marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | Not ready for end users / self contained install yet. Soon. | General project discussion | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
PthariensFlame[m] has quit [Server closed connection]
PthariensFlame[m] has joined #asahi
marvin24 has joined #asahi
Supra has joined #asahi
<Supra>
uhm
Supra has quit []
marvin24_ has quit [Ping timeout: 480 seconds]
skipwich has quit [Server closed connection]
skipwich has joined #asahi
refi64 has quit [Server closed connection]
refi64 has joined #asahi
nyx_o has quit [Quit: WeeChat 3.3]
Guest563 is now known as go4godvin
coder_kalyan has quit [Server closed connection]
coder_kalyan has joined #asahi
handlerug has quit [Server closed connection]
handlerug has joined #asahi
WhyNotHugo has quit [Server closed connection]
WhyNotHugo has joined #asahi
raiaq has quit [Server closed connection]
raiaq has joined #asahi
tdmm has quit [Server closed connection]
tdmm has joined #asahi
jleightcap has quit [Server closed connection]
jleightcap has joined #asahi
artemist has quit [Server closed connection]
artemist has joined #asahi
mblancom has joined #asahi
mblancom has left #asahi [#asahi]
<VinDuv>
Glanzmann: I’m very confused. After you said that the tune2fs on loop0 didn’t work, I retried on my side, and it failed. Today I tried it again, and it worked.
<ThomasGlanzmann>
vinduv: thank you. so the trick is to mount it twice and change the uuuid on the second time I'll adopt the bootstrap.sh of Debian and my rc.local file to do the same
<ThomasGlanzmann>
metadata_csum_seed Allow the file system to store the metadata checksum seed in the superblock, enabling the administrator to change the UUID of a file system using the metadata_csum feature while it is mounted.
<ThomasGlanzmann>
from the manage
ThomasGlanzmann has quit []
<VinDuv>
Glanzmann: Mounting it twice doesn’t make a difference in my case; I tried it to see if it would make a difference, but it didn’t
Thomas_Glanzman has joined #asahi
<Thomas_Glanzman>
vinduv for me it did not work but I'll investigate later why and let you know what was it.
<Thomas_Glanzman>
that was with a single mount.
Thomas_Glanzman has quit []
tdwebryv^ has quit [Remote host closed the connection]
cde[m] has quit [Server closed connection]
cde[m] has joined #asahi
___nick___ has joined #asahi
___nick___ has quit []
___nick___ has joined #asahi
ndut2 has joined #asahi
* ndut2
hey yah all skraito here what's up , i am here
* ndut2
brb
* ndut2
after sabbath day guys i join you all at 1:48 more hour rest well ... . is your sabbath day soon , at friday night to saturday night ... .
* ndut2
get the fuck off libera none of your concern here ... .
* ndut2
even rasberry pi is stupid guys , i would rather buy samsung a70 for 180 usd dollar and play tinkering from that ... .
* ndut2
all just code in c++ , c and the rest is just header , the rest think yourself ... .
* ndut2
to be honest you though you all code the kernel ? i just laugh ... . your just coding delusion ... . think for yourself why We say that ... .
* ndut2
if not the deal here i don't even care asshole ... . just get the fuck off libera and all frog ... . if you want just ban and k-line me ... . i got better thing to do rather than teaching kids with big ego or proud ... .
* ndut2
that usb encryption is one more thing super nice to for your key to turn on computer ... .
* ndut2
the password make it at my 0day c++ password with 7 string incase you forgot it
* ndut2
the password make it at my 0day c++ password with 7 string incase you forgot it <--- or just use the word and hash it
* ndut2
the password make it at my 0day c++ password with 7 string incase you forgot it <--- or just use the word and hash it think why i say so
* ndut2
you can hash it with time ... . the rest is your imagination ... .which mean other thing like your whatever favourite number or even letter or whatever ... .
* ndut2
haha is my old memory why i am the best in crypto now ... . i fail secondary 5 times lol ... . therefore The Fear of The Lord is The Beginning of wisdom ... . therefore be humble ... .
