[Mesa-dev] [PATCH] radv: emit correct last export when Z/stencil export is enabled
Dave Airlie
airlied at gmail.com
Tue Nov 8 06:24:45 UTC 2016
From: Dave Airlie <airlied at redhat.com>
I was getting a random GPU hang in the renderpass simple tests,
it turns out sometimes radv emitted the wrong thing "last".
This fixes the logic to emit Z/stencil last if they occur,
and not mark a color output as last. Also this relies on the
Z/STENCIL being the first two fragment outputs, which they are
so yay.
Fixes: dEQP-VK.renderpass.simple.color_depth (random hangs)
Cc: "13.0" <mesa-stable at lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied at redhat.com>
---
src/amd/common/ac_nir_to_llvm.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/amd/common/ac_nir_to_llvm.c b/src/amd/common/ac_nir_to_llvm.c
index 9b2663e..c8ee784 100644
--- a/src/amd/common/ac_nir_to_llvm.c
+++ b/src/amd/common/ac_nir_to_llvm.c
@@ -4378,12 +4378,10 @@ handle_fs_outputs_post(struct nir_to_llvm_context *ctx,
for (unsigned i = 0; i < RADEON_LLVM_MAX_OUTPUTS; ++i) {
LLVMValueRef values[4];
- bool last;
+
if (!(ctx->output_mask & (1ull << i)))
continue;
- last = ctx->output_mask <= ((1ull << (i + 1)) - 1);
-
if (i == FRAG_RESULT_DEPTH) {
ctx->shader_info->fs.writes_z = true;
depth = to_float(ctx, LLVMBuildLoad(ctx->builder,
@@ -4393,10 +4391,14 @@ handle_fs_outputs_post(struct nir_to_llvm_context *ctx,
stencil = to_float(ctx, LLVMBuildLoad(ctx->builder,
ctx->outputs[radeon_llvm_reg_index_soa(i, 0)], ""));
} else {
+ bool last = false;
for (unsigned j = 0; j < 4; j++)
values[j] = to_float(ctx, LLVMBuildLoad(ctx->builder,
ctx->outputs[radeon_llvm_reg_index_soa(i, j)], ""));
+ if (!ctx->shader_info->fs.writes_z && !ctx->shader_info->fs.writes_stencil)
+ last = ctx->output_mask <= ((1ull << (i + 1)) - 1);
+
si_export_mrt_color(ctx, values, V_008DFC_SQ_EXP_MRT + index, last);
index++;
}
--
2.5.5
More information about the mesa-dev
mailing list