r600g, radeonsi: fix primitives-generated query with disabled streamout

Buffers are disabled by VGT_STRMOUT_BUFFER_CONFIG, but the query only works
if VGT_STRMOUT_CONFIG.STREAMOUT_0_EN is enabled.

This moves VGT_STRMOUT_CONFIG to its own state. The register is set to 1
if either streamout or the primitives-generated query is enabled.

However, the primitives-emitted query is also incremented, so it's disabled
by setting VGT_STRMOUT_BUFFER_SIZE to 0 when there is no buffer bound.

This fixes piglit:
  ARB_transform_feedback2/counting with pause
  EXT_transform_feedback/primgen-query transform-feedback-disabled

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>

4 files changed, 7 insertions, 10 deletions

diff --git a/src/gallium/drivers/r600/evergreen_state.c b/src/gallium/drivers/r600/evergreen_state.c
index dca7c58d415..18c2ee59b87 100644
--- a/src/gallium/drivers/r600/evergreen_state.c
+++ b/src/gallium/drivers/r600/evergreen_state.c
@@ -2248,9 +2248,7 @@ static void cayman_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_value(cb, 0); /* R_028A3C_VGT_GROUP_VECT_1_FMT_CNTL */
 	r600_store_value(cb, 0); /* R_028A40_VGT_GS_MODE */
 
-	r600_store_context_reg_seq(cb, R_028B94_VGT_STRMOUT_CONFIG, 2);
-	r600_store_value(cb, 0); /* R_028B94_VGT_STRMOUT_CONFIG */
-	r600_store_value(cb, 0); /* R_028B98_VGT_STRMOUT_BUFFER_CONFIG */
+	r600_store_context_reg(cb, R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0);
 
 	r600_store_context_reg_seq(cb, R_028AB4_VGT_REUSE_OFF, 2);
 	r600_store_value(cb, 0); /* R_028AB4_VGT_REUSE_OFF */
@@ -2809,9 +2807,7 @@ void evergreen_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_value(cb, 0);
 	r600_store_value(cb, 0);
 
-	r600_store_context_reg_seq(cb, R_028B94_VGT_STRMOUT_CONFIG, 2);
-	r600_store_value(cb, 0); /* R_028B94_VGT_STRMOUT_CONFIG */
-	r600_store_value(cb, 0); /* R_028B98_VGT_STRMOUT_BUFFER_CONFIG */
+	r600_store_context_reg(cb, R_028B98_VGT_STRMOUT_BUFFER_CONFIG, 0);
 
 	if (rctx->screen->b.has_streamout) {
 		r600_store_context_reg(cb, R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET, 0);
@@ -3488,6 +3484,7 @@ void evergreen_init_state_functions(struct r600_context *rctx)
 	r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4);
 	r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, evergreen_emit_vertex_fetch_shader, 5);
 	rctx->atoms[id++] = &rctx->b.streamout.begin_atom;
+	rctx->atoms[id++] = &rctx->b.streamout.enable_atom;
 	r600_init_atom(rctx, &rctx->vertex_shader.atom, id++, r600_emit_shader, 23);
 	r600_init_atom(rctx, &rctx->pixel_shader.atom, id++, r600_emit_shader, 0);
 	r600_init_atom(rctx, &rctx->geometry_shader.atom, id++, r600_emit_shader, 0);
diff --git a/src/gallium/drivers/r600/r600_hw_context.c b/src/gallium/drivers/r600/r600_hw_context.c
index 3d0da2c212f..3a3b3d58c15 100644
--- a/src/gallium/drivers/r600/r600_hw_context.c
+++ b/src/gallium/drivers/r600/r600_hw_context.c
@@ -311,7 +311,7 @@ void r600_begin_new_cs(struct r600_context *ctx)
 		ctx->gs_rings.atom.dirty = true;
 	}
 	ctx->vertex_shader.atom.dirty = true;
-
+	ctx->b.streamout.enable_atom.dirty = true;
 
 	if (ctx->blend_state.cso)
 		ctx->blend_state.atom.dirty = true;
diff --git a/src/gallium/drivers/r600/r600_pipe.h b/src/gallium/drivers/r600/r600_pipe.h
index e5bb2b883d5..6d627e5adee 100644
--- a/src/gallium/drivers/r600/r600_pipe.h
+++ b/src/gallium/drivers/r600/r600_pipe.h
@@ -37,7 +37,7 @@
 #include "util/u_double_list.h"
 #include "util/u_transfer.h"
 
-#define R600_NUM_ATOMS 72
+#define R600_NUM_ATOMS 73
 
 /* the number of CS dwords for flushing and drawing */
 #define R600_MAX_FLUSH_CS_DWORDS	16
diff --git a/src/gallium/drivers/r600/r600_state.c b/src/gallium/drivers/r600/r600_state.c
index 6d89e6ce88b..e6254c69a2c 100644
--- a/src/gallium/drivers/r600/r600_state.c
+++ b/src/gallium/drivers/r600/r600_state.c
@@ -2316,8 +2316,7 @@ void r600_init_atom_start_cs(struct r600_context *rctx)
 	r600_store_context_reg(cb, R_028AA0_VGT_INSTANCE_STEP_RATE_0, 0);
 	r600_store_context_reg(cb, R_028AA4_VGT_INSTANCE_STEP_RATE_1, 0);
 
-	r600_store_context_reg_seq(cb, R_028AB0_VGT_STRMOUT_EN, 3);
-	r600_store_value(cb, 0); /* R_028AB0_VGT_STRMOUT_EN */
+	r600_store_context_reg_seq(cb, R_028AB4_VGT_REUSE_OFF, 2);
 	r600_store_value(cb, 1); /* R_028AB4_VGT_REUSE_OFF */
 	r600_store_value(cb, 0); /* R_028AB8_VGT_VTX_CNT_EN */
 
@@ -3027,6 +3026,7 @@ void r600_init_state_functions(struct r600_context *rctx)
 	r600_init_atom(rctx, &rctx->stencil_ref.atom, id++, r600_emit_stencil_ref, 4);
 	r600_init_atom(rctx, &rctx->vertex_fetch_shader.atom, id++, r600_emit_vertex_fetch_shader, 5);
 	rctx->atoms[id++] = &rctx->b.streamout.begin_atom;
+	rctx->atoms[id++] = &rctx->b.streamout.enable_atom;
 	r600_init_atom(rctx, &rctx->vertex_shader.atom, id++, r600_emit_shader, 23);
 	r600_init_atom(rctx, &rctx->pixel_shader.atom, id++, r600_emit_shader, 0);
 	r600_init_atom(rctx, &rctx->geometry_shader.atom, id++, r600_emit_shader, 0);