marcan 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-stream #asahi-offtopic | Keep things on topic | Logs: https://alx.sh/l/asahi
nsklaus has quit [Quit: WeeChat 3.3]
<RealityVoid> I used this to apply all patches, seems to work well-ish
<RealityVoid> for i in ../AsahiLinux-PKGBUILD/linux-apple/*.patch; do echo Applying patch from $i && patch -p1 <$i; done
klaus has joined #asahi
klaus has quit []
<RealityVoid> It seems to build and load.
<RealityVoid> Is that debian installer usable for you guys when using it from the pycon? Just a pin undersatanding what it's doing.
<RealityVoid> Let's say I want to get the keyboard running, how would one go about that?
<RealityVoid> Like.. the proper way. Something that could be mainlined or something.
nsklaus has joined #asahi
kenzie has quit [Quit: The Lounge - https://thelounge.chat]
_chamomile has quit [Ping timeout: 480 seconds]
chamomile has joined #asahi
yuyichao_ has joined #asahi
<cth451_desktop> I'm not using Debian (and thus not using the installer)
<cth451_desktop> I have provisioned a root filesystem elsewhere and dumped that to a thumb drive
aquijoule_ has joined #asahi
<cth451_desktop> The goal was to boot from iSCSI (which I had originally planned for 10GE) but alas
richbridger has quit [Ping timeout: 480 seconds]
psykose has quit [Remote host closed the connection]
psykose has joined #asahi
PhilippvK has joined #asahi
phiologe has quit [Ping timeout: 480 seconds]
kov has quit [Quit: Coyote finally caught me]
palmer_ has joined #asahi
palmer_ has quit [Remote host closed the connection]
marvin24_ has joined #asahi
marvin24 has quit [Ping timeout: 480 seconds]
phire_ has joined #asahi
phire is now known as Guest6421
phire_ is now known as phire
Guest6421 has quit [Ping timeout: 480 seconds]
Glanzmann has joined #asahi
<Glanzmann> RealityVoid: I used the following procedure to install Debian based on what allyssa did. https://ab34.de/u/asahi.txt
<Glanzmann> That is on the mini. I than installed xorg, firefox, fvwm2, you name it and was able to work from it for a week.
<Glanzmann> Of course you could also boot via iscsi or nvme-over-tcp.
Dcow has joined #asahi
Dcow has quit [Ping timeout: 480 seconds]
chamomile has quit [Ping timeout: 480 seconds]
chamomile has joined #asahi
Dcow has joined #asahi
Dcow has quit [Ping timeout: 480 seconds]
chamomile has quit [Ping timeout: 480 seconds]
tertu has quit [Ping timeout: 480 seconds]
tertu has joined #asahi
aquijoule_ is now known as richbridger
<jannau> cth451_desktop: please copy the m1n1 output and the kernel boot log to a pastebin
<jannau> there is a slim chance that https://github.com/AsahiLinux/m1n1/pull/125 helps
richbridger has quit []
richbridger has joined #asahi
richbridger has quit [Remote host closed the connection]
richbridger has joined #asahi
Dcow has joined #asahi
<Glanzmann> Does someone have a tree that works ond a macbook air including keyboard, wifi and framebuffer x?
inertia has quit []
<cth451_desktop> jannau: m1n1 output on the chainloader with PR125 applied - https://pastebin.aosc.io/paste/jWSM9QL1vQDwJjpB2Ol2TA
<j_ey> [ 1.307838] systemd[1]: Listening on Journal Audit Socket.
<j_ey> err
<j_ey> this pastebin is terrible for selecting text..
<j_ey> [ 2.090276] r8152 3-1:1.0: Direct firmware load for rtl_nic/rtl8153a-3.fw failed with error -2
<cth451_desktop> Yeah this is an ethernet dongle on USB-C
<cth451_desktop> The firmware wasn't added to the initrd
<cth451_desktop> but this should be independent from the 10GE issue
richbridger has quit [Remote host closed the connection]
<cth451_desktop> Not entirely sure whether this is related, but for some reason the NVMe SSD itself is not present under lspci
<marcan> it's not PCI
<jannau> the problem is pcie-apple 690000000.pcie: /soc/pcie@690000000/pci@2,0 link didn't come up
<jannau> the device config seem readable later though. I suspect the pci init is not sufficient for the 10ge variant
<marcan> do we have an ADT from that variant?
<marcan> I'm wondering how link bonding works
<marcan> I assume the 10ge variant has more than a 1x pcie link
<marcan> I think all the other devices are 1x?
