<winter>
tpw_rules: I believe idevicerestore works just fine on Windows as well, though I'm not sure how to actually compile it (I guess with mingw or somehow get MSVC to do it...)
<tpw_rules>
yeah there was a rumor to that effect in nixpkgs
xfim has quit [Quit: leaving]
xfim has joined #asahi
guillaume_g has joined #asahi
_bolt has quit [Remote host closed the connection]
<ElvishJerricco[m]>
Well, I managed to get m1n1 building with Nix with a fairly simple derivation, as long as I build on aarch64 linux or x86_64 linux (cross compile). Still having issues with darwin, but it probably stems from nixpkgs stuff
<ElvishJerricco[m]>
I diff'ed hexdumps of m1n1.bin as built on aarch64 and as cross built on x86_64, and there are 5698 lines of diff. For an 832K binary, I'd say that's pretty good as far as reproducibility across build architectures go
dubiousness has quit [Remote host closed the connection]
derzahl has quit [Ping timeout: 480 seconds]
xfim_ has joined #asahi
xfim has quit [Ping timeout: 480 seconds]
xfim_ has quit [Quit: leaving]
xfim has joined #asahi
<xfim>
A couple of days ago I was here trying to set up Asahi on the Mac Studio Pro. With the generous help of jannau I went through different steps. I have the installer.tar.gz and was able to do the partitioning. I have been able to crosscompile u-boot, but I am stuck with building m1n1, because it seems that there is no way to add bootlogo_*.png to the bootloader. I've even opened an issue at github with more
<_jannau_>
xfim: I saw the issue and it looks strange. The error is the one make would report if the files did not exist
petterroea has joined #asahi
<povik>
you need to fetch the submodule don't you
<xfim>
But the files do not exist, right? I can't see any file named bootlogo_128.o, as the Makefile expects in line 74. Is it probably something very silly, I am sure.
<xfim>
I read Makefile and on line 74 it says that there are two files that should be compiled, namely bootlogo_128.o bootlogo_256.o.
<xfim>
I can find data/bootlogo_128.png and the 256 version.
<xfim>
aaaaaaaaand it is working now. I owe you yet another one, jannau
<xfim>
Probably that should have been very obvious to most, but when you mention to clone it, I didn't thought about having to load all submodules.
<mps>
oh yes, I forgot that I added artwork tarball download for alpine m1n1 pkg
<_jannau_>
argh, data/bootlogo_{128,256}.png are symlinks into the artwork submodule
<xfim>
So the exciting time is coming. I am putting all of it into boot.bin and then I will move on to the MacStudio to continue with it. Can I simply copy it to one of the partitions, or do I need either the USB a-c cable or the m1n1 proxy (whatever this last one means)? I am afraid I will need more help today. Thank you all!
<_jannau_>
usb-c cable is more convient if it doesn't work but replacing the file on the fat32 partition in mac os should work
bps has joined #asahi
the_lanetly_052 has joined #asahi
<povik>
hm, what's that rin20xx wrote here yesterday
<povik>
doesn't the arch kernel package have the current 'asahi' branch?
<povik>
that should have the i2c fix, so the flakiness should be gone for good
<_jannau_>
povik: no, the arch package was not updated after the rebase so it doesn't have the i2c fix
<xfim>
I don't really know which partition is the fat32 one. I can seed /dev/disk0 and /dev/disk3.
<xfim>
In /dev/disk0, the internal, I have #0 GUID partition scheme, #1 Apple_APFS_ISC, #2 the Apple_APFS container disk3 where I assume lies the MacOS. Then free space, and finally #3 Apple_APFS_recovery.
<xfim>
I would say it should be, discarding the others, #1, disk0s1, with 524.3 Mb. Would that be correct?
jluthra has quit [Remote host closed the connection]
jluthra has joined #asahi
ciggi has joined #asahi
millenialhacker has joined #asahi
the_lanetly_052__ has joined #asahi
the_lanetly_052 has quit [Ping timeout: 480 seconds]
the_lanetly_052__ has quit [Ping timeout: 480 seconds]
dost^ has quit [Remote host closed the connection]
dost^ has joined #asahi
<xfim>
I'm scared! So I have generated boot.bin so far, and I understand that I have to add it to the FAT32 partition. But there was none, so I decided to go with the installer and select the UEFI environment only.
