ChanServ changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | General project discussion | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
<brentr123[m]>
I’m assuming eventually this will be more user friendly? Like it won’t have crazy techy terms for people who wanna install Linux but don’t understand computer terminology?
<marcan>
eventually, but the steps at the very end (the reboots and running the shell script) are pretty much required
<marcan>
the rest of the process can have an easy automatic mode
<phire>
So the "press and hold power button" step will stay around?
<brentr123[m]>
Well yes
<brentr123[m]>
Because that is required to enter recovery mode
<marcan>
that's the entire point of the security design
<brentr123[m]>
And i think it’s impossible to automate that?
<marcan>
you are not allowed to run custom code without physically asserting your presence like that
riker77 has quit [Ping timeout: 480 seconds]
<phire>
oh... that makes total sense then
<marcan>
it would be nice though if we could elide the paired recovery requirement to avoid having to split the process into two steps
<marcan>
in fact that's a new requirement
<marcan>
right *now* with <12.0 it could be done in one step from 1TR
riker77 has joined #asahi
<marcan>
but alas that is going away
<phire>
might want to explain the "why" as part of the instructions
<marcan>
it would also be nice if we could set the new volume as bootable without a reboot, to avoid having to tell people to force shut down the boot loop
<marcan>
those are things that I could see improving in the future, but we depend on apple
yuyichao has joined #asahi
artemist has quit [Ping timeout: 480 seconds]
PhilippvK has joined #asahi
phiologe has quit [Ping timeout: 480 seconds]
artemist has joined #asahi
<alyssa>
🍎
kov has quit [Quit: Coyote finally caught me]
kov has joined #asahi
marvin24_ has joined #asahi
marvin24 has quit [Ping timeout: 480 seconds]
<chadmed>
phire: thats probably something that should go in the developer area/extended documentation. its probably best to keep the how-to guide as simple and uncluttered as possible to prevent confusing users who just want linux on their mac and dont care how it works
<chadmed>
if you pack basic instructions with heaps of technical tidbits you run the risk of confusing novice users who will then come looking for support for extremely basic things they would not have screwed up if they were given clear instructions. just look at the gentoo irc for an example of this
<j_ey>
it includes some changes to use sven's clock gate driver. even though sven said that driver/bindings will have to change, I thought it was worthwhile doing anyway
<pipcet[m]>
ooh, I think I'm going to bake myself an everything-asahi-can-do-now kernel :-)
<j_ey>
I probably need to re-test since I did a bit of rebasing and cleanup
<j_ey>
also there seemed to have been a double delay in the spi driver that I removed, I'm hoping it wasn't intentional
<pipcet[m]>
stan mentioned some timing problems with the SPI driver, particularly on MBAs
<pipcet[m]>
and I believe the keyboard SPI controller needs a delay both before and after transmission, during which !CS is held low.
<pipcet[m]>
I've had timing issue with the touchpad, but the only timing issue I had with the keyboard is when I changed the delay to 100 ms rather than 100 us, and it worked then, just slowly :P
<j_ey>
I haven't yet booted to anything that I can test the touchpad with
<pipcet[m]>
gpm? :-)
<j_ey>
never heard of that! "gpm - a cut and paste utility and mouse server for virtual consoles
<j_ey>
"
<pipcet[m]>
it's debian package gpm if you just want to include it in a debootstrap
<pipcet[m]>
i mean, you can also include an xserver that way but I only got x to work by making the framebuffer x8r8g8b8
<j_ey>
I'm just using busybox so far, maybe it's simple enough to include/build. but I should look at debboostrap eventually anyway
<pipcet[m]>
I set up github CI to build a debootstrap image, need to change that so it actually includes that in published "releases" (but it's probably too large for m1n1 to handle without my other patches)
<j_ey>
does debboostrap build anything? i mean, it uses binary packages for most things?
<pipcet[m]>
AFAIUI, debootstrap downloads and unpacks .debs, then remembers which of those were included so the native system can run "stage2", which runs the setup scripts.
<pipcet[m]>
what I'm doing right now is booting the stage2 system with init=/bin/bash, which is annoying because only a few things work. The right thing to do is to run it with the right init, which will set up a debian system in RAM and give you a login prompt (that part works) and allow you to log in (that part doesn't, because passwd isn't set up)
tomtastic has joined #asahi
<marcan>
interesting, I can SSH into macOS during an upgrade, during the post-reboot 15 minute progress bar thing
<marcan>
and sw_vers says it's already on the new version
<marcan>
so what is it *doing*?
<ar>
not sure, but the time it takes macos to update amazes me
<ar>
it falls just short of me suspecting it's actually compiling significant parts of the system
<mini>
isn't that the point where it does the APFS 'sealing' process?
yuyichao has quit [Ping timeout: 480 seconds]
<alyssa>
pipcet[m]: you around?
