Bug 104281

Summary: black / corrupted screen when resuming from S3 [AMDGPU]
Product: DRI Reporter: Carlo Caione <carlo>
Component: DRM/AMDgpuAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium CC: harry.wentland, jordan.lazare
Version: XOrg git   
Hardware: Other   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
journal_suspend_resume_corrupted_screen
none
journal corrupted screen amd-staging-drm-next none

Description Carlo Caione 2017-12-15 16:27:42 UTC
Created attachment 136199 [details]
journal_suspend_resume_corrupted_screen

Hardware is AMD Bristol Ridge chipset with AMD A10-9620P RADEON R5 SoC.

00:01.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Carrizo (rev ca)
03:00.0 Display controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 699f (rev ff)

Kernel is latest Linus master (4.15.0-rc3) compiled with CONFIG_DRM_AMD_DC_PRE_VEGA=y

When resuming from suspend most of the times the screen is black / corrupted. I can still SSH into the system.

In attachment the full log, including suspend + resume.
Comment 1 Carlo Caione 2017-12-15 16:29:08 UTC
The issue is not limited to this laptop model. We have at least 3 other ACER models affected by this.
Comment 2 Daniel Drake 2017-12-18 08:21:01 UTC
Suspend/resume is working OK on these machines with DC disabled, so it seems to be a DC regression.
Comment 3 Daniel Drake 2017-12-21 12:28:54 UTC
Need to do more testing to be sure, but it appears that this is not reproducible on the development branch https://cgit.freedesktop.org/~agd5f/linux/log/?h=amd-staging-drm-next
Comment 4 Harry Wentland 2017-12-21 14:57:42 UTC
Thanks, Daniel. The fix should land for 4.16 then.
Comment 5 Carlo Caione 2018-01-02 13:47:04 UTC
Created attachment 136493 [details]
journal corrupted screen amd-staging-drm-next

Unfortunately I just verified that this is still reproducible on the development branch (agd5f/amd-staging-drm-next), it's just a bit more difficult.

In attachment the log captured when the system is resumed and the screen is corrupted.
Comment 6 Carlo Caione 2018-01-26 10:17:06 UTC
Hey Harry,
any idea about this? This is still reproducible today on the agd5f/amd-staging-drm-next branch.

Thanks,
Comment 7 Alex Deucher 2018-01-26 23:08:18 UTC
Does it work with drm-next-4.17-wip?  There was a non-drm bug that broke S3 in the kernel what was not fixed until rc6.  amd-staging-drm-next is still based on rc4.
Comment 8 Carlo Caione 2018-01-28 16:43:23 UTC
(In reply to Alex Deucher from comment #7)
> Does it work with drm-next-4.17-wip?  There was a non-drm bug that broke S3
> in the kernel what was not fixed until rc6.  amd-staging-drm-next is still
> based on rc4.
Thank you Alex. I tried drm-next-4.17-wip and I wasn't able to reproduce again the issue. Do you have by any change a pointer to the bug/fix?

thanks!
Comment 9 Alex Deucher 2018-01-29 15:18:08 UTC
The commit that broke it was:

commit ca37e57bbe0cf1455ea3e84eb89ed04a132d59e1 (refs/bisect/bad)
Author: Andy Lutomirski <luto@kernel.org>
Date:   Wed Nov 22 20:39:16 2017 -0800

    x86/entry/64: Add missing irqflags tracing to native_load_gs_index()

I'm not sure what commit fixed it.
Comment 10 Harry Wentland 2018-04-24 18:53:52 UTC
Marking resolved as reported to work fine on drm-next-4.17-wip. If this is still an issue feel free to reopen.

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.