemd-boot, but I cant find docs to install manually
<Glanzmann>
beacon[m]: Do you have filevault?
<Glanzmann>
beacon[m]: Are you running the installer from 1tr or from macos?
<silverocean[m]>
mps: fair point!
<j`ey>
hellosway[m]: you can install the uefi option, and then create a usb stick or something, but it sounds more painful than doing the asahi minimal install and modifying it
<beacon[m]>
Glanzmann: wdym 1tr
<beacon[m]>
oh one true uh no from macos
<nicolas17>
hellosway[m]: replicating what the script is doing by hand would be a pain, there's more steps than you think
<j`ey>
hellosway[m]: or you can modify the installer such that it downloads your zip file instead of the one from the repo
<nicolas17>
and no I don't think it's documented as a guide
<Glanzmann>
beacon[m]: Okay, than I would try to wipe the incomplete installation and try to run the installer again and see if you get the same error.
<Glanzmann>
beacon[m]: In order to do so you need to execute the following script as root: curl -L https://alx.sh/wipe-linux | sudo sh
<beacon[m]>
k cool thank you
<hellosway[m]>
j`ey: I think I will do this, but for systemd-boot the ESP has to be at /boot
<Glanzmann>
beacon[m]: Let me know if it helped, if not maybe send marcan an email and attach the pbot, but hopefully it works.
<hellosway[m]>
and the installer mounts it at /boot/efi
kjm99 has joined #asahi
<hellosway[m]>
nicolas17: It cant be that bad, you just copy the commands into a shell script, change hwat you want, then run it
<tpw_rules>
ftr systemd-boot works fine under nixos
csd123 has joined #asahi
FarukAydn[m] has joined #asahi
<FarukAydn[m]>
Is there any way to use asahi alpha without wired keyboard/mouse?
<j`ey>
with a USB BT adapter
<j`ey>
or over ssh :P
<Quixote[m]>
Please, Can I download the linux-headers from this: 5.17.0-rc7-asahi-next-20220310-12131-gbb583ca91538
<FarukAydn[m]>
but how am i going to configure bt options if I don't have a wired set?
<FarukAydn[m]>
usb bt is cheap enough solution though
<FarukAydn[m]>
can be done over ssh i guess?
<j`ey>
USB BT might Just Work
itisyeetimetoday[m] has joined #asahi
<j`ey>
Quixote[m]: you need those specific versions?
<itisyeetimetoday[m]>
I just wanted to ask, is it possible to install asahi on a external SSD?
<j`ey>
Quixote[m]: kernel headers should be backwards compatible
<j`ey>
itisyeetimetoday[m]: not fully, you still need 3GB of stuff on the internal nvme
<Quixote[m]>
j`ey: yeah bc is the kernel I have right now, I want to build wireguard
<itisyeetimetoday[m]>
j`ey: thank you, that actually pretty good. Does the script support external install?
<j`ey>
itisyeetimetoday[m]: not yet
<j`ey>
Quixote[m]: I think that should just work with the normal linux headers package
<itisyeetimetoday[m]>
Also, I don't see an unofficial install guide for Ubuntu, has anyone installed ubuntu yet?
<Glanzmann>
Quixote[m]: Are you running Debian?
<Quixote[m]>
Glanzmann: yes
<Quixote[m]>
> <@_oftc_Glanzmann:matrix.org> Quixote: Are you running Debian?
<Quixote[m]>
* yes, your version
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<byrantech[m]>
but min total size is 585 gb? reading marcan's article it should only occupy ~53 gbs right
<tpw_rules>
what is the bug?
