Bug 56038

Summary: "r100_ring_test] *ERROR*" messages sometimes on boot with ati radeon 9000 AGP ( r200 driver, dri2 KMS enabled )
Product: DRI Reporter: Barto <mister.freeman>
Component: DRM/RadeonAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED NOTABUG QA Contact:
Severity: normal    
Priority: medium    
Version: XOrg git   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg output
none
dmesg error with radeon.agpmode=-1 none

Description Barto 2012-10-16 17:12:21 UTC 
Created attachment 68625 [details]
dmesg output

randomly ( it's very rare, 90% of the time I don't have this error message ) I can see these error messages on boot :

[    1.516518] [drm] Loading R200 Microcode
[    1.517953] [drm] radeon: ring at 0x00000000E0001000
[    1.665941] [drm:r100_ring_test] *ERROR* radeon: ring test failed (scratch(0x15E4)=0xCAFEDEAD)
[    1.665996] [drm:r100_cp_init] *ERROR* radeon: cp isn't working (-22).
[    1.666043] radeon 0000:01:00.0: failed initializing CP (-22).
[    1.666084] radeon 0000:01:00.0: Disabling GPU acceleration
[    1.812455] [drm:r100_cp_fini] *ERROR* Wait for CP idle timeout, shutting down CP.
[    1.812600] [drm] radeon: cp finalized
[    1.812629] [drm] radeon: cp finalized

but the boot process continues and I see no problem on KDE4.9.2 and 3d games,

this bug occurs since the update of ati-dri 9.0 and xorg packages 1.13 on my archlinux 32 bits installation, before that I had never seen this kind of error message

my system :

- archlinux 32 bits
- laptop pentium 4 2.4 ghz 1,2 Gb ram
- video card: ati radeon 9000 agp ( M9 rv250, r200 driver, dri2 enabled with kms early start )
- SIS645DX chipset on motherboard

You can fin an attachment ( dmesg output ) with this message,

this bug may related to another ( startx problems ) :

https://bugs.freedesktop.org/show_bug.cgi?id=55982
Comment 1 Michel Dänzer 2012-10-17 15:19:38 UTC 
(In reply to comment #1)
> randomly ( it's very rare, 90% of the time I don't have this error message )
> I can see these error messages on boot :

Does radeon.agpmode=-1 (or not passing radeon.agpmode on the kernel command line in the first place) prevent it from happening?


> but the boot process continues and I see no problem on KDE4.9.2 and 3d games,

After the cascaded failure shown in the attached dmesg output, there is no 3D or even 2D hardware acceleration...
Comment 2 Barto 2012-10-17 22:03:27 UTC 
(In reply to comment #1)
> Does radeon.agpmode=-1 (or not passing radeon.agpmode on the kernel command
> line in the first place) prevent it from happening?
> 

If I set the "radeon.agpmode" to "-1" I get low video performance and sometimes the error message ( ring test *error* ),

If I remove the "radeon.agmode" setting I have no problem, the agp setting is set automatically to 4x ( according to the dmesg output ), 

but as the bug occurs randomly ( 10% of the time ) I don't know if removing the "radeon.agmode" setting can fix the problem


> After the cascaded failure shown in the attached dmesg output, there is no
> 3D or even 2D hardware acceleration...

yes you are right, when I type glxinfo I can see that dri2 is not used when the bug occurs
Comment 3 Barto 2012-10-17 22:10:58 UTC 
Created attachment 68727 [details]
dmesg error with radeon.agpmode=-1
Comment 4 Barto 2012-10-17 22:12:08 UTC 
(In reply to comment #2)
> If I set the "radeon.agpmode" to "-1" I get low video performance and
> sometimes the error message ( ring test *error* ),

in fact it's not the same error with "radeon.agpmode=-1", "ring test is ok" but there is a different error :

 12.013412] [drm:radeon_ib_ring_tests] *ERROR* radeon: failed testing IB on GFX ring (-35).
[   12.013458] [drm:radeon_device_init] *ERROR* ib ring test failed (-35).
[   12.016565] [drm] Panel ID String: 1024x768                
[   12.016571] [drm] Panel Size 1024x768

you can see the attachement for more details ( dmesg output )
Comment 5 Barto 2012-10-27 09:30:28 UTC 
the bug seems to be gone since I flashed the bios of my laptop, so I consider this problem is solved

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.