<xfim>
I thought that at some point I would be able to change the file from my newly generated one, but the installer has arrived to the end, and I am in the screen where it says that the system will shut down.
<xfim>
And that I have to wait for 15 seconds, etc...
<j`ey>
xfim: you can replace it later
<j`ey>
macOS can read/write FAT
<xfim>
Because I have not changed the boot.bin that the installer has probably added with my own "boot.bin" for MacStudio, does this mean that I will not be able to boot correctly?
<xfim>
Can I change it later?
<xfim>
I am in the abyss and freaked at leaving it as a brick at this point.
<xfim>
OK, j`ey, I can replace it later.
<xfim>
So should I proceed anyway and see what happens?
<j`ey>
it won't be bricked anyway
<xfim>
Or you would recommend, now that I am in MacOS and the installer is not yet rebooting, to mount the FAT32 partition and change the boot.bin?
<j`ey>
carry on with the installation
<xfim>
Not bricked because I still have the option of MacOS vs my other installation, right?
<j`ey>
yup
<xfim>
OK, here we go.
<j`ey>
btw markan just said he may add studio support to the installer tomorrow..
<_jannau_>
you should be able to break out of the installer with STRC-C without reboot
<j`ey>
but you're here now
<marcan>
ElvishJerricco[m]: I would expect cross and native to produce the same binaries if you use the same GCC version and options, but there are likely differences in default options related to the gcc build config
<_jannau_>
still useful to have someone to test the m1 mqax mac studio dts beforehand
<marcan>
not 100% sure though, maybe gcc is worse at this than I expect
<xfim>
Oh, oh! This changes things... Interesting. I'm here anyway, but let me first check the partitios as they are right now, after the installer has done its work.
<xfim>
I want to take a note about everything.
<xfim>
I will be super happy to be the tester.
<xfim>
But I am not sure if I can be of much assistance giving feedback. I'm willing to learn, though!
<xfim>
/partitios/partitions/
<marcan>
it would be nice if folks tested the asahi rebase kernel, that will probably be pushed this week if nothing is broken
<marcan>
that has smc rebased on top of the new rtkit stuff etcv
<marcan>
*etc
<marcan>
basically that plus the t6k2 stuff from _jannau_ should be the next significant update to the installer/repos
<j`ey>
i have tested the rebase (with my own config + alpine), worked well
<mps>
marcan: some branch or latest git commit?
<mps>
j`ey: ah, could you post url if you have it
<j`ey>
mps: just the latest asahi branch
<marcan>
asahi branch
<mps>
I would like to use git commit and prepare upgrade for alpine
<xfim>
But I am not sure if I can be of much assistance giving feedback. I'm willing to learn, though!
<marcan>
also, does anyone actually have the M1 Max Mac Studio? :-)
<mps>
last commit is 21 ago
<j`ey>
yup
<j`ey>
marcan: xfim does
<xfim>
marcan: Yes, that is me
<marcan>
ah cool
<marcan>
then testing that will be very helpful too :)
<xfim>
Base one, 32 RAM, 500 GB SSD
<xfim>
It is now closed. I'm waiting 15*4, just to double-double check
<marcan>
I'm thinking of making a testing repo where I push things before prod, and making the images build off of that (but replaced with the regular repo at the end of the build process, with the expectation that those images + repo are promoted to stable simultaneously)
<mps>
I'm now building toolchains for riscv, when it finish I will upgrade and build kernel
<marcan>
I'll be back later, got some stuff to do tonight, but I'll spend tomorrow doing another round of updates and figuring out the DTs
<marcan>
I'll rebase the kernel onto the latest rc of that branch (shouldn't break anything)
<marcan>
then I can push to a testing repo / installer and let people test it, and promote it to stable on friday or something
<marcan>
(well, as "stable" as asahi gets anyway)
<j`ey>
marcan: 5.18 was released
<marcan>
thought so, so yeah, based on that then
<cmalvi[m]>
marcan: any Asahi related news in this release?
<marcan>
mostly mac studio support, general things getting upstreamed, bunch of bugfixes
<xfim>
OK, here I am, on "Running proxy..."
<marcan>
I'll also merge in the configurable display stuff for m1n1
<xfim>
What's next, then?