* ndut2
don't use vpn and the rest prince , is interpol problem , you just create trouble for yourself and them to protect you ... . i reveal to you that wireless encryption ... .
* ndut2
you should first connect to wireless without encryption ... . than go to their https , fill in the form with your phone or ic ... . than get the password than connect to encrypted wireless ... .
* ndut2
you still sure there is privacy ? don't be fool ... . if not why the whole country has proxy ... . of course they know what's going on behind the door ... . even you need ic for your internet ... . brb working ... .
ndut2 was banned on #asahi by marcan [*!*@180.241.169.97]
ndut2 was kicked from #asahi by marcan [ndut2]
m42uko_ is now known as m42uko
VinDuv has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
VinDuv has joined #asahi
<Glanzmann>
marcan[m]: I did not understand why you change the uuid of the esp partition? Should it not be already unique due to the creation under macos/1tr?
bisko has quit [Read error: Connection reset by peer]
alicela1n has quit [Server closed connection]
<Glanzmann>
VinDuv: Your script works for me as well, I'm now rebuilding the Debian artefacts and testing it.
alicela1n has joined #asahi
bisko has joined #asahi
<marcan>
Glanzmann: the distro has to *find* the ESP to begin with, and the creation script sets a static UUID
<marcan>
I talked about this on the stream
<marcan>
if I make it dynamic then I need to tell the OS somehow what the UUID is so it can mount it
the_lanetly_052 has joined #asahi
<Glanzmann>
marcan: I see, is it possible to assign a random uuid from the asahi installer for the esp parition? Because I would like to use that for Debian, because I can identify the esp parttion by search for the uuid of the ext4 in grub.cfg.
<Glanzmann>
But if it is not possible, than I will just assign a new uuid like you did.
<Glanzmann>
marcan: Never mind, I'll also assign a new random uuid to the esp.
the_lanetly_052 has quit [Remote host closed the connection]
<marcan>
Glanzmann: it is possible, setting a fixed ID is optional in the installer
<marcan>
but I don't see how that would be much better than the hack of looking for the ext4 uuid in the grub.cfg in the ESP
<marcan>
I mean, how that hack would be much better than what I'm doing with the fixed initial UUID + that dd hack to change it
<kettenis>
Did I see correctly that you set the volume ID in the VFAT filesystem instead of the UUID in the GPT?
<Glanzmann>
marcan: I agree. I already adopted my code, currently testing it. And for me it is just a single extra line.
<Glanzmann>
kettenis: I use Debian grub, Debian grub can search for an ext4 by fs uuid. So I use that.
<Glanzmann>
Because Debian does it by default.
<marcan>
kettenis: you're thinking the PARTUUID, and yes, we're talking about FS-resident UUIDs, not PARTUUIDs
<marcan>
we can't set PARTUUIDs in the installer
<marcan>
diskutil won't let us and there's no way to bypass it to my knowledge
<marcan>
so those are just going to end up random
<Glanzmann>
marcan: Also they might change if any disk operations are done with diskutil due to reordering partitions.
<marcan>
I sure hope a partition reorder does not change PARTUUIDs, that would be one nasty bug
<Glanzmann>
Okay, maybe I misunderstood, but I thought that this happened once for me, but maybe I'm just confusing things.
<Glanzmann>
But I don't care because debian doesn't use them anywhere.
<mps>
so, macos doesn't have 'sgdisk'?
<kettenis>
heh, to me a UUID is an 128-bit integer that is insanely formatted such that everybody gets its endianness wrong ;)
arctan has joined #asahi
JuniorJPDJ has quit [Server closed connection]
JuniorJPDJ has joined #asahi
arctan has left #asahi [#asahi]
<Glanzmann>
That is not funny. Debian grub apparently can't read ext4fs that have metadata_csum_seed. Back to square one.
<mps>
Glanzmann: and (iirc) after that running 'grub-install' embeds gpt number to bootaa64.efi
<Glanzmann>
marcan: Debian has 2.04-20.