<pipcet[m]>
I am, but I won't have access to the M1 for an hour or so :-)
<tophevich[m]>
alyssa: is this already using a dcp driven fb based on mini-'dcp-clusterfuck'?
<alyssa>
no, this is iBoot
<alyssa>
^framebuffer
<tophevich[m]>
but looks good anyways, gz! :)
<alyssa>
:p
Andalu30 has quit [Remote host closed the connection]
Andalu30 has joined #asahi
<marcan>
OK, here goes. First "public" installer. All testing is appreciated. Note: for development only. This isn't ready for end users in any way, shape, or form. It's not even hosted or built properly, just running off of my personal server.
<marcan>
TL;DR: use macOS 11.4, normal boot or recovery, doesn't matter, and run `curl -L https://mrcn.st/alxsh | sh`. You need unpartitioned space or a blank APFS partition (>2.5G).
<marcan>
It only installs m1n1, not a distro.
<marcan>
git repo isn't even up, but I guess it's time to do that
<steev>
do you want people to test > 11.4, or is that known to not work?
<marcan>
sorry, >=11.4 works
<marcan>
but I actually haven't tested anything <11.5.2
<marcan>
I just claim to support 11.4 (and can install using that version for the stub), so please let me know if it works installing *from* 11.4 :-)
<marcan>
I've tested 11.5.2 and 12.0 beta5
<marcan>
The options you get will depend on what version you're running and what environment you're in
Andalu30 has quit [Remote host closed the connection]
<tophevich[m]>
the export IPSW in install.sh is no long used, no ?
<tophevich[m]>
*longer
Andalu30 has joined #asahi
<tophevich[m]>
it implies it could still be set from "IPWS=custom install.sh" could still be a thing
<tophevich[m]>
wow, I guess I could try to write sentences that make any sort of sense
<tophevich[m]>
what I mean is if I only look at install.sh, it implies I could provide the IPWS and that indeed I had to as I have no access to the default, but you ignore it in the pyton part entirely
<marcan>
ah yeah, it isn't
<j_ey>
I see there's BIT() and GENMASK(), but is there anything for non-contiguous masks like (5 << 1)?
<marcan>
j_ey: don't think so
<marcan>
tophevich[m]: removed
<j_ey>
this code looks ugly with a mix of BIT, GENMASK and manually <<s
<alyssa>
marcan: does this fetch 12.x firmwares from a base 11.x system?
<marcan>
no, you cannot use a version newer than what you currently have installed
<marcan>
because 11.x SFR can't boot 12.x, and because 11.x macos doesn't know how to make 12.x bootable either
<marcan>
it'll filter the list of options to whatever will work for you
<alyssa>
got it
<svenpeter[m]>
j_ey: BIT(1) | BIT(3) maybe?
<alyssa>
marcan: also hello from asahi linux :>
<marcan>
:>
<j_ey>
svenpeter[m]: i guess i have to do more grepping to see if that's what's sued
<j_ey>
*used
<marcan>
it does claim to allow you to install 11.5.2 from 11.5, because I *think* SFR didn't change there (at least iBoot didn't), but I haven't tried that
<marcan>
it also claims to allow you to install 12.0 beta5 from any 12.0 beta, which is again untested, caveat emptor, etc
<marcan>
if anyone is e.g on 11.5 or 11.5.1, or an older 12.0 beta, that would be a good test to do
<marcan>
(in particular if you've never upgraded beyond that, so SFR is old too)
<tophevich[m]>
* marcan: don't know if I'm over sensitive there, but when I see an rm -rf with glob I get anxious, I like to do `rm -rf -- *glob*` to make sure no funny business with files/dirs with the name of options happens: https://github.com/AsahiLinux/asahi-installer/blob/main/build.sh#L57
<tophevich[m]>
hm, interessting to see what happens with a matrix edit in irc
<marcan>
it gets sent again
<jn>
unfortunately it makes it quite inconvenient to spot the difference
<tophevich[m]>
yeah, I just had a look in the irclogs, but with a asterix as prefix
<tophevich[m]>
*an
<tophevich[m]>
contribution guidlines are the same for all projects under the AsahiLinux umbrella right?
<marcan>
yes
<marcan>
well, except linux which has is own rules (e.g. real names only for commits)
Andalu30 has quit [Remote host closed the connection]
<tophevich[m]>
ok, I feel silly to create that PR, but I guess it helps to get my feet wet
ihaveamac has quit [Quit: fail]
ihaveamac has joined #asahi
<marcan>
:)
* marcan
sleeps
<jn>
good night :)
ifthenelse has quit [Quit: Ping timeout (120 seconds)]
ifthenelse has joined #asahi
apetresc has quit [Ping timeout: 480 seconds]
rbenua_ has joined #asahi
alyssa_ has joined #asahi
nkaretnikov has quit [Ping timeout: 480 seconds]
nkaretnikov has joined #asahi
alyssa has quit [Read error: Connection reset by peer]
Techcable has quit [Read error: Connection reset by peer]