OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

nir: Check against combined alignment in nir_lower_mem_access_bit_sizes

Browse Source 
Checking against align_mul is insufficient if align_offset > 0.  We need
to check against the combined alignment instead.

Fixes: 2e2d7803c7 ("nir: Add a load/store bit size lowering pass")
Reviewed-by: M Henning <drawoc@darkrefraction.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/21524>
This commit is contained in:
Faith Ekstrand
2023-02-24 14:26:44 -06:00
committed by Marge Bot
parent 352fee176c
commit 4b06b1a7c5
1 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/compiler/nir/nir_lower_mem_access_bit_sizes.c
View File

@@ -82,6 +82,7 @@ lower_mem_load(nir_builder *b, nir_intrinsic_instr *intrin,
const unsigned bytes_read = num_components * (bit_size / 8); const unsigned bytes_read = num_components * (bit_size / 8);
const uint32_t align_mul = nir_intrinsic_align_mul(intrin); const uint32_t align_mul = nir_intrinsic_align_mul(intrin);
const uint32_t align_offset = nir_intrinsic_align_offset(intrin); const uint32_t align_offset = nir_intrinsic_align_offset(intrin);
const uint32_t align = nir_intrinsic_align(intrin);
nir_src *offset_src = nir_get_io_offset_src(intrin); nir_src *offset_src = nir_get_io_offset_src(intrin);
const bool offset_is_const = nir_src_is_const(*offset_src); const bool offset_is_const = nir_src_is_const(*offset_src);
assert(offset_src->is_ssa); assert(offset_src->is_ssa);
@@ -96,7 +97,7 @@ lower_mem_load(nir_builder *b, nir_intrinsic_instr *intrin,
assert(util_is_power_of_two_nonzero(requested.align_mul)); assert(util_is_power_of_two_nonzero(requested.align_mul));
if (requested.num_components == num_components && if (requested.num_components == num_components &&
requested.bit_size == bit_size && requested.bit_size == bit_size &&
requested.align_mul <= align_mul) requested.align_mul <= align)
return false; return false;
/* Otherwise, we have to break it into chunks. We could end up with as /* Otherwise, we have to break it into chunks. We could end up with as
@@ -208,6 +209,7 @@ lower_mem_store(nir_builder *b, nir_intrinsic_instr *intrin,
const unsigned bytes_written = num_components * byte_size; const unsigned bytes_written = num_components * byte_size;
const uint32_t align_mul = nir_intrinsic_align_mul(intrin); const uint32_t align_mul = nir_intrinsic_align_mul(intrin);
const uint32_t align_offset = nir_intrinsic_align_offset(intrin); const uint32_t align_offset = nir_intrinsic_align_offset(intrin);
const uint32_t align = nir_intrinsic_align(intrin);
nir_src *offset_src = nir_get_io_offset_src(intrin); nir_src *offset_src = nir_get_io_offset_src(intrin);
const bool offset_is_const = nir_src_is_const(*offset_src); const bool offset_is_const = nir_src_is_const(*offset_src);
assert(offset_src->is_ssa); assert(offset_src->is_ssa);
@@ -225,7 +227,7 @@ lower_mem_store(nir_builder *b, nir_intrinsic_instr *intrin,
assert(util_is_power_of_two_nonzero(requested.align_mul)); assert(util_is_power_of_two_nonzero(requested.align_mul));
if (requested.num_components == num_components && if (requested.num_components == num_components &&
requested.bit_size == bit_size && requested.bit_size == bit_size &&
requested.align_mul <= align_mul && requested.align_mul <= align &&
writemask == BITFIELD_MASK(num_components)) writemask == BITFIELD_MASK(num_components))
return false; return false;