marcan changed the topic of #asahi to: Asahi Linux: porting Linux to Apple Silicon macs | Not ready for end users / self contained install yet. Soon. | General project discussion | GitHub: https://alx.sh/g | Wiki: https://alx.sh/w | Topics: #asahi-dev #asahi-re #asahi-gpu #asahi-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
<zv>
hey so the GCC compile farm is preparing an announcement about the new M1 hardware being made available to the public, and we are also planning to coordinate an announcement with LWN to spread the news. I know Asahi is planning a blog post soon--any idea when that will drop? It will be good to have that happen first so everyone gets the newest information.
<zv>
marcan: Glanzmann: ^
zamadatix has joined #asahi
al3xtjames has quit [Quit: Ping timeout (120 seconds)]
al3xtjames has joined #asahi
zamadatix has quit [Quit: Leaving]
<chadmed>
id imagine it would coincide with the alpha release of the installer, which is happening very soon. marcan's been tying up some release engineering loose ends and adding some nice-to-have functionality to m1n1 so that distros/users can upgrade it, u-boot and their dt as they please without changing the boot policy every time
yuyichao_ has quit [Ping timeout: 480 seconds]
Gue___________________________ has joined #asahi
Gues__________________________ has quit [Ping timeout: 480 seconds]
yuyichao_ has joined #asahi
riker77_ has joined #asahi
<matthewayers[m]>
Does anyone have any recommendations for a desktop environment? I'm a little tired of GNOME and I haven't quite decided to go back to KDE yet.
<zv>
lxqt?
riker77 has quit [Ping timeout: 480 seconds]
riker77_ is now known as riker77
<matthewayers[m]>
I'll look into it
minkuu[m] has joined #asahi
<minkuu[m]>
Hello :)
<minkuu[m]>
I’m a noob but I’d like to learn pls teach me
<chadmed>
the wiki is a good place to start
<emergenz[m]>
<minkuu[m]> "I’m a noob but I’d like to learn..." <- also I think the book linux device drivers might be a good starting point
<emergenz[m]>
for the fundamentals
<bluetail[m]>
<matthewayers[m]> "Does anyone have any recommendat..." <- Debian-11.2+cinnamon
phiologe has joined #asahi
<minkuu[m]>
<chadmed> "the wiki is a good place to..." <- I will definitely look!
PhilippvK has quit [Ping timeout: 480 seconds]
kov has quit [Quit: Coyote finally caught me]
kov has joined #asahi
marvin24 has joined #asahi
marvin24_ has quit [Ping timeout: 480 seconds]
<bluetail[m]>
with debian linux on the mac, I should theoretically have more privacy and if properly configured, more security.
<bluetail[m]>
Planning in addition to what I use already to use the mac mini also as 'pihole' https://pi-hole.net/
<bluetail[m]>
But mostly it will be a media server with a couple of services (locally)
<zv>
> more privacy and if properly configured, more security. // as opposed to?
<bluetail[m]>
bare macOS. Don't they share at least some structural information of what files I have et cetera? Prism Peak org states that Apples ecosystem is part of NSA's surveillance network.
<bluetail[m]>
I dont have proof for the opposite.
<zv>
not really a debate I want to get into but I would not take Linux or any distribution thereof to be secure if I'm not aware of and in control of every single aspect of the system.
<bluetail[m]>
Doesnt that go for any system? I'm just trying to make the best out of my privacy and security. Of course I wouldn't want a Windows OS for that purpose..
<zv>
I mean, if I'm exposing something to the Internet I'd probably go for a mainframe
<bluetail[m]>
I don't get that reasoning. How does a piece of hardware relate to operating systems?
<bluetail[m]>
Wouldn't you just try to grant as little privilege as possible and don't use internet if possible?
<zv>
mainframe + accompanying operating system + bulletproof configuration. anyway off-topic so I'll sit this out.
<bluetail[m]>
#_oftc_#asahi-offtopic:matrix.org
gladiac has joined #asahi
<marcan>
zv: I'd love to get a release out this weekend/monday, though it depends a bit on how well this weekend goes with tying up loose ends
<marcan>
but given chainloading is in place I'm less worried about fixing things after the fact, so that should make things happen Pretty Soon Now
<marcan>
it says 4 cores / 4 threads though, that sounds wrong? it should be 8/8 unless you're completely ignoring the E-cores
Gue___________________________ has quit [Quit: Textual IRC Client: www.textualapp.com]
<zv>
marcan: yeah we'll fix the 4/4 tomorrow. I think baptiste is pulling from /proc/cpuinfo and not getting the weird output format for this big/little configuration
<Glanzmann>
p
<zv>
this weekend or early next week is fine, just be in touch please.