<tpw_rules>
no, because you have time machine snapshots
<tpw_rules>
OS upgrades can create them even if you've never turned on time machine
<byrantech[m]>
got it tysm
<tpw_rules>
but the article linked in the announcement post i believe goes over how to delete them if you want more free space
byrantech[m] is now known as hello9999901[m]
<hello9999901[m]>
yeah i thought i didn't have any
r0ni has joined #asahi
<hello9999901[m]>
just checked and i do
<hello9999901[m]>
my bad; thanks for the help
kreechy has quit [Ping timeout: 480 seconds]
<hellosway[m]>
How can you change ther installer to use /boot for the ESP
idaadi[m] has joined #asahi
kreechy has joined #asahi
<idaadi[m]>
Did anyone have issues deleting the partitions created by the asahi installer ? do we use some other room for such questions and install/remove issues? thanks
testingtestest has quit [Remote host closed the connection]
kevanf has joined #asahi
<idaadi[m]>
thanks tpw_rules i will try it now , i tried via the diskutil gui and totally failed :)
<Mayel[m]>
👋 Anyone installed Alpine with Asahi yet?
<tpw_rules>
Mayel[m]: mps has been working on that
* j`ey
has alpine on his m1 too
csd123 has joined #asahi
<Mayel[m]>
Thanks, any issues to be aware of?
<j`ey>
not really
kreechy has quit [Ping timeout: 480 seconds]
<mps>
Mayel[m]: today one user (enchiladasconpinguino[m]) installed alpine
<mps>
with yesterday alpha release announce looks like I have finally to write guide for installing alpine
<j`ey>
Mayel[m]: the main thing is setting up the boot flow now
<idaadi[m]>
if i try a reinstall , should i use that script to cleanup policies or not really needed ?
<j`ey>
Mayel[m]: and the kernel
<hellosway[m]>
<j`ey> "hellosway: https://github.com/..."; <- I am trying to install from archlinux arm, without first boot scripts, so I can unhardoce the uuids
<mps>
I had a hope that I will have to write guide and scripts in May
<hellosway[m]>
and dont have to do that
<Mayel[m]>
Cool, I'll keep an eye out for the guide when ready then :)
rikkaa has quit [Quit: Connection closed for inactivity]
<mps>
Mayel[m]: you have to follow first steps in asahi installer and stop at install asahi distro (forgot exact option name)
abbe has joined #asahi
<mps>
Mayel[m]: and then you will have to run scripts which installs bootable alpine on usb disk, boot from it and do rest
<j`ey>
hellosway[m]: that's where the boot partition is setup, if you dont use that script you cn setup fstab yourself
<mps>
will ask enchiladasconpinguino[m] to post me changes for script which install alpine on usb disk
<Mayel[m]>
thx :)
kevanf has quit [Quit: Page closed]
<mps>
Mayel[m]: today I made linux-asahi kernel apk (ugly but works fine)
<hello9999901[m]>
Error: -69521: Your APFS Container resize request is below the APFS-system-imposed minimal container size (perhaps caused by APFS Snapshot usage by Time Machine)
<hello9999901[m]>
i already deleted all my tms
<mps>
I will ask Glanzmann tomorrow to explain at which point to stop in asahi installer
<hello9999901[m]>
what could be the possible cause?
<nicolas17>
hello9999901[m]: how much free space do you have now?
<j`ey>
mps: dont you just pick the UEFI option in the installer?
<hello9999901[m]>
nicolas17: 619 gbs
<mps>
j`ey: I forgot, but think yes
<hello9999901[m]>
nicolas17 that should be way more than enough right
<mps>
j`ey: would be nice if someone can tell where exactly to stop, or write which steps are needed to get this
<j`ey>
I dont think you need to stop at aall, UEFI just creates an ESP partition and installs u-boot and thats it
<mps>
but I'm now drinking red wine with friends, so I'm somewhat 'out-of-the-task'
mrguppy02[m] has joined #asahi
<mps>
j`ey: thanks, will look Glanzmann video tomorrow (I hope) and try to resurrect what I did last time
<Glanzmann>
mps: Its like all my other videos, just with Fedora this time. :-)
<mps>
Glanzmann: I guess so, but thanks for confirming
<nicolas17>
may want to add that -L :)
<mrguppy02[m]>
This might be dumb, but is my xattr the only one that does not have the -x flag? I do not have any other brew packages to replace xattr. Am I doing something wrong?
kreechy has quit [Ping timeout: 480 seconds]
Raito_Bezarius has quit [Ping timeout: 480 seconds]
csd123 has quit []
<Glanzmann>
mrguppy02[m]: It might be that the kernel config option is missing.