<mps>
I mean running debian grub-install
<Glanzmann>
mps: For the debian grub, not. The debian grub has a patch which picks up a grub.cfg from /EFI/debian/grub.cfg that searches for the boot or root partition by using the fs uuid and than loads the real grub.cfg
<marcan>
Glanzmann: that patch went in like days after 2.06 was released, and arch doesn't have that patch on top, so no
<marcan>
blame grub for not making a release in half a year :p
<Glanzmann>
hrhr.
<marcan>
I'm just going to turn off metadata checksums
<marcan>
I trust Apple SSDs enough for that
<Glanzmann>
marcan: Tell me what will you do now? Will you patch grub or will you write a script that runs in the initrd?
<marcan>
we can revisit, maybe go the grub patch route, in the future
<Glanzmann>
marcan: Any new thoughts on how to chain u-boot to the esp?
<Glanzmann>
I would like to keep metadata checksumming enabled, so I'll file a bug report against debians grub (one more patch can't hurt, can it?). And also use the Debian infrastructure to change the uuid in the initrd on first boot.
<marcan>
I'll look at that next, might take a couple days though, weekend got unexpetectedly busy with other stuff and I'm also feeling a bit sick today
<Glanzmann>
marcan: Get well soon. I'm excited what you come up with.
<marcan>
and as long as you stick to the installer framework and what it can do, you can do it however you want :p
<marcan>
the ESP is all yours, we support arbitrary files, and setting the UUID or not
<Glanzmann>
marcan: My plan is to get things in shape that you will include it. :-)
<Glanzmann>
marcan: I understand your reasoning, so I already change the uuid of the efs. But I'm stuck on this metadata checksum thing. Lets see if I can come up with something.
grgy has left #asahi [.]
<marcan>
might want to get some buy-in from other debian folks though
<marcan>
I don't mind listing unofficial stuff, but I don't want to call it "Debian" if it isn't official, same as I'm calling our (very minor) ALARM remix something else
<marcan>
could split up the images into two categories, "official" and "contrib" or somesuch
<Glanzmann>
Fine with me, so one Debian guy approached me want to get the d-i in shape for m1. But before that can happen the kernel needs to be upstream.
<Glanzmann>
marcan: Fine with me.
<Glanzmann>
Maybe some Debian guy takes over at some point once m1 Linux gets more traction.
<marcan>
well, I'd hope distros are interested in talking *before* kernels get upstream, so they can get everything else they need ready, unless they want to be a year behind everyone else
<marcan>
but yeah, I think I'll just have a "user-contributed" category, and then anything that at least works and isn't doing something utterly insane is fair game there
<Glanzmann>
marcan: For Debian that is almost nothing. I already looked at the code. The only thing that needs to be changed is that two flags need to be passed by the installer to the grub-install (removable media path; don't try to set nvram) And than the Debian installer works out of the box expect the kernel.
<marcan>
yeah, sounds about right
<marcan>
plus something to find the right ESP
<Glanzmann>
marcan: The debian guy wanted to include the firmware into firmware-nonfree. I told him it is not allowed to redistribute it and never heard back.
<marcan>
heh
<kettenis>
Glanzmann: U-Boot's EFI implementation does indicate that runtime support for EFI variables isn't there
<marcan>
does it? I saw EFI variables...
<kettenis>
it supports them at boot time
<Glanzmann>
kettenis: I see. The issue is that it tries to set the nvram variables, fails and than bails out.
<kettenis>
grub itself, or grub-install running in the kernel?
<Glanzmann>
But I already posted all my findings to the debian-boot mailing list and also asked for help and filed a bug (after they told me I should do so).
<Glanzmann>
kettenis: grub-install running in the kernel.
<kettenis>
right
<kettenis>
not sure if the Linux kernel advertises the situation (not supporting persistent EFI variables) to userland
<Glanzmann>
Neither do I, but the debian guy told me it is fine to detect the model and than set some flags so that it does no longer happen.
