marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | "Does XXX work yet?": https://alx.sh/fs | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-alt #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
Moprius has joined #asahi
Moprius has quit [Remote host closed the connection]
hightower3 has joined #asahi
jeisom has quit [Ping timeout: 480 seconds]
hightower2 has quit [Ping timeout: 480 seconds]
AnuthaDev_ has quit []
as400 has quit [Remote host closed the connection]
as400 has joined #asahi
hex-spell has quit [Ping timeout: 480 seconds]
gabuscus has quit []
hex-spell has joined #asahi
bluetail has quit [Quit: The Lounge - https://thelounge.chat]
bluetail has joined #asahi
gabuscus has joined #asahi
marvin24 has joined #asahi
DannyB has joined #asahi
marvin24_ has quit [Ping timeout: 480 seconds]
cylm has quit [Ping timeout: 480 seconds]
hex-spell has quit [Ping timeout: 480 seconds]
dylanchapell has quit [Read error: Connection reset by peer]
jason1923 has joined #asahi
jason1923 has quit [Quit: Lost terminal]
dylanchapell has joined #asahi
DannyB has quit [Ping timeout: 480 seconds]
ec0 has left #asahi [#asahi]
jason1923 has joined #asahi
jason1923 has quit []
crabbedhaloablut has joined #asahi
balrog has quit [Ping timeout: 480 seconds]
balrog has joined #asahi
johnbchron has joined #asahi
johnbchron has quit []
<marcan> kidplayer666: tags are not all stable releases. we decide which tags are stable releases. if a tag is not pushed to stable it means it's not stable.
<marcan> our process is I tag a random kernel I think is worth *testing* and then it goes into developer channels and we find problems (as we did this time) and fix them
<kidplayer666> Oh ok thanks
eroc1990 is now known as Guest2567
eroc1990 has joined #asahi
<marcan> in particular, any time there are new drivers or new platform support involved, you'll see multiple cycles of that
<marcan> stuff only gets pushed quickly when it's just bugfixes or trivial changes we know won't break things
<espo>
lena6 has quit [Remote host closed the connection]
<kidplayer666> Just because I saw the “verified” tag on a bunch of older kernels
<marcan> that just means the signature is correct
<marcan> it has nothing to do with stability
<kidplayer666> That got into the fedora compiling shortly thereafter
<kidplayer666> Oh ok
<kidplayer666> Thanks
<kidplayer666> Not this week I can main Linux then :P
<marcan> last few kernels were fixing some snafus we had because we *didn't* do the dev cycle properly for Fedora and ended up in a bad situation, which is why we're doing that now, and why this never hit ALARM (where we were always doing separate dev repos and those releases were never promoted)
<marcan> Fedora itself upstream has this process, we didn't for downstream Fedora Asahi repos, hence the fuckup
<marcan> now we have a solution for that too
armaine has joined #asahi
armaine has quit [Remote host closed the connection]
<marcan> the flows are very different for Fedora and ALARM because for ALARM I just DIY everything, while for Fedora we are using Fedora infra for everything (COPR), so we need to work with that.
armaine has joined #asahi
<marcan> (one of the problems we have is kernel builds take half a day due to the builders being undersized, which we're trying to get fixed)
armaine has joined #asahi
Guest2567 has quit [Ping timeout: 480 seconds]
<kidplayer666> I noticed on the kernel builds page the long building time of some kernels
<marcan> we're building 4 kernels per build, on ridiculously undersized EC2 instances, so yeah, that doesn't end well
<marcan> if we get the changes we want pushed through, that will reduce build times to ~1 hour, which is about the same as on an M1 Ultra, so that's on par with local builds
<armaine> hello, ive been trying to enable speakers by changing the dtb located at /usr/lib/modules/6.5.0-asahi-3-1-edge-ARCH/dtbs/t8103-j293.dtb by converting it into dts then replacing it with the modified dtb. however speakers aren't still enabled. I tried running update-m1n1 as well but it still doesnt work
<marcan> armaine: no.
<marcan> just don't.
armaine has quit []
<kidplayer666> marcam:
<kidplayer666> marcan: thanks
<marcan> that machine doesn't even have a speaker profile yet
armaine has joined #asahi
armaine has quit []
<marcan> I'm going to make this very clear: unless you have a calibration microphone, DSP and audio tuning experience (on top of general Linux experience), willingness to possibly blow up your speakers, and are interested in joining the currently 2-person team taking care of this stuff in Asahi, there will be *no* support or help enablin the speakers until they are officially supported. Just wait. Please.
