[Mesa-dev] [Mesa-stable] [PATCH 1/2] i965: Skip allocating UNIFORM file storage for uniforms of size 0.

[Ian Romanick]

On 09/11/2014 10:07 PM, Kenneth Graunke wrote:
> Samplers take up zero slots and therefore don't exist in the params
> array, nor are they included in stage_prog_data->nr_params.  There's no
> need to store their size in param_size, as it's only used for dealing
> with arrays of "real" uniforms (ones uploaded as shader constants).
> 
> We run into all kinds of problems trying to refer to the uniform storage
> for variables that don't have uniform storage.  For one, we may use some
> other variable's index, or access out of bounds in arrays.  In the FS
> backend, our extra 2 * MaxSamplerImageUnits params for texture rectangle
> rescaling paper over a lot of problems.  In the VS backend, we claim
> samplers take up a slot, which also papers over problems.
> 
> Instead, just skip allocating storage for variables that don't have any.

At least until we get to bindless, this seems right.  Are there any bad
interactions with ARB_gpu_shader5 dynamic index of sampler arrays?

> Signed-off-by: Kenneth Graunke <kenneth at whitecape.org>
> Cc: mesa-stable at lists.freedesktop.org
> ---
>  src/mesa/drivers/dri/i965/brw_fs_visitor.cpp   | 6 +++---
>  src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp | 6 +++---
>  2 files changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
> index 5d2e7c8..2d5318a 100644
> --- a/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_fs_visitor.cpp
> @@ -109,10 +109,10 @@ fs_visitor::visit(ir_variable *ir)
>         * ir_binop_ubo_load expressions and not ir_dereference_variable for UBO
>         * variables, so no need for them to be in variable_ht.
>         *
> -       * Atomic counters take no uniform storage, no need to do
> -       * anything here.
> +       * Some uniforms, such as samplers and atomic counters, have no actual
> +       * storage, so we should ignore them.
>         */
> -      if (ir->is_in_uniform_block() || ir->type->contains_atomic())
> +      if (ir->is_in_uniform_block() || type_size(ir->type) == 0)
>           return;
>  
>        if (dispatch_width == 16) {
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> index 1e823da..d504e2e 100644
> --- a/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp
> @@ -1023,10 +1023,10 @@ vec4_visitor::visit(ir_variable *ir)
>         * ir_binop_ubo_load expressions and not ir_dereference_variable for UBO
>         * variables, so no need for them to be in variable_ht.
>         *
> -       * Atomic counters take no uniform storage, no need to do
> -       * anything here.
> +       * Some uniforms, such as samplers and atomic counters, have no actual
> +       * storage, so we should ignore them.
>         */
> -      if (ir->is_in_uniform_block() || ir->type->contains_atomic())
> +      if (ir->is_in_uniform_block() || type_size(ir->type) == 0)
>           return;
>  
>        /* Track how big the whole uniform variable is, in case we need to put a
>