Summary: | [NVCF] X loses resources after suspend/resume | ||
---|---|---|---|
Product: | xorg | Reporter: | Daniel Barkalow <barkalow> |
Component: | Driver/nouveau | Assignee: | Nouveau Project <nouveau> |
Status: | RESOLVED FIXED | QA Contact: | Xorg Project Team <xorg-team> |
Severity: | normal | ||
Priority: | medium | ||
Version: | unspecified | ||
Hardware: | x86-64 (AMD64) | ||
OS: | Linux (All) | ||
Whiteboard: | |||
i915 platform: | i915 features: |
Description
Daniel Barkalow
2014-10-26 01:33:07 UTC
Could you try booting with nouveau.config=PCE1=0 And seeing if that helps? This sounds like some resource copies aren't working properly (either on suspend or on resume) and some GF116's have been known to be silently missing their second copy engine (or we're not doing something to properly enable or we're not looking in the right place for the disable or ... who knows). That seems to work, so I can just use that. Thanks! Let me know if you'd like me to test patches to handle this hardware automatically. Well, first I'd like to get information... basically we need to figure out how to figure out that the second copy engine is busted. (Or, conversely, how to un-bust it.) Can you install envytools (https://github.com/envytools/envytools) and let's start with some easy stuff... as root, get the output of: nvapeek 22400 200 I get: 00022400: 00000000 00000000 00000000 00000002 00022410: 00000000 00000000 30000000 00000000 00022420: 00000000 00000000 00000800 00000000 00022430: 00000001 00000004 00000003 00000000 ... After further investigation it appears that no GF116's actually contain a second "regular" copy engine, so it's been disabled completely. 3.18 should fix it and the patch is being backported to the stable trees. |
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.