External monitors not recognized Ubuntu 25.10 after updating to kernel 6.17

Framework Laptop 13 Linux
1

“Ubuntu 25.10 (Questing Quokka), debian

Updated kernel from linux-modules-6.14.0-33-generic to 6.17.0-5-generic

BIOS: 03.16

Framework Laptop 13 model are you using? Framework Laptop 13 model are you using? AMD Ryzen™ 7040 Series, AMD Ryzen 7 7840U w/ Radeon 780M Graphics (Phoenix, gfx1103).

After updating the kernel from 6.14.0-33-generic to 6.17.0-5-generic, the external monitor is no longer recognized for HDMI and usb-c ports.

sudo dmesg -w output after manually unplugging HDMI-cable and it in again:

```[ 2224.262644] usb 7-1: USB disconnect, device number 11
[ 2224.488541] EXT4-fs (sda1): unmounting filesystem 08f01082-460c-4d05-9d4c-868fe15b188f.
[ 2224.784702] [UFW BLOCK] IN=enx9cbf0d0086f0 OUT= MAC= SRC=192.168.133.123 DST=239.255.255.250 LEN=635 TOS=0x00 PREC=0x00 TTL=1 ID=45031 DF PROTO=UDP SPT=35434 DPT=3702 LEN=615
[ 2224.785168] [UFW BLOCK] IN=docker0 OUT= MAC= SRC=172.17.0.1 DST=239.255.255.250 LEN=635 TOS=0x00 PREC=0x00 TTL=1 ID=41865 DF PROTO=UDP SPT=52273 DPT=3702 LEN=615
[ 2225.650587] ucsi_acpi USBC000:00: unknown error 256
[ 2225.650604] ucsi_acpi USBC000:00: GET_CABLE_PROPERTY failed (-5)
[ 2226.574847] usb 7-1: new full-speed USB device number 12 using xhci_hcd
[ 2226.735632] usb 7-1: New USB device found, idVendor=32ac, idProduct=0002, bcdDevice= 0.00
[ 2226.735649] usb 7-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 2226.735656] usb 7-1: Product: HDMI Expansion Card
[ 2226.735662] usb 7-1: Manufacturer: Framework
[ 2226.735667] usb 7-1: SerialNumber: 11AD1D009CC7401227250B00
[ 2226.822605] hid-generic 0003:32AC:0002.0016: hiddev0,hidraw0: USB HID v1.11 Device [Framework HDMI Expansion Card] on usb-0000:c3:00.4-1/input1
[ 2226.937529] hid-generic 0003:32AC:0002.0017: hiddev0,hidraw0: USB HID v1.11 Device [Framework HDMI Expansion Card] on usb-0000:c3:00.4-1/input1```

Gemini A explains as follows, manually unplugging HDMI-cable and reconnecting it again:

Timestamp Log Message Analysis
[ 2224.262644] usb 7-1: USB disconnect, device number 11 The system registers that a device (the Expansion Card) has been disconnected.
[ 2225.650587] ucsi_acpi USBC000:00: unknown error 256 This is the first critical error. ucsi_acpi stands for the USB Type-C Subsystem Interface ACPI driver. An unknown error here suggests a problem in the kernel’s handling of the USB-C port.
[ 2225.650604] ucsi_acpi USBC000:00: GET_CABLE_PROPERTY failed (-5) This is the second, most serious error. The kernel fails to retrieve the properties of the cable/connection. This process is essential for activating the DisplayPort Alternate Mode over USB-C. The -5 (IO error) indicates that the low-level communication is failing.
[ 2226.735632] usb 7-1: New USB device found, idVendor=32ac, idProduct=0002... The system later recognizes the card, but only as a generic USB device (idProduct=0002 is the Framework HDMI Expansion Card).
[ 2226.735662] usb 7-1: Product: HDMI Expansion Card The card is recognized as “HDMI Expansion Card,” but the video functionality is still not activated due to the preceding ucsi_acpi errors.

Plugging in a usb-c cable directly from monitor in usb-c port also not working, trying to revert to previous kernel but that one is removed from the system, and is also removed from the official repository…

2

And also adjusting brightness of laptopscreen is not possible, resolution options no longer possible, scaling is worse, etc…

3

FIX summary (for Framework forum bug report) with help from GROK AI:

Root cause: Installing ROCm 7.0.2 via the noble (Ubuntu 24.04) AMD repo on Ubuntu 25.10 (questing) replaced Ubuntu’s open-source amdgpu driver with workaround proprietary modules. Although this worked fine for kernel 6.14.x after update to 6.17 —> failed to build DKMS modules for kernel 6.17 → amdgpu never loaded → GPU disabled → no brightness, no HDMI/USB-C detection, llvmpipe fallback.

Fix:

sudo apt purge --autoremove 'amdgpu.*' 'rocm.*' 'mesa-amdgpu.*' 'xserver-xorg-video-amdgpu'
sudo rm /etc/apt/sources.list.d/amdgpu* /etc/apt/sources.list.d/rocm*
sudo reboot

→ Ubuntu’s stock amdgpu + Mesa took over → everything works again on 6.17.0-6, no kernel downgrade needed.

Warning: Never use noble AMD/ROCm repos on questing. Wait for official support, hopefully coming soon!

Bug was not kernel 6.17 — it was the incompatible ROCm stack, although it worked with the previous kernel 6.14 with a workaround ;-(
4

You should not be installing amdgpu-dkms for rocm on Ryzen. The documentation explicitly says to install with --no-dkms.

5

I did not install with amdgpu-dkms the rocm work around install was with amdgpu --no dkms, and the video-problems were not there untill the update to the newer kernel 6.17.x. So it was working all very well with the previous kernel 6.14.x…