Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
millenialhacker has joined #asahi-dev
Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
Catyre has quit [Ping timeout: 480 seconds]
bard[m] has joined #asahi-dev
Catyre has joined #asahi-dev
Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
<chadmed>
marcan: does asahi-fwextract do anything different to the extant update-vendorfw script that would necessitate an ebuild?
<chadmed>
or would i be better off just writing a little openrc service that runs update-vendorfw on boot
Catyre has quit [Ping timeout: 480 seconds]
Catyre has joined #asahi-dev
derzahl has quit [Remote host closed the connection]
WindowPain has quit [Quit: ZNC 1.8.2+deb2~bpo10+1 - https://znc.in]
Catyre has quit [Remote host closed the connection]
millenialhacker has joined #asahi-dev
Catyre has joined #asahi-dev
PhilippvK has joined #asahi-dev
WindowPain has joined #asahi-dev
skipwich_ has quit []
skipwich has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
phiologe has quit [Ping timeout: 480 seconds]
kov has quit [Quit: Coyote finally caught me]
<marcan>
they are different things
<marcan>
asahi-fwextract isn't really intended to keep existing in the long term
<marcan>
once we extract all firmwares it will not longer be needed
<marcan>
*no longer
<marcan>
jannau: m1n1 should already reset NVMe on shutdown, why would we need to do it twice?
<marcan>
maybe udelay(1) is not enough?
<marcan>
oh wait, I bet I know
<marcan>
it's called ANS now, not ANS2.
<marcan>
so the m1n1 reset does not work
<marcan>
since it was by name
<marcan>
chadmed: update-vendorfw *already* runs on boot
<marcan>
that one just copies firmwares
<marcan>
asahi-fwextract *recreates* the source file for that copy, it has actual logic
<marcan>
update-vendorfw puts the firmwares in the right place after an external entity (installer) updates them, e.g. due to a stub upgrade
<marcan>
asahi-fwextract just does what the installer does to extract them in the first place, because the installer isn't complete yet and I didn't want to have to tell people to go back to the installer just for this
millenialhacker has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
<marcan>
ok, boots now and the trackpad works but... the first time setup stuff failed? got me stuck in an empty login screen, heh
<marcan>
ohh calamares needs a rebuild
<marcan>
ruh roh, different issue now
<marcan>
u-boot explodes with USB devices connected
<marcan>
scanning bus usb@502280000 for devices... Device NOT ready
<marcan>
elr: 000000000001d5d0 lr : 0000000000028440 (reloc)
<marcan>
ah wait, this is because it's a storage device that's not ready? bug then I guess (I have a multithing connected that includes card readers)
<marcan>
kettenis: might want to look at this ^
<marcan>
probably a null deref or something when USB storage devices are connected and not ready
<chadmed>
marcan: yea ik it runs on boot on asahi but i havent gotten around to setting up an openrc service to do so on gentoo just yet so i probably should for when these fw upgrades roll out
<marcan>
yes
<nicolas17>
earlier I saw a merge request in KDE and recognized chadmed's username... before even reading the description which clearly mentioned asahi ^^
<chadmed>
heh yeah ive got a few weeks off at the moment so i thought id start paying off the enormous tech debt i owe
<chadmed>
i hacked basic DT support into kinfocenter last night but its pretty ugly so idk if itll fly upstream
<chadmed>
its no uglier than the dmidecode stuff i guess but...
<nicolas17>
hm slighly concerning that we don't even compile that file in GitLab CI
<nicolas17>
oh, I guess we do as part of android
<chadmed>
it works on x86 machines, kinfocenter shows my macbook pro's dmi information
<nicolas17>
sorry I meant solid's cpuinfo_arm.cpp
<chadmed>
oh right
<chadmed>
well it complained about a syntax error i made but maybe that was the android pass
<nicolas17>
hm did it?
