[Mesa-dev] [PATCH 43/70] i965: Move all the render preamble together

[Chris Wilson]

Rather than split the render batch setup between two hooks, coalesce it
into a single callback. To simplify this, move some of the state
dirtying from the start to the finish hook hook.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 src/mesa/drivers/dri/i965/brw_batch.h         |  6 ++---
 src/mesa/drivers/dri/i965/brw_context.c       | 36 ++++++++++++++++-----------
 src/mesa/drivers/dri/i965/intel_batchbuffer.c | 16 +-----------
 3 files changed, 26 insertions(+), 32 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_batch.h b/src/mesa/drivers/dri/i965/brw_batch.h
index bef544d..bbfb736 100644
--- a/src/mesa/drivers/dri/i965/brw_batch.h
+++ b/src/mesa/drivers/dri/i965/brw_batch.h
@@ -229,7 +229,7 @@ inline static void brw_batch_maybe_flush(struct brw_batch *batch)
 void intel_batchbuffer_save_state(struct brw_batch *batch);
 void intel_batchbuffer_reset_to_saved(struct brw_batch *batch);
 
-void intel_batchbuffer_emit_render_ring_prelude(struct brw_batch *batch);
+void brw_batch_start_hook(struct brw_batch *batch);
 
 #define USED_BATCH(batch) ((uintptr_t)((batch)->map_next - (batch)->map))
 
@@ -262,8 +262,8 @@ intel_batchbuffer_require_space(struct brw_batch *batch, GLuint sz,
     */
    batch->ring = ring;
 
-   if (unlikely(prev_ring == UNKNOWN_RING && ring == RENDER_RING))
-      intel_batchbuffer_emit_render_ring_prelude(batch);
+   if (unlikely(prev_ring == UNKNOWN_RING))
+      brw_batch_start_hook(batch);
 }
 
 int brw_batch_begin(struct brw_batch *batch,
diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index 2126d68..95e545c 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -1502,20 +1502,8 @@ void brw_batch_start_hook(brw_batch *batch)
 {
    struct brw_context *brw = container_of(batch, brw, batch);
 
-   /* If the kernel supports hardware contexts, then most hardware state is
-    * preserved between batches; we only need to re-emit state that is required
-    * to be in every batch.  Otherwise we need to re-emit all the state that
-    * would otherwise be stored in the context (which for all intents and
-    * purposes means everything).
-    */
-   if (!batch->hw_ctx)
-      brw->ctx.NewDriverState |= BRW_NEW_CONTEXT;
-
-   brw->ctx.NewDriverState |= BRW_NEW_BATCH;
-
-   brw->state_batch_count = 0;
-
-   brw->ib.type = -1;
+   if (batch->ring != RENDER_RING)
+      return;
 
    /* We need to periodically reap the shader time results, because rollover
     * happens every few seconds.  We also want to see results every once in a
@@ -1527,6 +1515,8 @@ void brw_batch_start_hook(brw_batch *batch)
 
    if (INTEL_DEBUG & DEBUG_PERFMON)
       brw_dump_perf_monitors(brw);
+
+   brw_perf_monitor_new_batch(brw);
 }
 
 /**
@@ -1579,6 +1569,24 @@ void brw_batch_finish_hook(brw_batch *batch)
    brw_perf_monitor_finish_batch(brw);
 
    brw->cache.bo_used_by_gpu = true;
+
+   brw->state_batch_count = 0;
+
+   brw->ib.type = -1;
+
+   if (brw->use_resource_streamer)
+      gen7_reset_hw_bt_pool_offsets(brw);
+
+   /* If the kernel supports hardware contexts, then most hardware state is
+    * preserved between batches; we only need to re-emit state that is required
+    * to be in every batch.  Otherwise we need to re-emit all the state that
+    * would otherwise be stored in the context (which for all intents and
+    * purposes means everything).
+    */
+   if (!batch->hw_ctx)
+      brw->ctx.NewDriverState |= BRW_NEW_CONTEXT;
+
+   brw->ctx.NewDriverState |= BRW_NEW_BATCH;
 }
 
 void brw_batch_report_flush_hook(struct brw_batch *batch,
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index 49c1c27..0d5821f 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -168,15 +168,6 @@ do_batch_dump(struct brw_context *brw)
    }
 }
 
-void
-intel_batchbuffer_emit_render_ring_prelude(struct brw_batch *batch)
-{
-   struct brw_context *brw = container_of(batch, brw, batch);
-
-   /* We may need to enable and snapshot OA counters. */
-   brw_perf_monitor_new_batch(brw);
-}
-
 /**
  * Called when starting a new batch buffer.
  */
@@ -186,8 +177,6 @@ brw_new_batch(struct brw_context *brw)
    /* Create a new batchbuffer and reset the associated state: */
    drm_intel_gem_bo_clear_relocs(brw->batch.bo, 0);
    intel_batchbuffer_reset(brw);
-
-   brw_batch_start_hook(&brw->batch);
 }
 
 static void
@@ -344,9 +333,6 @@ brw_batch_flush(struct brw_batch *batch, struct perf_debug *info)
       drm_intel_bo_wait_rendering(brw->batch.bo);
    }
 
-   if (brw->use_resource_streamer)
-      gen7_reset_hw_bt_pool_offsets(brw);
-
    /* Start a new batch buffer. */
    brw_new_batch(brw);
 
@@ -443,7 +429,7 @@ int brw_batch_end(struct brw_batch *batch)
 
          batch->ring = ring;
          if (ring == RENDER_RING)
-            intel_batchbuffer_emit_render_ring_prelude(batch);
+            brw_batch_start_hook(batch);
 
          batch->repeat = true;
          longjmp(batch->jmpbuf, 1);
-- 
2.5.0