<xfim>
"ERROR: Kernel found but no devicetree for apple,j375c available"
<cmalvi[m]>
marcan: Thank you :)
<j`ey>
xfim: I havent been following, but this may be where you boot back into 1TR and replace the boot.bin?
dost^ has quit [Ping timeout: 480 seconds]
<xfim>
Sorry j`ey, what do you mean by "back into 1TR". Sorry, but what is 1TR?
<j`ey>
1TR is the mode you get into with the long hold of the power button
<cmalvi[m]>
any plans to write a changelog for the kernel release?
<xfim>
OK, got it now. But from the menu, how can I replace the boot.bin that I have in my usb drive to the fat partition? I have three options. First is the installer, and I can't do it from that, right? Then I have MAcOS, which I could use, and the third? Is it possible to switch such a file from within ther?
<j`ey>
xfim: you can open a terminal, utilities -> terminal
<xfim>
Utilities sends me to "Recover from time machine", "reinstall monterey", "safari" and "disk utility". But from a menu I have been able to find a terminal. Let me continue then. Now I have to learn how to mount volumes on MacOS, because I imagine that 'mount' will not do.
<j`ey>
mount works, but also `diskutil mount`
<xfim>
Yes, it does. BTW, do you want an anecdote? I make dvorak the main keyboard during the installation of MacOS. But switched to qwerty when exploring this first couple of days.
<xfim>
And now in the recovery, as well as in the password introduction, it was understanding dvorak, and kept me saying that the password was wrong...
<j`ey>
:-)
<xfim>
OK, now I have to guess what is thepartition type for fat32. Is it not vfat?
<xfim>
Oh, no, it is already mounted. Good.
<j`ey>
it should be /Volumes/EFI
<xfim>
Is it possible that the file that I have to replace is /Volumes/Asahi/Finish Installation.app/Contents/Resources/boot.bin, with 848K, and mine is 1.1 Mb?
<j`ey>
possibly, it has more/bigger DTBs
kov has joined #asahi
<mps>
j`ey: did you upgraded m1n1+u-boot with new DTBs when upgraded kernel
<xfim>
The destination partition is mounted read only, and when trying to mount it with "mount -o rw", mount complains that it is unknown. Any clue?
<_jannau_>
mps: you need to update the dtb for cpufreq
<j`ey>
xfim: hm, unmount and try diskutil mount
<_jannau_>
xfim: I think mount -w /mount_point works
<xfim>
I've done it with '-u -w'
<mps>
_jannau_: thanks
<xfim>
I really must stop asking directly and try to find my own answers first before bothering people.
<j`ey>
xfim: you're not annoying..... yet :p
<xfim>
OK :-) But do you promise to tell when I do?
<j`ey>
xfim: hehe
<xfim>
OK, I have rebooted again to 1TR and now I have the "Running proxy...", but there are several more messages that don't look very well.
<xfim>
Still "Kernel found but no devicetree for apple,j375c available"
<xfim>
"No valid payload found"
<xfim>
And before that 4 different "Unknown variable" followed by two and a half terminal lines of "??????|?a7??\?' etc...
<mps>
upgraded kernel. uname -a says => Linux m1 5.18.0-rc5-1-asahi #2-Alpine SMP PREEMPT Mon, 23 May 2022 10:41:40 +0000 aarch64 GNU/Linux
<mps>
this is on mbp 2020
<mps>
for now no issues
<j`ey>
xfim: thats a weird error
<xfim>
So I am not certain I'm somewhere. It certainly says "Running proxy..." but nothing less, and can't type anything. Besides, once in proxy, if everything had went well, what would be the next step? Configure for "stage 1" as in https://github.com/AsahiLinux/docs/wiki/m1n1:User-Guide?
<j`ey>
m1n1 doesnt accept keyboard input directly, hence no typing
<j`ey>
xfim: you built the installer yourself right? and used some branch of jannau's for m1n1 or?
<xfim>
Yes.
<xfim>
Let me upload the picture somewhere.
<_jannau_>
asahi m1n1 should be fine, u-boot needs to be my branch for the dtb
<xfim>
Shall I simply just wait for tomorrow's new version of the installer, at this point?
<marcan>
are you replacing the stage2 boot.bin with one containing u-boot? if so that is wrong
<marcan>
you need to replace the one in the EFI partition
<marcan>
so /Volumes/EFI, not /Volumes/Asahi
<j`ey>
I think xfim replaced the one in /Voluemes/EFI via 1TR just now
<marcan>
if the partition was mounted readonly that sounds like /Volumes/Asahi, which will be mounted readonly by default
<xfim>
Sorry, 5 minutes and back
<marcan>
the "unknown variable" stuff is expected if you install a m1n1 built by concatenating payloads yourself (you need to append some NULs); if it happens with stage2 I don't know what that is about though. I think alyssa also saw that. Must be a bug somewhere but I don't see how it could happen...
