correction: the FHD monitor is not connected via HDMI, but DVI

please find attached the dmesg output including some stack-traces.

Created attachment 140392 [details]
dmesg-log + strack traces

I am using amdgpu after a rather severe performance regression was introduced  with commit8b3a257851905ff444d981e52938cbf2b36ba830 which hits glamor for certain workloads pretty hard. 

AMDGPU didn't suffer that much from this change: https://bugs.freedesktop.org/show_bug.cgi?id=105171

What mode are you trying to run on your monitor?  Can you attach your xorg log if you are using X?  Might be related to bug 106959.

Does amdgpu.dc=0 avoid the problem? DC will be disabled by default again for Kaveri.

Created attachment 140472 [details]
xorg log file with 4.17.3

amdgpu.dc=0 did the trick - behaviour is back to normal.

I guzess it is not related to  bug 106959. - kaveri doesn't support HDMI2, so my refresh rate was 30Hz even with the old display stack.

Looking at your dmesg output, it looks like you might have hit this bug/issue: bug 105880

"[dc] No support for LVDS or VGA connectors (Cannot find any crtc or sizes)"

AMD/Harry Wentland/Alex Deucher have commits in the mainline kernel to disable the DC code by default for Bonaire, Kabini, Kaveri, and Mullins as certain hardware interfaces are not supported in the DC code. Your cpu (from dmesg) is an AMD A8-7650K Radeon R7, codename Kaveri.

The clue in dmesg output is the following errors:

[   14.186318] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:1! type 0 expected 3
[   14.186458] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:3! type 0 expected 3
[   14.186592] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:4! type 0 expected 3
[   14.186725] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:5! type 0 expected 3

You might try forcing amdgpu.dc=0 in the kernel command line to ensure you use the old driver code.

graaskaeg

(In reply to Pontus Gråskæg from comment #9)
> Looking at your dmesg output, it looks like you might have hit this
> bug/issue: bug 105880
> 
> "[dc] No support for LVDS or VGA connectors (Cannot find any crtc or sizes)"
> 
> AMD/Harry Wentland/Alex Deucher have commits in the mainline kernel to
> disable the DC code by default for Bonaire, Kabini, Kaveri, and Mullins as
> certain hardware interfaces are not supported in the DC code. Your cpu (from
> dmesg) is an AMD A8-7650K Radeon R7, codename Kaveri.

Unless you are using VGA or LVDS connectors this is a different issues.  That does not seem to be the case here (HDMI and DVI are supported in the DC code).

This is likely a duplicate of bug 106940.

Does reverting f0c0761b38ac30b04d4fed436ff10e894ec0e525 help?

OK, I was misled by the errors in the dmesg log:

[   14.186318] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:1! type 0 expected 3
[   14.186458] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:3! type 0 expected 3
[   14.186592] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:4! type 0 expected 3
[   14.186725] [drm:construct [amdgpu]] *ERROR* construct: Invalid Connector ObjectID from Adapter Service for connector index:5! type 0 expected 3

Apologies for the misconception.

graaskaeg

-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/amd/issues/436.