<Glanzmann>
mrguppy02[m]: I had for very long time that not included in the linux kernel.
<Glanzmann>
for debian.
<Glanzmann>
Only fixed it 2 days ago.
<tpw_rules>
i'm pretty sure mrguppy02[m] is talking about xattr under macos
<mrguppy02[m]>
Yep.
<tpw_rules>
you are on 12.3, right/
<mrguppy02[m]>
Yep
kreechy has joined #asahi
<mrguppy02[m]>
and the script just tries to do -wx and there's no -x flag for my xattr
<beacon[m]>
Glanzmann: hi i tried doing a fresh restall from earlier (this is the preboot folder path question) and it seems to still not be working
heth has joined #asahi
<mrguppy02[m]>
I created an issue on GitHub for the installer a day ago related to it if anyone wants to see the logs
<Glanzmann>
beacon[m]: Damn it.
<Glanzmann>
beacon[m]: Same error message?
<beacon[m]>
yeah
<j`ey>
mrguppy02[m]: why are you sure it's the -x?
Raito_Bezarius has joined #asahi
<mrguppy02[m]>
That's what I get from subprocess.CalledProcessError: Command '['xattr', '-wx', 'com.apple.FinderInfo',
<nyanpasu64>
is there a fast way to switch between oses on bootup? holding power is slow and awkward
<mrguppy02[m]>
Because it tries do to xattr -wx com.apple.FinderInfo .... /Volumes/Asahi right? , and that's where it breaks
<j`ey>
can you try run that command manually?
<mrguppy02[m]>
and if I try to xattr -x a file in a terminal on my own it tells me -x is unrecognised
<mps>
nyanpasu64: no, limited by apple boot system
<mrguppy02[m]>
or just do it
<mrguppy02[m]>
as in running the command
<mrguppy02[m]>
at first I thought I had had some xattr thing in my usr/bin folder
<nyanpasu64>
yikes
<nyanpasu64>
so you can't have grub boot macos?
<j`ey>
no
<mrguppy02[m]>
that does not work on intel Macs does it?
<mrguppy02[m]>
there would be dsmos freaking out
<mrguppy02[m]>
why would it work on AS
<mrguppy02[m]>
having it show up in startup disks in sys pref could be interesting
<AdwyzzOLEDEdition[m]>
nyanpasu64: no, grub is after u-boot
<nyanpasu64>
and neither u-boot nor grub can boot macos?
<AdwyzzOLEDEdition[m]>
nope
<AdwyzzOLEDEdition[m]>
at least afaik it shouldnt be possible
<nicolas17>
afaik u-boot and grub each take the environment one step farther from what macos expects and closer to a "normal" aarch64 system
<mrguppy02[m]>
you would have to basically hackintosh it to have it boot from grub
<nyanpasu64>
:(
<mrguppy02[m]>
because from what I understand macOS does not boot from EFI on arm
<AdwyzzOLEDEdition[m]>
it doesnt
<gsora[m]>
I wonder if uefi is a necessity for some reason or we could boot directly off uboot
<mrguppy02[m]>
I need to find and try to understand what causes my -x issue
<j`ey>
gsora[m]: you can
<mrguppy02[m]>
I genuinely just
<nyanpasu64>
am reminded of surface rt hacking
<gsora[m]>
Cool!
