Re: [PATCH] tools/null_state/gen9: Send atleast one valid component in VF state

[Ben Widawsky <benjamin.widawsky@intel.com>]

On Wed, Aug 05, 2015 at 11:13:46AM +0300, Mika Kuoppala wrote:
> Ben Widawsky <benjamin.widawsky@intel.com> writes:
> 
> > On Fri, Jul 31, 2015 at 04:27:07PM +0100, Arun Siluvery wrote:
> >> A programming restriction exists for this instruction, atleast one component
> >> of one valid vertex element must be enabled.
> >> 
> >> Cc: Ben Widawsky <benjamin.widawsky@intel.com>
> >> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> >> Signed-off-by: Arun Siluvery <arun.siluvery@linux.intel.com>
> >> ---
> >>  tools/null_state_gen/intel_renderstate_gen9.c | 7 ++++++-
> >>  1 file changed, 6 insertions(+), 1 deletion(-)
> >> 
> >> diff --git a/tools/null_state_gen/intel_renderstate_gen9.c b/tools/null_state_gen/intel_renderstate_gen9.c
> >> index 6f808f8..b3766ea 100644
> >> --- a/tools/null_state_gen/intel_renderstate_gen9.c
> >> +++ b/tools/null_state_gen/intel_renderstate_gen9.c
> >> @@ -440,7 +440,12 @@ int gen9_setup_null_render_state(struct intel_batchbuffer *batch)
> >>  	/* Vertex buffers */
> >>  	gen8_emit_vertex_buffers(batch);
> >>  	gen8_emit_vertex_elements(batch);
> >> -	OUT_CMD(GEN9_3DSTATE_COMPONENT_PACKING, 5);
> >> +
> >> +	OUT_BATCH(GEN9_3DSTATE_COMPONENT_PACKING | 3);
> >> +	OUT_BATCH(1);
> >> +	OUT_BATCH(0);
> >> +	OUT_BATCH(0);
> >> +	OUT_BATCH(0);
> >>  
> >>  	OUT_BATCH(GEN6_3DSTATE_VF_STATISTICS | 1 /* Enable */);
> >>  
> >
> > Like I said on IRC, I don't think this does anything unless you set bit 9 of
> > 3DSTATE_VF.0. Also, I believe you should be setting 0xf, not 1 since you'd want
> > to use all 4 components.
> >
> > So I'm not really sure what we're aiming to do. If you make it 0xf, and add a
> > comment about how this command doesn't do anything because we're not setting up
> > 3DSTATE_VF packing, it's:
> > Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
> >
> > I'm still more in favor of dropping the command altogether, that too would be:
> > Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
> 
> We are trying here just to satisfy the minimum requirements the bspec
> states.
> 

Right, except I am saying at least set minimal state which makes sense. Packing
1 component of the vertex element makes little sense (though doing this at all
makes even less sense).

> The bspec also states that with gen9+, the 3d pipeline state
> doesn't need to be initialized. We should seek confirmation
> about this and if it is so, remove the null state for gen9+.
> 

That would be super terrific.

> -Mika
> 
> >
> > -- 
> > Ben Widawsky, Intel Open Source Technology Center
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Ben Widawsky, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx