intel: Only do frame throttling at glFlush time when using frontbuffer.

This is the hack for input interactivity of frontbuffer rendering
(like we do for backbuffer at intelDRI2Flush()) by waiting for the n-2
frame to complete before starting a new one.  However, for an
application doing multiple contexts or regular rebinding of a single
context, this would end up lockstepping the CPU to the GPU because
every unbind was considered the end of a frame.

Improves WOW performance on my Ironlake by 48.8% (+/- 2.3%, n=5)

diff --git a/src/mesa/drivers/dri/intel/intel_context.c b/src/mesa/drivers/dri/intel/intel_context.c
index 9c222c7b4857e7b0b68727718d2cc5220dc15b2b..d183d275e73d1f213209339dde1fd44faea029df 100644 (file)
--- a/src/mesa/drivers/dri/intel/intel_context.c
+++ b/src/mesa/drivers/dri/intel/intel_context.c
@@ -565,7 +565,8 @@ intel_glFlush(struct gl_context *ctx)
 
    intel_flush(ctx);
    intel_flush_front(ctx);
-   intel->need_throttle = GL_TRUE;
+   if (intel->is_front_buffer_rendering)
+      intel->need_throttle = GL_TRUE;
 }
 
 void