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
DarkShadow44 has quit [Ping timeout: 480 seconds]
<jevinskie[m]> Merry Christmas! I made a thing with m1n1: standalone xnu kernel loading with builtin DT patching. Working on adding APFS support to kettenis’s u-boot so you don’t have to concatenate the XNU kernel to m1n1. :) https://mobile.twitter.com/jevinskie/status/1474533162608996352
DarkShadow44 has joined #asahi
tpw_rules has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nobodynada has joined #asahi
aleasto has quit [Quit: Konversation terminated!]
nobodynada has quit [Ping timeout: 480 seconds]
Techcable has quit [Quit: ZNC - https://znc.in]
Techcable has joined #asahi
cth451_desktop has quit [Quit: Quitting]
Emantor has quit [Quit: ZNC - http://znc.in]
Emantor has joined #asahi
DreamOtter has joined #asahi
<DreamOtter> Howdy
tpw_rules has joined #asahi
yuyichao has joined #asahi
phiologe has quit [Ping timeout: 480 seconds]
DreamOtter has quit [Ping timeout: 480 seconds]
marvin24 has joined #asahi
marvin24_ has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
nobodynada has joined #asahi
DreamOtter has joined #asahi
<Glanzmann> NightRaven[m]: I updated the guide to include the right config. Also there is now a faster way to bootstrap. I incorporate that.
<Glanzmann> Saves you the usb stick roundtrip.
DreamOtter has quit [Ping timeout: 480 seconds]
nobodynada has quit [Ping timeout: 480 seconds]
<Glanzmann> marcan: wifi take3 still does not work on air. Same behaviour. See the device, can assoicate, no traffic. If I find out the reason, I let you know.
<Glanzmann> amw: I tried your kernel config. What software do you use for wifi? wpa_supplicant? Do you have a personal wpa3 hotspot?
nobodynada has joined #asahi
DreamOtter has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
darkapex1 is now known as darkapex
nobodynada has quit [Ping timeout: 480 seconds]
yuyichao has quit [Ping timeout: 480 seconds]
joske has joined #asahi
<joske> Glanzmann: wifi is working perfectly on my air
<Glanzmann> joske / amw / marcan: The issue with my wifi is that I lack the kernel support see: https://pbot.rmdir.de/e5Af1EW6W1ZguU4biXct4A
<Glanzmann> So this was pebkac all long. I'll use the debian default config and add the m1 specific options and hope that than I have finnaly all necessary kernel modules.
<Glanzmann> joske: Btw. I also still have a problem with the touchpad. It is detecting it as mouse. As a result I don't have the middle and right buttons and the scrolling: ttps://pbot.rmdir.de/oMpVkDPLl5qva8jHANfE3A
<Glanzmann> mps: Thank you for pointing me to iwd, that way it was easy to debug. :-)
<joske> Glanzmann: it does not work with your config. I diffed many times with mps' config bit could not find what is the problem. I just used mps' config
<mps> joske: and you changed PAGE_SIZE to 16K I presume
<mps> my current config is here https://tpaste.us/0weE
<joske> mps: his config has 4k pages, maybe that is the problem, I never changed that in his config
<mps> if it is 4K it will not work as jannau pointed yesterday
<mps> I mean, internal wifi will not work
gabuscus has quit [Ping timeout: 480 seconds]
caef^ has quit [Remote host closed the connection]
martinstewart has joined #asahi
gabuscus has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
MajorBiscuit has joined #asahi
martinstewart has quit [Ping timeout: 480 seconds]
<mps> do M1 machines have DesignWare SPI controler/device, anyone knows
<j`ey> it seems custom, but samsung derived
<mps> hm, custom will not be easy to reverse engineered
<j`ey> SPI is already done
<j`ey> SPI is used by the keyboard/trackpad
DreamOtter has quit [Ping timeout: 480 seconds]
<mps> j`ey: yes I know this but thought maybe there are some other devices on SPI bus
DreamOtter has joined #asahi
<j`ey> there's a (8MB?) flash with some of the firmware on spi too
<marcan> and Touch ID, and the display TCON
<marcan> the SPI controller is done, there isn't really anything left to do there
<j`ey> oh right 'mesa'
<j`ey> whats TCON?
