Bug 198713 - AMD DC crashes when computing clocks/detecting freesync
Summary: AMD DC crashes when computing clocks/detecting freesync
Status: NEW
Alias: None
Product: Drivers
Classification: Unclassified
Component: Video(DRI - non Intel) (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: drivers_video-dri
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-02-07 15:59 UTC by Jon
Modified: 2018-06-13 13:53 UTC (History)
5 users (show)

See Also:
Kernel Version: 4.15.1-7
Subsystem:
Regression: No
Bisected commit-id:


Attachments
dmesg output (138.85 KB, text/plain)
2018-02-07 15:59 UTC, Jon
Details
kernel config from /boot (195.11 KB, text/x-mpsub)
2018-02-07 16:00 UTC, Jon
Details
dmesg for kernel 4.17 showing warnings with traces (169.14 KB, text/plain)
2018-06-13 13:53 UTC, Jon
Details

Description Jon 2018-02-07 15:59:10 UTC
Created attachment 274053 [details]
dmesg output

Fury X  w/ BenQ XL2730Z connected with DisplayPort

amdgpu.dc=1 is NOT set.

See attached dmesg for 12 traces related to DC.

It looks to be crashing whilst detecting/computing clocks when freesync is detected. I've not set it to be enabled.
Comment 1 Jon 2018-02-07 16:00:03 UTC
Created attachment 274055 [details]
kernel config from /boot
Comment 2 Mike Lothian 2018-02-07 16:17:57 UTC
You don't need amdgpu.dc=1 set, it's already enabled in the .config that you provided:

CONFIG_DRM_AMD_DC = y
CONFIG_DRM_AMD_DC_PRE_VEGA = y

As a workaround try amdgpu.dc=0
Comment 3 Jon 2018-02-07 17:12:46 UTC
(In reply to Mike Lothian from comment #2)
> As a workaround try amdgpu.dc=0

Well spotted. Workaround has stopped the traces as you'd expect. It's also stopped the weird blue flicker I was getting in KDE with OGL3 compositor (different DC bug).

I hope the above is useful for getting PRE_VEGA out of experimental!


Missing from original report, it's 3x BenQ XL2730Z, all display port.
Comment 4 Nicholas Johnson 2018-06-11 08:28:42 UTC
I did not find this when posting my bug. This might be the same thing. Except mine is two whole kernel versions newer, and it is still doing it. AMD do not seem to be taking Linux seriously....

https://bugzilla.kernel.org/show_bug.cgi?id=199917

Could somebody who knows the ropes around here poke the AMD driver team for us? Thank you!
Comment 5 Andrey Grodzovsky 2018-06-12 20:45:19 UTC
It's not a crash but a warning over invalid reg fields access - here 
https://elixir.bootlin.com/linux/v4.15.1/source/drivers/gpu/drm/amd/display/dc/dm_services.h#L132

I see you are working with 4.15 kernel, can you take the latest stable from here - https://www.kernel.org/ and see if problem goes away ? 

For quicker response time you can try submitting tickets to Free Desktop Bugzilla - 
https://bugs.freedesktop.org/


Andrey
Comment 6 Nicholas Johnson 2018-06-13 02:28:37 UTC
If you are referring to my comment, mine was with 4.17 kernel (latest stable release). I was just commenting on this one (4.15) because it looks like the same bug, and I am highlighting that they still have not fixed it in 4.17.
Comment 7 Andrey Grodzovsky 2018-06-13 13:42:09 UTC
Well, the dmesg attached is also from 4.15 kernel so I assumed it's all 4.15. Anyway, I will Harry from our Display team to take a look at this.
Comment 8 Jon 2018-06-13 13:53:02 UTC
Updated to 4.17 as per (indirect) request :)


Linux mudkip.farm 4.17.1-6-default #1 SMP PREEMPT Tue Jun 12 09:55:31 UTC 2018 (e721478) x86_64 x86_64 x86_64 GNU/Linux
Comment 9 Jon 2018-06-13 13:53:55 UTC
Created attachment 276531 [details]
dmesg for kernel 4.17 showing warnings with traces

Note You need to log in before you can comment on or make changes to this bug.