<mrguppy02[m]>
surface rt was more norma
<nyanpasu64>
except m1 macbooks are a far more attractive target than 9 year old tegras
<nyanpasu64>
and has more funding and more progress
<gsora[m]>
So as soon as uboot is available it’s standard then
<mrguppy02[m]>
that booted off of uefi
<AdwyzzOLEDEdition[m]>
gsora[m]: we currently cant and afaik that requires more work/isnt possible and imo theres no reason to do it
<mrguppy02[m]>
the whole Microsoft arm system is uefi
<mrguppy02[m]>
from windows phone 8 to now
<j`ey>
AdwyzzOLEDEdition[m]: we can, Im doing it (booting the kernel from u-boot)
<AdwyzzOLEDEdition[m]>
theres no reason to try skipping u-boot
<j`ey>
AdwyzzOLEDEdition[m]: skip grub they meant, not skip u-boot
<AdwyzzOLEDEdition[m]>
ooh i'm stupid
<mps>
gsora[m]: create extlinux/extlinux.conf on ESP, thats all
<j`ey>
and you can skip uboot too, with chainloading
<nicolas17>
yes grub can be skipped, "GRUB does not need any patches to work on Apple Silicon machines, as it relies entirely on EFI services to do its job. It is up to the OS distribution to decide what to use here; GRUB is merely an example."
<gsora[m]>
AdwyzzOLEDEdition[m]: Oh I didn’t mean that, I meant use uboot kernel loading features and boot off that
<gsora[m]>
Without grub
<nyanpasu64>
i have no clue how assembly programming, bootloaders, drivers, work, i'm just a quiche eater, albeit a bit more meaty than usual (compiled languages rather than js)
<AdwyzzOLEDEdition[m]>
j`ey: well yeah then cant you use EFISTUB for that since after u-boot it's just plain EFI, right?
<j`ey>
AdwyzzOLEDEdition[m]: exactly
___nick___ has quit [Ping timeout: 480 seconds]
<mps>
gsora[m]: with extlinux.conf (boot.scr) you don't need grub
<AdwyzzOLEDEdition[m]>
nyanpasu64: i just spend hours lurking and reading wikis
<gsora[m]>
mps: Yeah, and that’s cool af
<gsora[m]>
essentially once you get off m1n1 it’s like a common arm sbc, except for drivers ofc
<mps>
gsora[m]: I boot with this for long, few days ago switched back to grub again
* Glanzmann
hits bed. n8
<j`ey>
gsora[m]: note that extlinux.conf as far as I can tell, does not boot with the UEFI protocol. some features later may depend on UEFI runtime services
<mps>
Glanzmann: sleep well
<bluetail[m]>
going to sleep too
<Glanzmann>
mps: Thanks, you too. n8.
SamHep0803[m] has joined #asahi
<Quixote[m]>
Glanzmann: I installed your linux headers but doesnt work. Your link is for version: 20220310-00002 and my version is 20220310-12131
<gsora[m]>
j`ey: I wonder what would it be honestly
<Quixote[m]>
* headers but I thinl doesnt work., * is 20220310-12131 .
<tpw_rules>
gsora[m]: note that if u-boot or grub tries to load a device tree everything will explode
<gsora[m]>
j`ey: I can’t imagine what feature might need uefi explicitly
<gsora[m]>
tpw_rules: M1n1 does that right?
<sven>
sleep
<tpw_rules>
gsora[m]: yes, m1n1 loads and modifies the device tree
<j`ey>
gsora[m]: ^ PSCI sleep, cores off
<gsora[m]>
Huh
<gsora[m]>
tpw_rules: ah I see
<gsora[m]>
Well one less thing do do with uboot I guess
<SamHep0803[m]>
hi, i'm currently having an issue with the installer where when i go into 1TR, bputil fails with the error: "Failed to create local policy". i tried doing "csrutil clear" and rebooting but that hasn't fixed it
<gsora[m]>
j`ey: Ah alright!
<SamHep0803[m]>
hope this is the right place to ask lol
<Quixote[m]>
j`ey: do you know if I can download the right linux-headers version? I need exactly this: 5.17.0-rc7-asahi-next-20220310-12131
<Quixote[m]>
I try with 5.17.0-rc7-asahi-next-20220310-00002 but doesnt work
<j`ey>
Quixote[m]: im still not sure why you need that exact version, the headers should be backwards compatible
<Quixote[m]>
j`ey: When I build I get this: Module build for kernel 5.17.0-rc7-asahi-next-20220310-12131-gbb583ca91538 was skipped since the
<Quixote[m]>
kernel headers for this kernel does not seem to be installed.
