[Mesa-dev] [PATCH v2 1/4] i965: Track last location of bo used for the batch

[Chris Wilson]

Borrow a trick from anv, and use the last known index for the bo to skip
a search of the batch->exec_bo when adding a new relocation. In defence
against the bo being used in multiple batches simultaneously, we check
that this slot exists and points back to us.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Kenneth Graunke <kenneth at whitecape.org>
Cc: Jason Ekstrand <jason.ekstrand at intel.com>
---
 src/mesa/drivers/dri/i965/brw_bufmgr.h        |  5 +++++
 src/mesa/drivers/dri/i965/intel_batchbuffer.c | 14 ++++++++++++--
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_bufmgr.h b/src/mesa/drivers/dri/i965/brw_bufmgr.h
index 56ec206d30..1305183a4d 100644
--- a/src/mesa/drivers/dri/i965/brw_bufmgr.h
+++ b/src/mesa/drivers/dri/i965/brw_bufmgr.h
@@ -86,6 +86,11 @@ struct brw_bo {
    uint64_t offset64;
 
    /**
+    * Index of this buffer inside the batch, -1 when not in a batch.
+    */
+   unsigned int index;
+
+   /**
     * Boolean of whether the GPU is definitely not accessing the buffer.
     *
     * This is only valid when reusable, since non-reusable
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index 36faa7ae6f..752e9b389b 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -516,12 +516,20 @@ throttle(struct brw_context *brw)
    }
 }
 
+#define READ_ONCE(x) (*(volatile __typeof__(x) *)&(x))
+
 static void
 add_exec_bo(struct intel_batchbuffer *batch, struct brw_bo *bo)
 {
    if (bo != batch->bo) {
-      for (int i = 0; i < batch->exec_count; i++) {
-         if (batch->exec_bos[i] == bo)
+      unsigned int index = READ_ONCE(bo->index);
+
+      if (index < batch->exec_count && batch->exec_bos[index] == bo)
+         return;
+
+      /* May have been shared between multiple active batches */
+      for (index = 0; index < batch->exec_count; index++) {
+         if (batch->exec_bos[index] == bo)
             return;
       }
 
@@ -554,6 +562,7 @@ add_exec_bo(struct intel_batchbuffer *batch, struct brw_bo *bo)
    validation_entry->rsvd1 = 0;
    validation_entry->rsvd2 = 0;
 
+   bo->index = batch->exec_count;
    batch->exec_bos[batch->exec_count] = bo;
    batch->exec_count++;
    batch->aperture_space += bo->size;
@@ -598,6 +607,7 @@ execbuffer(int fd,
       struct brw_bo *bo = batch->exec_bos[i];
 
       bo->idle = false;
+      bo->index = -1;
 
       /* Update brw_bo::offset64 */
       if (batch->exec_objects[i].offset != bo->offset64) {
-- 
2.11.0