<zv>
my goal is to help y'all as much as possible
<matthewayers[m]>
Glanzmann: I was able to get everything up and running with the installer, but just out of curiosity, do you have a list of the packages you included with Debian available anywhere? There are some packages on there I was really surprised to see and I’m not sure if they were part of the install source files or if a desktop environment decided to give them to me.
<Glanzmann>
I included some extra packages in order to be able to fix problems if anything does not go smoothly.
<Glanzmann>
Oh and firmware-linux.
<matthewayers[m]>
Glanzmann: Got it, thanks! Also, do you know how the system would've gained Wordpress themes? I looked through the packages earlier and noticed that Wordpress was installed. I don't see it in the file you sent but thought I'd ask to be sure.
<Glanzmann>
matthewayers[m]: No, I don't install wordpress.
<matthewayers[m]>
I figured. I'm guessing KDE or one of its dependencies decided to do that then.
<matthewayers[m]>
* I figured, just wanted to be sure since it was a bit odd to see. I'm
the_lanetly_052 has joined #asahi
<j`ey>
zv: can you link to the CI/builder/whatever page?
ChaosPrincess has quit [Ping timeout: 480 seconds]
<Glanzmann>
j`ey: I found it, you do a whois on zv than there is hostname you go there.
<marcan>
zv: also fwiw, apparently Amazon has servos pressing the power switch for their Mac Minis in the cloud so... yeah :-)
<marcan>
not the most remote management friendly machines
<marcan>
though I hope to work out a way to handle the LOM features of the 10G model with FOSS in the future
<marcan>
oh, that was already mentioned, sorry
<marcan>
also, if you have a PDU/switch and you want to be ultra safe, you could blacklist the macsmc-reboot module (which you might not even have in that kernel); reboot will still work using the watchdog driver as a fallback, but clean shutdown will be impossible at least from linux, which I guess is a good thing for you :)
<marcan>
that module also lets you (when loaded) configure power restore on AC loss via sysfs, so you could set it once then blacklist it
<matthewayers[m]>
<marcan> "zv: also fwiw, apparently Amazon..." <- Sounds like they became an enterprise customer for Switchbot :)
<VinDuv>
IIRC macvdmtool reboot is able to start up a MBP from normal$ shutdown
<Glanzmann>
I'm quiet excited what amazon does with the m1 mini, it is in beta now but probably will be released soon.
<Glanzmann>
How they manage to boot it.
<VinDuv>
not sure if it works with the mini though
<jannau>
it doesn't work on the mini
<VinDuv>
aw
<marcan>
I've gotten it to work on the mini in the past, but there seem to be different kinds of shutdown state; it was a long time ago and I don't remember exactly how that worked
<marcan>
if it can't be made to work from a shutdown state coming from cold AC power up though, it's not very useful unfortunately
<Glanzmann>
marcan: Which file controls the power loss action?
<marcan>
I'm not booted but just grep /sys for macsmc-reboot, it has an obvious name
<zorun>
thanks for your work on Asahi, and thanks zv for connecting the two projects!
<Glanzmann>
Nice. :-)
<Glanzmann>
zn / zorun: Btw. you're currently using a 4k page size kernel, if you're using it mainly for compiles, maybe you should switch to a 16k kernel.
<marcan>
zv was asking about our release to coordinate with the announcement :')
<Glanzmann>
marcan: We need to release right now.
<marcan>
uh... no :p
<zorun>
it's most an internal announcement to our users (although we do publish it on the website)
<zorun>
we were thinking of synchronizing for a larger announcement, talk to LWN, etc
<zorun>
mostly*
<zorun>
Glanzmann: ah, interesting. we probably need more hardware to provide several combinations of kernel parameters and userspace ;)
<marcan>
makes sense!
<zorun>
if I understand correctly, 16K is the default for the hardware, but 4K is the default for the kernel on aarch64?