<Quixote[m]>
I guess is for the version but maybe I'm wrong
<Quixote[m]>
> <@_oftc_j`ey:matrix.org> Quixote: im still not sure why you need that exact version, the headers should be backwards compatible
<Quixote[m]>
* When I build I get this: Module build for kernel 5.17.0-rc7-asahi-next-20220310-12131-gbb583ca91538 was skipped since the kernel headers for this kernel does not seem to be installed.
<j`ey>
Quixote[m]: did you try install the headers from Glanzmann's deb anyway/
<Quixote[m]>
j`ey: yes, it was first I did
<Quixote[m]>
I have already installed those headers.deb
kreechy has quit [Ping timeout: 480 seconds]
<SamHep0803[m]>
<SamHep0803[m]> "hi, i'm currently having an..." <- also tried doing csrutil disable and its still not working, not sure what to do here
<nilsi[m]>
Installation was super smooth but I got a linux noob question here. I got a message "this package is not supported by your architecture" for a bunch of programs I tried to install with jay from aur. Is there any way to get around this or they simple don't work? I tried installing teams, slack and bitwarden.
<j`ey>
nilsi[m]: some packages might not be available, some might just not be packaged properly/yet
<nicolas17>
nilsi[m]: aren't all 3 proprietary? doesn't surprise me there's no aarch64 binaries for them
<nicolas17>
also, for apps based on Chromium/Electron, even with aarch64 they won't work on Asahi yet
<mrguppy02[m]>
Ok so /usr/bin/xattr does have the -x , but the xattr command lacks it
<j`ey>
mrguppy02[m]: so what does: which xattr show?
<j`ey>
nicolas17: they work if you use a 4k kernel, which asahi isnt "officially" shipping yet
<mrguppy02[m]>
so , I tried rerunning it and removing the part of the stub.py that enables the icon , because I do not care for it
<nilsi[m]>
Okay, thanks. Then I know it wasn't something i did wrong. Most of them I get going in Firefox anyway.
<mrguppy02[m]>
did not work, probably more needed removing
<j`ey>
what was the error?
<mrguppy02[m]>
if I run a terminal with sudo xattr -x it brings me to the -h page
<nicolas17>
j`ey: ah right, I suppose building your own 4k kernel is still an option :)
<mrguppy02[m]>
if I run a terminal with /usr/bin/xattr it's fine
<tpw_rules>
mrguppy02[m]: what do you get if you do `which xattr1?
<tpw_rules>
s/1/`/ sorry
<mrguppy02[m]>
and if I ask the terminal where xattr is it only points me to /usr/bin and a xattr1
<gsora[m]>
Hoping jemalloc patch will ne upstreamed soon
<gsora[m]>
It’s a deal breaker for sure
<mrguppy02[m]>
in usr/share/man/man1/xattr.1
<mrguppy02[m]>
but if I try going there it does not exist. is it some sort of broken symlink?
<mrguppy02[m]>
it's annoying because I have to remove the container the script makes every time
<AdwyzzOLEDEdition[m]>
gsora[m]: honestly i didn't think this many apps had issues with 16K pages
kreechy has joined #asahi
<mrguppy02[m]>
should I try running it in advanced mode or something
<gsora[m]>
AdwyzzOLEDEdition[m]: Me neither, it’s absurd
<j`ey>
AdwyzzOLEDEdition[m]: not many do. it's just jemalloc and chromium that are the biggest ones
<AdwyzzOLEDEdition[m]>
also because can't you just make 4 "virtual pages" inside a 16K page and have all your logic stay the same? it's a bit of a hack but it would sork
<nicolas17>
AdwyzzOLEDEdition[m]: Chromium has issues with 16K pages, and (too) many apps are web apps running in Electron, that's most of the problems I think...
