ChannelMojo: Handle when ChannelMojo outlives ChannelMojoHost

content/browser/renderer_host/render_process_host_impl.h

Index: content/browser/renderer_host/render_process_host_impl.h
diff --git a/content/browser/renderer_host/render_process_host_impl.h b/content/browser/renderer_host/render_process_host_impl.h
index 92d1e4d1b78802735102f58033d32dd093bc8679..f2a2fdcab957bcb99f96a614df08b791fcfe1d12 100644
--- a/content/browser/renderer_host/render_process_host_impl.h
+++ b/content/browser/renderer_host/render_process_host_impl.h
@@ -22,6 +22,7 @@
 #include "content/public/browser/render_process_host.h"
 #include "ipc/ipc_channel_proxy.h"
 #include "ipc/ipc_platform_file.h"
+#include "ipc/mojo/ipc_channel_mojo_host.h"
 #include "mojo/public/cpp/bindings/interface_ptr.h"
 #include "ui/gfx/gpu_memory_buffer.h"
 
@@ -34,10 +35,6 @@ namespace gfx {
 class Size;
 }
 
-namespace IPC {
-class ChannelMojoHost;
-}
-
 namespace content {
 class AudioRendererHost;
 class BrowserCdmManager;
@@ -266,7 +263,8 @@ class CONTENT_EXPORT RenderProcessHostImpl
 
   // A host object ChannelMojo needs. The lifetime is bound to
   // the RenderProcessHostImpl, not the channel.
-  scoped_ptr<IPC::ChannelMojoHost> channel_mojo_host_;
+  scoped_ptr<IPC::ChannelMojoHost, IPC::ChannelMojoHostDeleter>

[viettrungluu, 2014/09/25 19:38:34]

In general, this seems fragile/awkward to me.

It seems to me that part of the purpose of ChannelMojoHost should be to hide
details about ChannelMojo, and so ChannelMojoHost should just live on the main
thread (or any thread, really).

It suggests to me that instead of ChannelMojo talking directly to
ChannelMojoHost, it should talk to some private object owned by the
ChannelMojoHost (and whose lifecycle you can modify appropriately).

The cleanest way of doing something like this would be to have ChannelMojo have
some ChannelMojo::Delegate interface that it talks to (and which has certain
lifetime requirements), and for the ChannelMojoHost to have some private object
that implements this.

[Hajime Morrita, 2014/09/25 22:00:03]

Thanks for the suggestion! This totally makes sense.

+      channel_mojo_host_;
 
   // True if fast shutdown has been performed on this RPH.
   bool fast_shutdown_started_;