svga: document magic number of 8 render targets per batch

Grab the comments from commit message b84b7f19dfdc0 to explain
what the code is doing.

diff --git a/src/gallium/drivers/svga/svga_state_framebuffer.c b/src/gallium/drivers/svga/svga_state_framebuffer.c
index 8232c32..4254404 100644 (file)
--- a/src/gallium/drivers/svga/svga_state_framebuffer.c
+++ b/src/gallium/drivers/svga/svga_state_framebuffer.c
@@ -33,6 +33,18 @@
 #include "svga_debug.h"
 
 
+/*
+ * flush our command buffer after the 8th distinct render target
+ *
+ * This helps improve the surface cache behaviour in the face of the
+ * large number of single-use render targets generated by EXA and the xorg
+ * state tracker.  Without this we can reference hundreds of individual
+ * render targets from a command buffer, which leaves little scope for
+ * sharing or reuse of those targets.
+ */
+#define MAX_RT_PER_BATCH 8
+
+
 /***********************************************************************
  * Hardware state update
  */
@@ -56,7 +68,7 @@ emit_framebuffer( struct svga_context *svga,
    for(i = 0; i < PIPE_MAX_COLOR_BUFS; ++i) {
       if (curr->cbufs[i] != hw->cbufs[i] ||
           (reemit && hw->cbufs[i])) {
-         if (svga->curr.nr_fbs++ > 8)
+         if (svga->curr.nr_fbs++ > MAX_RT_PER_BATCH)
             return PIPE_ERROR_OUT_OF_MEMORY;
 
          ret = SVGA3D_SetRenderTarget(svga->swc, SVGA3D_RT_COLOR0 + i, curr->cbufs[i]);