<chadmed>
yeah i had a . instead of a , in the array somehow despite the fact that i built the branch to test it, and it told me that i ruined the code quality
<nicolas17>
ah looks like cppcheck caught it (because it tests all ifdef combinations)
<chadmed>
ah righto
<nicolas17>
we do need to add a proper ARM build then...
Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
<marcan>
making another kernel...
<marcan>
I *think* with the m1n1 changes and other things this will finally work on M2 (with proper keyboard detection even)
Catyre has quit [Ping timeout: 480 seconds]
<marcan>
sven: bluetooth works on the funky revision in the M2pro :)
Catyre has joined #asahi-dev
<nicolas17>
soon apple will release the larger chip and make all this confusing again (M2 Pro chip or M2 MacBook Pro?)
<marcan>
yup...
<sven>
nice!
<marcan>
jannau: I think the trackpad in practice works because the init is slow enough that by the time it requests the firmware the root FS is mounted
<sven>
i wonder if you’d notice if we sent the wrong calibration data though
<marcan>
(because it takes ~1 second for the trackpad to init)
<marcan>
sven: *shrug*
<sven>
fair enough :D
<marcan>
this release is for the news cycle, "we have M2 support oh yeah and bluetooth", it's gated behind expert mode for a reason :p
<sven>
:)
<sven>
Guess that really just leaves T2
<marcan>
yup
<nicolas17>
do ARM MacBook Pro/Air still have the 'fake' trackpad click that they simulate with haptic feedback?
<chadmed>
yeah
<marcan>
(that's handled by default in firmware, but can be taken over by the OS)
<marcan>
building new asahi images, hopefully this works because I really don't have time to spend on this the rest of the day...
Catyre has quit [Remote host closed the connection]
nicolas17 has quit [Quit: Konversation terminated!]
<marcan>
images uploading, alx.sh/dev should work in a few minutes
<marcan>
going to grab lunch...
Catyre has joined #asahi-dev
doggkruse has quit [Ping timeout: 480 seconds]
doggkruse has joined #asahi-dev
<jannau>
marcan: trackpad was consistently broken with the rootfs on usb and modules but no firmware inside the initramfs
<jannau>
bluetooth was very unstable on my m2 and detected only 1/2 of the devices
<jannau>
installer worked on the m2 macbook pro but touchpad is not working on first boot
doggkruse has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jannau>
touchpad worked 3 out of 5 boots including the first boot
<jannau>
install finished on the mac studio as well
Catyre has quit [Remote host closed the connection]
<sven>
unstable Bluetooth sounds a bit like wrong calibration data
<jannau>
I see lots of "unknow advertising packet type:" and devices missing from the scan which show on my x86 notebook
<jannau>
minor annoyance with the dev installer: it uses "asahi" as package repo. this prevents us from providing updates but I guess it necessary if we want to promote dev to stable at some point
<sven>
the unknown packet type thing should be harmelsa
<jannau>
marcan: both the desktop and minimal os images work
<sven>
*harmless
<sven>
i also see that when scanning
<sven>
the missing devices is strange though
millenialhacker has joined #asahi-dev
minecrell9 has joined #asahi-dev
mrkajetanp has joined #asahi-dev
skrzyp has quit [Remote host closed the connection]
dabaum|2er0 has joined #asahi-dev
skrzyp has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
the_lanetly_052__ has joined #asahi-dev
mps has quit [Ping timeout: 480 seconds]
<marcan>
jannau: the package repo thing is intentional
<marcan>
for that reason
<marcan>
and yeah, trackpad does not work on first/some boots
<marcan>
"good enough for experimental"
<marcan>
it works, it proves M2 works, bugs happen, whatever
<marcan>
ok, then, please test on M1 and friends :)
<marcan>
sven: soundflora* tested her AirPods and said they had periodic glitches, but that could be any number of things
<marcan>
I did test some speakers myself and they sounded fine but I wasn't playing super close attention
<marcan>
anyway, all this is acceptable for an experimental-gated platform
<sven>
I’ve tested with my Bose headphones and they were fine on M1 fwiw
<sven>
maybe I should test with my AirPods as well
<marcan>
(note: this was on M2)
<sven>
wouldnt surprise me if apple made Broadcom add some, uh, surprise features for AirPods
<marcan>
oh they definitely do weird magic for airpods, but I don't think the controller would care about that unless the software tells it to
<marcan>
(e.g. the magic pairing etc)
<marcan>
pretty sure apple writes this firmware btw, or part of it
<marcan>
(based on whatever broadcom provides)
<marcan>
it's got tons of AWDL/etc stuff in it
<marcan>
(the AirDrop stuff)
<sven>
maybe that explains why the ipc interface is so cursed
<sven>
they had to start with Broadcom’s mess and instead if cleaning it up just added more hacks on top
<marcan>
yeah, see, I think Apple wrote 100% of the macOS driver these days, for wifi
<marcan>
and I guarantee the reason for that was that broadcom's reference code is a massive cursed mess
<marcan>
(you've seen the android thing, right?)
