Bug 25398

Summary: DRI not working with HD3870
Product: DRI Reporter: Christopher Schramm <freedesktop>
Component: DRM/RadeonAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: major    
Priority: medium    
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Xorg log with radeonhd from current git
none
Xorg log with radeon from current git none

Description Christopher Schramm 2009-12-02 09:53:32 UTC
I've got a Radeon HD3870 which I'm not able to use with DRI on my Debian system. First I didn't get anything but a black screen when starting X, until I changed Driver to radeonhd.

With Debian's radeonhd without DIR I get a working X but that's the only configuration that works.

Not working configurations include:

- Explicitly enabling DRI/EXA in xorg.conf

- Using Debian's radeon driver (since it has DRI enabled by default)

- Using current git version of xf86-video-ati

- Using current git version or official version 1.3.0 release of radeonhd driver which also brings default DRI


With all those setups I only get a black screen - sometimes with some fragments of what should show up - and a very juddering mouse cursor. I can't do nothing but reset the computer to get out of there. Using ssh I was able to figure out that Xorg uses 100% of the CPU.

As you can see in the log file attached, radeonhd gives continuous "(WW) RADEONHD(0): DRMCPIdle: DRM CP IDLE returned BUSY!" messages.

Any ideas how to debug?
Comment 1 Christopher Schramm 2009-12-02 09:54:48 UTC
Created attachment 31666 [details]
Xorg log with radeonhd from current git
Comment 2 Alex Deucher 2009-12-04 06:29:30 UTC
Can you attach your failing radeon log?
Comment 3 Christopher Schramm 2009-12-04 07:17:15 UTC
Created attachment 31742 [details] [review]
Xorg log with radeon from current git
Comment 4 Alex Deucher 2009-12-04 07:24:08 UTC
Does disconnecting the monitor on DVI-1 (the Panasonic) help?  There was a bug in the xserver that did not handle field/frame size properly in some cases which resulted in 1920x540 rather than 1080i.
Comment 5 Christopher Schramm 2009-12-04 23:31:04 UTC
No, that didn't change anything (with radeon from git). In the radeon log attached, I can see the bug you mentioned, when both are connected:

(II) RADEON(0): Output DVI-1 using initial mode 1920x540
(II) RADEON(0): Output DVI-0 using initial mode 1920x1080

But DVI-1 is the Panasonic which I don't use at the moment. I tried to get it running with xrandr yesterday, but I had the problem, that it had no modes. And adding and activating a new mode manually didn't show anything, although xrandr stated it would be active. Maybe that's somehow related to the DRI issues, but I think it's a whole different problem.

By the way: If upgraded the kernel to 2.6.32, so dri module version is now 1.31.0.

What could I try to update? The whole xorg system?

Or is there some way to get some more debug information about the DRMCPIdle problem?
Comment 6 Marek Olšák 2014-01-23 11:24:50 UTC
I'm pretty sure this has been fixed with KMS. I've got Radeon HD3870 too.

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.