<SamHep0803[m]>
<SamHep0803[m]> "hi, i'm currently having an..." <- i found this one stackoverflow question that said that csrutil clear would work but i tried it and it didn't so i'm not too sure what to do at this point
<j`ey>
SamHep0803[m]: you should make an issue and attach the logs
<SamHep0803[m]>
j`ey: the thing is, i've had this issue before when messing around with SIP trying to get things to build on my mac
<SamHep0803[m]>
but i can't quite remember how i fixed it
FilipJeretina[m] has joined #asahi
hspak has quit []
kreechy has quit [Ping timeout: 480 seconds]
hspak has joined #asahi
<beacon[m]>
hey guys, does anyone know why this is happening
<j`ey>
beacon[m]: make a github issue and include the full logs
<hello9999901[m]>
just booted asahi
<hello9999901[m]>
works great! thanks to everyone for making it happen
kjm99 has quit [Remote host closed the connection]
kjm99 has joined #asahi
giskard has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
kjm99 has quit [Remote host closed the connection]
mefff has quit [Remote host closed the connection]
<mps>
got APFS mounted with linux
msteffen has joined #asahi
<nicolas17>
mps: cool, what did you use?
<mps>
nicolas17: build linux-apfs-rw, insmod and ` mount -o cknodes,vol=3 /dev/nvme0n1p2 /mnt/`
<mps>
I don't understand macos volumes so I just 'throw dice' :)
<nicolas17>
yeah I'll have to try that, when mounting an iOS rootfs, apfs-fuse gave I/O errors reading small files (and I once got it to crash running 'find')
<nicolas17>
I found a "go-apfs" command line tool that can extract individual files out of an image (no fuse mounting) and it turned out to fail with *the same set of small files*
<mps>
nicolas17: I can read some random files but don't dare to write anything
<nicolas17>
but at least I had better luck reporting the bug to go-apfs
<mps>
will create APFS on usb disk and test if write works
jeffmiw has joined #asahi
<nicolas17>
... wait, linux-apfs-rw is a kernel driver? I thought it was fuse too!
<mps>
nicolas17: yes, kernel driver
jelly has quit [Remote host closed the connection]
dantedrac[m] has joined #asahi
jelly has joined #asahi
<povik>
mps: you were on the right track with diffing the dts to explain different controls
<ducc[m]>
now that asahi alpha has come out, is there going to be a more lengthened cadence to progress reports, or will we see them more often?
<povik>
j313 is missing sound-name-prefix = "Jack"; in jack_codec
<povik>
that's why your controls are named differently, without the jack prefix, but other than that they are the same
<povik>
(it's not intentional)
<mps>
povik: you mean j293 missing Jack prefix
<gsora[m]>
Wait, so sound works but it’s not included in alpha?
<povik>
yeah, right
<povik>
j293 i meant
<mps>
povik: jack doesn't work when speakers are enabled?
<mps>
at least I can't get it when speakers are enabled
<povik>
as i said, it's should, it just moves to hw:0,1 which lot of userspace ignores by default
<nicolas17>
gsora[m]: afaik internal speakers kinda work but there's a risk of damaging them, so they're not yet enabled in alpha
<povik>
mps: what about, ehm, speaker-test -D hw:0,1 when both are enabled?
<mps>
povik: sounds complicated to use
<povik>
well it's not done
<povik>
we are missing UCM among other things
<povik>
the kernel side should work though
<mps>
ok, will test tomorrow
<mps>
too late is now
kreechy has quit [Ping timeout: 480 seconds]
<mps>
povik: thanks for explanation
<gsora[m]>
nicolas17: Ah this is why the asahi-audio repo asks you to leave speakers at -60
<nicolas17>
gsora[m]: in fact chadmed considered updating a readme to explain how to enable them manually, and then said "or maybe not, i dont want people yelling at me that "i" blew up their speakers"
<povik>
mps: no problem, let me know any issues you come across
<mps>
povik: sure
<gsora[m]>
Lol yeah makes sense to have them disabled right now
<mps>
some of us fried speakers with drivers on some other machines, someone even melted plastic
bgentil[m] has joined #asahi
<gsora[m]>
What the hell, really?
