[Mesa-dev] [PATCH 19/21] intel/nir: Use the correct indirect lowering masks in link_shaders
Timothy Arceri
tarceri at itsqueeze.com
Fri Nov 3 11:04:47 UTC 2017
On 29/10/17 05:36, Jason Ekstrand wrote:
> Previously, if we were linking a vec4 VS with a SIMD8/16 FS, we wouldn't
> lower indirects on the fragment shader which is wrong. Instead of using
> a single indirect mask, take advantage of our new little helper.
It took me a while to actually see the bug here, but yes this was silly
oversight. I suspect this might be the cause of [1].
17-19:
Reviewed-by: Timothy Arceri <tarceri at itsqueeze.com>
Would be great if you could push these asap.
[1] https://bugs.freedesktop.org/show_bug.cgi?id=103537
>
> Cc: mesa-stable at lists.freedesktop.org
> ---
> src/intel/compiler/brw_nir.c | 10 ++++------
> 1 file changed, 4 insertions(+), 6 deletions(-)
>
> diff --git a/src/intel/compiler/brw_nir.c b/src/intel/compiler/brw_nir.c
> index 9a085ea5..4d7d01a 100644
> --- a/src/intel/compiler/brw_nir.c
> +++ b/src/intel/compiler/brw_nir.c
> @@ -677,16 +677,14 @@ brw_nir_link_shaders(const struct brw_compiler *compiler,
> NIR_PASS_V(*producer, nir_lower_global_vars_to_local);
> NIR_PASS_V(*consumer, nir_lower_global_vars_to_local);
>
> - nir_variable_mode indirect_mask = (nir_variable_mode) 0;
> - if (compiler->glsl_compiler_options[(*producer)->info.stage].EmitNoIndirectTemp)
> - indirect_mask = nir_var_local;
> -
> /* The backend might not be able to handle indirects on
> * temporaries so we need to lower indirects on any of the
> * varyings we have demoted here.
> */
> - NIR_PASS_V(*producer, nir_lower_indirect_derefs, indirect_mask);
> - NIR_PASS_V(*consumer, nir_lower_indirect_derefs, indirect_mask);
> + NIR_PASS_V(*producer, nir_lower_indirect_derefs,
> + brw_nir_no_indirect_mask(compiler, (*producer)->info.stage));
> + NIR_PASS_V(*consumer, nir_lower_indirect_derefs,
> + brw_nir_no_indirect_mask(compiler, (*consumer)->info.stage));
>
> const bool p_is_scalar =
> compiler->scalar_stage[(*producer)->info.stage];
>
More information about the mesa-dev
mailing list