<marcan> *enabling
<marcan> if this stuff were as easy as "turning it on" it would've been turned on already.
armaine has joined #asahi
armaine has quit []
compassion17850 has joined #asahi
r0ni has quit [Ping timeout: 480 seconds]
compassion1785 has quit [Ping timeout: 480 seconds]
compassion17850 is now known as compassion1785
<yuka> So with the latest kernel update my camera is working, but I haven't updated the firmware in a long time (haven't booted and updated my macOS in a long time either), so it seems I don't have the calibration fixes.
<yuka> Now, I booted macOS and clicked update in the system settings, and now I have macOS 14.0. Can I do an asahi-fwupdate with that, or do I have to dongrade to some 13.x?
<yuka> (device is a Macbook Pro 14" 2021, M1 Pro)
<janneg> you need to add /usr/bin/appleh13camerad from recovery/macos to /asahi/all_firmware.tar.gz on the ESP, run asahi-fwupdate and reboot
<marcan> sbin I think
<janneg> yes, sbin
ydalton has joined #asahi
<ydalton> why does nobody learn, just wait for proper speaker safety instead of damaging your devices :P
<chadmed> ill play devils advocate for a second, it has been like two years and we've been semi-regularly teasing the work we're doing on it
<chadmed> so i get peoples excitement
<chadmed> but also yeah, keep in mind that m arcan permanently destroyed an m2 air in literally less than 5 seconds
<ydalton> chadmed: i saw that video of yours of the m2 macbook air 15 inch with music :)
<ydalton> that you brought up in 6 hours
<chadmed> yeah, worth mentioning though that being able to iterate that quickly relies on work ive been doing in the background for like 18 months
<chadmed> so its not like we could have just done that at any time in the past
<ydalton> yeah, you need practice and shit
<chadmed> also that was basically just an any% speedrun, ive worked on it since then and its much better
<chadmed> no not even practice, i ran front of house and foldback at a 5000 capacity venue for 10 years i dont need to practice :P
<chadmed> all that time was just getting the infrastructure into pipewire and wireplumber and figuring out the best way to express the required dsp
<chadmed> i ended up writing a shitty psychoacoustic bass plugin because we werent happy with any of the viable alternatives
<chadmed> just stuff like that
<ydalton> wasn't most of the time spent installing gentoo
<chadmed> for this specific machine yeah, in fact it wasnt even the install
<chadmed> it was making my gentoo bootstrapping stuff play nicely with a fedora kernel
<ydalton> normally i'd expect it to just work? what happened exactly
<chadmed> #asahi-alt
ydalton has quit [Remote host closed the connection]
ydalton has joined #asahi
hightower3 has quit [Ping timeout: 480 seconds]
zalyx0 has quit [Quit: later alligator]
chadmed has quit [Remote host closed the connection]
zalyx0 has joined #asahi
jeisom has joined #asahi
Slowlydev_ has joined #asahi
Slowlydev_ has left #asahi [#asahi]
zalyx0 has quit []
Slowlydev has joined #asahi
Slowlydev has quit []
ydalton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.0.92)]
chadmed has joined #asahi
zalyx0 has joined #asahi
cyrinux has quit []
cyrinux has joined #asahi
cyrinux has quit []
cyrinux has joined #asahi
cyrinux has quit []
cyrinux has joined #asahi
Deewiant has quit [Remote host closed the connection]
Deewiant has joined #asahi
espo has quit [Quit: WeeChat 4.0.5]
espo has joined #asahi
hex-spell has joined #asahi
<stintel> is it safe to upgrade macOS to 14.0?
<janneg> stintel: yes
<espo> should be i run macOS 14.0 and asahi and didn't have any problems
<stintel> thanks!
<janneg> leio: you want to update m1n1 to v1.4.2
darkapex has quit [Ping timeout: 480 seconds]
<leio> I do? :)
<marcan> really stupid M2 Ultra bugfix
<marcan> not sure what the impact is :)
Moprius has joined #asahi
cy8aer has quit [Remote host closed the connection]
Moprius has quit [Remote host closed the connection]
cy8aer has joined #asahi
jeisom has quit [Ping timeout: 480 seconds]
ydalton has joined #asahi
<ydalton> i don't get what appleh13camerad is. what is it running on? the webcam?
hspak101539703712 has quit [Ping timeout: 480 seconds]
<ChaosPrincess> on host cpu
<ChaosPrincess> it contains calibration blobs that it uploads to the camera
<ydalton> oh, is it a baremetal executable?\
<j`ey> it's a macOS executable
<ydalton> i don't get it, what part of the boot process is it executed?
