Bug 81888 - [dri3] Freeze when screen switches back on
Summary: [dri3] Freeze when screen switches back on
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Chris Wilson
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-07-29 22:18 UTC by Mike Lothian
Modified: 2015-09-11 13:48 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:

Attachments
Dmesg (125.37 KB, text/plain)
2014-07-29 22:18 UTC, Mike Lothian 		no flags Details
Xorg log (39.73 KB, text/plain)
2014-07-29 22:18 UTC, Mike Lothian 		no flags Details
View All

Description Mike Lothian 2014-07-29 22:18:13 UTC 
Created attachment 103658 [details]
Dmesg

When the screen switches on after being off (DPMS?) lots of kernel warnings are produced and the screen freezes - the cursor still moves though

Before I could switch VTs several times and toggle compositing to bring back the session. With the latest stack I now have to restart X

I was using xorg-server from git but I've switched back to 1.16 - dri3 is enabled

I'm using linus's tree (drm-next won't boot at the moment and I've not bisected yet) 

libdrm and mesa are both from git as is xf86-video-intel

This is a sandybridge machine
Comment 1 Mike Lothian 2014-07-29 22:18:58 UTC 
Created attachment 103659 [details]
Xorg log
Comment 2 Chris Wilson 2014-07-30 06:05:18 UTC 
Disable DRI3, report back.
Comment 3 Chris Wilson 2014-07-30 06:07:25 UTC 
The warnings are interesting, but unrelated to the lack of display after resume.
Comment 4 Mike Lothian 2014-07-30 06:55:47 UTC 
I'm recompiling mesa without dri3 is there a way I can tell X not to use it?
Comment 5 Mike Lothian 2014-07-30 22:34:59 UTC 
I can confirm that compiling mesa without dri3 fixes the issue
Comment 6 Mike Lothian 2014-07-30 22:35:30 UTC 
Or more accurately works around the issue
Comment 7 Chris Wilson 2014-09-10 16:05:10 UTC 
Time to dig deeper, apparently. Could you reproduce this with an --enable-debug=full build so that we can see the what the waits are that causing the stall.
Comment 8 Mike Lothian 2015-09-11 13:48:53 UTC 
Sorry I didn't notice the reply to this last year and...

my laptop has went to the big enrichment centre in the sky

So closing this bug

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.