[AMDGPU][CIK] Unable to hibernate functionality is not stable

Submitted by Shawn Starr

Assigned to Default DRI bug account

Link to original bug (#98513)

Description

Something regressed for CIK on AMDGPU, I am unable to hibernate anymore, hangs laptop when it attempts to freeze().

added AMDgpu bugzilla labels

Shawn Starr said:

Tried to hibernate again, total deadlocking of laptop.

Shawn Starr said:

Bisected results confirmed:

[spstarr@segfault linux]$ git bisect bad
86f8c599b09c916f9aad30563271440dbd79213a is the first bad commit
commit 86f8c599b09c916f9aad30563271440dbd79213a
Author: Rex Zhu Rex.Zhu@amd.com
Date: Mon Oct 3 20:46:36 2016 +0800

drm/amdgpu: when suspend, set boot state instand of disable dpm.

fix pm-hibernate bug, when suspend/resume, dpm start failed.

Signed-off-by: Rex Zhu Rex.Zhu@amd.com
Acked-by: Christian König christian.koenig@amd.com
Reviewed-by: Alex Deucher alexander.deucher@amd.com
Signed-off-by: Alex Deucher alexander.deucher@amd.com

:040000 040000 5537c5937b203abdb5c48c5a0dc8dc5f6a3d00f2 7f048011c861d8c35b922fd94f089cd07d670f60 M drivers
[spstarr@segfault linux]$ git bisect log
git bisect start

good: [1001354ca34179f3db924eb66672442a173147dc] Linux 4.9-rc1

git bisect good 1001354ca34179f3db924eb66672442a173147dc

bad: [2f5945e707b5dffbf12444ad8bea079626ad30ec] drm/amdgpu: add the interface of waiting multiple fences (v4)

git bisect bad 2f5945e707b5dffbf12444ad8bea079626ad30ec

good: [61d0a04d6f5b2122f88aacbc4b1716e571961660] Merge tag 'topic/drm-misc-2016-10-24' of git://anongit.freedesktop.org/drm-intel into drm-next

git bisect good 61d0a04d6f5b2122f88aacbc4b1716e571961660

good: [e4ab73a13291fc844c9e24d5c347bd95818544d2] drm/i915: Respect alternate_ddc_pin for all DDI ports

git bisect good e4ab73a13291fc844c9e24d5c347bd95818544d2

bad: [415282b15e15c2d7fb18e29c5b554cc7f4ff5c52] drm/amdgpu: disable dpm before turn off clock when vce idle.

git bisect bad 415282b15e15c2d7fb18e29c5b554cc7f4ff5c52

good: [392f0c775c80de0eae4c07227cc220015df70abc] drm/amdgpu/gfx8: cache rb config values

git bisect good 392f0c775c80de0eae4c07227cc220015df70abc

good: [472259f02657ef99cba2a64832ccadad8e3baabe] drm/amd/amdgpu: re-factor debugfs wave reader

git bisect good 472259f02657ef99cba2a64832ccadad8e3baabe

good: [f1e68a7cf582b41d6da1dd15b9f0bfb9057c1164] drm/amdgpu/atom: remove a bunch of unused functions

git bisect good f1e68a7cf582b41d6da1dd15b9f0bfb9057c1164

good: [585ffd65441a4aea7e762d17f7a248d07cd1c9ac] drm/ttm: fix coding style in ttm_bo_driver.h

git bisect good 585ffd65441a4aea7e762d17f7a248d07cd1c9ac

good: [8ed8147abc7cf1f689245deb316aabfe2f657ade] drm/amdgpu: use failed label to handle context init failure

git bisect good 8ed8147abc7cf1f689245deb316aabfe2f657ade

bad: [3f767e3d076dd2a24a614917c8f0b05d8d82b90b] drm/amdgpu: just not load smc firmware if smu is already running

git bisect bad 3f767e3d076dd2a24a614917c8f0b05d8d82b90b

bad: [86f8c599b09c916f9aad30563271440dbd79213a] drm/amdgpu: when suspend, set boot state instand of disable dpm.

git bisect bad 86f8c599b09c916f9aad30563271440dbd79213a

first bad commit: [86f8c599b09c916f9aad30563271440dbd79213a] drm/amdgpu: when suspend, set boot state instand of disable dpm.

Shawn Starr said:

Is this code path checking Disk hibernation aka freeze() / raw()? If this this just spend to RAM, then this might explain why this fails for me.

Shawn Starr said:

As noted on IRC 4.10-rc2 w/ drm-next-4.10-wip and 4.10-rc3 w/ drm-next-4.11-wip do not require this commit to be reverted.

Hibernate (freeze), Resume (thaw) seem to work fine here, keeping ticket open for further testing.

Shawn Starr said:

This broke in either Linus rc4 or drm-next-4.11-wip changes, will bisect this week

Shawn Starr said:

This currently works in 4.12.0-rc7+ w/ drm-next-4.13-wip, will try latest and close if works.

This issue hasn't had any activity since 2019-11-19. The AMD driver stack changes rapidly and contains lots of shared code across products so it's possible that it has already been fixed. Please upgrade to a current stable kernel and userspace stack and try again. If you still experience this issue with the latest driver stack, please capture relevant logging and open a new issue referring back to this one.

closed