[fixed in linux-firmware 20240130.5217b76b] ath11k/QCNFA765 breaks suspend on laptop with amdgpu errors "MES failed to response msg=14" / "failed to reg_write_reg_wait"
Description
My Thinkpad P14s always fails to suspend.
The suspend kinda starts (screen turns off), but then immediately resumes, and these errors show up in dmesg:
[ 54.594789] [drm:mes_v11_0_submit_pkt_and_poll_completion.constprop.0 [amdgpu]] *ERROR* MES failed to response msg=14
[ 54.595098] [drm:amdgpu_mes_reg_write_reg_wait [amdgpu]] *ERROR* failed to reg_write_reg_wait
Here's a color-highlighted dmesg, for readability & context:
Relationship to ath11k
Unloading ath11k fixes the issue: after I modprobe -r ath11k_pci ath11k , suspend & resume work as expected (suggested here)
Additional info
- Thinkpad P14s Gen 4 AMD, with up-to-date BIOS 1.16
- CPU lshw:
AMD Ryzen 7 PRO 7840U w/ Radeon 780M Graphics - Wireless lspci:
Qualcomm Technologies, Inc QCNFA765 Wireless Network Adapter (rev 01) - Package versions:
- Arch Linux fully up-to-date as of Feb 10 with core,extra,multilib
- All versions of kernel 6.7: 6.7.4-arch1-1, 6.7.3-arch1-2, 6.7.3-arch1-1, 6.7.2-arch1-2, 6.7.2-arch1-1, 6.7.1-arch1-1
- 6.6 also affected (this laptop has always had the issue since I bought it)
- linux-firmware 20240115.9b6d0b08-2
- Bug reports upstream / in other distros:
- RedHat Bug 2262577 - kernel-6.7.4 broken suspend (QCNFA765 ath11k) is a similar bug. It's not identical, though. For me, suspend fails and immediately resumes, for them suspend causes a deadlock: "Suspend causes deadlock. Screen goes black but does not turn on. Keyboard lights are on. Caps Lock does not respond suggesting deadlock. Nothing is logged to journal."
- RedHat bug also links to maybe an issue in linux-firmware: https://bugzilla.redhat.com/show_bug.cgi?id=2262577#c17
- RedHat bug also links to DRM/AMD bug #3132 - Asus Zephyrus G14 GA402 - Suspend not working reliably since Kernel 6.6.8
Steps to reproduce
- Suspend, using the GNOME Shell action
Edited by Ronan Jouchet