<kettenis>
what I'm trying to say is that there is a way to find this out and do the right thing instead of controlling this by passing stuff along from the install script
<VinDuv>
How does grub finds the ESP “normally”? Even on a x86 PC you can have multiple disks so multiple ESP
<marcan>
Glanzmann needs to learn the art of doing the right thing instead of the quickest hack :p
<kettenis>
the issue isn't unique to Apple M1 systems; basically all arm64 SBCs have the same issue
<Glanzmann>
kettenis: Got your message, I'll find out and ask them to fix it, if the infrastructure is there.
<marcan>
Glanzmann: you know you can also invite them here, right? ;-)
<marcan>
playing a game of telephone isn't terribly efficient
<Glanzmann>
marcan: Sure, maybe the guy was already here once.
<Glanzmann>
marcan: I'll invite him.
mlq_ has joined #asahi
<kov>
btw was there someone interested in doing a fedora installer build? I remember it being mentioned in a report maybe
<Glanzmann>
kov: Do it, if you want. I thought about doing it in order to onboard Linus.
mlq has quit [Ping timeout: 480 seconds]
<kov>
Glanzmann, it was you then? I would like to but honestly out of my depth at the moment, d-i is so much simpler xD
<kov>
maybe I'll end up learning but I was hoping this person existed and could provide some guidance at least
<Glanzmann>
kov: I already had a quick look at it, it might be as simple as dding an image and installing grub.
<kov>
yeah, I mean I have installed fedora already
<Glanzmann>
On the m1?
<kov>
yeah
<Glanzmann>
Well then ... :-)
<kov>
I'd like to get the installer up thought to make it simpler to do
<kov>
and to be able to more easily make the filesystem encrypted
<kov>
that bit is very tricky to get right
<kettenis>
that's why I think running the normal OS installer makes most sense for distributions that provide one
<tpw_rules>
^^
<mps>
I agree
<Tramtrist>
Hopefully I can get the slackware installer working 🙃
<Glanzmann>
kov: I don't know about other distributions, but on Debian it would be two commands and the rest the Debian infrastructure takes care of.
<j`ey>
Tramtrist: I know my coworker will be interested in that too
<kov>
Glanzmann, for rebuilding the installer?
<Tramtrist>
I plan on getting my first ever mac when the new Air comes out
<Tramtrist>
Will be a fun project if someone hasnt figured it out by then
<mps>
Tramtrist: I'm lazy to try alpine installer (I have fine working system) but I think it should work out-of-the-box with addition to install firmware from ESP
<Tramtrist>
👍
The_DarkFire_[m] has quit [Server closed connection]
<mps>
if I know python I would adapt these needed parts, but I don't
The_DarkFire_[m] has joined #asahi
<Glanzmann>
kov: I meant enable luks on a manual installation in Debian.
<mps>
i should look more deeply what are 'strong point' of using EFI with grub over plain extlinux.conf to decide what path to make for alpine
<Tramtrist>
for me it was the framebuffer on ryzen hardware.. but that was all
<mps>
my experience with EFI is only about two years ago installing syslinux (boot loader) on x86_64 old macbook
<kov>
Glanzmann, oh, what would those commands be?
<Glanzmann>
marcan: Disabling metadata_csum works, I just tested it, but I'm not happy with it, so maybe I find a way to do it the right way.
<kov>
Glanzmann, you do that before the partition step and the installation properly sets up crypttab and initrd?
<Glanzmann>
Normally I do that after partitioning and software raid but before lvm.
<Glanzmann>
On the m1, no need for a software raid.
<Glanzmann>
initrd is automatically done by debian unless you want to be able unlock it via ssh.
cyb3rn3rd[m] has quit [Server closed connection]
<kov>
cool
cyb3rn3rd[m] has joined #asahi
<kov>
yeah, if the fedora installer can be booted it will be even easier hehe
<kov>
but I had to do everything manually when I did it, it's not great
<Glanzmann>
kov: I see.
<Glanzmann>
kov: Debian live system to your rescue. :-)
<kov>
well, it's still manual with debian live system
<Glanzmann>
Yep, but you have all the tools and when some are missing, you can just install them.
<kov>
(it is what I used when I did it! though I booted it tethered using d-i's initrd)
<Glanzmann>
kov: Oh, good to know. :-)
<Glanzmann>
So the d-i initrd and the initrd for the live system are totally different.