<zorun>
ideally we provide what will end up being the "standard" configuration for people using the same hardware / software
chengsun has joined #asahi
<zorun>
but providing a configuration that exposes bugs is also interesting :)
<j`ey>
4k is what distributions default to, since its the page size on x86 and what most Arm CPUs support
<sven>
there’s a decent chance that internal announcement will end up on hacker news anyway ;)
<j`ey>
The reason we've used 16K for a while is that linux doesnt support a page size mismatch betwwen CPU and IOMMU
<zorun>
ok, sorry if it puts pressure on you then ^^ we were quite excited to have it up and running without major issues
<sven>
nah, no worries :-)
<sven>
I’m happy about everyone who uses our work
<marcan>
in the end we're going to offer both options and other distros probably should too
<espo>
morning :)
chengsun_ has quit [Ping timeout: 480 seconds]
<matthewayers[m]>
Morning :) I haven't slept since Wednesday night (US Eastern Time)
<marcan>
zorun: IIRC someone benchmarked compiles at 20% faster on 16K, which is kind of ridiculous and deserves further research
<marcan>
OTOH there is software out there that does silly things like assume compile time page size == runtime page size
<marcan>
both have merits either way; 16K is always going to be faster to some extent even if we figure out a solution for that large gap, and 4K is the only way to e.g. emulate x86 software sanely
<arnd>
marcan: I did those benchmarks last year and mentioned the numbers here. I don't remember if it was 20% or a bit less, but it was definitely more than I had expected
<maz>
marcan: it is the sort of speedup you'd see if 4kB resulted in only using a quarter of the cache...
<marcan>
ha
<marcan>
that is... a good point
<marcan>
but would apple really do that?
<marcan>
for rosetta?
<marcan>
I don't buy it
<arnd>
maz: when I did microbenchmarks, everything seemed to indicate that it was using the full 128KB of L1 and it had hte same latency regardless of page size
<marcan>
it is of course possible that rosetta pokes some other magic sysregs to make it not suck at 4K
<arnd>
but it's possible that there is something else going on that makes the microbenchmarks look better than real-world workloads
<mps>
arnd: I benchmarked few times build kernel and some other software
<mps>
with 4K it is around 20-25% slower
<_jannau_>
I saw a difference of 10% between 4k and 16k with kcbench (linux 5.15 defconfig) on the m1 max
<mps>
build some other programs it is about 15% slower
<arnd>
I think when I tested the same on EC2, the 16KB numbers were not quite as good. IIRC something like 18% speedup for 64K pages and 15% for 16KB pages
<marcan>
ah, so it's not just apple
<mps>
so all in all it varies between 10 and 25 percents
<marcan>
this seems pretty extreme though
<arnd>
I may also be misremembering things
<_jannau_>
and I think axboe saw around 15% difference
<marcan>
like either there is some big cache difference, or TLB thrashing really is a lot worse with 4K
<arnd>
the flip side is that a lot of hardware just doesn't support 16KB, and the extra memory overhead for 64KB pages is crazy.
<arnd>
j`ey: ah, thanks for digging that out
<maz>
I remember running axboe's io_uring test, and consistently getting a 20% difference between a 4k guest and a 16k guest, the host system using 16k.