<marcan>
but they're not quite at the point of writing their own firmware yet
<sven>
that out-of-tree android broadcom driver?
<marcan>
yeah
<sven>
yeah, it's, uh, great
<marcan>
I think at least ever since Skywalk Apple just have a fully bespoke, IOKit/C++ macOS brcmfmac driver
<marcan>
pretty sure they used broadcom reference code before, but no longer
<sven>
interesting
<sven>
hm, had some trouble to get the airpods to connect here (but that's common without apple magic) but sounds seems to be fine
MajorBiscuit has joined #asahi-dev
the_lanetly_052___ has joined #asahi-dev
the_lanetly_052__ has quit [Ping timeout: 480 seconds]
millenialhacker has joined #asahi-dev
the_lanetly_052 has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
firefox317 has joined #asahi-dev
<firefox317>
marcan: i tested a system update (pacman -Syu) with asahi-dev packages on a m1 air, and got the following crash in dmesg: https://pastebin.com/b5gX0Wre
the_lanetly_052___ has quit [Ping timeout: 480 seconds]
<sven>
bluetooth firmware is missing and that's "only" a WARN_ON so the system shouldn't die after that
<sven>
but let me see what causes that WARN
<firefox317>
sven: oh yeah i see, I thought doing a pacman -Syu with asahi-dev would cause the firmware to install somehow, but I guess that is incorrect
<firefox317>
sven: The system is not crashing tho, its just the bluetooth driver, the rest still works okay
<sven>
i think i just forgot to add a call pci_free_irq_vectors which causes that warn_on
<sven>
the bluetooth driver also doesn't crash
<sven>
it just doesn't load
<kettenis>
marcan: the USB code in u-boot isn't great
<kettenis>
but I'm not sure I'm the one to improve it...
<firefox317>
Ah interesting that a WARN_ON causes such a stracktrace to be printed, but yeah I see whats happening
<marcan>
the firmware does get installed with a pacman upgrade
<firefox317>
hmm, maybe just a reboot from the terminal caused it to not properly shutdown or something
<marcan>
paste a ls -al /boot/efi/asahi /boot/efi/vendorfw
<marcan>
looks like the update didn't work, do you have the `asahi-meta` package installed?
<firefox317>
aha, that one was not installed, not sure why I did not have that package. I think I did a regular install when the first beta came out
<marcan>
you probably uninstalled a dependency at some point and accidentally uninstalled the meta in the process
<firefox317>
could be yeah, lets try a reboot and see if it works :)
firefox317 has quit [Quit: Page closed]
firefox317 has joined #asahi-dev
<marcan>
(either that or you installed *before* the public announcement, since those meta packages were a very last minute addition)
<firefox317>
yeah I was playing with asahi before the public release, so might have been that. It found the firmware now :)
<marcan>
yay!
