AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

intel/fs: Fix bug in lower_simd_width while splitting an instruction which was already split.

Browse Source 
This seems to be a problem in combination with the lower_regioning
pass introduced by a future commit, which can modify a SIMD-split
instruction causing its execution size to become illegal again.  A
subsequent call to lower_simd_width() would hit this bug on a future
platform.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
This commit is contained in:
Francisco Jerez
2018-12-07 14:15:50 -08:00
parent 812ede088f
commit bc781a0323
1 changed files with 5 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/intel/compiler/brw_fs.cpp
+5 -2
View File
@@ -5669,8 +5669,10 @@ needs_src_copy(const fs_builder &lbld, const fs_inst *inst, unsigned i)
static fs_reg
emit_unzip(const fs_builder &lbld, fs_inst *inst, unsigned i)
{
assert(lbld.group() >= inst->group);
/* Specified channel group from the source region. */
const fs_reg src = horiz_offset(inst->src[i], lbld.group());
const fs_reg src = horiz_offset(inst->src[i], lbld.group() - inst->group);
if (needs_src_copy(lbld, inst, i)) {
/* Builder of the right width to perform the copy avoiding uninitialized
@@ -5759,9 +5761,10 @@ emit_zip(const fs_builder &lbld_before, const fs_builder &lbld_after,
{
assert(lbld_before.dispatch_width() == lbld_after.dispatch_width());
assert(lbld_before.group() == lbld_after.group());
assert(lbld_after.group() >= inst->group);
/* Specified channel group from the destination region. */
const fs_reg dst = horiz_offset(inst->dst, lbld_after.group());
const fs_reg dst = horiz_offset(inst->dst, lbld_after.group() - inst->group);
const unsigned dst_size = inst->size_written /
inst->dst.component_size(inst->exec_size);