<marcan> then again on the M1 Mini with 3 physical ports, how would they do that?
<marcan> I guess 10ge would just about saturate a PCIe 3.0 link so maybe they do only need one lane
<mps> is tree https://github.com/jannau/linux/tree/apple-m1-exp_v5.16-rc1 ok starting point for kernel, and add keyboard patches on it?
<j_ey> yep
<mps> j_ey: thanks
<j_ey> (most of the patches in the apple-m1-exp_v5.16-rc1 tree are for the pro/max, but dont hurt anyway)
<jannau> marcan: the j274 template ADT has lan-10g and some "*-10g" properties in pci-bridge2
<jannau> it has maximum-link-speed = 4 so I assume it uses 4 PCIe 3 links
<marcan> I think you're confusing link speed with link width
<marcan> maximum-link-speed = 4 probably means PCIe Gen4
<marcan> I'm pretty sure t8103 only has 3 lanes total and t600x only has 4 lanes total
<marcan> (which also means the 16 bound in your patch isn't quite right :))
<marcan> yeah, definitely only 3 lanes on t8103
<jannau> makes sense
<marcan> so all links are 1x on t8103 and t600x (at least on current devices; not sure if they might be able to bond links into one root port, maybe? but not on current devices)
<marcan> it's entirely possible that Gen4 needs more initialization than we currently do
<marcan> OTOH, I would *expect* both sides to fall back to an older gen if things go wrong during link training
<marcan> so it's a bit weird that it wouldn't work at all
kov has joined #asahi
aleasto has joined #asahi
<kettenis> function-pcie_port_control = 49:PrtC(0x3b)
<kettenis> that is present on lan-10gb but not on lan-1gb
gabuscus has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
gabuscus has joined #asahi
<jannau> the pcie usb controller has "function-pcie_port_control = 49:PrtC(0x35)" and works out of the box
<kettenis> true
Glanzmann has quit [Quit: EOF]
gladiac is now known as Guest6447
gladiac has joined #asahi
gladiac has quit []
gladiac has joined #asahi
Guest6447 has quit [Ping timeout: 480 seconds]
gladiac is now known as Guest6448
gladiac has joined #asahi
Guest6448 has quit [Ping timeout: 480 seconds]
aleasto has quit [Remote host closed the connection]
gladiac is now known as Guest6452
gladiac has joined #asahi
Guest6452 has quit [Ping timeout: 480 seconds]
nsklaus has quit [Quit: WeeChat 3.3]
klaus has joined #asahi
Dcow has quit [Quit: My Mac Mini has gone to sleep. ZZZzzz…]
<mps> now that I got boot macbook and can see nvme disk with partitions, how and where I can install linux root FS on it
<mps> on nvme0n1p5 which is vfat and size is 198GB?
<marcan> jannau: pull/125 does nothing because Linux currently ignores the max speed prop, see -dev :-)
<marcan> but if the link isn't up then it isn't up
<marcan> ISTR maz mentioned something about PERSTN?
<marcan> ah yes, that one
<marcan> you can try this to fix the link speed problem, but I doubt that will make it work (it should still negotiate at Gen1): https://github.com/AsahiLinux/linux/commit/35ee181fdfdea4c5c87b2365c6f3de707ddef745
<jannau> the adt even has "t-refclk-to-perst" and "perst-to-config" both set to 100
<jannau> cth451_desktop: ^^^
<jannau> I don't have a 10g ethernet mac mini
<marcan> there is an i2c device associated with the 10ge
<marcan> aqc-a on i2c3
<marcan> this has to do with the BMC type functionality that is supposedly supported on these macs maybe?
<marcan> hm, the datasheet says Gen2/Gen3
<marcan> so maybe indeed it does refuse to train at Gen1? if so try my patch
<jannau> maybe it seems to have an additional mac address, whatever "extlom" is
<marcan> for the BMC
<marcan> lom = "lan on motherboard"
<marcan> this is server terminology
<marcan> the second MAC would be for the BMC
<marcan> oh sorry, "lights out management"
<marcan> silly overloaded terms
<kettenis> scary
<marcan> supposedly it talks IPv6/TLS using the onboard CPU and can turn the computer on and off
<marcan> presumably it is provisioned via that i2c link
<kettenis> so maybe they did the sane thing and disabled it by default?
<kettenis> (I firmly believe that only LOM/BMC solution that use a dedicated network port are acceptable)
<marcan> apparently it even does mDNS, heh
<jn> oh, dreaded NC/SI...
