Bug 106291 - amdgpu fails GPU reset when resuming from suspend
Summary: amdgpu fails GPU reset when resuming from suspend
Status: RESOLVED OBSOLETE
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: 2015-10-19 20:13 UTC by universaledge97
Modified: 2016-05-01 16:35 UTC (History)
0 users

See Also:
Kernel Version: 4.2.3
Subsystem:
Regression: No
Bisected commit-id:

Attachments
output of "dmesg | grep amdgpu" (10.37 KB, application/octet-stream)
2015-10-19 20:13 UTC, universaledge97 		Details
output of dmesg (175.93 KB, text/x-log)
2015-10-20 14:06 UTC, universaledge97 		Details
dmesg from failed 4.3rc6 resume (72.29 KB, text/x-log)
2015-10-25 03:18 UTC, universaledge97 		Details
dmesg from successful 4.2.3 resume (79.00 KB, text/x-log)
2015-10-25 03:19 UTC, universaledge97 		Details
dmesg 4.3.0 - Xorg hang (110.83 KB, application/octet-stream)
2015-11-11 17:26 UTC, universaledge97 		Details
Show Obsolete (1) Add an attachment (proposed patch, testcase, etc.)

Description universaledge97 2015-10-19 20:13:28 UTC 
Created attachment 190561 [details]
output of "dmesg | grep amdgpu"

Rarely, when resuming from suspend, amdgpu will fail to reset the GPU, causing X.org to crash. This does have any particular cause I can identify. Any variety of programs could be open when this error occurs. Attached is a log of the output of "dmesg | grep amdgpu", although I have the full ouput from dmesg if needed.

Running XFX R9 285 Tonga
Comment 1 Michel Dänzer 2015-10-20 09:50:22 UTC 
Please attach the full dmesg output.
Comment 2 universaledge97 2015-10-20 14:06:19 UTC 
Created attachment 190631 [details]
output of dmesg

here is the full dmesg log
Comment 3 Michel Dänzer 2015-10-21 01:29:53 UTC 
Does this still happen with a 4.3(-rc) kernel?
Comment 4 universaledge97 2015-10-25 03:18:51 UTC 
Created attachment 191041 [details]
dmesg from failed 4.3rc6 resume
Comment 5 universaledge97 2015-10-25 03:19:50 UTC 
Created attachment 191051 [details]
dmesg from successful 4.2.3 resume
Comment 6 universaledge97 2015-10-25 03:20:25 UTC 
Sorry for the delay; not much time to debug when in college.

I am having an issue with suspend on 4.3rc6, but instead of amdgpu failing to reset, I simply experience an uninteractive black screen on resume everytime. Not sure if the two issues are in any way related. Accessing the machine through SSH, I was able to grab dmesg output when it happened.

For comparison, I also grabbed dmesg from a successful 4.2.3 suspend-resume.
Comment 7 Michel Dänzer 2015-10-27 03:43:19 UTC 
The log from 4.3-rc6 doesn't show any trace of the amdgpu driver. Did you accidentally disable it when compiling that kernel?
Comment 8 universaledge97 2015-10-27 15:53:38 UTC 
It turned out that the amdgpu kernel module was not included in compilation. I normally use the linux-mainline package from the AUR (I run Arch Linux) to install rc kernels, where, for some reason, the amdgpu module was not enabled to compile by default. I now have amdgpu built into 4.3rc7, and will report back if the issue is still present.
Comment 9 universaledge97 2015-11-11 17:26:38 UTC 
Created attachment 192801 [details]
dmesg 4.3.0 - Xorg hang

I ended up having to reinstall Arch Linux after my gpg directory was corrupted when the system auto-suspended in the middle of a package upgrade.

Running Linux 4.3.0 from Arch Linux testing repo: I'm not getting the GPU reset bug anymore; rather Xorg just randomly hangs indefinitely after waking from suspend, resulting in an uninteractive black screen. Curiously, there's no Xorg crash log. The dmesg log provides a call trace from before the hang.
Comment 10 universaledge97 2016-05-01 16:35:34 UTC 
Running 4.5.1 stable, no trace of the former problem. I guess it was resolved at some point.
Note You need to log in before you can comment on or make changes to this bug.