[Mesa-dev] [PATCH 01/10] r600g, radeonsi: don't append to streamout buffers that haven't been used yet
Marek Olšák
maraeo at gmail.com
Sun Feb 1 09:36:58 PST 2015
From: Marek Olšák <marek.olsak at amd.com>
The FILLED_SIZE counter is uninitialized at the beginning, so we can't use it.
Instead, use offset = 0, which is what we always do when not appending.
This unexpectedly fixes spec/ARB_texture_multisample/sample-position/*.
Yes, the test does use transform feedback.
Cc: 10.3 10.4 <mesa-stable at lists.freedesktop.org>
---
src/gallium/drivers/radeon/r600_pipe_common.h | 1 +
src/gallium/drivers/radeon/r600_streamout.c | 4 +++-
2 files changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/gallium/drivers/radeon/r600_pipe_common.h b/src/gallium/drivers/radeon/r600_pipe_common.h
index 6224668..46a6bf3 100644
--- a/src/gallium/drivers/radeon/r600_pipe_common.h
+++ b/src/gallium/drivers/radeon/r600_pipe_common.h
@@ -294,6 +294,7 @@ struct r600_so_target {
/* The buffer where BUFFER_FILLED_SIZE is stored. */
struct r600_resource *buf_filled_size;
unsigned buf_filled_size_offset;
+ bool buf_filled_size_valid;
unsigned stride_in_dw;
};
diff --git a/src/gallium/drivers/radeon/r600_streamout.c b/src/gallium/drivers/radeon/r600_streamout.c
index c44f0f2..bc8bf97 100644
--- a/src/gallium/drivers/radeon/r600_streamout.c
+++ b/src/gallium/drivers/radeon/r600_streamout.c
@@ -237,7 +237,7 @@ static void r600_emit_streamout_begin(struct r600_common_context *rctx, struct r
}
}
- if (rctx->streamout.append_bitmask & (1 << i)) {
+ if (rctx->streamout.append_bitmask & (1 << i) && t[i]->buf_filled_size_valid) {
uint64_t va = t[i]->buf_filled_size->gpu_address +
t[i]->buf_filled_size_offset;
@@ -302,6 +302,8 @@ void r600_emit_streamout_end(struct r600_common_context *rctx)
* buffer bound. This ensures that the primitives-emitted query
* won't increment. */
r600_write_context_reg(cs, R_028AD0_VGT_STRMOUT_BUFFER_SIZE_0 + 16*i, 0);
+
+ t[i]->buf_filled_size_valid = true;
}
rctx->streamout.begin_emitted = false;
--
2.1.0
More information about the mesa-dev
mailing list