[drm:init_ring_common] *ERROR* render ring head not reset to zero

[drm:init_ring_common] *ERROR* render ring head not reset to zero

[Thomas Meyer]

This happens in 2.6.36-rcX:

[    0.523130] [drm] Initialized drm 1.1.0 20060810
[    0.523159] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[    0.523165] i915 0000:00:02.0: setting latency timer to 64
[    0.555999] [drm] detected 63M stolen memory, trimming to 32M
[    0.556065] i915 0000:00:02.0: irq 40 for MSI/MSI-X
[    0.556077] [drm] set up 32M of stolen space
[    0.556247] [drm:init_ring_common] *ERROR* render ring head not reset to zero ctl 00000000 head 02001000 tail 00000000 start 02001000
[    0.556320] [drm:init_ring_common] *ERROR* render ring head forced to zero ctl 00000000 head 00000000 tail 00000000 start 02001000

Bug or feature?

Re: [drm:init_ring_common] *ERROR* render ring head not reset to zero

[Chris Wilson]

On Mon, 4 Oct 2010 22:31:20 +0200, Thomas Meyer <thomas@m3y3r.de> wrote:
> [    0.556247] [drm:init_ring_common] *ERROR* render ring head not reset to zero ctl 00000000 head 02001000 tail 00000000 start 02001000
> [    0.556320] [drm:init_ring_common] *ERROR* render ring head forced to zero ctl 00000000 head 00000000 tail 00000000 start 02001000
> 
> Bug or feature?

Hardware feature, software bug. Could also be a feature of the software to
report a hardware bug...

At any rate, it shouldn't be happening and you are not alone at finding
such errors in your logs.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre

Re: [drm:init_ring_common] *ERROR* render ring head not reset to zero

[Sriram]

Chris Wilson <chris <at> chris-wilson.co.uk> writes:

> 
> On Mon, 4 Oct 2010 22:31:20 +0200, Thomas Meyer <thomas <at> m3y3r.de> wrote:
> > [    0.556247] [drm:init_ring_common] *ERROR* render ring head not reset to
zero ctl 00000000 head
> 02001000 tail 00000000 start 02001000
> > [    0.556320] [drm:init_ring_common] *ERROR* render ring head forced to
zero ctl 00000000 head 00000000
> tail 00000000 start 02001000
> > 
> > Bug or feature?
> 
> Hardware feature, software bug. Could also be a feature of the software to
> report a hardware bug...
> 
> At any rate, it shouldn't be happening and you are not alone at finding
> such errors in your logs.
> -Chris
> 


Hi Chris,

Please explain what this error means & why it is happening...

Re: [drm:init_ring_common] *ERROR* render ring head not reset to zero

[Chris Wilson]

On Sun, 5 Dec 2010 06:28:25 +0000 (UTC), Sriram <sri.kmb@gmail.com> wrote:
> Please explain what this error means & why it is happening...

We wrote a zero into the register (whilst the ring is completely idle,
supposedly anyway) and read back a non-zero value. We believe the silicon
is doing evil things, we workaround it by writing the value again, and
continue on our merry way.

The dance only gets more complex as the hardware evolves.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre