[Mesa-dev] [PATCH 2/2] i965: Drop the memcmp for finding duplicated CURBE uploads.

[Eric Anholt]

At this point, the extra copy of the data and memcmp are as expensive as
just re-uploading.

Note: now that we'll always upload, and brw_constant_buffer watches
BRW_NEW_BATCH anyway, we don't need to explicitly unref the old curbe_bo
at batch reset time.

No significant performance difference on glamor copywinwin10 (n=55),
despite that test having a 98% hit rate on the cache.
---
 src/mesa/drivers/dri/i965/brw_context.c       |  8 --------
 src/mesa/drivers/dri/i965/brw_context.h       | 14 --------------
 src/mesa/drivers/dri/i965/brw_curbe.c         | 25 ++-----------------------
 src/mesa/drivers/dri/i965/intel_batchbuffer.c |  5 -----
 4 files changed, 2 insertions(+), 50 deletions(-)

diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c
index 39dd582..ee5c7b5 100644
--- a/src/mesa/drivers/dri/i965/brw_context.c
+++ b/src/mesa/drivers/dri/i965/brw_context.c
@@ -751,11 +751,6 @@ brwCreateContext(gl_api api,
    brw->prim_restart.enable_cut_index = false;
    brw->gs.enabled = false;
 
-   if (brw->gen < 6) {
-      brw->curbe.last_buf = calloc(1, 4096);
-      brw->curbe.next_buf = calloc(1, 4096);
-   }
-
    ctx->VertexProgram._MaintainTnlProgram = true;
    ctx->FragmentProgram._MaintainTexEnvProgram = true;
 
@@ -819,9 +814,6 @@ intelDestroyContext(__DRIcontext * driContextPriv)
 
    drm_intel_bo_unreference(brw->curbe.curbe_bo);
 
-   free(brw->curbe.last_buf);
-   free(brw->curbe.next_buf);
-
    drm_intel_gem_context_destroy(brw->hw_ctx);
 
    if (ctx->swrast_context) {
diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index b84b2a2..0c942b1 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1230,20 +1230,6 @@ struct brw_context
       drm_intel_bo *curbe_bo;
       /** Offset within curbe_bo of space for current curbe entry */
       GLuint curbe_offset;
-
-      /**
-       * Copy of the last set of CURBEs uploaded.  Frequently we'll end up
-       * in brw_curbe.c with the same set of constant data to be uploaded,
-       * so we'd rather not upload new constants in that case (it can cause
-       * a pipeline bubble since only up to 4 can be pipelined at a time).
-       */
-      GLfloat *last_buf;
-      /**
-       * Allocation for where to calculate the next set of CURBEs.
-       * It's a hot enough path that malloc/free of that data matters.
-       */
-      GLfloat *next_buf;
-      GLuint last_bufsz;
    } curbe;
 
    /**
diff --git a/src/mesa/drivers/dri/i965/brw_curbe.c b/src/mesa/drivers/dri/i965/brw_curbe.c
index b776bdc..0db3b94 100644
--- a/src/mesa/drivers/dri/i965/brw_curbe.c
+++ b/src/mesa/drivers/dri/i965/brw_curbe.c
@@ -188,11 +188,11 @@ brw_upload_constant_buffer(struct brw_context *brw)
    gl_clip_plane *clip_planes;
 
    if (sz == 0) {
-      brw->curbe.last_bufsz  = 0;
       goto emit;
    }
 
-   buf = brw->curbe.next_buf;
+   buf = intel_upload_space(brw, bufsz, 64,
+                            &brw->curbe.curbe_bo, &brw->curbe.curbe_offset);
 
    /* fragment shader constants */
    if (brw->curbe.wm_size) {
@@ -246,27 +246,6 @@ brw_upload_constant_buffer(struct brw_context *brw)
       for (i = 0; i < sz*16; i+=4)
 	 fprintf(stderr, "curbe %d.%d: %f %f %f %f\n", i/8, i&4,
                  buf[i+0], buf[i+1], buf[i+2], buf[i+3]);
-
-      fprintf(stderr, "last_buf %p buf %p sz %d/%d cmp %d\n",
-              brw->curbe.last_buf, buf,
-              bufsz, brw->curbe.last_bufsz,
-              brw->curbe.last_buf ? memcmp(buf, brw->curbe.last_buf, bufsz) : -1);
-   }
-
-   if (brw->curbe.curbe_bo != NULL &&
-       bufsz == brw->curbe.last_bufsz &&
-       memcmp(buf, brw->curbe.last_buf, bufsz) == 0) {
-      /* constants have not changed */
-   } else {
-      /* Update the record of what our last set of constants was.  We
-       * don't just flip the pointers because we don't fill in the
-       * data in the padding between the entries.
-       */
-      memcpy(brw->curbe.last_buf, buf, bufsz);
-      brw->curbe.last_bufsz = bufsz;
-
-      intel_upload_data(brw, buf, bufsz, 64,
-                        &brw->curbe.curbe_bo, &brw->curbe.curbe_offset);
    }
 
    /* Because this provokes an action (ie copy the constants into the
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index cba81bc..71dc268 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -224,11 +224,6 @@ brw_finish_batch(struct brw_context *brw)
    if (brw->batch.ring == RENDER_RING)
       brw_perf_monitor_finish_batch(brw);
 
-   if (brw->curbe.curbe_bo) {
-      drm_intel_bo_unreference(brw->curbe.curbe_bo);
-      brw->curbe.curbe_bo = NULL;
-   }
-
    /* Mark that the current program cache BO has been used by the GPU.
     * It will be reallocated if we need to put new programs in for the
     * next batch.
-- 
2.0.0.rc2