Bug 26185 - [M96==RV730][KMS] Black screen after resume from RAM
Summary: [M96==RV730][KMS] Black screen after resume from RAM
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Radeon (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact:
Depends on:
Reported: 2010-01-23 05:13 UTC by Rafał Miłecki
Modified: 2010-01-26 03:24 UTC (History)
0 users

See Also:
i915 platform:
i915 features:

dmesg (65.28 KB, text/plain)
2010-01-23 05:14 UTC, Rafał Miłecki
no flags Details
regs dump before suspend (152.06 KB, text/plain)
2010-01-23 05:14 UTC, Rafał Miłecki
no flags Details
regs dump after resume (152.06 KB, text/plain)
2010-01-23 05:15 UTC, Rafał Miłecki
no flags Details
diff -u init3.after.log init3.before.log (14.34 KB, text/plain)
2010-01-23 05:19 UTC, Rafał Miłecki
no flags Details

Description Rafał Miłecki 2010-01-23 05:13:37 UTC
Resume on this machine works fine when I boot with init=/bin/bash or if I boot to init 3 and blacklist radeon module.

However when booting into init 3 with radeon module loaded resume works partially.

When resuming from RAM I can see console for half a second and then it disappears. In that half a second backlight seems to be lower as well, which is fine for me, but maybe will give you some idea?
Comment 1 Rafał Miłecki 2010-01-23 05:14:02 UTC
Created attachment 32783 [details]
Comment 2 Rafał Miłecki 2010-01-23 05:14:35 UTC
Created attachment 32784 [details]
regs dump before suspend
Comment 3 Rafał Miłecki 2010-01-23 05:15:06 UTC
Created attachment 32785 [details]
regs dump after resume

got it by ssh
Comment 4 Rafał Miłecki 2010-01-23 05:19:12 UTC
Created attachment 32786 [details]
diff -u init3.after.log init3.before.log
Comment 5 Rafał Miłecki 2010-01-23 05:21:36 UTC
Is this possible that just after resume radeon loads everything fine, but then some other module overwrites it's work?

On the other hand keep in mind that blacklisting radeon makes resume work fine... Weird.

Can I provide more info for debugging it?
Comment 6 Rafał Miłecki 2010-01-23 05:22:20 UTC
Oh, what I use is:

commit c8c15ff1e90bfc4a2db1ba77a01b3b2783e723fc
Author: Jerome Glisse <jglisse@redhat.com>
Date:   Mon Jan 18 13:01:36 2010 +0100

drm/radeon: r6xx/r7xx possible security issue, system ram access
Comment 7 Rafał Miłecki 2010-01-25 09:43:37 UTC
Hm, the same bug happens on different notebook with RV620... Maybe it's something in openSUSE that causes this? Both run this distribution.
Comment 8 Rafał Miłecki 2010-01-26 03:24:01 UTC
The problem is that pm-suspend (s2ram?) uses quriks even for KMS. It tries to restore GPU state in KMS, where it's already handled by KMS GPU driver.

Forcing pm-suspend to do not use quirks fixes problem.

See bug #26235 for details

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.