Make SpeechRecognition per RenderFrame instead of per RenderView.

content/public/browser/speech_recognition_session_context.h

Index: content/public/browser/speech_recognition_session_context.h
diff --git a/content/public/browser/speech_recognition_session_context.h b/content/public/browser/speech_recognition_session_context.h
index 94931894e68ed6dca5fcf4ec9a00dd0301bd1e18..00a1d06d459fe543004603c87298d2d0ff1db43d 100644
--- a/content/public/browser/speech_recognition_session_context.h
+++ b/content/public/browser/speech_recognition_session_context.h
@@ -25,25 +25,24 @@ struct CONTENT_EXPORT SpeechRecognitionSessionContext {
   ~SpeechRecognitionSessionContext();
 
   int render_process_id;
-  int render_view_id;
   int render_frame_id;
 
-  // Browser plugin guest's render view id, if this context represents a speech
+  // Browser plugin guest's render frame id, if this context represents a speech
   // recognition request from an embedder on behalf of the guest. This is used
   // for input tag where speech bubble is to be shown.
   //
   // TODO(lazyboy): Right now showing bubble from guest does not work, we fall
   // back to embedder instead, fix this and use
-  // embedder_render_process_id/embedder_render_view_id similar to Web Speech
+  // embedder_render_process_id/embedder_render_frame_id similar to Web Speech
   // API below.
-  int guest_render_view_id;
+  int guest_render_frame_id;
 
-  // The pair (|embedder_render_process_id|, |embedder_render_view_id|)
+  // The pair (|embedder_render_process_id|, |embedder_render_frame_id|)
   // represents a Browser plugin guest's embedder. This is filled in if the
   // session is from a guest Web Speech API. We use these to check if the
   // embedder (app) is permitted to use audio.
   int embedder_render_process_id;
-  int embedder_render_view_id;
+  int embedder_render_frame_id;
 
   int request_id;