[7.0.0-1-cachyos] - amdgpu driver broken on Thinkpad T14Gen2

Issues & Assistance
1

Hello there!

I installed CachyOS two days ago and the latest Linux kernel version 7 doesn’t want to boot to the desktop. The LTS-Kernel version 6.18.22-1-cachyos-lts is booting to the desktop as expected.

Via ssh I collected the logs using the cachyos-bugreport.sh. The following excerpt of dmesg seems to be related:

[   11.324424] amdgpu 0000:07:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[   11.431303] amdgpu 0000:07:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[   11.551977] amdgpu 0000:07:00.0: [drm] *ERROR* dc_dmub_srv_log_diagnostic_data: DMCUB error - collecting diagnostic data
[   11.813560] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1122: core_link_write_dpcd (DP_DOWNSPREAD_CTRL) failed
[   11.833019] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1127: core_link_write_dpcd (DP_LANE_COUNT_SET) failed
[   11.852559] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1155: core_link_write_dpcd (DP_LINK_BW_SET) failed
[   12.620564] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1122: core_link_write_dpcd (DP_DOWNSPREAD_CTRL) failed
[   12.640018] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1127: core_link_write_dpcd (DP_LANE_COUNT_SET) failed
[   12.659557] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1155: core_link_write_dpcd (DP_LINK_BW_SET) failed
[   13.163239] wlan0: authenticate with <mac-address-redacted> (local address=<mac-address-redacted>)
[   13.324278] wlan0: send auth to <mac-address-redacted> (try 1/3)
[   13.326499] wlan0: <mac-address-redacted> denied authentication (status 30)
[   13.476832] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1122: core_link_write_dpcd (DP_DOWNSPREAD_CTRL) failed
[   13.496286] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1127: core_link_write_dpcd (DP_LANE_COUNT_SET) failed
[   13.515831] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1155: core_link_write_dpcd (DP_LINK_BW_SET) failed
[   13.606297] wlan0: No basic rates, using min rate instead
[   13.615001] wlan0: authenticate with <mac-address-redacted> (local address=<mac-address-redacted>)
[   13.750910] wlan0: send auth to <mac-address-redacted> (try 1/3)
[   13.872995] wlan0: authenticate with <mac-address-redacted> (local address=<mac-address-redacted>)
[   13.873006] wlan0: send auth to <mac-address-redacted> (try 1/3)
[   13.911682] wlan0: authenticated
[   13.913386] wlan0: associate with <mac-address-redacted> (try 1/3)
[   14.023645] wlan0: associate with <mac-address-redacted> (try 2/3)
[   14.127637] wlan0: associate with <mac-address-redacted> (try 3/3)
[   14.239675] wlan0: association with <mac-address-redacted> timed out
[   14.389384] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1122: core_link_write_dpcd (DP_DOWNSPREAD_CTRL) failed
[   14.408919] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1127: core_link_write_dpcd (DP_LANE_COUNT_SET) failed
[   14.428380] amdgpu 0000:07:00.0: [drm] *ERROR* dpcd_set_link_settings:1155: core_link_write_dpcd (DP_LINK_BW_SET) failed

You can find the whole log at this pastebin: cae4be4

Cheers! :clinking_beer_mugs:

2

en mi caso arranca normal, qué podria enviar para que se pueda ver que en mi caso funciona bien?

3

I have the Thinkpad T14 Gen2 with the Model number 20XL with the latest BIOS version 1.33.
Is this your case as well?

4

I just found out my issue is related to this topic: Ran into a frozen black screen issue with latest update - #10 by SoarSpeed

EDIT: The workaround mentioned in the linked thread does actually work: When the boot animation of cachyOS :penguin: finished, close the lid of the laptop and wait until it went to sleep, then open up the lid again for the login screen to appear on the screen :laughing:

I uploaded a new pastebin with the workaround: 0fb6c53

Lets see when this gets fixed upstream :man_shrugging:

5

Same issue here.

I had to live edit the cmdline parameters in the Limine selection menu and add the nomodeset parameter for the system to load. For a more permanent solution, I installed an LTS kernel using CachyOS Kernel Manager and selected that instead after a reset.

Seems like the amdgpu module is broken on the latest kernel.

6

Same issue here on T14Gen2.

7

Same issue here on an HP OMEN 16-c0xxx.

  • iGPU: AMD Cezanne [Radeon Vega Series] @ 07:00.0
  • dGPU: NVIDIA GA104M [GeForce RTX 3070 Mobile / Max-Q] @ 01:00.0
  • Driver: nvidia-open-dkms 595.58.03

Kernel 7.0.1-1-cachyos gives a black screen. dmesg shows:

amdgpu 0000:07:00.0: amdgpu: MODE2 reset
amdgpu 0000:07:00.0: amdgpu: asic reset on init failed
amdgpu 0000:07:00.0: amdgpu: Fatal error during GPU init
amdgpu 0000:07:00.0: probe with driver amdgpu failed with error -95

Kernel 6.19.10-1-cachyos (from the non-v3 repo) works perfectly — amdgpu loads without issues.

Confirming this is a regression in 7.0.

8

Marking this as fixed as with the latest 7.0.3-1-cachyos the screen works as expected again :smiley: