[Mesa-dev] [PATCH] nvc0: do not bind input params at compute state init on Fermi

[Ilia Mirkin]

This seems surprising... could I convince you to trace a test that
executes both a graphics and compute pipeline, which both use
(different) uniforms?

Anyways, this patch is fine for now, this is Reviewed-by: Ilia Mirkin
<imirkin at alum.mit.edu>

On Sat, Oct 17, 2015 at 12:19 PM, Samuel Pitoiset
<samuel.pitoiset at gmail.com> wrote:
> It looks like binding a constant buffer on compute overwrites the 3D
> state. To avoid that, we already re-bind all the 3D constant buffers
> after launching a compute grid but this is not enough.
>
> Binding the constant buffer of input parameters for the compute state at
> initialization corrupts the 3D constant buffers, and it's just useless
> to bind it because this is not needed until we really launch a grid.
>
> This fixes some piglit regressions related to interpolation tests
> introduced in "nvc0: enable compute support by default on Fermi".
>
> Fixes: 00d6186 (nvc0: enable compute support by default on Fermi)
>
> Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
> ---
>  src/gallium/drivers/nouveau/nvc0/nvc0_compute.c | 8 --------
>  1 file changed, 8 deletions(-)
>
> diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_compute.c b/src/gallium/drivers/nouveau/nvc0/nvc0_compute.c
> index 96d753c..e33af04 100644
> --- a/src/gallium/drivers/nouveau/nvc0/nvc0_compute.c
> +++ b/src/gallium/drivers/nouveau/nvc0/nvc0_compute.c
> @@ -105,14 +105,6 @@ nvc0_screen_compute_setup(struct nvc0_screen *screen,
>     PUSH_DATAh(push, screen->text->offset);
>     PUSH_DATA (push, screen->text->offset);
>
> -   /* bind parameters buffer */
> -   BEGIN_NVC0(push, NVC0_COMPUTE(CB_SIZE), 3);
> -   PUSH_DATA (push, screen->parm->size);
> -   PUSH_DATAh(push, screen->parm->offset);
> -   PUSH_DATA (push, screen->parm->offset);
> -   BEGIN_NVC0(push, NVC0_COMPUTE(CB_BIND), 1);
> -   PUSH_DATA (push, (0 << 8) | 1);
> -
>     /* TODO: textures & samplers */
>
>     return 0;
> --
> 2.5.3
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev