intel/fs: Use shuffle_from_32bit_read for 64-bit FS load_input

As the previous use of shuffle_32bit_load_result_to_64bit_data
had a source/destination overlap for 64-bit. Now a temporary destination
is used for 64-bit cases to use shuffle_from_32bit_read that doesn't
handle src/dst overlaps.

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
This commit is contained in:
Jose Maria Casanova Crespo
2018-06-09 11:46:11 +02:00
parent 8003ae87f4
commit 71b319a285

View File

@@ -3353,6 +3353,7 @@ fs_visitor::nir_emit_fs_intrinsic(const fs_builder &bld,
unsigned base = nir_intrinsic_base(instr); unsigned base = nir_intrinsic_base(instr);
unsigned comp = nir_intrinsic_component(instr); unsigned comp = nir_intrinsic_component(instr);
unsigned num_components = instr->num_components; unsigned num_components = instr->num_components;
fs_reg orig_dest = dest;
enum brw_reg_type type = dest.type; enum brw_reg_type type = dest.type;
/* Special case fields in the VUE header */ /* Special case fields in the VUE header */
@@ -3368,6 +3369,7 @@ fs_visitor::nir_emit_fs_intrinsic(const fs_builder &bld,
*/ */
type = BRW_REGISTER_TYPE_F; type = BRW_REGISTER_TYPE_F;
num_components *= 2; num_components *= 2;
dest = bld.vgrf(type, num_components);
} }
for (unsigned int i = 0; i < num_components; i++) { for (unsigned int i = 0; i < num_components; i++) {
@@ -3376,10 +3378,8 @@ fs_visitor::nir_emit_fs_intrinsic(const fs_builder &bld,
} }
if (nir_dest_bit_size(instr->dest) == 64) { if (nir_dest_bit_size(instr->dest) == 64) {
shuffle_32bit_load_result_to_64bit_data(bld, shuffle_from_32bit_read(bld, orig_dest, dest, 0,
dest, instr->num_components);
retype(dest, type),
instr->num_components);
} }
break; break;
} }