From the trace this is triggered from a userspace IOCTL to remove a
framebuffer.  This causes drm core to process atomic_remove_fb() as
work, and this will use the drm driver to do atomic modesets.

One of the things that happens for this is going to be recalculating MST
DSC configs (if appropriate).  You can see more about this in
compute_mst_dsc_configs_for_link().

Returning -ENOSPC I would guess happens as part of
drm_dp_mst_atomic_check() meaning the atomic check failed.

I think this should be debugged in two ways.

1) Try a mainline 6.13 kernel.  I know for a fact we had a variety of
MST/DSC fixes that have gone in since 6.11.0.  Most of them are in
stable 6.11.y and 6.12.y, but the Canonical kernel isn't tracking
stable.

Unfortunately; it seems that the Canonical kernel PPA is effectively
dead, so you'll have to compile this yourself.

2) Turn on dynamic debug for DRM (drm.debug=0x106 on kernel command
line) and turn up log buffer length (log_buf_len=50M).  Then there will
be some more messages that come in (for example the debug message in
drm_mode_rmfb_work_fn()) and compute_mst_dsc_configs_for_link().  This
will help give more hints where the -ENOSPC is coming from.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2095100

Title:
  kernel panic when setting application in fullscreen

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2095100/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to