<marcan>
if you have the desktop install, also do asahi-desktop-meta
<marcan>
that will pull in all the other randomness to make it just work
<firefox317>
Ah thanks, I was just about to ask that indeed, because I saw that the bluetooth.service was not enabled, which I guess that package does xd
<marcan>
yup, and also pulls in the KDE support stuff
<marcan>
the minimal install as usual is "roll it yourself" for everything not specific to asahi
firefox317 has quit [Quit: Page closed]
firefox317 has joined #asahi-dev
<firefox317>
fwiw installing that package did not actually enable the bluetooth service, but that might have been something on my side again.. I will just manually enable it :)
firefox317_ has joined #asahi-dev
firefox317 has quit []
<marcan>
ah, installing it wouldn't
<marcan>
upgrading it would
<marcan>
that was intentional
<marcan>
it is upgraded separately when building images
<marcan>
*enabled separately
<marcan>
so I only made it enable it specifically on upgrade from v1 to v2, for existing users
<firefox317_>
okay, that makes sense :) sound from the bluetooth JBL speaker sounds good!
<marcan>
:D
<sven>
:)
mps has joined #asahi-dev
<chadmed>
ok there we go bt is working now on gentoo
<chadmed>
updated m1n1 and got the bluetooth firmware out of macos
derzahl has joined #asahi-dev
<chadmed>
my crappy little taotronics buds sound surprisingly good all things considered
<chadmed>
nowhere near as good as the built in speakers do but yknow, good enough
<kettenis>
marcan, jannau: should we still reset ANS/ANS2 from U-Boot as well even if m1n1 is now fixed?
<marcan>
I don't think so, because if ANS2 has been put into quiesced mode instead, that would probably break it instead
<marcan>
when we get the handoff from iBoot it's in hibernate mode with the reset already applied (or some dance that puts it into an equivalent state), so I think we should keep the reset the responsibility of the shutting down party, since we know it's not something that works every time indiscriminately
<marcan>
and then that keeps the same ABI
<marcan>
however we should still do the message+CPU wakeup (which u-boot already does), since that seems to be a reliable way of bringing ASCs up regardless of whether they are in hib or quiesce mode, and that's good (macOS can't handle that, it has to be told in the ADT what the current state is)
the_lanetly_052__ has joined #asahi-dev
the_lanetly_052 has quit [Ping timeout: 480 seconds]
<dottedmag>
marcan: is pointing out typos useful at this stage of draft, or you're going to go through it anyway and correct these?
<marcan>
I'm too tired to find typos, so yeah, please make a list
<marcan>
need to pack for a flight...
mps_ has quit [Remote host closed the connection]
mps has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
<chadmed>
whats the go upstream with exposing stuff like serial numbers and the like in the devicetree? there used to be docs on it but all of it seems to have vanished
<chadmed>
did it become unacceptable to do that?
firefox317_ has quit [Remote host closed the connection]
<chadmed>
i know the rasperry pi firmware inserts /serial-number into the fdt and it seems to have been the canonical way of doing that in 2015 or so
mps has quit [Ping timeout: 480 seconds]
<dottedmag>
marcan: Checked the whole text, only one typo spotted: "We’ve decide to take the plunge"
<marcan>
fixed, thanks!
<marcan>
chadmed: I don't see how it's any different from DMI, which also exposes serial numbers. was it ever bad?
<marcan>
MAC addresses also count as unique IDs and those certainly go in the DT
<dottedmag>
marcan: wouldn't it be useful to have a link to a page with current hardware support status?
<marcan>
we don't have a canonical one, I'd make one but I'm out of time, I just need to get this out...
