gpu: Thread-safe command buffer state lookup.

content/common/gpu/client/context_provider_command_buffer.cc

Index: content/common/gpu/client/context_provider_command_buffer.cc
diff --git a/content/common/gpu/client/context_provider_command_buffer.cc b/content/common/gpu/client/context_provider_command_buffer.cc
index 74bbf375caece9a56c5883ca887dfb2db98dbfb7..a1339f20aeb3b8c0c10c3f935e8f7550d2dfb0c2 100644
--- a/content/common/gpu/client/context_provider_command_buffer.cc
+++ b/content/common/gpu/client/context_provider_command_buffer.cc
@@ -292,9 +292,9 @@ bool ContextProviderCommandBuffer::BindToCurrentThread() {
       return false;
     }
 
-    if (command_buffer_->GetLastError() != gpu::error::kNoError) {
+    if (command_buffer_->GetLastState().error != gpu::error::kNoError) {
       DLOG(ERROR) << "Context dead on arrival. Last error: "
-                  << command_buffer_->GetLastError();
+                  << command_buffer_->GetLastState().error;
       return false;
     }
 
@@ -454,6 +454,8 @@ bool ContextProviderCommandBuffer::OnMemoryDump(
   if (support_locking_)
     hold.emplace(context_lock_);
 
+  gles2_impl_->OnMemoryDump(args, pmd);

[ericrk, 2016/12/08 01:59:39]

Seems like this will fix crbug.com/638862, can you close that out when this
lands? Thanks!

+
   context_thread_checker_.DetachFromThread();
   SkiaGpuTraceMemoryDump trace_memory_dump(
       pmd, gles2_impl_->ShareGroupTracingGUID());