<marcan> the display panel controller
<marcan> (on laptops)
<marcan> (and imacs)
<j`ey> (the only thing let to do with SPI is get it upstream :P)
Dcow has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
<Glanzmann> mps: Thank you.
Dcow has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
DreamOtter has quit [Ping timeout: 480 seconds]
<mps> Glanzmann: you are welcome
DreamOtter has joined #asahi
the_lanetly_052 has quit [Remote host closed the connection]
n1c[m] has joined #asahi
aleasto has joined #asahi
off^ has joined #asahi
Dcow has joined #asahi
MajorBiscuit has quit [Ping timeout: 480 seconds]
Dcow has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
MajorBiscuit has joined #asahi
MajorBiscuit has quit []
MajorBiscuit has joined #asahi
aleasto has quit [Quit: Konversation terminated!]
aleasto has joined #asahi
Arahael has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
<sven> yeah, with 4K any pcie dma allocations will just fail
<sven> the patch set actually already works fine, it just needs a bit more work because the upstream IOMMU API has a slightly odd level of abstraction
<sven> and it doesn’t have support for swiotlb for untrusted devices because that api unsurprisingly can’t align buffers to more than PAGE_SIZE
DreamOtter has quit [Ping timeout: 480 seconds]
<sven> but that’s all doable. Just needs some more work :)
DreamOtter has joined #asahi
hckr has joined #asahi
joske has quit [Read error: Connection reset by peer]
joske has joined #asahi
Major_Biscuit has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
DreamOtter has joined #asahi
MajorBiscuit has quit [Ping timeout: 480 seconds]
joske has quit [Ping timeout: 480 seconds]
joske has joined #asahi
<marcan> sven: remind me again why pcie fails but platform works?
<sven> because the pcie darts have no hardware bypass mode
<sven> if you boot with 4K the usb darts are just forced to bypass mode
hckr has quit [Quit: hckr]
<marcan> ah, it was just bypass, right
<sven> yeah. and I don’t think I’d feel comfortable with the WiFi being in bypass mode anyway
<sven> right now it’ll even see the entire pages of its buffers
<sven> i’d prefer if it just goes through swiotlb or that subpage protection
n1c has joined #asahi
n1c_ has joined #asahi
n1c_ has quit [Remote host closed the connection]
n1c_ has joined #asahi
n1c has left #asahi [WeeChat 3.4]
n1c_ has quit [Remote host closed the connection]
n1c has joined #asahi
rando97 has joined #asahi
rando97 has quit []
joske has quit [Ping timeout: 480 seconds]
joske has joined #asahi
___nick___ has joined #asahi
___nick___ has quit []
___nick___ has joined #asahi
yuyichao has joined #asahi
Dcow_ has quit [Read error: Connection reset by peer]
Dcow has joined #asahi
n1c has quit [Quit: WeeChat 3.4]
Dcow_ has joined #asahi
martinstewart has joined #asahi
martinstewart has quit [Remote host closed the connection]
c10l2 is now known as c10l
NicolasOury[m] has joined #asahi
Dcow_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Dcow_ has joined #asahi
joske has quit [Remote host closed the connection]
aleasto has quit [Ping timeout: 480 seconds]
DreamOtter has quit [Ping timeout: 480 seconds]
c10l has quit [Quit: The Lounge - https://thelounge.chat]
c10l has joined #asahi
darkapex1 has joined #asahi
c10l has quit [Quit: The Lounge - https://thelounge.chat]
darkapex has quit [Ping timeout: 480 seconds]
<Glanzmann> mps: With your kernel config I got the wireless running. Thank you again. Strange thing is I don't get it running using wpa-supplicant but iwd works a lot and I get why you prefer it. ;-)
<mps> Glanzmann: I think you should have '-Dnl80211' parameter for wpa_supplicant if you want to use it with brcmfmac
c10l has joined #asahi
<Glanzmann> mps: Normally I have but for this notebook I revereted back to a simple config. Maybe that's it. Let me try.
<mps> but using wpa_s if you can use iwd is 'painful' imo
<Glanzmann> mps: I agree. Also iwd told me that I'm too stupid compile a kernel and set what the problem is and it is indeed fast. I like it. I consider switching.