<gsora[m]>
I know earbuds can literally fry and blow up, but not due to bad drivers…
<gsora[m]>
(Not bad as in “badly written” mind you, just in their infancy)
<mps>
gsora[m]: I did it on rk3399 chromebook and m a r c a n did on older chromebook also
kreechy has joined #asahi
<psykose>
w i d e m a r c a n
<psykose>
i still chuckle at that
<mps>
because drivers didn't had limited power output
<nicolas17>
psykose: ｍａｒｃａｎ
<mps>
psykose: write 'm p s'
<psykose>
m p s
<nicolas17>
(fullwidth)
<mps>
hah, good, not highlighted
<gsora[m]>
Ahhh right rk chromebooks we’re a thing
<mps>
I usually used n_ick when don't like to highlight someone
<povik>
any clues to the 'audio causes stuck video playback' issue would be helpful
<povik>
especially if there's something in the kernel log
<mps>
povik: didn't had this
<povik>
haven't seen that myself (since very early admac driver)
<nicolas17>
mps: what about this :P
<mps>
nicolas17: ah, teach me the 'trick'
<psykose>
zerowidth space
<nicolas17>
yup
<mps>
how?
<psykose>
you copy it from somewhere
<psykose>
and just paste it in the middle
<psykose>
between the letters
<psykose>
unicode rendering will just not have it
<mps>
povik: I had stuck video earlier (in October/November) with usb audio card
<nicolas17>
my IRC client has Insert -> Special Character that lets me search the whole unicode... still not particularly convenient for this though :P
<psykose>
if you look at it in like.. busybox vi it will still show up
<nicolas17>
I sent "mp<zero width space>s: what about this"
<_alice>
oh i see
<nicolas17>
:)
<psykose>
:)
<psykose>
heh, even a tty renders the zw space fine for me
<AdwyzzOLEDEdition[m]>
<povik> "any clues to the 'audio causes..." <- it happened to me a while ago (not on a mac). you just need an app opening an audio device as an ALSA plugin exclusively and then any other app on my pipewire server will hang waiting to open the device. i don't know if it's the same thing but it seems like it is
<mps>
AdwyzzOLEDEdition[m]: I use dmix in asoundrc, so nothing bloks
<mps>
blocks*
<AdwyzzOLEDEdition[m]>
hmm then i have no clue
<AdwyzzOLEDEdition[m]>
but it might be related, because there aren't that many things that just hang video playback like that
Major_Biscuit has quit [Ping timeout: 480 seconds]
andrewp has joined #asahi
<winter>
<gsora[m]> Hoping jemalloc patch will ne upstreamed soon <-- i thought there wasn't a jemalloc patch at all?
andrewp has quit [Quit: Page closed]
<gsora[m]>
winter: I remember reading about a patch set to fix the page size issue, but I might’ve read wrong
<gsora[m]>
Yeah I read wrong, libunwind and emacs have patches ready
<nicolas17>
the alpha release blog post mentions patches sent to fix the 16K issue, and jemalloc isn't listed as 'patch sent'
<j`ey>
there isn't a patch for that, but you can build jemalloc with a flag..
<j`ey>
jemalloc doesnt seem to want to make ita runtime thing
<gsora[m]>
So a custom jemalloc PKGBUILD could fix this issue?
<M0x8FF[m]>
how can i make macOS the default boot option?
<M0x8FF[m]>
just from the boot picker / recoveryOS?
<j`ey>
I think recoveryOS has a startup disk application?
<tpw_rules>
you can also hold the option key in the boot picker and the "Continue" button will change to "Always Use"
<gsora[m]>
TIL!
<tpw_rules>
i only learned that from the release post lol
<clover[m]>
i was hoping to be able to run chromium inside of docker but that doesnt seem to work either