https://bugzilla.kernel.org/show_bug.cgi?id=211425 Andreas (icedragon.aw@xxxxxx) changed: What |Removed |Added ---------------------------------------------------------------------------- Kernel Version|5.15.5 |5.15.10 --- Comment #24 from Andreas (icedragon.aw@xxxxxx) --- In general the blocking bug still occurs until/with the current mainline kernel version 5.15.10. With kernel 5.15.10 I could observe additionally some new warnings messages before the "BUG: workqueue ..." errors (like last message) starting: [drm] Fence fallback timer expired on ring gfx [drm] Fence fallback timer expired on ring gfx BUG: workqueue lockup - pool cpus=7 node=0 flags=0x0 nice=-20 stuck for 47s! Showing busy workqueues and worker pools: ... BUT I found a new hardware-patch solution! In my case the monitor accepts only over the DP-to-DP link connection 4K@60Hz. Over a HDMI-to-HDMI (same computer and monitor) I got only the 4K@30Hz (not a cable limitation, it is a monitor limit). But if I connect both cables at the same time, means DP-to-DP 4K@60Hz, and in addition a HDMI-to-HDMI cable (but this screen connection disabled from the desktop session control manager) - I could avoid the black screen completely (without any limitations and stable over reboots)!!! -> In this double cable configuration (HDMI disabled) the screen (DP connection) always recovers after some seconds (no hung up and no 2x20sec delay) - I was not able to trigger the error behaviour any more and I could also not observe any of the above error messages. -> But with this stable double cable configuration I observe some new warning messages (more than one times): [drm] Unknown EDID CEA parser results ... Maybe this helps a little bit for progress ... If someone like, I could make more tests ... -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
