Added GL_CHROMIUM_trace_marker feature as well as gpu_toplevel markers.

webkit/common/gpu/context_provider_in_process.cc

Index: webkit/common/gpu/context_provider_in_process.cc
diff --git a/webkit/common/gpu/context_provider_in_process.cc b/webkit/common/gpu/context_provider_in_process.cc
index b02d961d70cf109cdf434197fe6dc7b86551df41..289814d757b1c2da9a6b7124b563d2130456fcaf 100644
--- a/webkit/common/gpu/context_provider_in_process.cc
+++ b/webkit/common/gpu/context_provider_in_process.cc
@@ -102,7 +102,8 @@ bool ContextProviderInProcess::BindToCurrentThread() {
 
   std::string unique_context_name =
       base::StringPrintf("%s-%p", debug_name_.c_str(), context3d_.get());
-  context3d_->pushGroupMarkerEXT(unique_context_name.c_str());
+  context3d_->traceBeginCHROMIUM("gpu_toplevel",
+                                 unique_context_name.c_str());

[Ken Russell (switch to Gerrit), 2014/12/17 23:56:10]

I'll have to defer to Victor's review on whether it's correct that this will be
called once per context.

[vmiura, 2014/12/18 02:38:52]

I think this is fine.

 
   lost_context_callback_proxy_.reset(new LostContextCallbackProxy(this));
   return true;
@@ -194,6 +195,7 @@ void ContextProviderInProcess::OnLostContext() {
 
 bool ContextProviderInProcess::DestroyedOnMainThread() {
   DCHECK(main_thread_checker_.CalledOnValidThread());
+  context3d_->traceEndCHROMIUM();

[Ken Russell (switch to Gerrit), 2014/12/17 23:56:10]

Similarly it's not obvious to me that calling this once at the end of the
context's lifetime is correct.

[vmiura, 2014/12/18 02:38:52]

It is not necessarily needed, as the End is implied after the last command
executed on a context.

Ken is it safe to call context3d_->* from this method?

[Ken Russell (switch to Gerrit), 2014/12/18 19:03:56]

Wait a second. Looking more closely, this method is a query, not a notification
that the context was destroyed on the main thread. I think calling
traceEndCHROMIUM is semantically incorrect here.

[vmiura, 2014/12/18 19:30:27]

Can we just remove this call David, as GL_CHROMIUM_trace_marker spec implies
that this is not needed.

[David Yen, 2014/12/18 21:25:59]

Done.

 
   base::AutoLock lock(destroyed_lock_);
   return destroyed_;