<Glanzmann>
The live system is just the same image that I dd in a initrd with one addition: sudo -- ln -s lib/systemd/systemd init
<Glanzmann>
That starts systemd in the initrd.
mlq_ has quit [Ping timeout: 480 seconds]
* Glanzmann
just noticed that debian grub seems to have to different modes when installing. One is patching the grub binary with the partition, the other one is using the grub.cfg. I have not yet found out when Debian uses one method or the other one.
mlq_ has joined #asahi
nyx_o has joined #asahi
mlq_ has quit [Ping timeout: 480 seconds]
the_lanetly_052 has joined #asahi
Techcable has quit [Ping timeout: 480 seconds]
<Glanzmann>
marcan: Will you ship for alarm a kernel with 4kb page size or 16kb? Will you provide a repository for custom build packages like the kernel? How will the update process look like if device tree need to be updated?
bisko has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Glanzmann>
2 o'clock in the morning in tokyo, have a good night sleep.
bisko has joined #asahi
<j`ey>
yeah there's a repo, he set that up on the stream
bisko has quit []
bisko has joined #asahi
nehsou^ has joined #asahi
<Glanzmann>
I see.
<jannau>
currently just a kernel pkg and a modified mkinitcpio which we can hopefully drop if it's accepted either upstream or in alarm
<j`ey>
and that kernel pkg is just a random placeholder
<jannau>
it is certinly not random? it's a working asahi kernel
mlq_ has joined #asahi
<j`ey>
sorry not random, but he said it wasnt the latest branch
<jannau>
did I miss a kernel rebase?
<j`ey>
oh no, it is the right version. I dunno then, maybe I misunderstood what was said
bisko has quit [Read error: Connection reset by peer]
bisko has joined #asahi
<Glanzmann>
I see.
mlq has joined #asahi
mlq_ has quit [Ping timeout: 480 seconds]
bisko has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bisko has joined #asahi
bisko has quit []
bisko has joined #asahi
eroux has joined #asahi
eroux has quit []
ciggi_ has joined #asahi
ciggi has quit [Ping timeout: 480 seconds]
mlq_ has joined #asahi
mlq has quit [Ping timeout: 480 seconds]
the_lanetly_052 has quit [Ping timeout: 480 seconds]
bluetail[m] has joined #asahi
<bluetail[m]>
Hello. I am planning to installing Asahi-Linux on my M1 Mac Mini once it is ready. It should serve Network drives (SMB) and some other services such as a small webserver (Jellyfin). Also, running Asahi-Linux would fix the output on my external Monitor. macOS has a different display output than Windows or Linux. This claim can be measured... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/LLLEOPuPftfdqhljIvQIpoQI)
<bluetail[m]>
* Hello. I am planning to installing Asahi-Linux on my M1 Mac Mini once it is ready. It should serve Network drives (SMB) and some other services such as a small webserver (Jellyfin). Also, running Asahi-Linux would fix the output on my external Monitor. macOS has a different display output than Windows or Linux. This claim can be measured... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/canbsnPLTyBSwrKlEnHTuwkv)
<tpw_rules>
linux still uses all the same display output stuff on macos
<tpw_rules>
s/on/as/
<bluetail[m]>
tpw_rules: How come my Display is 'faulty' through macOS and working on Windows 10, given the iMac 2019 Windows 10 Bootcamp case?
<bluetail[m]>
Monitor misbehaving only on macOS is a LG-24GM79G.
<bluetail[m]>
It doesn't matter whether I use DisplayPort, USB-C, or even an Powered DisplayLink Hub for 300 Bucks.
<kloenk>
What does misbehaving mean?
<kloenk>
My monitor shows a black bar on the left side. But changing resolution to something other and back fixes it.
