[Bug 208947] amdgpu DisplayPort won't recognize all display modes after 5.9 merges

bugzilla-daemon at bugzilla.kernel.org bugzilla-daemon at bugzilla.kernel.org
Fri Aug 21 19:52:38 UTC 2020


https://bugzilla.kernel.org/show_bug.cgi?id=208947

--- Comment #16 from Coleman Kane (ckane at colemankane.org) ---
Pulling the call / reassignment of "status" added in that commit seems to
revert the undesirable behavior, including in the latest "staging-testing" code
that I've tested on.

My hardware situation seems to be that the "native" resolution of my DP display
(3840x2160 at 60Hz) is also advertised as "supported" by the RX 580m on my laptop,
but for whatever reason the GPU is unable to actually drive that resolution. I
don't know enough about the hardware to speculate as to why this is, though.
Using the "video=" option to the kernel, I can force a functioning video mode,
which seems to be consistent with the intent of that variable, when this call
is commented out. However, if the call to check_link_loss_status happens, then
the inconsistent behavior occurs that I described above where it randomly might
work, might only give me 1024x768, or might never display on my DP-connected
monitor regardless of the resolution chosen.

That said, I wonder if it makes sense to wrap this call with another kernel
command-line boolean variable, so the behavior can be disabled in the event
someone has a misbehaving GPU or monitor like I do - unless you think there
might be a graceful way to auto-identify this type of defect. Not sure how the
auto-selection works for KMS console resolutions, but it would be helpful too
if it detected that an attempt at a high refresh rate didn't work, it
automatically tried lower refresh rates.

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


More information about the dri-devel mailing list