<marcan> yes, it does that too
eichin has quit [Remote host closed the connection]
jkkm has quit [Remote host closed the connection]
eric_engestrom has quit [Remote host closed the connection]
cptcobalt has quit [Remote host closed the connection]
nkaretnikov has quit [Remote host closed the connection]
steev has quit [Remote host closed the connection]
Vaughn has quit [Remote host closed the connection]
jabashque has quit [Remote host closed the connection]
sorear has quit [Remote host closed the connection]
brinly has quit [Remote host closed the connection]
taowa has quit [Remote host closed the connection]
stblassitude has quit [Remote host closed the connection]
austriancoder has quit [Ping timeout: 480 seconds]
esden has quit [Ping timeout: 480 seconds]
philpax has quit [Ping timeout: 480 seconds]
tom-w has quit [Ping timeout: 480 seconds]
robher has quit [Ping timeout: 480 seconds]
kendfinger has quit [Ping timeout: 480 seconds]
Chainsaw has quit [Ping timeout: 480 seconds]
Ariadne has quit [Ping timeout: 480 seconds]
ZLSA has quit [Ping timeout: 480 seconds]
sjg1 has quit [Ping timeout: 480 seconds]
rann has quit [Ping timeout: 480 seconds]
ovf has quit [Ping timeout: 480 seconds]
arnd_ has quit [Ping timeout: 480 seconds]
daniels has quit [Ping timeout: 480 seconds]
maennich has quit [Ping timeout: 480 seconds]
Dcow has joined #asahi
sorear has joined #asahi
stblassitude has joined #asahi
saintdev_ has joined #asahi
taowa has joined #asahi
tardyp has joined #asahi
brinly has joined #asahi
enomem has joined #asahi
guan has joined #asahi
NekomimiScience has joined #asahi
mustache_ride___ has joined #asahi
aeroraptor has joined #asahi
jabashque has joined #asahi
ovf has joined #asahi
arnd_ has joined #asahi
kendfinger has joined #asahi
esden has joined #asahi
robher has joined #asahi
sjg1 has joined #asahi
daniels has joined #asahi
steev has joined #asahi
rann has joined #asahi
eichin has joined #asahi
Vaughn has joined #asahi
cptcobalt has joined #asahi
jkkm has joined #asahi
maennich has joined #asahi
nkaretnikov has joined #asahi
Chainsaw has joined #asahi
philpax has joined #asahi
Ariadne has joined #asahi
ZLSA has joined #asahi
eric_engestrom has joined #asahi
austriancoder has joined #asahi
<jannau> sigh, I missed that j293 has i2c4
tom-w has joined #asahi
arahael3 has quit [Ping timeout: 480 seconds]
<sven> i2c should also be a separate commit from cd321x imho
<kettenis> also some of the diffs had v2 in the subject and others didn't...
<kettenis> sven: I see no trace of the endpoint number for the SMC in the ADT
<kettenis> so I assume the macos driver just knows the endpoint number?
maor26 has joined #asahi
thunfisch has quit [Remote host closed the connection]
thunfisch has joined #asahi
chamomile has joined #asahi
<sven> kettenis: yes. afaict most endpoint numbers are hardcoded inside the drivers
thunfisch has quit [Remote host closed the connection]
thunfisch has joined #asahi
thunfisch is now known as Guest6468
thunfisch has joined #asahi
Guest6468 has quit [Ping timeout: 480 seconds]
<mps> is it possible to kernel cmdline when creating chainload image? I saw that linux.py have this option (-b) but can't find is it possible for chainload
<mps> s/to kernel/to add kernel/
thunfisch has quit [Remote host closed the connection]
thunfisch has joined #asahi
<j_ey> cat m1n1.macho <(echo 'boot-args=here are boot args') linux.gz dtb.dtb etc
<mps> j_ey: many thanks
<mps> could I repeat my question about nvme partition on which to install (actually copy from usb) linux root FS
<mps> looks like it is nvme0n1p5 in my case, but I'm not sure
chamomile has quit [Ping timeout: 480 seconds]
<mps> it is vfat fs 'formated' and I think it should be 'reformated' as ext4, though not sure will that make a mess for macos
chamomile has joined #asahi
<kettenis> mps: nobody can answer that question with just that snippet of information
gladiac is now known as Guest6469
gladiac has joined #asahi
<kettenis> all I can say that a normal macos installation does not have any vfat partitions
<mps> kettenis: ok, so I have to risk and see results :)
<kettenis> ok, I've got u-boot working with something that resembles sven's mailbox/ANS2 bindings
<mps> kettenis: I followed quickstart and created these two vfat partitions, one for /boot and one for /root, iiuc
<kettenis> then you should have one small and one large vfat partition
<mps> yes, that is
Guest6469 has quit [Ping timeout: 480 seconds]
<mps> I'll risk after I finish with kernel upgrade
<cth451_desktop> jannau: The ethernet card came up with the two patches applied
<cth451_desktop> ("Configure link speed properly" and "Reset the port for 100ms on probe")
<kettenis> if nvme0n1p4 is your small vfat partition and nvme0n1p5 is the large vfat partition then nvme0n1p5 is almost certainly your root partition
<cth451_desktop> Currentlly operating at GE speed. I need to move the mini into the rack for 10G connectivity and will try that later.