<ChaosPrincess> it only runs under macos, the linux side just rips out the calibration out of it and doesnt use the code
<ydalton> oh, it just takes out data from that executable and the linux driver just uploads it, right?
<ChaosPrincess> yes
<ydalton> ah, i see
<ydalton> wait, where in the driver is it doing that? i'm reading the source but i can't seem to find it
<ydalton> i'm looking in drivers/media/apple/isp
<marcan> the firmware extractor does that
<ydalton> oh okay
<j`ey> also: isp_ch_load_setfile
<j`ey> (in linux)
<ydalton> which file is it in?
<j`ey> drivers/media/platform/apple/isp/isp-cam.c
<ydalton> github's search function is plain garbage XD
<ydalton> i see it now
<chadmed> browsing code in github is a waste of time
<psykose> it used to be a lot better :'(
<chadmed> just pull down the tree with --depth=1 and fire up a real editor
<psykose> nowadays the search just misses tokens..
<ChaosPrincess> psykose: it was always garbo for kernel-sized repos
<ydalton> grepping the tree is much better
<chadmed> "sorry you cant search forks" piss off
<psykose> for kernel sized sure
<psykose> but these days you can have any random app, and the search will just not find a string
<psykose> (sometimes)
<ChaosPrincess> also fuck that mba asshole who decided to lock search behind login
<psykose> ya
<ydalton> agreed
<ydalton> maybe to prevent spamming
<ydalton> but that excuse is overused
<ydalton> does C in the linux kernel support zeroing out strcuts with {0}?
<ydalton> s/strcuts/structs
<psykose> yes
<ydalton> hm, but i see memset being used sometimes
<ydalton> would like to see a elixir bootlin-style asahi linux kernel source explorer!
<ydalton> maybe i could set that up myself
<ChaosPrincess> isnt it usually on code that was written before kzalloc became a thing
<psykose> i thought struct zero init was a toggle?
<psykose> it can totally be off
<psykose> or maybe that was just allocs
<ydalton> unless you stack allocate or heap allocate
<ydalton> kzalloc is a heap allocator no?
<ChaosPrincess> heap
<ydalton> but i mean if you just stack allocate a struct you would either do memset or {0}
<psykose> the lack of consistency isn't surprising
<ydalton> {0} is kind of a new thing right?
<ydalton> "relatively" new
<psykose> {0} is c99
<psykose> = {}; (same thing) is c23
<ydalton> ah
<ydalton> and also this thing
<ydalton> struct file_operations fops = { .open = foo_open, .read = foo_read };
<ydalton> that's a dime a dozen in the kernel
<ydalton> that's also c99 i think
darkapex has joined #asahi
<j`ey> the kernel builds as C11 now!
<j`ey> very modern
<psykose> yep, also c99
<ydalton> interesting
<ydalton> yet a lot of kernel code i see still declares variables at the beginning of scope even though it's not required anymore
<ydalton> that's just for readability's sake right?
<ChaosPrincess> its kernel code style
flom84 has joined #asahi
<ydalton> is it? i don't see it in this document https://www.kernel.org/doc/html/latest/process/coding-style.html
<ydalton> unless i'm blind lol
<ChaosPrincess> i think checkpatch.pl complains, and people on mailing lists did too when i accidentally forgot that
<j`ey> also it's the style of code because 90% of code was written before that was allowed
flom84 has quit [Quit: Leaving]
<ydalton> makes sense
<chadmed> its silly, we ran into some odd behaviour in the opp-microwatt handling code because someone had done int ret; and not int ret = 0;
<ydalton> wait, how?
<chadmed> it was being initialised to a nonzero value
<ydalton> wouldn't a function set it to 0 or non zero?
<ydalton> like ret = foo();
flom84 has joined #asahi
<chadmed> yeah
<ydalton> if(ret) return ret; or a goto somewhere
<ydalton> i don't see where that would cause issues
<chadmed> it was somewhere in opp.c iirc
<chadmed> or it might have been inside the function that was returning ret
<chadmed> something like that
<chadmed> but initialising ret to 0 fixed it
<ydalton> interesting
<stintel> marcan: https://discussion.fedoraproject.org/t/webcam-users-manual-action-needed-for-full-image-quality/92027 shouldn't mv /path/to/appleh13camerad . be cp instead of mv ?
<ydalton> don't think so
<ydalton> that's where you've moved it to a temporary location
<stintel> right
<ydalton> not directly from macOS
* stintel should read everything, sorry for the noise
<ydalton> XD
* stintel considers more coffee
jeisom has joined #asahi
DannyB has joined #asahi
dcow has joined #asahi
<mort_> what is the current status on speakers by the way? Is speakersafetyd pretty much solid but the kernel interlock mechanism is lacking, or is the project not confident in speakersafetyd itself yet?