<marcan>
(stage1 will add 4 zeroes when launching stage2 to avoid that)
<_jannau_>
I doubt xfim attached an "chosen.asahi..." variable so the replacement of the EFI boot.bin should be fine
<marcan>
I mean if he replaced the one in the step2 bundle and did nothing else, then that's a no-op
<marcan>
since that never got installed
<marcan>
sorry. I meant step2, not stage2
<mps>
for me stage2 m1n1 workded without "chosen.asahi..." variable
<mps>
though I added it later
<marcan>
stage2 does *not* need the chosen.asahi stuff since stage1 puts it there if it has it itself, which is the correct way to do it
<mps>
not in my case, I had to add it in stage2 m1n1
<marcan>
then you installed your stage1 wrong
<mps>
marcan: I used official installer
<marcan>
and the variable was not getting passed through?
<marcan>
you didn't actually do anything by replacing that
<marcan>
mps: so you append the variable first then the other payloads?
<mps>
yes
<marcan>
it's starting to sound like the gzip thing doesn't work to find the next payload. can you send me your combined new-uboot.bin with everything?
<j`ey>
xfim: ah sorry, it has to be the EFI/FAT partition
<marcan>
ah, okay, I just realized what's broken with the variable code
<marcan>
okay, so that explains the garbage output sometimes
<marcan>
it does not explain the chosen thing not working though
<mps>
will be afk for some time
the_lanetly_052 has joined #asahi
<xfim>
So, even worst: see picture https://ibb.co/h2yhDLB. "Unable to read file ubootefi.var", and "failed to find boot partition `run distro_bootcmd` to boot from any partition.
<marcan>
that sounds like the chosen issue mps is talking about
<j`ey>
thats progress!
<marcan>
wait no it does look for the ESP
<marcan>
then it just has nothing installed
<marcan>
so it's working fine
<j`ey>
xfim: you havent installed anything yet, right?
<j`ey>
anything = any distro
<marcan>
untested, just pushed ad659daa which should fix the variable garbage spam (but it should've been harmless anyway, does not explain not picking up the chosen var)
<marcan>
I need to head out, talk later :)
<j`ey>
ciao
<jannau>
xfim: you said you had the uefi / u-boot only installed. so that looks like everything works as long as you don't have usb keyboard/mouse connected
<xfim>
No, no, I have not installed anything. But the partition table looks like a puzzle!
<xfim>
jannau: I do have a keyboard and a mouse connected
<xfim>
So what would be the next thing to do now?
<xfim>
I have to admit that I am completely lost right now...
<jannau>
xfim: to one of the usb-c ports on the back? those are one supported ones
<xfim>
Yes, the ones in the back
<xfim>
In the front I have a dodocool docking station with the usb disk that has allowed the boot.bin to be inserted.
<jannau>
xfim: the usb type-c ones? the two type-a ports are not supported
<xfim>
Oh, sorry, not the type-c ones, but the plain old type-a
<xfim>
Ah, now I understand then. Would them being connected to the docking through the usb-c change the situation?
<xfim>
If not, then the machine is unusable, right?
<jannau>
I wouldn't expect the front usb ports to work under linux and certainly not under u-boot
<xfim>
So then let me try the usb-c from the back connected to a docking and then plug in the keyboard+mouse to that docking through usb-a. Should I?
<jannau>
I'm not sure how the usb-c ports on the front of the m1 max mac studio are implemented
<j`ey>
xfim: the next step is to make a bootable USB drive
<j`ey>
so you can start installing
<xfim>
From any distro of my choice, right?
<jannau>
xfim: that should work if the docking station is not a thunderbolt docking station
<xfim>
But if I can't have a keyboard, how am I doing to install anything?
<j`ey>
you could do it remotely, if you get wifi/ssh working (does ethernet work OOTB for the mac studio max?)
<xfim>
jannau: so then this one should be fine as well.
<xfim>
Perfect then, sorry I was very scared with so many "failed".
