i965: Convert brw->batch.is_blit to a BLT_RING/RENDER_RING enum.
Passing BLT_RING or RENDER_RING to batchbuffer functions is a lot more
obvious than passing true or false.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
diff --git a/src/mesa/drivers/dri/i965/brw_blorp.cpp b/src/mesa/drivers/dri/i965/brw_blorp.cpp
index 4cbcebe..e555f46 100644
--- a/src/mesa/drivers/dri/i965/brw_blorp.cpp
+++ b/src/mesa/drivers/dri/i965/brw_blorp.cpp
@@ -210,7 +210,7 @@
intel_batchbuffer_emit_mi_flush(brw);
retry:
- intel_batchbuffer_require_space(brw, estimated_max_batch_usage, false);
+ intel_batchbuffer_require_space(brw, estimated_max_batch_usage, RENDER_RING);
intel_batchbuffer_save_state(brw);
drm_intel_bo *saved_bo = brw->batch.bo;
uint32_t saved_used = brw->batch.used;
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 4a08986..7576a22 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -861,6 +861,11 @@
drm_intel_bo *bo;
};
+enum brw_gpu_ring {
+ RENDER_RING,
+ BLT_RING,
+};
+
struct intel_batchbuffer {
/** Current batchbuffer being queued up. */
drm_intel_bo *bo;
@@ -879,7 +884,7 @@
#define BATCH_SZ (8192*sizeof(uint32_t))
uint32_t state_batch_offset;
- bool is_blit;
+ enum brw_gpu_ring ring;
bool needs_sol_reset;
struct {
diff --git a/src/mesa/drivers/dri/i965/brw_draw.c b/src/mesa/drivers/dri/i965/brw_draw.c
index 7b33b76..1da45df 100644
--- a/src/mesa/drivers/dri/i965/brw_draw.c
+++ b/src/mesa/drivers/dri/i965/brw_draw.c
@@ -392,7 +392,7 @@
* we've got validated state that needs to be in the same batch as the
* primitives.
*/
- intel_batchbuffer_require_space(brw, estimated_max_prim_size, false);
+ intel_batchbuffer_require_space(brw, estimated_max_prim_size, RENDER_RING);
intel_batchbuffer_save_state(brw);
if (brw->num_instances != prims[i].num_instances) {
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index babe9ea..672bc02 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -260,19 +260,18 @@
if (!brw->intelScreen->no_hw) {
int flags;
- if (brw->gen < 6 || !batch->is_blit) {
- flags = I915_EXEC_RENDER;
+ if (brw->gen >= 6 && batch->ring == BLT_RING) {
+ flags = I915_EXEC_BLT;
} else {
- flags = I915_EXEC_BLT;
+ flags = I915_EXEC_RENDER;
}
-
if (batch->needs_sol_reset)
flags |= I915_EXEC_GEN7_SOL_RESET;
if (ret == 0) {
if (unlikely(INTEL_DEBUG & DEBUG_AUB))
brw_annotate_aub(brw);
- if (brw->hw_ctx == NULL || batch->is_blit) {
+ if (brw->hw_ctx == NULL || batch->ring != RENDER_RING) {
ret = drm_intel_bo_mrb_exec(batch->bo, 4 * batch->used, NULL, 0, 0,
flags);
} else {
@@ -401,10 +400,10 @@
void
intel_batchbuffer_data(struct brw_context *brw,
- const void *data, GLuint bytes, bool is_blit)
+ const void *data, GLuint bytes, enum brw_gpu_ring ring)
{
assert((bytes & 3) == 0);
- intel_batchbuffer_require_space(brw, bytes, is_blit);
+ intel_batchbuffer_require_space(brw, bytes, ring);
__memcpy(brw->batch.map + brw->batch.used, data, bytes);
brw->batch.used += bytes >> 2;
}
@@ -613,7 +612,7 @@
intel_batchbuffer_emit_mi_flush(struct brw_context *brw)
{
if (brw->gen >= 6) {
- if (brw->batch.is_blit) {
+ if (brw->batch.ring == BLT_RING) {
BEGIN_BATCH_BLT(4);
OUT_BATCH(MI_FLUSH_DW);
OUT_BATCH(0);
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.h b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
index cabbb69..40b0a29 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.h
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
@@ -43,7 +43,8 @@
* intel_buffer_dword() calls.