<arnd>
ok, so only 10% improvement with 16KB pages, or 12% with 64KB pages
<arnd>
and 16% on M1, which can be a combination of the differences in L1 cache behavior and TLB layout between M1 and N1
<arnd>
There is likely also a difference between VA_BITS_36/39/47/48
<arnd>
the default for both is 3-level tables (39 and 47), but comparing two-level 16KB tables (36 bits) against four-level 4K tables (48 bits) would skew this even further towards the 16KB pages
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<j`ey>
matthewayers[m]: lets move it here, that's more for dev stuff
<j`ey>
matthewayers[m]: sometimes it comes up, sometimes it doesn't. if you type 'ip a' into a terminal, can you see wlan0 or the enlsplslp0 whatever interface
darkapex has joined #asahi
darkapex has quit [Remote host closed the connection]
darkapex has joined #asahi
kajiryoji has quit []
kajiryoji has joined #asahi
amarioguy has joined #asahi
kajiryoji has quit []
kajiryoji has joined #asahi
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ChaosPrincess has quit [Remote host closed the connection]
amarioguy has quit [Ping timeout: 480 seconds]
* kov
saw the discussion in -dev, a few comments:
<kov>
I've tested both Debian and Fedora, both are working quite well, I have only very very very rarely seen the WiFi issue, I have brcmfmac compiled as a module, both on Debian and Fedora with Gnome
<kov>
Gnome had 2 big issues since I started using it on M1, first on a VM, then on actual hardware; there was a crash on startup on a gnome-shell component caused by code that only worked on amd64, I fixed that months ago, and then anything using webkitgtk would crash with the 16k kernel (evolution, builder, online accounts, etc), that has been fixed as well
<opticron>
I've never seen the wifi issue on debian on the 14" m1m
<kov>
yesterday I did a full reinstall of fedora on my Air, no issues found so far, both of my fixes made it to fedora 35 already, so all reasonably recent distributions and DEs should work out of the box with asahi
<kov>
oh, all of that using the default for fedora which is NetworkManager, I have tested offline upgrades which use the UEFI infrastructure, kernel upgrades etc, everything seems to be working
<kov>
for debian I'd recommend using at least testing, probably sid would be a better bet tbh, as it is more likely to have fixes such as the webkit one in the short term
<agraf>
arnd: wrt <arnd> the flip side is that a lot of hardware just doesn't support 16KB
<agraf>
arnd: How hard would it be to have 4kb granule fallback code in Linux if PAGE_SIZE is 16kb?
<agraf>
arnd: so that you could build a kernel that has PAGE_SIZE=16kb, but still runs with 4kb granule configs
<j`ey>
agraf: unrelated but parallels hit the same issue as qemu on the ID_AA64ISAR2 register issue :D
<arnd>
agraf: not worth it, the mm code is built around knowing the page table layout at compile time, trying to make it a runtime thing would add a ton of complexity and runtime overhead where we don't want it
<arnd>
the leaf pages would be fine, but the higher-up page tables would not
<agraf>
j`ey: Hah :)
<agraf>
arnd: I wasn't thinking of actually making anything a runtime thing. You would basically create alternatives for the fault handler and PTE modification code. I guess a lot of that is super deeply engrained on arm64 though :/
<agraf>
arnd: So my thinking was that the Linux internal page tables would still be the 16kb format basically
amarioguy has joined #asahi
<agraf>
arnd: and then you would do alternatives for pte removal/modification and tlb invalidations. I agree that it's probably going to become very messy very quickly and is probably not worth it
<agraf>
Especially since 16kb doesn't give you any functional benefit. It's not like you get more PA bits like with 64kb
<arnd>
yes, that was exactly my point: shadow page tables are really expensive to do
<agraf>
Point taken :)
<arnd>
you could probably do it on powerpc (with hashed tables) or on loongarch (with software tlb), but not on architectures where Linux accesses the hardware page tables directly
<Glanzmann>
kov: Did I understand that correctly, you submitted patches for fedora for the 16k issues that you found and they applied them?
<_jannau_>
no, for gnome and webkitgtk, patches were submitted to upstream projects and fixes are in fedora 35
<Glanzmann>
jannau: I see.
<Glanzmann>
kov: Can you point me to this fixes, than I open a bug with debian to pull them in.
yuyichao_ has quit [Ping timeout: 480 seconds]
joske has quit [Remote host closed the connection]
nico_32 has quit [Remote host closed the connection]
nico_32 has joined #asahi
ChaosPrincess has joined #asahi
yuyichao_ has joined #asahi
bisko has quit [Read error: Connection reset by peer]
<kov>
Glanzmann, that is already on the 2.34.6 stable release
<kov>
the gnome one was unrelated and is from a bunch of months ago, I have not seen any crashes in my testing with debian's gnome so I assume the fix is already in
<matthewayers[m]>
Please excuse my primitive pastebin
<j`ey>
so does ping work?
<j`ey>
you have an IP it seems
<matthewayers[m]>
Ping fails and when I try to curl something like Google it fails to connect to server
<j`ey>
what about pinging your routers IP?
<tohatsu[m]>
Hi guys. I want to boot Debian from USB on Macbook M1. Is it possible to do without internal linux partition?