<mps>
0 usb storage devices found
<jannau>
7 USB devices looks good. should there have been a storage usb device?
<xfim>
Let me recap about the options: use alpine to boot to a live scenario so that I can start installing the distro of my choice
<mps>
yes, if usb storage is connected
<xfim>
mps and jannau: Yes, for this last boot, I removed the usb disk, so none should appear.
<xfim>
If I boot from a live usb with alpine, how would I proceed. From the 1TR must I still go for "Asahi", and then how will I be able to run the live alpine?
<jannau>
yes, the live alpine usb device should boot automatically
<mps>
xfim: u-boot should boot usb by default if it can't find linux on nvme
<Sobek[m]>
So, basically the Asahi entry on the apple boot loader will hand off to m1n1, which then hands-off to U-boot, and then U-boot does its thing
<jannau>
is the keyboard working now?
<xfim>
jannau: hard to tell if it works. On one point, because of the "=>" prompt, I don't seem to have anything going on, but if I try and press CTRL+ALT+delete I see some "IIIIII" in the screen, so I would say that there is connection between the keyboard and the computer.
<jannau>
I think there should be echo on the u-boot prompt
<xfim>
So there is none. At least none that I can say for sure. Have you seen the picture ( https://ibb.co/QPbRMn7)? I would not say that this is the "u-boot" prompt, isn't it?
<xfim>
I am rebooting with alpine right now
<j`ey>
that picture is the uboot prompt
<mps>
xfim: it is u-boot prompt
<xfim>
It seems it can find the usb and is trying to boot from that: https://ibb.co/VVHYkV2
<xfim>
But "No partition table - usb 0", "Couldn't find partition usb 0:1"
<xfim>
Ah, maybe it is because I have used dd if=alpine* of=/dev/sdb, instead of of=/dev/sdb1?
<mps>
xfim: no, you did right
<mps>
it needs to be dd-ed to 'whole usb/disk'
<mps>
xfim: you have ERROR, look at it
<mps>
I mean in u-boot screen
<mps>
failure to add disk device usb...
<xfim>
Yes, I see the error. But I can see that even with the failure to add disk device
<xfim>
It tries to boot from usb disk, and identifies the size and all, but not the partition table
<xfim>
So I don't really know what I can make out of it?
<mps>
hm, could you check on other machine disk state
<xfim>
Is the usb support for the usb mass storage disk not really there?
<mps>
i.e. does it have two partitions, one ESP and one ext4
<xfim>
mps: Yes, in fact I am trying it with another disk, just to rule out that.
<mps>
xfim: this image is used by some people to install alpine and by some to fix unbootable machines
<mps>
but I think you are first trying on mac studio
<mps>
and I don't know is the u-boot and dtbs are updated for mac studio
<mps>
jannau could tell
<_jannau_>
I haven't tested usb storage with u-boot on the mac studio but it should not differ from the macbook pros especially for the m1 max model
<mps>
actually one who build u-boot for mac studio
<xfim>
OK, let me try with another usb disk, dd if=X of=/dev/sdb
<_jannau_>
the dts/dtb is coming from me especially for mac studio. u-boot would not load without mactching dtb
<mps>
for me safe 'mode' is plain usb-c to usb-a adapter
<mps>
not those with hub
<xfim>
OK, I have keyboard control and all, but I am still at u-boot, and nothing has happened with the USB disk with alpine.
<xfim>
Still 1 storage device found.
<xfim>
Still Device recognized, capacity correctly guessed, "is not current device", but "No partition table - usb 0", and "couldn't find partition usb 0:1"
<mps>
strange, disk have partition table?
<mps>
sorry for stupid question but did you unziped image
<xfim>
mps: no, no unziped image
<mps>
ah
<mps>
xz -d m1-usb-alpine-install.img.xz
<xfim>
Oh, come on!
<mps>
and then `dd if=X of=/dev/xxx`
<xfim>
I simply assumed that xz decompression would be there. Sorry
<xfim>
Sorry!
<mps>
and then `dd if=m1-usb-alpine-install.img of=/dev/xxx`
<mps>
ok, np
<xfim>
Even, now that I check, the tutorial says to uncompress! Sorry
<mps>
heh, we all make mistakes and overlook something sometimes
<_jannau_>
or `xz -c -d m1-usb-alpine-install.img.xz | dd of=/dev/xxx` unless you plan copy the image multiple times
<xfim>
Done. I'm on alpine. Thank you all!