<chadmed>
yeah ive added it to m1n1 in my local branch already and its working nicely, but i want it to play nice with upstream software so i wanna know where in the DT to put that sorta thing
<marcan>
oh, I guess someone made it then :)
<marcan>
needs to be updated though
<kettenis>
chadmed: looks like that got list somehow with the move of some stuff into the device tree specification and dt-schema
<_jannau_>
marcan: USB-A ports on all mac studio models have the same firmware problem as the m1 max front ports
<kettenis>
u-boot will still set the property if a serial# environment variable exists
<marcan>
_jannau_: ah, right, forgot about those
<marcan>
_jannau_: fixed
<chadmed>
kettenis: ive got m1n1 grabbing it from the ADT and putting it at /serial-number currently, kinfocenter happily picks it up with my DT support changes
<chadmed>
would it make more sense to somehow pass it to u-boot another way and let u-boot do it?
<kettenis>
no, I think letting m1n1 do this is fine
the_lanetly_052__ has quit [Remote host closed the connection]
<chadmed>
sweet
Catyre has joined #asahi-dev
<_jannau_>
marcan: "never break backwards with older firmware" misses compatibility
<marcan>
fixed, thanks
<j`ey>
marcan: cool post!
<_jannau_>
marcan: you could mention in the second last DCP paragraph the this changes allows to turn the display off and on
<marcan>
ah, yes!
<marcan>
_jannau_: done
<marcan>
oddly enough, this trip is going to be the first time *I* will be using Asahi as a daily driver, since I'm only bringing Macs
<marcan>
we'll see how that goes
<marcan>
currently installing the proper release on my M1Pro (what I've been using as a build environment grew out of a manual install and isn't very kosher)
<chadmed>
for how far along we are its honestly surprisingly useable
<chadmed>
if we had gamma correction for blue light filtering at night id never feel the need to boot back into macos again
<chadmed>
and of course as i say that the alsa preallocation thing comes back :/
<marcan>
if I have free time (big, big if) I might start messing around with your audio stuff
<marcan>
during the trip, I mean
<marcan>
otherwise, after that
<j`ey>
big.IF
<dottedmag>
chadmed: Don't Wayland compositors have it? Or do they punt to DRI gamma calls (that are obviously not available yet)?
<dottedmag>
*some Wayland
<marcan>
I think most night shift stuff expects to be able to tweak gamma tables directly
<chadmed>
kwin (the only one worth using ;)) relies on gamma LUTs from CRTCs
<chadmed>
this seems to be the case even in xorg nowadays too unfortunately
<dottedmag>
Messing with gammas is probably not going to work well once CMS protocol lands in Wayland and compositors will start to take care about it.
<marcan>
the only Asahi machine I've been using with regularity is oddly enough the iMac, and that one has been pretty janky but it's the worst case (big screen, small CPU, and I mostly use it to preview youtube streams so the software rendering really shows there)
<dottedmag>
But that's in future anyway.
<marcan>
(using in the desktop sense)
<marcan>
though I see the Mac Studio becoming my storage server in the near future
<marcan>
considering buying 2 10Gb Mac Minis and turning it into a proper cluster... but I might wait for the M2 Mac Mini to see how that goes
<marcan>
well, storage and build server of course
<chadmed>
ive been trying to look for a "dumb" thunderbolt drive cage that only exposes a bunch of disks so i can eventually just use dmraid but theyre all like $3000 and have crap implementations of hardware raid in them
<maz>
marcan: I'm halfway there too. need to move the backups over to the Studio.
<marcan>
nice!
<marcan>
chadmed: I want USB3 ones for flexibility (don't really care about perf for spinning rust), and there are a few, but they all seem to have pretty bad performance even by USB3 standards
<marcan>
and questionable behavior on hotplug of the bays
<maz>
BTW, if anyone has recommendations for a good USB-SATA enclosure that would take 4 3.5 HDs, I'm all ears.
<marcan>
okay... I see we're all in the same boat here.
<maz>
;-)
<marcan>
for what it's worth, I'm starting to lean towards "bunch of single-drive cases" and hubs... I think that's the only thing that's guaranteed not to screw over other drives when one dies, and not to bottleneck as long as the hub is kosher...
