Collection of Issues on Fedora 40/41 with FW13 (AMD)

Important:

This post has been turned into a Wiki, meaning it can be edited by anyone.
If you encountered an issue with your AMD Framework Laptop 13 running the latest supported version of Fedora and you think it might affect many people, feel free to add on to this post. Otherwise you can also just post in the thread (which probably helps this to get visibility) and I will every once in a while update this post.


Hello everyone!

While using my Framework Laptop 13 (AMD) with Fedora 40 (now 41) for a while I came across a few issues. Some of them I could fix, with others I am still stuck so I thought
it might be a good idea to collect them in a single thread hoping that you can help me solve the rest and that others with similar problems find solutions all at once. After all it shouldn’t be a too uncommon scenario to install Fedora on your new Framework Laptop.
Some issues likely are not at all related to the Framework laptop and probably should be posted in a Fedora forum, but I would like to start here looking for help.

  • Framework Laptop 13 AMD 7840U
  • BIOS: 3.05
  • Fedora 40 (now 41)
  • Current Kernel: 6.8.5-301.fc40.x86_64 (with Fedora 41)
  • Gnome 46 (now 47)
  • high-res display

1. [Solved] Lock Screen not Enabled by default (Fedora 40)

By default there was no GUI option to lock the screen.
Closing the laptop lid also did not lock the screen.

This could be fixed by running:

gsettings set org.gnome.desktop.lockdown disable-lock-screen false

I am surprised that such an elementary function was hidden behind a command-line option. This seems like a strange choice by Fedora.

2. [Open] Docking Station

Using the docking station Lenovo ThinkPad Thunderbolt 3 Dock does not
charge the Framework 13.
Displays and Ethernet usually work, yet sometimes the displays do not work, but instead the laptop charges.
(This is using the back port on the right side)
It always either charges OR works as docking station. Usually it is the latter.

3. [Open] Display configuration not remembered

After unplugging the laptop from the dock the display configuration
is forgotten and has to be set every time.

$ xrandr --listmonitors
Monitors: 3
 0: +*eDP-1 1440/290x960/190+0+1600  eDP-1
 1: +DP-12 1440/600x2560/340+1440+0  DP-12
 2: +DP-14 1440/600x2560/340+2880+0  DP-14

Without any changes, a few days later the configuration
was remembered and this has never caused troubles since.
As the issue was somewhat frustrating at the time, I’d
still appreciate if someone has an idea what went wrong.
Maybe this can also help someone else.

4. [Solved] Scaling issues with some Flatpaks

This so far is the most frustrating issue and likely is Gnome/Wayland/Fedora/Flatpak issue instead of a Framework problem.
Some Flatpaks (e.g. Mattermost, Podman Desktop) are blurry on the Framwork’s display.
Most references to this online argue that fractional scaling is the reason for this, however
I set the scaling to 200% and still have the issue.
Setting the scaling to 100% solves the issue, but is too small to be convenient.

Setting
gsettings set org.gnome.mutter experimental-features "['scale-monitor-framebuffer']"
did not fix the issue (but fixed other issues I had with flatpaks scaling too big)

Using KDE instead of Gnome solves this, but I am just not a big KDE fan.

Solved within 1 minute after posting by inffy:

gsettings set org.gnome.mutter experimental-features '["scale-monitor-framebuffer", "xwayland-native-scaling"]'

5. [Solved] Scaling of mouse cursor (after upgrade to Fedora 41)

Maybe a new bug in Fedora 41 or Gnome 47?
In many applications (e.g. Settings) the mouse cursor is scaled to at least
twice its size and very blurry.
Changing the scaling to 100% (200%, etc.) doesn’t solve this issue.
Changing the cursor size in the accessibility settings has no effect.


Normal size (zoomed in)


Huge and blurry (same zoom level) as soon as hovering over some applications (file manager, settings)

Edit:
This was likely caused by me installing and uninstalling the KDE desktop environment.
Afterwards the Cursor was still set to the Breeze theme which did not behave well in Gnome.
Solved by setting cursor back to Adwaita in the Gnome tweaks.

6. [Closed] Input Source

Some input sources (e.g. US, international) are available in other distributions, but not in Fedora, even when setting show-all-sources to true:

gsettings set org.gnome.desktop.input-sources show-all-sources true

Solution:

Specific layouts can be added using:

gsettings set org.gnome.desktop.input-sources mru-sources "[('xkb', 'us+de_se_fi')]"
gsettings set org.gnome.desktop.input-sources sources "[('xkb', 'us+de_se_fi')]"

7. [Duplicate] Laptop becomes unresponsive

I can not reliably reproduce it and so far it has happened only twice in the
first two days.
The GUI became suddenly extremely slow with the mouse moving with a few mm/s
on the screen. Opening a terminal and rebooting via command-line solved this
issue but required a lot of patience.

This was seriously worrying me in the first days but never happened since.

Likely duplicate of Fedora KDE becomes suddenly slow

Suggested fix: amdgpu.dcdebugmask=0x10


Even though this is quite a list, I am otherwise very happy with my shiny new laptop.
If you have a solution or idea to any of these topics, please let me know and I will try it.
If problems are resolved that others encountered as well, maybe it would be worth it to add them to the Knowledge Base, even if they are caused by an officially supported distribution instead of Framework directly.