<mps> Glanzmann: you can use iwd with conman or NetworkManager if you need these gui front-ends
<Glanzmann> mps: Btw. about the tocuhpad, which is finally working. I forgot to install the mtrack module.
<Glanzmann> mps: I don't like NetworkManager and avoid wherever I can because it got to many times in my way.
<mps> hmm, I don't have mtrack installed. is it better than libinput
<Glanzmann> mps: Probably not, but I'm not sure. it is working for now but very sensitive and palm detection does not work that well.
<mps> ah same as libinput then
<mps> I disabled touch to click
<mps> xinput -set-prop 'MacBook Force Touch trackpad' 'libinput Tapping Enabled' 0
<Glanzmann> mps: The strange thing about libinput at least on debian bullseye it was detecting the trackpad as mouse and don't give me right and middle click.
<Glanzmann> I see.
<mps> Glanzmann: cat /etc/X11/xorg.conf.d/30-libinput.conf => https://tpaste.us/5n8M
<mps> this is my libinput driver config for xorg
<mps> going to lunch, cul
joske has joined #asahi
<Glanzmann> mps: Thank you.
DreamOtter has joined #asahi
c10l has quit [Quit: The Lounge - https://thelounge.chat]
c10l has joined #asahi
DreamOtter has quit [Ping timeout: 480 seconds]
c10l has quit [Quit: The Lounge - https://thelounge.chat]
c10l has joined #asahi
DreamOtter has joined #asahi
<Glanzmann> mps: Thank you again for input
<Glanzmann> .conf. That is much be
<Glanzmann> tter. In the beginning it was not working. But that was because the xorg in 'Debain stable bullseye does not have an up2date libinput xorg driver'. Now with testing it is working.
<Glanzmann> The returns above were the enabled Tapping. :-)
joske has quit [Quit: Quit]
DreamOtter has quit [Ping timeout: 480 seconds]
<marcan> pushed wifi/take5 with support for uploading TxCap blobs and WiFi calibration blobs. you need to update m1n1 (for the cal blob passthrough) and re-extract the firmware with the installer script (for the TxCap blobs)
<marcan> I wonder if this will change WiFi performance for anyone
c10l has quit [Quit: The Lounge - https://thelounge.chat]
c10l has joined #asahi
c10l has quit []
c10l has joined #asahi
<mps> marcan: how to measure performance? with iperf?
<mps> Glanzmann: you are welcome, and nice to hear that you fixed it
<marcan> mps: no idea, no idea what changes you might expect :-)
<marcan> could be range related, could only affect some bands/modes, etc
<marcan> "just try stuff"
<mps> marcan: ok
<Glanzmann> I'll try it and report.
<j`ey> marcan: must be 30 patches now..
<marcan> 33, apparently
<marcan> it's going to be fun firing that off
<mps> ok, will build it and test to see if there are differences, though it is quite good now
<marcan> which I will soon, can I still say merry xmas in the cover letter if it's on the 26th? :p
<Glanzmann> mps: Do you have any config that speeds up pointer speed? With 'mtrack' it was really sensitive. With libinput it is very slow. Something in the middle would be nice.
<j`ey> 26th is 'Boxing Day' in the UK
<mps> Glanzmann: I do with this in .xinitrc `xinput -set-prop 'MacBook Force Touch trackpad' 'libinput Accel Speed' 0.7`
<Glanzmann> mps: Thank you. That is much better.
<ar> mps: why not put that in xorg.conf.d?
<mps> ar: laziness, I guess :)
<mps> ar: jk, I have home dir which I run on different machines and set difference between machines in .xinitrc and this is one which differs on them
c10l has quit [Quit: The Lounge - https://thelounge.chat]
<mps> so I simply rsync home from one machine to another and I switch to another and have same settings, touchpad/keyboard tweaks and other things
<Glanzmann> mps: I also have an environmnt that I copy to all machines.
<Glanzmann> mps: Btw. Do you have something for firefox settings, because that ones I always do manually.
<mps> Glanzmann: `sed -i 's/gtk-font-name = Cousine ./gtk-font-name Cousine = 8/' .config/gtk-3.0/settings.ini` is mostly enough for firefox
<mps> and this `echo 'Xft.dpi: 140' | xrdb -merge` for DPI change
<Glanzmann> mps: I see. For the resolution I was happy I was thinking about all the other settings (basically I turn all crap off, blank start page, ublock origin, I don't care about cookies and delete entire history on startup)
<Glanzmann> Btw. I also have the xft.dpi setting. :-)
<mps> my $HOME/.mozilla is rsynced
<Glanzmann> mps: I see. and that is enough, good to know.