<marcan>
it won't be pretty but...
<chadmed>
the usb ones all look so dodgy which is why ive been avoiding them, im already sus enough about mechanical disks without having to worry about the interface i connect them to as well
<marcan>
yeah
<dottedmag>
marcan: nothing a duct tape or 3d-printed cage to keep them apart for better ventilation can't fix
<marcan>
if I decide to go thunderbolt, that would have to be DIY
<marcan>
some of those cheap reference TB3 bridge boards (I have one, it works), known good SATA controller
<marcan>
dottedmag: exactly, and I *do* have a 3D printer
<chadmed>
egpu enclosure -> pcie sata controller -> 4U disk shelf
<chadmed>
modern problems require modern solutions
<marcan>
also, I could just buy a bare SATA cage with SFF connectors and DIY the USB backend behind it, it would be prettier and give the same result
<marcan>
and then it would be upgradable to thunderbolt
<marcan>
do I actully put making this into a DIY board on my endless project list?
jluthra has quit [Ping timeout: 480 seconds]
<marcan>
(i.e. SFF connector, pile of USB3 SATA bridges, hub, all on one board)
<marcan>
... screw it, it goes on.
<marcan>
whether it'll ever get popped off, who knows
<chadmed>
i actually did think about attempting that, but fabricobbling together a tb4 to pcie interface doesnt sound like the easiest thing heh
<marcan>
note: does not provide 12V, you need to jury rig that on if you need it (it's meant for like NVMe drives and such that don't)
<marcan>
$150 is not that bad for the TB3 world
<chadmed>
oh wow thats exactly what i imagined in my head
<marcan>
I have one, it totally works
<marcan>
it's some intel reference board
<chadmed>
only 50 bucks for me too
<chadmed>
gotta love living so close to the supply chains
<marcan>
wait 50?
<marcan>
you sure that's not one of the accessories only?
<chadmed>
oh nvm yeah i hadnt selected the board in the options
<chadmed>
225 is still good compared to the $3000 or so that lacie et al ask for here
<marcan>
yup, exactly
<marcan>
and then you plug in your favorite HBA
<chadmed>
yep and i can just stuff it all into a custom unit
<marcan>
TBH since I want 3 of them, if I'm going this route (hi sven, thunderbolt when? :D) I might just go for cheap known-good multi-SATA boards. it's either that or LSI SAS2008, everything else is suspect.
<marcan>
yup
<chadmed>
yeah im just gonna grab whatever 6+ port sata controller i can get and use software raid since you cant really get "real" hw raid cards for reasonable money anymore
<chadmed>
(ones with big juicy caps)
<marcan>
SAS2008 is widely available, but I always use them in HBA mode anyway
<marcan>
I think these days random SATA controllers are good enough, as long as you check for benchmarks
<marcan>
in particular... I've had bad experiences with older JMicron
<chadmed>
that lsi card is cheaper than a generic asmedia sata card from a local computer hardware retailer
<chadmed>
and id prefer the sff connectors anyway so i can use a real disk shelf or something
<marcan>
look for IBM M1015
<marcan>
and yeah, the nice thing is these can all do the fancy SAS stuff, enclosures, etc
<marcan>
just make sure you crossflash a known good IT firmware on
<marcan>
MegaRAID firmware is terrible
<sven>
thunderbolt :<
<sven>
once I have that I can finally daily drive asahi :D
<marcan>
same :)
<chadmed>
is there finally an interface sven hates more than usb?
