Have upgraded to Ubuntu 25.10 and it seems the screen rotation in tablet mode has stopped working. Considering it’s due for release tomorrow I’d be surprised if the issue is fixed between now and then.
Perhaps a precaution before anyone upgrades, but any ideas? It was working fine with 25.04.
raised by another user days ago, linking for tracking:
i normally run Arch, but just downloaded and live booted the questing-desktop-amd64.iso from 2025-10-07 03:55 and the live environment at least, appears to switch to tablet mode and rotates the screen without any issues.
What i notice is that if i sit in the settings window for displays, the “orientation” option goes away when i rotate the screen open to flip it backwards, at that point the ability to rotate the screen kicks in.
Huh, you’re right, @jayasafunctionofe! I just tested with the final release and for me as well screen rotation works perfectly fine from the live USB. But once the installation is through it doesn’t work any more. Applying the `iio-sensor-proxy` doesn’t help either. How strange…
what desktop environment are you using?
i just did a clean install the same build of 25.10 (2025-10-07) using Gnome 49.
after install the display settings change when i flip the screen back, and rotation works fine, just like in the live boot environment.
i didn’t take any additional steps after install, i did not need to apply the iio-sensor-proxy changes.
Wow, that’s so strange! I just did a fresh install to be safe, and I definitely have tablet mode working on the live usb but not after install. Installing from the final release that is. (But afaik the final release is identical to the build from 10/07.)
And I’m using Gnome, everything stock.
FYI there are two drivers that can be loaded for the lid sensor, and one of them is “wrong.” In a default configuration, at least on Arch, the order on boot is de facto random. Anyhow, the way I “fixed” it was by adding pinctrl_tigerlake to my /etc/mkinitcpio.conf’s MODULES section and rebuilding initrd. This ensures the module is installed very early on, and nothing “claims” the devices it would have.
This isn’t a problem on some other distributions, because their kernels have the tigerlake driver compiled into the kernel. Also arguably the better solution is to blacklist the “wrong” module, but honestly I don’t even remember what it was.
Yes, thank you, @GenderShrapnel! That was it! I found the issue you mentioned in the framework documentation here. Loading the modules in the right order fixes the issue for me.
Thanks again, that’s great.
I don’t have this file and I’m unsure how to make this change permanent.
Why is this a problem in 25.10 and not 25.04?
That file doesn’t exist on Ubuntu, because Ubuntu 25.10 uses dracut instead of Arch’s mkinitcpio. On Ubuntu you can make the modules always load in the right order by running these two commands:
echo "force_drivers+=" pinctrl_tigerlake "" | sudo tee /etc/dracut.conf.d/fw12_tablet_mode_fix.conf
sudo update-initramfs -u
Then reboot.
Let me know, if this works for you, @20dogs!
I believe that a cleaner solution is to make an extra modprobe config in /etc/modprobe.d/soc_button_array.conf:
softdep soc_button_array pre: pinctrl_tigerlake
Honestly, I wonder if this not existing in the presets (see modprobe --showconfig | grep pinctrl) counts as a bug or something. I like this approach for being specific – forcing a module in the UKI/initramfs is a bit of a hack.
that’s similar to what i did on arch, i found this thread and disabled the soc_mod_array modules and added pinctrl_tigerlake to my mkinitcpio.conf
So it seems to correctly switch to tablet mode but the touch input is a bit jerky…odd…
I had this problem after upgrading to 25.10.
I got the screen rotation to work by enabling the Gnome shell auto-rotate extension but no on screen keyboard.
Following this thread earlier, loading the modules in the right order manually fixed the on screen keyboard.
I see a couple of alternatives for making this permanent - is there a Framework recommendation?