{^_^} has quit [Remote host closed the connection]
{^_^} has joined #nixos-aarch64
gchristensen has joined #nixos-aarch64
h0m1 has quit [Ping timeout: 260 seconds]
h0m1 has joined #nixos-aarch64
h0m1 has quit [Ping timeout: 264 seconds]
h0m1 has joined #nixos-aarch64
<samueldr>
danielrf[m]: long~ish shot, but are you interested and in a situation where you would make the changes required for `google-marlin` to get Wi-Fi support?
<samueldr>
though, uh, I do have a `google-sailfish` coming my way... I guess the changes are going to be the same so really it's not that important considering
<danielrf[m]>
definitely interested. I haven't followed work in mobile-nixos recently. Have you made progress on wifi there?
<samueldr>
only up to qcacld-2.0
<samueldr>
(I'm preparing the update post for tuesday)
<danielrf[m]>
I manually did some simple steps from the postmarketos wiki in the past and got it working in the past
<danielrf[m]>
so i doubt it would be too difficult
<samueldr>
I assumed qcacld-2.0 needed userland services, like qcacld-3.0
<samueldr>
but I was wrong
<samueldr>
I got it working on samsung-a5y17lte and oneplus-oneplus3 this month
<samueldr>
basically you'd need to get the firmware, which it's a coin flip if they're at themuppets or on a modem partition
<samueldr>
(I haven't looked yet for the google pixel family)
<samueldr>
oneplus is on a partition, samsung wasn't
<samueldr>
samsung's difference could come from the fact it's not a qualcomm SoC, only the modem is qualcomm
<danielrf[m]>
let me see if I had any notes...
<samueldr>
one weird issue on the oneplus device is how it takes about 2 minutes for the interface to come up
<samueldr>
*migh* be from the mixed-origin of the code
<samueldr>
might*
<samueldr>
transplanted the lineageos driver to the android-linux-stable repo
<danielrf[m]>
I thought I remember having to mount /vendor to get it to work
<danielrf[m]>
I only have one note in my TODO about wifi on marlin: "wlan driver unloads itself after system renamed the interface"
<samueldr>
since /vendor is added to the kernel's internal firmware list class path IIRC, you should be good to only mount it and enable the setup-wcnss service
<samueldr>
so for using the vendor partition, it'd need to be mounted at /lib directly, but more realistically, mount /vendor to /vendor, and bind mount /lib/firmware to /vendor/firmware
<samueldr>
that would make the vendor partition automatically work even if firmware classpath is updated by nixos (as it is)
* samueldr
takes notes
<samueldr>
I should do the same for the modem partition on addison and oneplus3
<samueldr>
rather than force a copy of the contents
<Ke>
if there is no cooperation on upstream side, nixos could just hardcode drivers for aarch64
<samueldr>
yes
<samueldr>
as long as it doesn't require _additional patching_
<samueldr>
if it's about enabling existing code, I don't see any issues
<Ke>
upstream could cooperate to define alternate sets also in addition to auto, like allUseful
<Ke>
no patches indeed
<samueldr>
yeah
<samueldr>
like, the previous rk3399 patches wouldn't be fit for NixOS, as IIRC they were implementation modifications
<Ke>
or allow appending to auto
<samueldr>
yes
<Ke>
which is not possible
nicoo has quit [Remote host closed the connection]
tilpner_ has joined #nixos-aarch64
nicoo has joined #nixos-aarch64
<LinuxHackerman>
Emil Karlson: did the hardware.opengl.package approach not work?