<cth451_desktop> 03:00.0 Ethernet controller: Aquantia Corp. Device 04c0 (rev 03)
<mps> kettenis: that is also my thinking
<sven> i think the M1 actually has ANS3 fwiw: +-o AppleANS3NVMeController <class AppleANS3NVMeController, id 0x1000003c2, registered, matched, active, busy 0 (159 ms), retain 20>
<kettenis> ADT says ANS2 in several places ;)
<sven> :D
<sven> ANS2.5 then i guess
<jannau> sart-ans is version 3
kenzie has joined #asahi
<sven> yeah, but that's different
Dcow_ has joined #asahi
<cth451_desktop> jannau: but the mac address seems unstable and keeps changing across reboots
<kettenis> the driver probably needs code to read the mac address from the device tree
<cth451_desktop> Yeah I'll hardcode a MAC address in NetworkManager for now
Dcow has quit [Ping timeout: 480 seconds]
kenzie has quit [Quit: The Lounge - https://thelounge.chat]
kenzie has joined #asahi
<cth451_desktop> [ 1.128309] atlantic 0000:03:00.0 (unnamed net_device) (uninitialized): MAC is invalid, will use random.
<mps> kettenis: installing root FS on this internal partition works
X-Scale has joined #asahi
<cth451_desktop> The device tree has the right MAC address populated so kettenis you could be right that the driver isn't looking at dt
X-Scale` has quit [Ping timeout: 480 seconds]
<lovesegfault> What's the status of the M1 Pro work?
<j_ey> pretty much on par with the original M1s
<lovesegfault> Nice, I'm tempted to get one to play with
<jannau> support currently only in the asahilinux tree and not mainline linux though
<matthewayers[m]> I have a M1 Max as my daily driver but I am eager to test some stuff out after I finish finals next month
<lovesegfault> I imagine the lack of wifi is the main drag right now
<j_ey> wifi works
<lovesegfault> Oh!
<lovesegfault> I thought it didn't work on the pro yet, that's awesome
<j_ey> Dunno if anyone has tried on the pro actually, jannau?
<j_ey> pro/max
<jannau> I haven't tried wifi on the m1 max yet
kenzie has quit [Quit: The Lounge - https://thelounge.chat]
kenzie has joined #asahi
X-Scale` has joined #asahi
<jevinskie[m]> marcan: as to why I want to boot xnu without the proxy: setting DT entries like debug-enabled is handy for macOS research and it would be nice to be able to use a custom DT config untethered from the proxy loader
X-Scale has quit [Ping timeout: 480 seconds]
<jevinskie[m]> I’m making progress on it :)
kenzie has quit [Quit: The Lounge - https://thelounge.chat]
<marcan> I don't think anyone has tried wifi on the pro, but it should work with that horrible patch...
<marcan> I just really don't want to rely on it so I'm using that excuse to make me rewrite it
<marcan> ... as soon as I send the PMGR series :p
<marcan> cth451_desktop: glad to hear it works :)
<cth451_desktop> Quickly devising a dirty patch to nudge atlantic to read MAC from dt if hw reports an invalid one
kenzie has joined #asahi
kenzie has quit []
kenzie has joined #asahi
RealityVoid has quit [Remote host closed the connection]
RealityVoid has joined #asahi
yrlf has quit [Quit: The Lounge - https://thelounge.chat]
<cth451_desktop> The patch to make 10GE card respect the MAC address supplied by DT https://pastebin.aosc.io/paste/TXJ-Rn8GhOBN8548XDQb0w
yrlf has joined #asahi
maor26 has quit [Ping timeout: 480 seconds]
palmer_ has joined #asahi
palmer_ has quit [Remote host closed the connection]