bjoernager/rust
bjoernager/rust
1
Fork 0
Code Issues Pull requests Activity

Rollup merge of #118864 - farnoy:masked-load-store-fixes, r=workingjubilee

Browse source 
Fix alignment passed down to LLVM for simd_masked_load

Follow up to #117953

The alignment for a masked load operation should be that of the element/lane, not the vector as a whole

It can produce miscompilations after the LLVM optimizer notices the higher alignment and promotes this to an unmasked, aligned load followed up by blend/select - https://rust.godbolt.org/z/KEeGbevbb
This commit is contained in:
Jubilee 2023-12-12 18:48:51 -08:00 • committed by GitHub
commit a33f1a3d3a
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

2
compiler/rustc_codegen_llvm/src/intrinsic.rs
View file

@ -1568,7 +1568,7 @@ fn generic_simd_intrinsic<'ll, 'tcx>(
// Alignment of T, must be a constant integer value:
let alignment_ty = bx.type_i32();
let alignment = bx.const_i32(bx.align_of(values_ty).bytes() as i32);
let alignment = bx.const_i32(bx.align_of(values_elem).bytes() as i32);
// Truncate the mask vector to a vector of i1s:
let (mask, mask_ty) = {