<tpw_rules>
well the iMac 2019 is a completely different hardware architecture
<bluetail[m]>
macMini m1, iMac 2019:
<bluetail[m]>
1. Luminosity (brightness) is too high, I can see the scanlines
<bluetail[m]>
2. Monitor flickers randomly for a split second
<tpw_rules>
and especially regarding how the hdmi signal is formatted and generated on m1 macs, both linux and mac use the same hardware and firmware to do it, so i would say it's unlikely that your problem will be solved by asahi
<bluetail[m]>
I tried iShades, but that did not preserve color accuracy if that makes any sense. I couldn't find the right values
<sarah[m]>
iPhones and macbooks are litterally the most locked down devices there are
<TellowKrinkle[m]>
iOS has issues with sideloading but macOS does not
<TellowKrinkle[m]>
Plenty of reasons to choose Linux over macOS but I don't think that's one of them
<bluetail[m]>
sarah[m]: I wished I could play old emulated games on my iPad pro 12.9 2020
<tpw_rules>
yeah, you can turn off all the signing and sideload blocking stuff trivially. as evidenced by the fact that you can load linux on it
<sarah[m]>
Oh wait my bad, sideloading is possible in macos. I was thinking of iphone. Sorey
<sarah[m]>
sorry
<sarah[m]>
I use windows, so I mixed that up. No direct experience with macos
<sarah[m]>
bluetail: Well the new macbooks have entirely new displays
<bluetail[m]>
sarah[m]: How are they different? They are still not matte
<sarah[m]>
Very few laptops have matte displays
<sarah[m]>
They have minileds now
<bluetail[m]>
I don't like mini-leds. I can see the blooming
<sarah[m]>
But Apple has the 2nd best anti-reflective coating. The dell xos 2-in-1 has the best.
<sarah[m]>
I also dislike blooming
<sarah[m]>
sarah[m]: source: i once read a review that claimed this. Don't remember where, but they did lab tests to check this.
<bluetail[m]>
I never got used to 27" monitors like the iMac has.
<bluetail[m]>
Now looking up 4k 144hz monitors and couldn't find one between 24 and 25"
<sarah[m]>
Oleds are the most beautiful, but burn-in can be an issue
<bluetail[m]>
I don't know. When I had a large monitor, I always had more neck movement
<bluetail[m]>
sarah[m]: Wait till 2024... I heard then there will be really great monitors around...
<paciento[m]>
bluetail: as far as I'm aware there are 2160p monitors between 23.5" and 25" but you won't find high-refresh variants of such in that niche. You can get 1440p monitors with that go up that high, but they are quite old.
<bluetail[m]>
paciento[m]: I am happy with the one I have. But the color output on the mac has too much luminosity.
bisko has quit [Remote host closed the connection]
<bluetail[m]>
I know for sure that the color in applications on macOS is evidently different.
<bluetail[m]>
There is a resource I cannot share. But I can say that defined colors for text in Buttons are not reliably measurable, using Safari.
<bluetail[m]>
Perhaps I should make a small comparision of my monitor, so that it will be more memorable.
<paciento[m]>
anyway, more in #_oftc_#asahi-offtopic:matrix.org
bisko has joined #asahi
ndut2 has joined #asahi
* ndut2
hi all how are ya ? skraito here ... . what you call me for ? ... .
ndut2 was kicked from #asahi by ChanServ [You are not permitted on this channel]
<mps>
sarah[m]: main reason I'm using M1 macbook is the display
<bluetail[m]>
I know, it's #_oftc_#asahi-offtopic:matrix.org
<bluetail[m]>
But you might take into consideration that you cannot compare a monitor image that was processed by macOS and then say 'I like their Display" and then compare it to anything that isn't macOS.
<bluetail[m]>
colors on macOS are more 'vivid' and 'brighter', using the same color profiles
<bluetail[m]>
s/brighter/darker/
<bluetail[m]>
It feels a bit like theres more 'hue'
<bluetail[m]>
* more 'hue' on macOS.
<rednaks[m]>
<bluetail[m]> "I wished I could play old..." <- have you tried web based emulators ?
<bluetail[m]>
Not yet. But thats a good idea actually. Nobody can hold back wasm :D
<bluetail[m]>
Thanks.
<bluetail[m]>
But I don't understand this color difference