*/
void intel_batchbuffer_data(struct brw_context *brw,
- const void *data, GLuint bytes, bool is_blit);
+ const void *data, GLuint bytes,
+ enum brw_gpu_ring ring);
bool intel_batchbuffer_emit_reloc(struct brw_context *brw,
drm_intel_bo *buffer,
@@ -101,14 +102,15 @@
}
static INLINE void
-intel_batchbuffer_require_space(struct brw_context *brw, GLuint sz, int is_blit)
+intel_batchbuffer_require_space(struct brw_context *brw, GLuint sz,
+ enum brw_gpu_ring ring)
{
- if (brw->gen >= 6 &&
- brw->batch.is_blit != is_blit && brw->batch.used) {
+ /* If we're switching rings, implicitly flush the batch. */
+ if (unlikely(ring != brw->batch.ring) && brw->batch.used && brw->gen >= 6) {
intel_batchbuffer_flush(brw);
}
- brw->batch.is_blit = is_blit;
+ brw->batch.ring = ring;
#ifdef DEBUG
assert(sz < BATCH_SZ - BATCH_RESERVED);
@@ -118,9 +120,9 @@
}
static INLINE void
-intel_batchbuffer_begin(struct brw_context *brw, int n, bool is_blit)
+intel_batchbuffer_begin(struct brw_context *brw, int n, enum brw_gpu_ring ring)
{
- intel_batchbuffer_require_space(brw, n * 4, is_blit);
+ intel_batchbuffer_require_space(brw, n * 4, ring);
brw->batch.emit = brw->batch.used;
#ifdef DEBUG
@@ -146,8 +148,8 @@
void intel_batchbuffer_cached_advance(struct brw_context *brw);
-#define BEGIN_BATCH(n) intel_batchbuffer_begin(brw, n, false)
-#define BEGIN_BATCH_BLT(n) intel_batchbuffer_begin(brw, n, true)
+#define BEGIN_BATCH(n) intel_batchbuffer_begin(brw, n, RENDER_RING)
+#define BEGIN_BATCH_BLT(n) intel_batchbuffer_begin(brw, n, BLT_RING)
#define OUT_BATCH(d) intel_batchbuffer_emit_dword(brw, d)
#define OUT_BATCH_F(f) intel_batchbuffer_emit_float(brw, f)
#define OUT_RELOC(buf, read_domains, write_domain, delta) do { \
diff --git a/src/mesa/drivers/dri/i965/intel_blit.c b/src/mesa/drivers/dri/i965/intel_blit.c
index 0a03859..a47d02e 100644
--- a/src/mesa/drivers/dri/i965/intel_blit.c
+++ b/src/mesa/drivers/dri/i965/intel_blit.c
@@ -312,7 +312,7 @@
if (pass >= 2)
return false;
- intel_batchbuffer_require_space(brw, 8 * 4, true);
+ intel_batchbuffer_require_space(brw, 8 * 4, BLT_RING);
DBG("%s src:buf(%p)/%d+%d %d,%d dst:buf(%p)/%d+%d %d,%d sz:%dx%d\n",
__FUNCTION__,
src_buffer, src_pitch, src_offset, src_x, src_y,
@@ -427,7 +427,7 @@
__FUNCTION__,
dst_buffer, dst_pitch, dst_offset, x, y, w, h, src_size, dwords);
- intel_batchbuffer_require_space(brw, (8 * 4) + (3 * 4) + dwords * 4, true);
+ intel_batchbuffer_require_space(brw, (8 * 4) + (3 * 4) + dwords * 4, BLT_RING);
opcode = XY_SETUP_BLT_CMD;
if (cpp == 4)
@@ -461,7 +461,7 @@
OUT_BATCH(((y + h) << 16) | (x + w));
ADVANCE_BATCH();
- intel_batchbuffer_data(brw, src_bits, dwords * 4, true);
+ intel_batchbuffer_data(brw, src_bits, dwords * 4, BLT_RING);
intel_batchbuffer_emit_mi_flush(brw);