Bug 93050 - Amdgpu, Tonga "IO_PAGE_FAULT" and "[amdgpu]] *ERROR* amdgpu: ring 0 test failed" result in Kernel Panic
Summary: Amdgpu, Tonga "IO_PAGE_FAULT" and "[amdgpu]] *ERROR* amdgpu: ring 0 test fail...
Status: RESOLVED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/AMDgpu (show other bugs)
Version: DRI git
Hardware: x86-64 (AMD64) Linux (All)
: high critical
Assignee: Default DRI bug account
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-11-20 21:35 UTC by jk
Modified: 2017-06-25 18:34 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:

Attachments
Logfile (169.58 KB, text/plain)
2015-11-20 21:35 UTC, jk 		no flags Details
View All

Description jk 2015-11-20 21:35:35 UTC 
Created attachment 120000 [details]
Logfile

Switched from catalyst to amdgpu, Kernel 4.2 and Kernel 4.3 with powerplay fixes crash, wenn modesetting is enabled (which just has to be, to get everything working..) 
When booting without modesetting no Kernel Panic occurs, but just nothing else works really good...

I'm using Archlinux with mesa/llvm etc build from git.
Kernel 4.2 is Archlinux vanilla kernel, 4.3 is build from http://cgit.freedesktop.org/~agd5f/linux/?h=amdgpu-powerplay

Log files are attached.I've tried with disabled/enabled Iommu, it think this Log is with disabled Iommu.
Comment 1 Michel Dänzer 2015-11-26 08:39:54 UTC 
I also get the I/O page faults with my Tonga, but AFAICT only when it's the primary GPU, not when it's the secondary one. Also, the ring test doesn't fail and acceleration works fine for me in either case, so I think the I/O page faults and the ring test failure are two separate issues and should be tracked separately.
Comment 2 Alexander Tsoy 2016-07-29 16:55:58 UTC 
These I/O page faults errors are quite annoying and delay boot for about 5 seconds. Should I open a separate bug report?
Comment 3 stsp2 2017-03-15 13:58:39 UTC 
(In reply to Alexander Tsoy from comment #2)
> These I/O page faults errors are quite annoying and delay boot for about 5
> seconds. Should I open a separate bug report?

IMHO these should be reported.
I had such boot-up delays due to iommu page_fault flood
on HD 3870 and they are quite annoying.
Comment 4 Vedran Miletić 2017-03-20 15:28:26 UTC 
(In reply to jk from comment #0)
> Created attachment 120000 [details]
> Logfile
> 
> Switched from catalyst to amdgpu, Kernel 4.2 and Kernel 4.3 with powerplay
> fixes crash, wenn modesetting is enabled (which just has to be, to get
> everything working..) 
> When booting without modesetting no Kernel Panic occurs, but just nothing
> else works really good...
> 
> I'm using Archlinux with mesa/llvm etc build from git.
> Kernel 4.2 is Archlinux vanilla kernel, 4.3 is build from
> http://cgit.freedesktop.org/~agd5f/linux/?h=amdgpu-powerplay
> 
> Log files are attached.I've tried with disabled/enabled Iommu, it think this
> Log is with disabled Iommu.

jk, did situation improve with newer kernels?
Comment 5 jk 2017-06-25 18:34:13 UTC 
Fixed with newer Kernels

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.