https://bugzilla.kernel.org/show_bug.cgi?id=220640

            Bug ID: 220640
           Summary: Hardware: - GPU: AMD Radeon RX 6600 / 6600 XT / 6600M
                    (Navi 23) - Subsystem: XFX Limited [1eae:6505] -
                    Kernel driver: amdgpu - Kernel: 6.14.0-33-generic -
                    OS: Ubuntu 24.04 (upstream testing)  Issue: During
                    normal usage in GNOME, the GPU occasionally loses
           Product: Drivers
           Version: 2.5
          Hardware: All
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P3
         Component: Video(DRI - non Intel)
          Assignee: [email protected]
          Reporter: [email protected]
        Regression: No

Created attachment 308775
  --> https://bugzilla.kernel.org/attachment.cgi?id=308775&action=edit
kernel logs showing context lost events, GPU resets, and SMU/PSP messages.

Hardware:
- GPU: AMD Radeon RX 6600 / 6600 XT / 6600M (Navi 23)
- Subsystem: XFX Limited [1eae:6505]
- Kernel driver: amdgpu
- Kernel: 6.14.0-33-generic
- OS: Ubuntu 24.04 (upstream testing)

Issue:
During normal usage in GNOME, the GPU occasionally loses context, resulting in
visible system issues:

- White, sometimes noisy, screen (GPU output frozen)
- System freeze / unresponsive graphical session
- GPU reset occurs automatically
- Kernel log shows:

    amdgpu: The CS has cancelled because the context is lost. This context is
innocent.
    amdgpu 0000:09:00.0: GPU reset(2) succeeded!

BACO (Bus Active/Chip Off) for runtime power management is enabled.

Logs indicate repeated SMU / PSP resume sequences and GPU mode resets:

    amdgpu 0000:09:00.0: Using BACO for runtime pm
    amdgpu 0000:09:00.0: SMU is resumed successfully!

Steps to reproduce:
1. Boot Ubuntu 24.04 with kernel 6.14.0-33-generic.
2. Use system normally (GNOME shell active).
3. Intermittently, observe:
   - White / noisy screen
   - GPU reset in kernel logs
   - Temporary freeze of graphical session

Additional info:
- SMU driver interface version does not match firmware (driver 0x0f, fw 0x13),
but GPU resumes.
- GPU operates normally after reset.
- BACO active for runtime PM.

Impact:
- System instability: freezes, temporary loss of graphical output.
- Potential interference with GPU workloads.
- Annoying / disruptive for desktop usage.

Request:
- Investigate root cause of context loss leading to visible screen corruption.
- Verify interaction with BACO power management.
- Advise on firmware/driver fixes or workarounds.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

Reply via email to