<mps> yes, if you set DPI, icon size, cursor size and font size for gtk that is fine (for me) though not super perfect
c10l has joined #asahi
<Glanzmann> Have you tried, how long you can stay on battery with your macbook pro on Linux?
<mps> didn't tested this yet
<mps> ok, ready to install new m1n1 and test wifi5 branch
<mps> take5*
<Glanzmann> Oh I forgot to rebuild m1n1.
<Glanzmann> But it also works well with the old m1n1
<Glanzmann> What is strange about the wifi is, I have a 5ghz access point on my desk right next to the macbook air, but it goes with 2ghz. Both are on the same SSID though.
<Glanzmann> marcan: About wifi and 5ghz. Any idea what this is about? https://pbot.rmdir.de/Nxc_s27mdeCnifUlobPEbw
<mps> marcan: it looks somewhat faster with take5
<marcan> Glanzmann: well 5000 is not a valid wifi channel, so of course that won't work?
<marcan> 5GHz starts at 5035 for the weird channels nobody uses, 5160 more commonly
<mps> marcan: I can paste iperf result if you want to see
<marcan> mps: probably won't be scientific anyway, but good to know it seems to be at least the same or better with this
<mps> though 'iw wlan0 link' shows no change
<Glanzmann> mps: I want to see it, please paste.
<Glanzmann> marcan: I see, 5035 and 5160 also do not work for me.
<mps> actually iperf shows nearly doubled speed
<mps> Glanzmann: https://tpaste.us/vgm5 ,first two are from take4 and last two are from take5
<mps> but as marcan said this is not scientific test
<Glanzmann> mps: I see, I get better numbers than that: https://pbot.rmdir.de/0P5qtG4Le0l6qxh5Umxv2Q First is upload, second download.
<Glanzmann> mps: I see you also have only 2ghz. Any luck with 5ghz?
<Glanzmann> mps: When I boot macos and hold down the option key while klicken on the wifi symbol top right, I see 1200 Mbps rate. When I measure on macos I get over 100 MB/s.
<mps> Glanzmann: no, didn't ever tried 5GHz, quite satisfied with 2
<marcan> 5G works perfectly fine for me on the m1 pro
joske has joined #asahi
<Glanzmann> marcan: I see, strange.
<mps> Glanzmann: look at RoamThreshold5G param in 'man iwd.config'
<Glanzmann> mps: I set the follwoing: /etc/iwd/main.conf Now I'm getting half a gbit/s.
<joske> 5GHz wifi saturates my broadband at 285 Mbps
<Glanzmann> joske: Currently only one device can utilize my external network bandwidth fully. Wifi 5 Ghz: https://pbot.rmdir.de/A-6ZXamkmHhBUhzkEXbPXg macos is 30 MB/s faster.
<joske> Glanzmann: well, it saturates my WAN connection so macOS can not be faster ;-)
<joske> will test LAN speed
<Glanzmann> joske: I got a dark fibre from German Telekom recently. They sell 1000 Mbit/s. At some point they noticed that this is more than 1 Gbit/s which is around 930 Mbit/s. So they now ship optical termination points with a 2.5 Gbit/s RJ45 interface. By using that I get slightly 110% of Gbit. So approx 128 MB/s IPv4 and 125 MB/s IPv6.
<Glanzmann> This is downspeed only. Uplink I get around 30 MB/s IPv4 which is more than the 200 Mbit/s they promise.
<marcan> Glanzmann: https://mrcn.st/p/6JqrRs7u
c10l has quit [Quit: The Lounge - https://thelounge.chat]
<marcan> speedtest.net says about 170/200, which isn't bad for 5G with the AP one floor up
<Glanzmann> marcan: https://pbot.rmdir.de/jWw3BO--fqLptwFtS0JGAQ So I'm full speed. But something is slowing down. Because I know on macos I get 25MB/s more. But it is of course still more than I need.
<marcan> it's a different IP stack and driver and it's very hard to do scientific tests of wifi
<joske> anyone know what's the status of sound on the m1 air?