<xfim>
All seems well, except for networking not starting: direct firmware load for rtl_nic/rtl8153a-4.fw failed with error -2
<j`ey>
xfim: use wifi?
<_jannau_>
looks like missing firmware for the network adapter in the usb dock
<mps>
oh I thought add all firmware and usb network drivers but forgot
<mps>
but wifi works
<_jannau_>
the aquantia 10GB nic should work as well
<mps>
xfim: you can install wifi firmware by mounting ESP on /mnt and run shell script in /root/update-wendor-firmware
<mps>
I mean ESP from nvme
<mps>
(I was busy with next stable alpine release which should be done this week)
<xfim>
I am trying with the ethernet, first. And it does not ping, but it recognizes a device.
<xfim>
But when restarting the network service: can't open /var/run/udhcpc.eth0.pid, no such file or directory, although it seems that "starting networking" for eth0 goes well, and enfs with ok. But then ifconfig does not show it.
<xfim>
If i only add auto eth it shows, but when I add "iface eth0 inet dhcp
<xfim>
" to that file it does not show.
<xfim>
But that is not asahi, so I will have to deal with it myself, don't worry!
<xfim>
Thank you anyway.
<mps>
xfim: yes, this is not tested with ethernet, most users use wifi
<mps>
till I find time and make propel alpine iso img
<xfim>
I am going to try wifi then.
<xfim>
mps, you mentioned at 11:15 to mount ESP on /mnt, and then run the shell script.
<mps>
yes, described in guide
<xfim>
I have found the shell script, but what exactly is ESP? I guess the device should be mounted to /mnt/vendorfw for the script to work, right?
<xfim>
OK, let me try then.
<mps>
lsblk -f
<mps>
will show vfat partition on nvme
<mps>
mount /dev/nvme0n1pX /mnt
<mps>
sh /root/update-vendor-firmware
<mps>
then, umount /mnt
<mps>
modprobe brcmfmac
<mps>
and `/etc/init.d/iwd start` to start iwd
<mps>
cat `cat /proc/device-tree/chosen/asahi,efi-system-partition` will show PARTUUID of the ESP where m1n1+u-boot and firmware are
Guest489 is now known as svenpeter[m]
<xfim>
mps: all well. wifi working. So I confirm that eth it is not working, if it is useful.
<xfim>
One question: I can see that nvme0n1 has 5 partitions: p1 500 Mb, p2 74Gb, p3 2.3 Gb, p4 477M and p5 5Gb.
<xfim>
Then I imagine that I still have space not assigned in the hard drive. Therefore, should I would like to install the distro in the free space of the hard drive. Will I mess something if I join together the 74Gb with the free space? Is this partitioning scheme that asahi has done meaningful?
<_jannau_>
the 74GB are the macos installation. It's currently recommended to keep that
<_jannau_>
do not touch the first and and last partition (at the end of the nvme)
<_jannau_>
p3 is a stub macOS partition required for booting an alternate OS, p4 is the EFI system partition
<mps>
free space could be partitioned to more partition, I have 3, /, /home and /work
<_jannau_>
there should be ~400GB unpartitioned space between p4 and p5
<_jannau_>
do not rely on unordered partition numbers since diskutil will reorder them
<xfim>
I am going for a single / for ext4. I've worked many years with separate /home and / and with sparse hard disk space I prefer now not to risk it and be more flexible.
<xfim>
It's done. So now I guess I can start with the installation.
<xfim>
Will keep you posted about my success. I'm going afk for a while. Thank you all.
xfim has quit [Ping timeout: 480 seconds]
Gaspare has joined #asahi
millenialhacker has quit [Quit: Konversation terminated!]
dost^ has joined #asahi
cassepipe has joined #asahi
dost^ has quit [Remote host closed the connection]
Gaspare has quit [Quit: Gaspare]
Gaspare has joined #asahi
kaprests has quit [autokilled: Possible botnet activity. Mail support@oftc.net with questions. (2022-05-23 14:49:59)]
kaprests has joined #asahi
<tpw_rules>
how erased is the apple dfu erase? does it just recreate the partition table and freshly install macos or does it regenerate the encryption keys?
xfim has joined #asahi
LinuxM1 has joined #asahi
<tpw_rules>
the platform security guide says that the erase creates a new "owner identity key" and destroys any old one. there's also a "media key" which can be discarded by instant remote wipe. there's also the SKP stuff.
