Plumb ScheduleOverlayPlane into the GPU process.

gpu/command_buffer/service/gles2_cmd_decoder.cc

Index: gpu/command_buffer/service/gles2_cmd_decoder.cc
diff --git a/gpu/command_buffer/service/gles2_cmd_decoder.cc b/gpu/command_buffer/service/gles2_cmd_decoder.cc
index c3bdae5d5d94a78b9cfa80d8087581f367118ec0..34418359de216822dbef4615791e4f470e21e298 100644
--- a/gpu/command_buffer/service/gles2_cmd_decoder.cc
+++ b/gpu/command_buffer/service/gles2_cmd_decoder.cc
@@ -7526,6 +7526,13 @@ error::Error GLES2DecoderImpl::HandlePostSubBufferCHROMIUM(
   }
 }
 
+error::Error GLES2DecoderImpl::HandleScheduleOverlayPlaneCHROMIUM(
+    uint32 immediate_data_size,
+    const cmds::ScheduleOverlayPlaneCHROMIUM& c) {
+  NOTREACHED() << "Overlay supported isn't finished.";

[piman, 2014/04/02 20:09:38]

nit: NOTIMPLEMENTED. It can be reached, with a buggy or malicious client.

+  return error::kGenericError;

[piman, 2014/04/02 20:09:38]

Rather than generating a command buffer error, which will lose the context, can
you raise GL_INVALID_OPERATION? (use LOCAL_SET_GL_ERROR).

+}
+
 error::Error GLES2DecoderImpl::GetAttribLocationHelper(
     GLuint client_id, uint32 location_shm_id, uint32 location_shm_offset,
     const std::string& name_str) {