<marcan> > At some point they noticed that this is more than 1 Gbit/s which is around 930 Mbit/s. ????
<marcan> 1Gbit/s is 1000Mbit/s
<marcan> that's... how SI prefixes work
<marcan> I'm confused
<Glanzmann> marcan: Yes, but there is overhead.
c10l has joined #asahi
<Glanzmann> mac, ip, tcp, preamble ...
<marcan> oh, you mean TCP throughput vs wire speed
<marcan> anyway, I'm off to sleep
<Glanzmann> marcan: Have a good night sleep.
<Glanzmann> This is what I get on a 1 Gbit/s network: https://pbot.rmdir.de/dngslb7QzpvHF5wDz7K_DQ. This is what I get over my internet connection: https://pbot.rmdir.de/RAnZkoQccd-1RQ9rnwhYIQ
aleasto has joined #asahi
<mps> does audio works on air
<Glanzmann> mps: Not to my knowledge. But I think it is very close that the audio jack works. For the speakers is more work to do.
<Glanzmann> mps: povik knows more. :-)
<joske> when I add a new resolution and try to switch to it, macbook air locks up hard
<Glanzmann> joske: It is not yet supported.
<joske> clearly ;-)
<Glanzmann> joske: The dcp driver could do that. But it is currently not working.
<Glanzmann> joske: I assume that marcan will at some point write a new driver for m1n1.
<Glanzmann> and the kernel.
<Glanzmann> So probably a modesetting driving.
<Glanzmann> joske: But there is an old tree with dcp from alyssa which worked for me on the mini. But it was a few roundtrips until it worked for me.
<joske> Glanzmann: no biggie, just testing out what works and what doesn't
<Glanzmann> I see.
<joske> it's clearly not yet daily driver material
<joske> even for 'bleeding edge' people
<Glanzmann> joske: Now that wifi, keyboard and touchpard work. I could work with it.
<Glanzmann> And I think mps works already quite a while with it.
<joske> technically, yes, you could, but no sound, no chrome, no resolution change (native is a bit too small), no poweroff, hassle to reboot etc
<Glanzmann> And I just tested video performance. It is not perfect but for software rendering okay. I tried with mpv and a 1080p video.
<joske> if you want to develop for ARM64, it's great already I guess
<joske> the fastest ARM linux available by a long margin ;-)
<joske> does external display work? I guess not yet
<Glanzmann> joske: A friend of mine bought a mini, put a linux vm on macos on it and uses it as gitlub runner for over a year now.
<mps> Glanzmann: yes, I use it as workstation for about 2-3 weeks, not perfect but quite fine
<Glanzmann> joske: No, external monitor does not work and requires a lot of work if I understood Alyssa correctly.
<joske> I'd figured so
<Glanzmann> mps: I also worked the mini for a week or, was also good for me.
<joske> and suspend?
<joske> seems to do nothing for now
<Glanzmann> No way. But it boots very fast. :-)
zimsneexh has quit [Remote host closed the connection]
<joske> true
<joske> *if* you have another machine ;-)
reillyeon has quit [Quit: The Lounge - https://thelounge.github.io]
<Glanzmann> 5 seconds to desktop.
<joske> well, about the same as my intel desktop
<Glanzmann> joske: Yes, but I assume marcan will tackle smc soon and than you no nolonger need a second machine.
<joske> the BIOS takes much longer to initialize than booting
<joske> I think it boots in less than 2 seconds
<Glanzmann> joske: Wow. What distribution do you use?
<joske> i5 11600
reillyeon has joined #asahi
<joske> manjaro on the intel and M1
___nick___ has quit [Ping timeout: 480 seconds]
c10l has quit [Quit: The Lounge - https://thelounge.chat]
c10l has joined #asahi
Techcable has quit [Quit: ZNC - https://znc.in]
bdju has quit [Ping timeout: 480 seconds]
bdju has joined #asahi
alexjsteffen has joined #asahi
alexjsteffen has quit []
aleasto has quit [Ping timeout: 480 seconds]
joske has quit [Ping timeout: 480 seconds]
Major_Biscuit has quit []
MajorBiscuit has joined #asahi
DarkShadow44 has quit [Quit: ZNC - https://znc.in]