<tpw_rules>
and anti-replay keys
Manouchehri has joined #asahi
<tpw_rules>
but i can't find anything that directly states "if you erase-install, everything previous will be inaccessible"
<tpw_rules>
it does say that the "erase all content and settings" in system preferences is an un-replayable event which implies it updates that key
derzahl has joined #asahi
the_lanetly_052 has quit [Remote host closed the connection]
nicolas17 has joined #asahi
<j`ey>
I need to use 'chosen.boot-args' now, right?
<j`ey>
chosen.bootargs it is
<mipi[m]>
What is the current status wrt. rust? https://github.com/AsahiLinux/docs/wiki/Software-known-to-have-issues-with-16k-page-size suggests that it's fixed, but I still get errors because of the page size problem when I build rust packages. The native rust installation works, however, but I have to adjust the PKGBUILD (remove rust/cargo as make dependency).
<mps>
j`ey: isn't it set automatically
<j`ey>
mps: im booting using m1n1 proxy, needing to pass some extra kernel params, I got it working now
<_jannau_>
mps: no what would know how to set kernel boot cmd line, especially when using chainload.py
<Manouchehri>
Is it possible to install the latest macOS on an M1 without updating the firmware?
<nicolas17>
well, latest macOS drivers might expect latest firmware and not work otherwise...
<Manouchehri>
haha fair :p
<Manouchehri>
I just wanna use the new virtualization framework on my M1 Mini, but worried about losing the option to use Asahi later in the future.
LinuxM1 has quit [Ping timeout: 480 seconds]
<_jannau_>
marcan: I've tested it
Gaspare has quit [Quit: Gaspare]
<marcan>
Manouchehri: system firmware is backwards compatible
<marcan>
there is no way you would ever lose the ability to run asahi
<marcan>
since apple need to keep backwards compatibility with older versions of macOS too
<Manouchehri>
oh perfect :)
<marcan>
that warning is kind of wrong
<nicolas17>
what does it mean by "firmware" here? iBoot?
<nicolas17>
afaik coprocessor firmware is per-OS-install anyway...
cassepipe_ has joined #asahi
<marcan>
I removed it
<marcan>
nicolas17: only some of it
<marcan>
cio, tmu, iBoot1, SEP, ANS, SMC, PMU, and I think possibly a couple others are systemwide
<marcan>
the rest are per OS
<nicolas17>
huh so SEP-macOS has backwards compatibility?
<marcan>
it has to
<marcan>
SEP never downgrades, that'd be insane for security
<marcan>
and it handles global state all over the place
cassepipe has quit [Ping timeout: 480 seconds]
Gaspare has joined #asahi
<nicolas17>
that's interesting because on iOS they break compat often :P
<marcan>
because iOS only runs one OS
<marcan>
no such luxury on a Mac
<nicolas17>
yeah but still, they have a well-defined stable protocol to avoid breaking backwards compat, or the SEP knows how to talk old versions of the protocol... and they go out of their way to disable that on iOS?
<marcan>
I'm pretty sure they developed this for macs...
<marcan>
so maybe SEP will be less unstable going forward on iOS?
<nicolas17>
maybe yeah
<nicolas17>
admittedly one of the problematic compatibility areas is FaceID :)
<marcan>
:)
Gaspare has quit [Quit: Gaspare]
guillaume_g has quit []
cassepipe__ has joined #asahi
cassepipe__ has quit [Remote host closed the connection]
cassepipe__ has joined #asahi
cassepipe__ has quit [Remote host closed the connection]
freefiree has quit [autokilled: Please do not spam on IRC. Email support@oftc.net with questions. (2022-05-23 17:47:17)]
<mipi[m]>
OK, I see. If I find the time, I will test it and report back.
<mipi[m]>
That's great, but I do not use alpine
millenialhacker has joined #asahi
bisko has quit [Ping timeout: 480 seconds]
Gaspare has quit [Quit: Gaspare]
bisko has joined #asahi
Gaspare has joined #asahi
<AstrOwO[m]1>
wut
<xfim>
mps: I am using alpine to chroot into a Gentoo env. While asahi uses zstd compression algorithm for kernel modules, which compression algorithm is used in alpine? I need it in order to build gentoo with such capability.
cassepipe has joined #asahi
cassepipe__ has quit [Read error: Connection reset by peer]
Glanzmann has joined #asahi
<mps>
xfim: for linux-asahi kernel modules are not compressed
<mps>
so 'None'
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<xfim>
mps: excellent, thank you. So far, so good and Gentoo is preparing everything. But I will have to leave the rest for tomorrow eventually.
