Bug 38977 - KMS unusable on R710
Summary: KMS unusable on R710
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-07-05 06:31 UTC by Micael Dias
Modified: 2012-04-25 13:46 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Kernel 2.39 log (with KMS) (125.63 KB, patch)
2011-07-05 06:31 UTC, Micael Dias
no flags Details | Splinter Review
kernel 3.0-git log (without KMS) (98.74 KB, patch)
2011-07-05 06:32 UTC, Micael Dias
no flags Details | Splinter Review
Picture illustrating screen corruption (275.42 KB, image/jpeg)
2011-07-05 06:41 UTC, Micael Dias
no flags Details

Description Micael Dias 2011-07-05 06:31:45 UTC
Created attachment 48773 [details] [review]
Kernel 2.39 log (with KMS)

When KMS starts, I get a screen full of garbage pixels that randomly updates itself when there are screen updates.
With kernel 2.39 computer doesn't lock and I'm able to access it through ssh, however with 3.0-git computer freezes during the boot process and I'm unable to get a log from the kernel.

I'm attaching a full log with KMS enabled (2.39) and with KMS disabled (3.0-git).

The system works fine without KMS.
Comment 1 Micael Dias 2011-07-05 06:32:41 UTC
Created attachment 48774 [details] [review]
kernel 3.0-git log (without KMS)
Comment 2 Micael Dias 2011-07-05 06:41:57 UTC
Created attachment 48775 [details]
Picture illustrating screen corruption

Note that during boot itself the image doesn't have this look of "vertical lines" but is completely random pixels, more like random noise.

I believe this "vertical line" pattern (ocurring after the desktop boots) is due to a browser window vertically maximized that I have running at startup.
Comment 3 Michel Dänzer 2011-07-07 07:30:24 UTC
Weird that it works with UMS but not with KMS...

Would radeon.no_wb=1 or radeon.gartsize=256 (or 128, ...) happen to help?
Comment 4 Micael Dias 2011-07-07 08:28:58 UTC
(In reply to comment #3)
> Weird that it works with UMS but not with KMS...
> 
> Would radeon.no_wb=1 or radeon.gartsize=256 (or 128, ...) happen to help?

Just tried with:
- radeon.no_wb=1
- radeon.gartsize=256
- radeon.no_wb=1 radeon.gartsize=128
- radeon.gartsize=128
- radeon.gartsize=64

None helped. dmesg still shows drm complaining about a ring test failed just like the logs attached.

I tried with another RV710 card (this time a "XFX 4550") (not working chip is a "EAH 4350") and the XFX card works fine. Very strange.

I have about one more week with this hardware so if there's something more I can do to help, I do have some time to spend on this and would enjoy doing so.
Comment 5 Alex Deucher 2012-04-25 06:12:19 UTC
Is this still an issue with a newer driver/kernel?  I think this commit may have fixed it:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=87463ff83bcda210d8f0ae440bd64d1548f852e7
Please reopen if it's still an issue.
Comment 6 Micael Dias 2012-04-25 13:46:27 UTC
I don't know if that's the commit that fixed it, but yes, it's working now.

Thanks


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.