ChaosPrincess has quit [Quit: ChaosPrincess]
ChaosPrincess has joined #asahi
DannyB has quit [Ping timeout: 480 seconds]
<chadmed> the daemon works, it just needs to be cleaned up a bit. the main blocker is the kernel interface right now
<ydalton> is there an ioctl implemented to shut off the speakers?
<j`ey> it just turns them down
r0ni has joined #asahi
DannyB has joined #asahi
<ydalton> will the daemon have a feature where i can issue a command to turn the speakers off?
<mort_> I'd guess you would want to use the normal mute functionality for that?
<ydalton> but i mean return to the state we are in where the speakers don't do anything
<mort_> you could unload the kernel module for the driver I guess
AnuthaDev has joined #asahi
<marcan> right now the speakers are hard-disabled, that's not a normal state
<marcan> the normal safe state will be speakers capped to a very low volume
<marcan> if you want them off just mute them
<chadmed> new pipewire feature request: tie a virtual sink's mute toggle to an ioctl :P
<chadmed> pretty sure id get kicked off g.fd.o if i asked for that
<espo> cam working :D, let us quickly turn it off again :D.
espo has quit [Quit: WeeChat 4.0.5]
espo has joined #asahi
DannyBATL has joined #asahi
DannyB has quit [Read error: No route to host]
<ydalton> does the kernel provide a uapi for haptic feedback? on macOS you could hold down a word on the screen and it would click harder
hightower2 has joined #asahi
flom84 has quit [Ping timeout: 480 seconds]
<chadmed> libinput kinda knows about haptic touchpads which it calls "forcepads" but nothing is really wired up
<ar> ydalton: iirc evdev provides some forcefeedback capabilities, though i'm not sure if evdev supports defining that for non-joysticks/gamepads
hightower2 has quit [Read error: Connection reset by peer]
ydalton has quit [Ping timeout: 480 seconds]
AnuthaDev has quit [Ping timeout: 480 seconds]
dcow has quit [Remote host closed the connection]
<ah-> and pretty nicely contained with the microvm that has fex and the rootfs etc. all ready to go
zalyx0 has quit [Quit: later alligator]
zalyx0 has joined #asahi
dcow has joined #asahi
hightower2 has joined #asahi
dcow has quit [Ping timeout: 480 seconds]
hightower3 has joined #asahi
hightower2 has quit [Read error: Connection reset by peer]
hightower4 has joined #asahi
darkapex has quit [Remote host closed the connection]
darkapex has joined #asahi
hightower4 has quit [Read error: Connection reset by peer]
hightower3 has quit [Ping timeout: 480 seconds]
jeisom has quit [Ping timeout: 480 seconds]
zalyx0 has quit [Quit: later alligator]
dcow has joined #asahi
DannyBATL has quit [Ping timeout: 480 seconds]
dcow has quit [Ping timeout: 480 seconds]
ydalton has joined #asahi
<ydalton> ah-: really? i could not log in for some reason
<blazra> ydalton: There is this patchset waiting on the LKML: https://lore.kernel.org/all/20221007182945.1654046-1-acz@semihalf.com/
<j`ey> a year old..
hightower2 has joined #asahi
<ydalton> damn
<ydalton> and hardly any input
<ydalton> pun intended
zalyx0 has joined #asahi
<blazra> If it is merged, we could use the new API for setting the haptic feedback strength and for the host click mode. Although, I am not sure how beneficial the host click mode is - there would have to be support for some "special" click in apps or at least in the DE.
dcow has joined #asahi
dcow has quit [Remote host closed the connection]
dcow has joined #asahi
dcow has quit [Ping timeout: 480 seconds]
<pitust> i write to the touchbar brightness, and the brightness doesn't change for some reason?
dcow has joined #asahi
<ah-> ydalton: yep, though I had to login via the phone app with the qr code
cyrinux has quit []
dcow has quit [Remote host closed the connection]
cyrinux has joined #asahi
dcow has joined #asahi
<ydalton> the code that you get via email didn't work
<ydalton> same with you?