<ivabus[m]>
I think it is not possible, because iBoot can't boot from external devices, so you will need at least bootloader partition
<jannau>
tohatsu[m]: depends what you understand as internal linux partition. the devices bootloader can not boot from external devices and need a 2.5G stub partition
<jannau>
the linux data partition (and linux boot loader partiton) can be on the usb device
<matthewayers[m]>
j`ey: I’ll try that in about an hour or so when I get out of class. Thanks for your help so far!
<tohatsu[m]>
ivabus: jannau thank you. Debian on USB stick persistent?
<tohatsu[m]>
from tutorial on GitHub
<mps>
jannau: isn't it possible to install m1n1+dtbs+u-boot on the macos, without touching internal disk
<mps>
I did something like this about year ago when tested corellium
<sven>
it’s still going to write that to the internal disk
<tohatsu[m]>
mps: that's why i was asking. corellium tutorial doesn't mention to partition internal disk
<jannau>
how does that not touch the internal ssd? Macos doesn't boot after that anymore and you have to go through 1tr to switch between linux and macos
<mps>
jannau: yes, that's true
leo60228 has quit [Ping timeout: 480 seconds]
<mps>
I think tohatsu[m] asked to not change disk partition
leo60228 has joined #asahi
<jannau>
I think so too but I suspect the goal behind that is to not touch the macos installation
<tohatsu[m]>
not to change disk partition and be able to boot from USB on another M1 device without problems
<mps>
tohatsu[m]: I think this is possible but I didn't tested for one year if there are some issues after I tested this
<jannau>
tohatsu[m]: every apple silicon device has to be prepared to use linux from the usb stick. you cann not prepare it on one device and than boot it on another one
<mps>
or I don't understand question
<tohatsu[m]>
the goal is to be able to boot from USB stick on any M1 device by entering 1tr and switching boot target manually
<mps>
tohatsu[m]: if you install m1n1+dtbs+u-boot on macos partition (i.e. not creating separate stub for linux) you can not then select boot device by simply entering 1TR
<bluetail[m]>
live boot stick? would be great as one could pre-install in persistence but never actually install a OS
yuyichao_ has quit [Quit: Konversation terminated!]
<tohatsu[m]>
mps: how do you choose then?
yuyichao has joined #asahi
<jannau>
I don't think this a practical solution if you intend to use te macos installation
george has joined #asahi
<kov>
tohatsu[m], mps don't think that's accurate, you'd have to enter the options, open a terminal and effectively remove the custom kernel (m1n1+uboot+...)
george is now known as Guest1894
<jannau>
you have to use 1tr and switch the security state every time you want to switch back to macos
<tohatsu[m]>
jannau: bputil -n for Mac Os and bash script to load macho for Linux ?
<kov>
just set up the stub partition tbh, it'll be safer and more practical
<jannau>
for installing m1n1+uboot one could use the same autorun installer trick we now use for the asahi installer
<kov>
then you just choose which volume to boot from instead of having to go into 1tr and opening a terminal
<Guest1894>
Does anyone know when the alpha installer is likely to be released? Saw on twitter that it would be soon
<jannau>
but you still have to enter the admin password every time you want switch between macos and linux (and back)
<j`ey>
tohatsu[m]: it's a lot of work to just not resize the internal disk by -2.5G :P
<jannau>
Guest1894: no, there a still a few release blocking things
creechy has joined #asahi
<Glanzmann>
matthewayers[m]: ping does not work as user on Debian, can you try as root. The reason is that the kernel is missing a feature.
<jannau>
it's not far off though, let's hope it's before marcan get's their m1 ultra mac studio
Guest1894 has quit []
<j`ey>
marcan: I learnt a new git command just now: git range-diff asahi/asahi-soc/prev..asahi asahi/asahi-soc/next..asahi/asahi good way to compare the new/old branches even after a rebase
<j`ey>
(where 'asahi' is my local not-yet-updated branch)
creechy has quit [Quit: WeeChat 3.4]
creechy has joined #asahi
<bluetail[m]>
didnt git change to recommending gh instead of git?
<george147>
Right, and by the looks of it doesn't have anywhere near as much hardware support
<j`ey>
they did an initial quick port, and abandoned it
<jannau>
one time quick publicity stunt of a company, obsolote by now
george147 has quit []
creechy has joined #asahi
creechy has quit []
creechy has joined #asahi
creechy has quit []
creechy has joined #asahi
arvind has joined #asahi
<arvind>
Hey is this the asahi linux group working on porting m1 to linux?
