Arch Linux on the Framework Laptop 16

Has anybody else seen weird behavior with the numpad and a yubikey? For some reason whenever I plug in my yubikey, the enter key and the + key on the numpad start to behave as if numlock is turned off. For those who don’t know, when numlock is off, + and enter are used to cycle the backlight states. This is becoming really annoying because whenever I plugin the yubikey, I usually need to enter a pin with it, but instead of the enter key submitting the pin, it just starts cycling the backlight. I can fix it by turning numlock off and then back on again, but I’d like to know if I need to add some kind of qmk config to fix this? Is this a firmware bug in the numpad? It only happens when I plugin a yubikey, not any other kind of usb device, though I’ve only tried a flashdrive. It also doesn’t matter which side I plug it into. I have 2 usb A modules, one on either side.

I think I may have figured out part of the problem. I was watching the journal logs and noticed that the yubikey was being detected like a keyboard, which makes sense because it does emulate a keyboard to enter a one time password. After I saw that, I tried plugging in another keyboard and it did the same thing the yubikey did, but the main difference is that the keyboard has a numlock LED. When I first plugged in the external keyboard, the numlock LED was on, but the enter key on the framework numpad was still cycling the backlight. If I cycle the numlock off and then back on again, the framework numpad starts behaving normally again. I don’t know if its the operating system or the numpad firmware that controls the behavior of the numpad. I still haven’t tried any QMK configuration, so maybe it can be fixed with that? I’m using KDE with a setting to turn numpad on when I login, but I don’t think it should change the keyboard behavior when a keyboard is plugged in.

1 Like

Anyone running Xorg still?

I’m having issues with my xorg.conf and I was wondering if anyone had a working one.

Trying to set up bspwm and it’s been a long time since I’ve done the initial Xorg setup.

My Xorg file generated with like 5 monitors and screens and when I try to open up a terminal or something it doesn’t show. I’m thinking maybe it’s hidden on a different ‘monitor/screen’ somewhere.

I think it might be something driver related. Xorg has a handful of lines saying
`AMDGPU(#): [KMS] Kernel modesetting enabled’

@Shahan_Studios I’m not running xorg anymore, but when I was I gave up on making my own config. I just deleted the config and let it auto configure. Have you tried that?
Note: renaming the config also works, and is safer.

Arch is running pretty smooth. Even fractional scaling on Wayland works fine.

Noticed only some minor things:

  1. Wifi throughput seems too low: although connection reports >1 gig link, e.g. Steam only reaches ~160 mbps when using local file transfer via 1gig / 10gig LAN backbone
  2. Firefox seems to randomly pick iGPU / dGPU (may also depende on power power profile or if on battery / AC power, need to observe it more)
  3. VRR is not reported correctly. See [ANNOUNCEMENT] Adaptive Sync / Freesync / VRR not working - #3 by Matt_Hartley
  4. PPD profile switching required custom udev rules. I’m guessing these are already present on Ubuntu and Fedora ?

EDIT:
Regarding harware monitoring:

  1. lm_sensors did not detect any sensor, but using ACPI fallback mostly works for temperatures
  2. CPU fan is not reported in ACPI “sensor”
  3. dGPU Fanspeed is not reported by amdgpu correctly: it is always 0 rpm, even in games
3 Likes

Letting Xorg auto configure worked. Thanks!

It was probably working before I thought it was broken and messed with it.

I tried using the conf generated by X -configure and that’s what gave me the multiple monitor / screens issue. :man_shrugging:

1 Like

Anyone else experiencing their WiFi cutting out occasionally?

The command that worked for me was this:

sudo usermod -a -G uucp $(whoami)

Followed by a reboot. Logging out might work too, and hopefully someone sharper than I am will come along and enlighten us all on how to effect this without needing to reboot or logout. I tried running

newgrp uucp

but was still unable to connect. After rebooting (and having a moment when on first boot
/boot/efi was not found, but it worked on a subsequent reboot), I am able to connect to the LED spacers.

I hope this helps.Have a great day!

Edit to add tagging @DRACONIUM

Edited again to note - updating the fstab to use the uuid for the /boot/efi partition rather than /dev/nvme1n1p1 seems to have resolved things, Noted for future setups. This may be obvious to more savvy users, and something in the back of my mind is saying that I have run across it previously, but this time I wrote it down. Maybe one day when I grow up I can be a real boy.

1 Like

This sadly didn’t work for me, but I’m actually running Bluefin. I hope this works for the Arch users here though!

Try using group dialout rather than uucp. It looks like that might be the one that you need.

I must be having a “skill issue”. I’ve done sudo usermod -aG dialout $(whoami), and, even after rebooting, the groups command does not show me as a member of the dialout group :confused:

My apologies, I didn’t dig in on that distro. It looks like it’s an offshoot (or whatever one wishes to call it) of silverblue. Duckduckgo came up with this, it may help: Troubleshooting :: Fedora Docs

Best of luck!

1 Like

i was on the discord a while back and was actually pointed to the udev rules i was talking about, let me see if i can dig it up again, it’s on framework’s github somewhere.

1 Like
2 Likes

Mine will take a bit to come up after a boot, but it does come up, I’ve tried looking at the logs, but didn’t necessarily see anything out of the ordinary aside from the interface changing names (sometimes it’s wlp5s0 and other times it’s wlp2s0 though I’m suspecting that’s actually with and without the GPU module installed)

Thanks, now I am able to run the commands to change and update the led spacers without needing to use sudo. I currently have it showing the clock, but am thinking about doing something to get the current outside temperature and display that. Lots of ideas, little skill, sigh.

1 Like

I haven’t tried without the gpu yet. Mine might have been IPv6 related. I forced the network to use IPv4 and haven’t seen a drop yet.

Don’t you need $user$ instead of (whoami)?
I can’t remember off hand if you need the brackets.

the $() bit is basically interpolating in the value of the whoami command, which just spits our your username