<ah-> yeah, it took ages for me to arrive and then it didn't work
<ydalton> guess i'm out of luck then
ydalton has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.0.92)]
<leio> janneg: marcan: that 1.4.2 doesn't work for me unfortunately - https://bpa.st/JEOA
<marcan> leio: that's a different issue, maybe your m1n1 stage 1 needs updating. poke janneg.
dcow has quit [Remote host closed the connection]
dcow has joined #asahi
dcow has quit [Remote host closed the connection]
dcow has joined #asahi
dcow has quit [Remote host closed the connection]
<leio> stage1 was updated too, then I didn't get anything; this is output after restoring stage1 to previous working and run_guest.py'ing the stage2
johnbchron has joined #asahi
<leio> previous working is janneg's m2_dptx_v13_5, only other difference is probably RELEASE=1 now while not before
jacksonchen666 has quit [Ping timeout: 480 seconds]
<johnbchron> looking at support for m2 pro macbook 14", looks like the installer is WIP but most of the other essential features work. does that mean there's a more involved process I can follow to install or I should just wait until installer is done?
dcow has joined #asahi
<johnbchron> device is more specifically "Mac14,9"
zzywysm has joined #asahi
johnbchron has quit [Quit: johnbchron]
jeisom has joined #asahi
dcow has quit [Remote host closed the connection]
dcow has joined #asahi
psykose has quit [Remote host closed the connection]
psykose has joined #asahi
dcow has quit [Ping timeout: 480 seconds]
dcow has joined #asahi
dcow has quit [Remote host closed the connection]
dcow has joined #asahi
dcow has quit [Remote host closed the connection]
<janneg> pitust: that's systemd preventing you from turning your display off, see https://github.com/WhatAmISupposedToPutHere/tiny-dfr/blob/master/etc/udev/rules.d/99-touchbar-backlight.rules how to prevent that
Stroller has joined #asahi
<janneg> leio, marcan: keeping the stage 1 m1n1 at a version which doesn't interact with the display is probably safer for now. stage 1 errors are invisible without hw-serial
<marcan> janneg: as long as stage1 doesn't outright crash/hang, backdoor proxy mode should get you a log from that
<janneg> leio: can you uncomment RTKIT_SYSLOG in config.h and #define DEBUG before the includes in src/afk.c
Stroller has quit []
hex-spell has quit [Quit: WeeChat 4.0.5]
<leio> janneg: now it booted fine instead
richyliu21 has joined #asahi
<janneg> comment the RTKIT_SYSLOG again, sounds like a timing issue and the 25ms are not enough for the studio
<leio> wasn't that about just not getting the display to light up, as opposed to complete m1n1 "sorry I give up" freeze?
<leio> but yeah, a minute; got my payloads over the other machine already for this :)
richyliu2 has quit [Ping timeout: 480 seconds]
richyliu21 is now known as richyliu2
<leio> now it just works, even after removing DEBUG as well...
<leio> janneg: been trying to hit it again to no success; I don't know what's up then; maybe it's too fast for stage1 only, but when stage1 of mine (which is your m2_dptx_v13_5 and does init display at 1080p) has done some setup already, it's fine in stage2 and the initial one I had failing (the paste), maybe it was something about not having shut down fully or who knows? Just can't reproduce anymore with stage2 even after going back to release tag
<leio> This is all with RELEASE=1 builds again too to be fast, given the delay suspicious; oh and compared to before, I did the csrutil disable dance that might not have been there before, thus stage1 didn't wait for proxy
<leio> I'll try emptying boot-args
lynndotpy has quit [Quit: bye bye]
lynndotpy has joined #asahi
<janneg> I can image that the 2nd stage failure you saw could be caused timout due to timing issues in the 1st stage
<leio> janneg: I can't reproduce anymore; the pastebin with the failure was already with the stage1 I have now again
<janneg> let me think of a different way to handle the timeouts, I don't think it makes sense to torment you with 1st stage reinstalls
<leio> I'm fine with it, as long as it doesn't do anything bad to SSD due to low-level writes or anything like that when done a lot :)
<leio> but yeah, next step would otherwise be to try to write it into stage1 again; which would replicate what people will get once it gets updated in the installer
<kettenis> that is what I get with CONFIG_RTKIT_SYSLOG
<kettenis> (that is stage2)
<janneg> kettenis: which m1n1 version is in stage 1? v1.3.3?
<kettenis> v1.3.3 indeed
AnuthaDev has joined #asahi
SalimTer- has joined #asahi
Z750 has quit [Quit: bye]
Z750 has joined #asahi
SalimTerryLi has quit [Ping timeout: 480 seconds]
hex-spell has joined #asahi
AnuthaDev has quit []
cylm has joined #asahi
ChaiTRex has quit [Remote host closed the connection]
ChaiTRex has joined #asahi
crabbedhaloablut has quit []
<pitust> janneg
<pitust> i can't see how the udev rules are an issue, i am *raising* the backlight brightness?
jeisom has quit [Remote host closed the connection]
jeisom has joined #asahi
dcow has joined #asahi