<marcan>
I wonder if by then the new iMac Pro will be out... that *could* become my main machine
<marcan>
chadmed: I mean he's dealing with bluetooth right now
<marcan>
which I've always maintained is USB, but worse
<marcan>
though that was before USB3/Type C
<marcan>
I think they're sort of tied again now
<sven>
Bluetooth was reasonable-ish because I could ignore all the low level crap
<sven>
and thunderbolt is so annoying because it’s part of the entire usb mess
<sven>
I have a rough idea how it works but I can’t really bring it up from Python/m1n1
the_lanetly_052 has joined #asahi-dev
<sven>
so I’ll have to untangle ATC and how the linux thunderbolt and typec subsystem work at the same time and make sure I don’t break dwc3 at the same time
Catyre has quit [Remote host closed the connection]
<marcan>
_jannau_: can you make sure that github comment I keep linking has the latest kernel/u-boot package links? I think what we have in the repo now is what will ship
<sven>
Bluetooth was fun compared to that: just implement that cursed ipc and hook into hci_recv and _send_frame :D
<marcan>
(unless something terrible happens)
Catyre has joined #asahi-dev
<kettenis>
someone should do a cheap NVMe-oF enclosure ;)
<kettenis>
marcan: is there time to get some U-Boot/OpenBSD stuff in that report?
<marcan>
if you write it up and I can just paste it in :)
<marcan>
probably won't matter for HDDs, but still
<marcan>
(~400MB/s ceiling, vs ~550 for an LSI, with an SSD)
<sven>
huh, is m2 support not included in the normal Mac is release or did my MacBook not update all the way?
<sven>
*macos release
<sven>
(ofc it enabled autocorrect with that update again *sigh*)
<marcan>
sven: it is not
zzywysm has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<marcan>
are you asking because the installer won't let you pick 12.4 on a non-m2 mac?
<marcan>
(even in expert mode)
<marcan>
because yes, it won't, for this reason
doggkruse_ has joined #asahi-dev
<marcan>
m2 12.4 is a super funky special version
<sven>
im asking because I need a BlueTool with m2 support to see which calibration data it wants :D
<marcan>
ahhh
<marcan>
sec
doggkruse has quit [Ping timeout: 480 seconds]
<sven>
I guess I can just grab it from the ipsw
<marcan>
is it in the ramdisk?
<marcan>
I guess it would be
<sven>
that's what I hope
<marcan>
or just give me a sec :p
<sven>
sure :D
zzywysm has joined #asahi-dev
Catyre has joined #asahi-dev
Catyre_ has quit [Ping timeout: 480 seconds]
<sven>
what's the m2 chip again? 4387 c2? c3?
<sven>
ah, no 43*78* b3 for the pro
alyssa has joined #asahi-dev
<alyssa>
re "Night Light" with DCP,
<sven>
looks like it also wants the beamforming calibration data
<alyssa>
there are 3 colour transforms in KMS. pre-gamma, colour transform matrix, and post-gamma.
<alyssa>
The middle one -- applying a linear (or affine? I forget) transformation to every pixel -- is easy with DCP
<alyssa>
I have a patch hardcoding such a matrix into my kernel for personal a11y reasons... would be easy to pipe into userspace properly if we had a consumer.
<alyssa>
It's less obvious how to set gamma curves, and at least GNOME's Night Light is based on gamma (not a linear transform)
<nicolas17>
sven: it's weird, there were 3 builds of 12.4, the original was M1-only, then they released one with M1 and M2, two days later they released another that is M2-only
mps has quit [Quit: Lost terminal]
<sven>
yup, b3 also needs that beamforming blob. that probably explains why bluetooth was unstable on the M2
<nicolas17>
also remember BuildManifest.plist can be downloaded from the same directory as the .ipsw, which is handy to quickly know contents and supported macs :)
mps has joined #asahi-dev
millenialhacker has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]
Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
Gaspare has quit [Quit: Gaspare]
Catyre has quit [Ping timeout: 480 seconds]
Catyre has joined #asahi-dev
MajorBiscuit has quit [Ping timeout: 480 seconds]
Catyre has quit [Remote host closed the connection]
Catyre has joined #asahi-dev
mps has quit [Remote host closed the connection]
millenialhacker has joined #asahi-dev
Gaspare has joined #asahi-dev
millenialhacker has quit [Ping timeout: 480 seconds]