fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2025-12-21 09:20:12 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

i965: add double support packing support to gs inputs

Browse source 
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
This commit is contained in:
Timothy Arceri 2016-06-10 14:23:58 +10:00
parent 20e935e6f6
commit 8b80e9c31d
1 changed files with 3 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
src/mesa/drivers/dri/i965/brw_fs_nir.cpp
View file

@ -2127,6 +2127,7 @@ fs_visitor::emit_gs_input_load(const fs_reg &dst,
} }
fs_reg tmp = fs_reg(VGRF, alloc.allocate(4), dst.type); fs_reg tmp = fs_reg(VGRF, alloc.allocate(4), dst.type);
tmp_dst = tmp; tmp_dst = tmp;
first_component = first_component / 2;
} }
for (unsigned iter = 0; iter < num_iterations; iter++) { for (unsigned iter = 0; iter < num_iterations; iter++) {
@ -2136,7 +2137,7 @@ fs_visitor::emit_gs_input_load(const fs_reg &dst,
unsigned read_components = num_components + first_component; unsigned read_components = num_components + first_component;
fs_reg tmp = bld.vgrf(dst.type, read_components); fs_reg tmp = bld.vgrf(dst.type, read_components);
inst = bld.emit(SHADER_OPCODE_URB_READ_SIMD8, tmp, icp_handle); inst = bld.emit(SHADER_OPCODE_URB_READ_SIMD8, tmp, icp_handle);
inst->regs_written = read_components; inst->regs_written = read_components * type_sz(tmp_dst.type) / 4;
for (unsigned i = 0; i < num_components; i++) { for (unsigned i = 0; i < num_components; i++) {
bld.MOV(offset(tmp_dst, bld, i), bld.MOV(offset(tmp_dst, bld, i),
offset(tmp, bld, i + first_component)); offset(tmp, bld, i + first_component));
@ -2158,7 +2159,7 @@ fs_visitor::emit_gs_input_load(const fs_reg &dst,
if (first_component != 0) { if (first_component != 0) {
inst = bld.emit(SHADER_OPCODE_URB_READ_SIMD8_PER_SLOT, tmp, inst = bld.emit(SHADER_OPCODE_URB_READ_SIMD8_PER_SLOT, tmp,
payload); payload);
inst->regs_written = read_components; inst->regs_written = read_components * type_sz(tmp_dst.type) / 4;
for (unsigned i = 0; i < num_components; i++) { for (unsigned i = 0; i < num_components; i++) {
bld.MOV(offset(tmp_dst, bld, i), bld.MOV(offset(tmp_dst, bld, i),
offset(tmp, bld, i + first_component)); offset(tmp, bld, i + first_component));