Skip to content

intel/fs: fix lower_simd_width for MOV_INDIRECT

What does this MR do and why?

intel/fs: fix lower_simd_width for MOV_INDIRECT

MOV_INDIRECT picks one lane from the src[0] and moves it to all lanes
in the destination. Even if we split the instruction in 2, all sources
should remain identical.

Noticed this while trying to use this instruction in SIMD32. All
current use cases are limited to SIMD8 shaders (or SIMD16 on Xe2).
That's why we've never seen the issue I think.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable

Merge request reports