<LinuxHackerman>
samueldr: did I hear xiaomi-tissot? :D
tilpner has quit [Ping timeout: 246 seconds]
tilpner_ is now known as tilpner
<samueldr>
LinuxHackerman: yeah
Darkmatter66 has quit [Ping timeout: 264 seconds]
<samueldr>
since it's WCNSS, it's trivial to enable wi-fi
<Ke>
linus.heckemann: I believe so, yes
<samueldr>
get the firmware files [the muppets may help you] package them as asus-z00t or motorola-addison does
<samueldr>
and then enable the service
<Ke>
linus.heckemann: when you have billions of generations at once, it's hard to figure out, what things you are using
<samueldr>
on my limited sample size of two devices, it just worked... but that's a well-known and documented method used by other distros too
<Ke>
but I saw mesa being rebuilt, so someone was invoking it
<Ke>
haven't rebooted since
<Ke>
linus.heckemann: for us adding radeonsi is not super urgent, as you probably also want the patch, that may not be wanted in nixpkgs
<LinuxHackerman>
you should be able to see if the radeonsi driver is available by checking if /run/opengl-driver/lib/dri/radeonsi_dri.so exists
<Ke>
sure
<Ke>
I don't apparently have dri at all there
<Ke>
that's odd
<Ke>
yes it works, but I had accidentally use mesa, not mesa.drivers
<Ke>
luckily I did not have to investigate, but saw it on first sight
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
<cptrbn>
samueldr: Got a 2nd pi4 over the weekend and solved my non-booting issues
<cptrbn>
Turns out it was the Pimoroni fan shim blocked booting for aarch64 images
<cptrbn>
I've not seen any suggestion anyone else has seen that
<cptrbn>
Still. I have a aarch64 pi for building on now
<cptrbn>
Now time to get my RPi0 images built
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-aarch64
<Ke>
and finally disabled that nastly firewall log in console
<Ke>
I don't know anymore, what does not work
<Ke>
I need to salvage some btrfs subvols that were not automaticallysent
<Ke>
not like anything is sent automatically in general, but I used it to move data
<angerman>
I have nix segfault me when evaluating an expression (only on aarch64), what are my options?
<archseer>
samueldr: The latest PR fails with `access to path '/nix/store/rv1kxz48j2lqiywkxgm81zrn68vvamvc-bczgjjc3x9syfi8zzyhvds2916pfpf0f-source/nixos/lib/eval-config.nix' is forbidden in restricted mode`
<archseer>
Inside modules/lib.nix
<archseer>
I forgot what the workaround for that was, I think you need to call toString on pkgs.path or something like that
cole-h has quit [Ping timeout: 256 seconds]
<LinuxHackerman>
angerman: my colleague says "macht der Herr nixUnstable? wenn ja, selber schuld."
<angerman>
LinuxHackerman: alle Macht dem Herren. Aber nein.
<angerman>
20.09
<LinuxHackerman>
you can use nixUnstable on 20.09 as well
<angerman>
LinuxHackerman: hmm yea, no nixUnstable here. Guess I'll go dig a bit deeper.
<LinuxHackerman>
Nice. I feel like 90% of the segfault bugs in nix were introduced between the latest stable release and unstable, but it's good to know stable has a few as well. :D
Darkmatter66 has joined #nixos-aarch64
pinage404[m] has quit [Quit: authenticating]
pinage404[m]1 has joined #nixos-aarch64
<angerman>
meh. :D back to cross compilation (where nix on x86_64 just works, lol)
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
monk has left #nixos-aarch64 ["Disconnected: Replaced by new connection"]
monk has joined #nixos-aarch64
zupo has joined #nixos-aarch64
Acou_Bass has quit [Ping timeout: 256 seconds]
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
<archseer>
samueldr: The latest PR fails with access to path '/nix/store/rv1kxz48j2lqiywkxgm81zrn68vvamvc-bczgjjc3x9syfi8zzyhvds2916pfpf0f-source/nixos/lib/eval-config.nix' is forbidden in restricted mode
<archseer>
Under lib.nix
<archseer>
I think you can replicate if you use the pure eval flag or if you test with my flake
<archseer>
I forgot what the workaround is but it might be calling toString on pkgs.path first
Acou_Bass has joined #nixos-aarch64
<archseer>
Ah sorry, my page didn't refresh, thought you were talking about yesterday's changes under 238
tilpner_ has joined #nixos-aarch64
<archseer>
I'll give it a quick test
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]