Thanks for reading so far,
Philipp

4 Likes

I think you also need the xwayland-scaling setting:

gsettings set org.gnome.mutter experimental-features '["scale-monitor-framebuffer", "xwayland-native-scaling"]'

2 Likes

Thanks a lot, this solved the issue!
I updated the post with your solution to keep this thread as clear as possible.

1 Like

It’s likely this amdgpu issue 680M (Rebrandt) laptop randomly becomes very slow to update the screen (#3647) · Issues · drm / amd · GitLab. See also Fedora KDE becomes suddenly slow

The “amdgpu.dcdebugmask=0x200” kernel param workaround works for now.

3 Likes

Thank you!
If this ever happens again, I at least know it’s an open issue and I’ll try your solution.

I’m having an issue with Fedora 41 (workstation edition) where after waking the computer from suspend, i get an authentication error on the gnome lock screen. I can move the mouse and a access the system menu in the top right corner, and i can swap between the clock on the lock screen and where my profile picture and password prompt are, but there is no profile image and i cannot type in the password prompt. below that is a flickering text that reads “Authentication Error”. i can’t switch to another tty. the only way forward seems to be to hard reboot the computer. things are okay until the next suspend.

Seems like the correct kernel parameter is instead “amdgpu.dcdebugmask=0x10”, do you mind fixing the post? There were previous discussion about 0x200 but the former has turned out to be the correct fix

Hi,
I found some solution for the keyboard layout. It worked for me.

gsettings set org.gnome.desktop.input-sources mru-sources "[('xkb', 'us+de_se_fi')]"
gsettings set org.gnome.desktop.input-sources sources "[('xkb', 'us+de_se_fi')]"
"

Solution found here:

EDIT:
Should not matter but I am using a different BIOS version. 3.06

Thanks a lot!
This is not exactly the layout I was looking for, but it works just as well.
Either way the method looks like it should work with any other layout, provided one figures out what they are called.

Thanks for the correction.
I would have liked to edit the post, but sadly the forum here is set up in a way that posts can only be edited for a certain amount of time.

Generally I think this type of trouble shooting information would be good to have either in the knowledge base or FAQ. Ideally vetted by framework’s Linux support.

1 Like

This thread has been turned into a Wiki now and I have edited the correction from @Tommaso_Seneci and the solution from @AntiMuon_Kouri.
The original post is now editable by anyone in case new problems and solutions appear.

I have an issue while i play games or i perform other stress task.
Basically the pc reboot itself, I tried to monitor temperature (under 85°-90°), Ram usage and Core usage (i notice some peak in the 11th core).

I’m on fedora 41 but this issue was present also in fedora 40.
Ram= Kingston FURY Impact PnP 32GB (2x16GB) 5600MT/S
SSD= Crucial T500 1TB PCIe Gen4 NVMe M.2
igpu enabled

Does anyone know how to successfully install the Symantec VIP Access client on a Framework 13 running fedora 41? I attempt to run
fedora:~$ wine /home/me/Downloads/VIPAccessSetup.exe
It fires up and the installation dialogue box pops up. I click “Next” and the installation begins but soon freezes. Output to the terminal session is lengthy with the last several lines as follows:
0234:err:environ:init_peb starting L"C:\Program Files (x86)\Symantec\VIP Access Client\VIPAppService.exe" in experimental wow64 mode
0184:fixme:msi:internal_ui_handler internal UI not implemented for message 0x0b000000 (UI level = 5)
0184:fixme:msi:internal_ui_handler internal UI not implemented for message 0x0b000000 (UI level = 5)
0248:fixme:msi:AutomationObject_GetIDsOfNames Unknown member L"strComputer", clsid {000c109e-0000-0000-c000-000000000046}
0248:err:ole:apartment_getclassobject DllGetClassObject returned error 0x80004002 for dll L"C:\windows\system32\wbem\wbemdisp.dll"
0248:err:ole:create_server class {172bddf8-ceea-11d1-8b05-00600806d9b6} not registered
0248:fixme:ole:com_get_class_object CLSCTX_REMOTE_SERVER not supported
0248:err:ole:com_get_class_object no class object {172bddf8-ceea-11d1-8b05-00600806d9b6} could be created for context 0x15
0248:fixme:wbemdisp:WinMGMTS_ParseDisplayName ignoring security settings
0248:err:sync:RtlpWaitForCriticalSection section 75E922EC “dlls/wbemprox/builtin.c: table.cs” wait timed out in thread 0248, blocked by 01f4, retrying (60 sec)

It never completes. There is a terminal session on the screen entitled “C:\users\me\AppData\Local\Temp\RarSFX3\install.exe”, but the screen is blank. ^C out of the wine command brings me back to a command prompt, but the “0248:err:sync:RtlpWaitForCriticalSection section 75E922EC “dlls/wbemprox/builtin.c: table.cs” wait timed out in thread 0248, blocked by 01f4, retrying (60 sec)” error messages continue. Does not close that terminal screen. Cleanup requires ps -ef | grep VIP and kill -9 of two leftover, hung processes.