The hardware only supports 0..15 constant buffers, but
R600_GS_RING_CONST_BUFFER has index 16, which isn't really a constant
buffer. Maybe that's related to GS stability issues somehow?
Marek
On Wed, Nov 19, 2014 at 7:18 AM, Dave Airlie wrote:
> Hey,
>
> I got piglit to run on my laptop rv635,
>
On 19 November 2014 17:01, Benjamin Bellec wrote:
> Hello,
> Did you tried with R600_DEBUG=nosb ?
It won't fix any of the GS tests since they don't use sb at all,
I don't think it fixed the multisample tests either, and I tried a piglit
run with nosb and I feel over somewhere else.
Dave.
> Reg
Hello,
Did you tried with R600_DEBUG=nosb ?
Regards.
Le 19 nov. 2014 07:18, "Dave Airlie" a écrit :
> Hey,
>
> I got piglit to run on my laptop rv635,
>
> two things seem to hang it
>
> geometry shader texel fetches (wierd incorrect values for 2D, hangs for
> 2DMS)
> and
> the clip scissor msaa b
Hey,
I got piglit to run on my laptop rv635,
two things seem to hang it
geometry shader texel fetches (wierd incorrect values for 2D, hangs for 2DMS)
and
the clip scissor msaa blits.
The main wierdness is that texelFetch gs sampler2D is busted,
you get wierd black lines in the middle of the re
