GPU process clients can perform a "shallow" finish.

cc/output/gl_renderer.cc

Index: cc/output/gl_renderer.cc
===================================================================
--- cc/output/gl_renderer.cc	(revision 207109)
+++ cc/output/gl_renderer.cc	(working copy)
@@ -2012,12 +2012,12 @@
   DrawQuadGeometry(
       frame, draw_matrix, rect, program->vertex_shader().matrix_location());
 }
+//
+//void GLRenderer::Finish() {
+//  TRACE_EVENT0("cc", "GLRenderer::finish");
+//  context_->finish();
+//}
 
-void GLRenderer::Finish() {
-  TRACE_EVENT0("cc", "GLRenderer::finish");
-  context_->finish();
-}
-
 void GLRenderer::SwapBuffers() {
   DCHECK(visible_);
   DCHECK(!is_backbuffer_discarded_);
@@ -2343,7 +2343,8 @@
         sync_point,
         finished_callback);
   } else {
-    resource_provider_->Finish();
+    // Can this be a flush?

[piman, 2013/06/21 02:31:10]

It can't because we're waiting on the the ReadPixels to be done on the GPU
process side (it's async when GL_PIXEL_PACK_TRANSFER_BUFFER_CHROMIUM != 0).
In the current version, it can be made a shallow flush.

In the future though, we want to have an async readpixels on the GPU process
side (using PBOs), which will need a sort of finish to complete synchronously,
or something.

+    // resource_provider_->Finish();
     finished_callback.Run();
   }