<j`ey>
yes
creechy has quit []
creechy has joined #asahi
<arvind>
Great. This might sound obvious but do you need a m1 in hand to work on asahi?
<j`ey>
probably
<tohatsu[m]>
big sur 11.6.3 won't boot Debian USB?
<tohatsu[m]>
does OS version even matter if boot with 1TR ?
<tpw_rules>
yes, you have to upgrade to 12.1
<tpw_rules>
or later
<Glanzmann>
marcan / j`ey: Any new additional features in the new asahi rebase?
<jannau>
performance counters
<marcan>
and a breaking change to aic2/t6000
<marcan>
needs dt update
<j`ey>
and a few of the s2idle fixes
<Glanzmann>
I see. I rebased the 4k patch (it had a small trival reject): tg.st/u/0001-4k-iommu-patch-2022-03-11.patch
<Glanzmann>
Well than I'll try this tree with the 'debian distro config' and see if it works for me.
<limegorilla[m]>
Hi everyone! Not too sure where to start, and I’m probably not skilled enough to do dev work, but I’m looking to help out where I can. I’m pretty good at diagnosing, running POC where the exact steps are unclear and clearly writing documentation for end users.
<limegorilla[m]>
Is there anything I could work on?
amarioguy has quit [Ping timeout: 480 seconds]
roxfan has quit [Remote host closed the connection]
roxfan has joined #asahi
<j`ey>
limegorilla[m]: wait for the alpha release and install and use that is the best thing probably
amarioguy has joined #asahi
<Glanzmann>
povik / marcan: For the release will the sound part updated that fixes this bug? apple-pmgr-pwrstate 23b700000.power-management:power-controller@2c0: PS mca1: Failed to reach power state 0xf (now: 0x24f) dmesg here: https://pbot.rmdir.de/ubs2XDJQNhhtPwDOmx4lSw
<Glanzmann>
povik: Or is there commit that I can cherrypick into the asahi branch that gets rid of it?
<matthewayers[m]>
I finally got Internet back up on my system. Lots of reconfiguring wpa_supplicant.conf
<milek7>
hmm, I did get crash on efi_create_mapping
<j`ey>
milek7: expected, just a warning
<mps>
tohatsu[m]: sorry I was afk, but you got explanation from good people here. hope you know all about your question now
marvin24 has joined #asahi
<milek7>
it also once hanged on boot, until watchdog rebooted it
<milek7>
'soft lockup cpu5' or something like that
marvin24_ has quit [Ping timeout: 480 seconds]
lennylxx[m] has joined #asahi
<lennylxx[m]>
just followed the guide to install Arch Linux, how am I supposed to turn on wifi?
<Glanzmann>
sven: When I build the dwc3 as modules, it does not detect for me usb. Is this is a known issue or am I doing something wrong?
<Glanzmann>
sven: Config: tg.st/u/config-debian-distro-kernel-2022-03-09-4k dmesg: https://pbot.rmdir.de/ubs2XDJQNhhtPwDOmx4lSw I also noticed that dwc3-pci kernel module was not loaded.
<sven>
no need for pci, dwc3 isn’t on that bus
<sven>
I haven’t tried it as a module but I believe it worked for marcan
<Glanzmann>
sven: Okay than I'll take my known good config and try to load it there as a module and report back.
<sven>
it doesn’t seem to probe for some reason
<Glanzmann>
Yep. This is the debian config with the asahi options enabled version.
<Glanzmann>
But I did not force things to be =y if a dependency is =m
<Glanzmann>
But I'll try if I can reprocue it with a known config just by setting dwc3 to m.
<marcan>
lennylxx[m]: wifi is broken, this is all alpha, don't expect anything to work, etc
<marcan>
no promises until the first official release
<marcan>
(which is coming soon, but not yet)
<milek7>
lennylxx[m]: install wpa_supplicant, iw, libnl, and then you can connect using iwctl (look at arch wiki for that)
<jannau>
would not expect a rebase to work. it's not 5.17-rc7 but linux-next-20220310. rebase should hvae massive problems since both linux-next and asahi are recreated. you will end up with tons of commits you're not really interested in but git chould not detect as unchanged