[Mesa-dev] [PATCH 2/4] i965: Split batch emission from relocation functions.

[Matt Turner]

So that everything writing to the batch between BEGIN_BATCH() and
ADVANCE_BATCH() goes through OUT_BATCH.

Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 src/mesa/drivers/dri/i965/intel_batchbuffer.c | 30 ++++++++++-------------
 src/mesa/drivers/dri/i965/intel_batchbuffer.h | 34 ++++++++++++++-------------
 2 files changed, 30 insertions(+), 34 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index d93ee6e..f82958f 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -393,11 +393,11 @@ _intel_batchbuffer_flush(struct brw_context *brw,
 
 /*  This is the only way buffers get added to the validate list.
  */
-bool
-intel_batchbuffer_emit_reloc(struct brw_context *brw,
-                             drm_intel_bo *buffer,
-                             uint32_t read_domains, uint32_t write_domain,
-			     uint32_t delta)
+uint32_t
+intel_batchbuffer_reloc(struct brw_context *brw,
+                        drm_intel_bo *buffer,
+                        uint32_t read_domains, uint32_t write_domain,
+                        uint32_t delta)
 {
    int ret;
 
@@ -411,16 +411,14 @@ intel_batchbuffer_emit_reloc(struct brw_context *brw,
     * case the buffer doesn't move and we can short-circuit the relocation
     * processing in the kernel
     */
-   intel_batchbuffer_emit_dword(brw, buffer->offset64 + delta);
-
-   return true;
+   return buffer->offset64 + delta;
 }
 
-bool
-intel_batchbuffer_emit_reloc64(struct brw_context *brw,
-                               drm_intel_bo *buffer,
-                               uint32_t read_domains, uint32_t write_domain,
-			       uint32_t delta)
+uint64_t
+intel_batchbuffer_reloc64(struct brw_context *brw,
+                          drm_intel_bo *buffer,
+                          uint32_t read_domains, uint32_t write_domain,
+                          uint32_t delta)
 {
    int ret = drm_intel_bo_emit_reloc(brw->batch.bo, 4*brw->batch.used,
                                      buffer, delta,
@@ -432,11 +430,7 @@ intel_batchbuffer_emit_reloc64(struct brw_context *brw,
     * case the buffer doesn't move and we can short-circuit the relocation
     * processing in the kernel
     */
-   uint64_t offset = buffer->offset64 + delta;
-   intel_batchbuffer_emit_dword(brw, offset);
-   intel_batchbuffer_emit_dword(brw, offset >> 32);
-
-   return true;
+   return buffer->offset64 + delta;
 }
 
 
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.h b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
index e58eae4..c0456f3 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.h
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
@@ -57,16 +57,16 @@ void intel_batchbuffer_data(struct brw_context *brw,
                             const void *data, GLuint bytes,
                             enum brw_gpu_ring ring);
 
-bool intel_batchbuffer_emit_reloc(struct brw_context *brw,
-                                       drm_intel_bo *buffer,
-				       uint32_t read_domains,
-				       uint32_t write_domain,
-				       uint32_t offset);
-bool intel_batchbuffer_emit_reloc64(struct brw_context *brw,
-                                    drm_intel_bo *buffer,
-                                    uint32_t read_domains,
-                                    uint32_t write_domain,
-                                    uint32_t offset);
+uint32_t intel_batchbuffer_reloc(struct brw_context *brw,
+                                 drm_intel_bo *buffer,
+                                 uint32_t read_domains,
+                                 uint32_t write_domain,
+                                 uint32_t offset);
+uint64_t intel_batchbuffer_reloc64(struct brw_context *brw,
+                                   drm_intel_bo *buffer,
+                                   uint32_t read_domains,
+                                   uint32_t write_domain,
+                                   uint32_t offset);
 static inline uint32_t float_as_int(float f)
 {
    union {
@@ -164,14 +164,16 @@ intel_batchbuffer_advance(struct brw_context *brw)
 #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 {		\
-   intel_batchbuffer_emit_reloc(brw, buf,			\
-				read_domains, write_domain, delta);	\
-} while (0)
+#define OUT_RELOC(buf, read_domains, write_domain, delta)                  \
+   OUT_BATCH(intel_batchbuffer_reloc(brw, buf, read_domains, write_domain, \
+                                     delta))
 
 /* Handle 48-bit address relocations for Gen8+ */
-#define OUT_RELOC64(buf, read_domains, write_domain, delta) do { \
-   intel_batchbuffer_emit_reloc64(brw, buf, read_domains, write_domain, delta);	\
+#define OUT_RELOC64(buf, read_domains, write_domain, delta) do {        \
+   uint64_t reloc64 = intel_batchbuffer_reloc64(brw, buf, read_domains, \
+                                                write_domain, delta);   \
+   OUT_BATCH(reloc64);                                                  \
+   OUT_BATCH(reloc64 >> 32);                                            \
 } while (0)
 
 #define ADVANCE_BATCH() intel_batchbuffer_advance(brw);
-- 
2.3.6