I tried an upstream kernel ( 4.4.12-040412-generic #201606011712 ). Attaching after suspend/resume no longer results in a crash.
The screen frequently fails to restore on the first try. dmesg has these lines in that case:
[drm:intel_mst_disable_dp [i915]] *ERROR* failed to update payload -22 [drm] GPU HANG: ecode 8:-1:0x00000000, reason: Kicking stuck wait on blitter ring, action: continue [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun
But, in an improvement from the Ubuntu kernel, I can unplug and try again with success.
I tried an upstream kernel ( 4.4.12- 040412- generic #201606011712 ). Attaching after suspend/resume no longer results in a crash.
The screen frequently fails to restore on the first try. dmesg has these lines in that case:
[drm:intel_ mst_disable_ dp [i915]] *ERROR* failed to update payload -22 cpu_fifo_ underrun_ irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun
[drm] GPU HANG: ecode 8:-1:0x00000000, reason: Kicking stuck wait on blitter ring, action: continue
[drm:intel_
But, in an improvement from the Ubuntu kernel, I can unplug and try again with success.