<weapon_of_nix[m]>
is there any way to access the asahi filesystem and make edits from macos?
<mps>
weapon_of_nix[m]: you mean ext4 FS?
<weapon_of_nix[m]>
yes
<weapon_of_nix[m]>
i broke something and i need to edit it because i can't boot
<nicolas17>
I don't know of any macOS driver for ext4, no
<mps>
there is paragon comercial driver for macos
<nicolas17>
does it support ext4 (not just ext3), and does it support ARM?
<mps>
I think so
<mps>
even there is 'trial' version if someone want to try and not buy
<mps>
weapon_of_nix[m]: you can read today backlog here also, to see how to use usb bootable and fix problem with iit
<mps>
s/iit/it/
<mps>
(I'm tired to retyoe it again now)
<mps>
retype*
<weapon_of_nix[m]>
thanks i'll take a look
<cmalvi[m]>
<nicolas17> "does it support ext4 (not just..." <- I have it. I can confirm it support but don’t expect any partitioning utility like gparted or similar
<Amy[m]>
hi, asahi linux seems to have created undeleteable partitions on my mac
<Amy[m]>
i would like to remove them but disk utility wont let me
<Amy[m]>
my mac storage has been reduced to 120gb now thanks to asahi linux
<mps>
but this is for latest asahi kernel and have speakers enabled
<mps>
so somewhat risky
<cmalvi[m]>
Yes I understand
<cmalvi[m]>
I prefer to not enable speakers for now
<cmalvi[m]>
No need speaker at work 😂
<mps>
actually speakers are enabled only for mbp 2020
<mps>
j293
<cmalvi[m]>
No need to hurry, is a great project that needs time
<cmalvi[m]>
For me is just incredible the actual state
<cmalvi[m]>
Using a 4K compiled kernel, all apps works except few not available for arm neither with building from source. Is a very promising state
Major_Biscuit has joined #asahi
MajorBiscuit has quit [Ping timeout: 480 seconds]
<ktz_[m]>
which distro are you using cmalvi ?
<cmalvi[m]>
ktz_[m]: I’ve installed the arch minimal, then gnome and built the asahi kernel with a 4K kernel patch (not created by me)
<cmalvi[m]>
All extension I use on my tower pc works also here
<ktz_[m]>
personally I found a great deal of packages in the arch repos not working/broken whilst almost all of them work out of the box properly with alpine
Gaspare has quit [Quit: Gaspare]
<cmalvi[m]>
ktz_[m]: Yes, some packages are missing in arch repo due to arm architecture, also in aur but some of these builds and run fine. It’s not straightforward and took me some time to replicate the exact setup I have on my tower pc but I finally did it
<ktz_[m]>
not to mention aur.. 80% of it is broken
<cmalvi[m]>
Also pamac-full works, because sometime I like to use it
<cmalvi[m]>
And pop-shell that I love
<ktz_[m]>
yes you're right, if you're settled already you better stay as is, just mentioning it fyi
<mps>
on alpine aarch64 is 'first class citizen' on arch it is not
<cmalvi[m]>
mps: Now I need to take my degree during this summer, maybe after I’ll take a look at it 😂
<cmalvi[m]>
But sure, I need to explore alpine
<mps>
arch alarm is good really, but missing a lot of packages
<cmalvi[m]>
mps: Yes I totally agree, but I took it like a challenge to improve my build skill and in general my Linux usage skill
<mps>
arch and alpine build systems are similar
<mps>
but imo everyone should use what best fit workflow for them
<cmalvi[m]>
Oh wait my I made me misunderstood. I want to say that those problems encourage me to improve
<cmalvi[m]>
s/my/maybe/
<cmalvi[m]>
mps: Sure, totally agree :)
<cmalvi[m]>
I’m started using Linux because windows fried two ssd during university lectures during quarantine in 2020
<cmalvi[m]>
Also windows like to break the bootloader every year for me
MajorBiscuit has joined #asahi
Major_Biscuit has quit [Ping timeout: 480 seconds]
cassepipe has quit [Remote host closed the connection]