blimp: Set up the CompositorDependencies for blimp in Chrome.

content/browser/renderer_host/context_provider_factory_impl_android.cc

 // Copyright 2016 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "content/browser/renderer_host/context_provider_factory_impl_android.h"
 
 #include "base/auto_reset.h"
 #include "base/command_line.h"
 #include "base/lazy_instance.h"
+#include "base/memory/ref_counted.h"
 #include "base/memory/singleton.h"
 #include "cc/output/context_provider.h"
 #include "cc/output/vulkan_in_process_context_provider.h"
 #include "cc/surfaces/surface_manager.h"
 #include "content/browser/gpu/browser_gpu_memory_buffer_manager.h"
 #include "content/browser/gpu/compositor_util.h"
 #include "content/browser/gpu/gpu_surface_tracker.h"
 #include "content/common/gpu/client/context_provider_command_buffer.h"
 #include "content/common/host_shared_bitmap_manager.h"
 #include "content/public/common/content_switches.h"
@@ skipping 45 matching lines @@
     : gpu_channel_factory_(gpu_channel_factory),
       in_handle_pending_requests_(false),
       in_shutdown_(false),
       surface_client_id_(0),
       weak_factory_(this) {
   DCHECK(gpu_channel_factory_);
 }
 
 ContextProviderFactoryImpl::~ContextProviderFactoryImpl() {
   in_shutdown_ = true;
-  if (!context_provider_requests_.empty())
+  if (!gpu_channel_requests_.empty())
     HandlePendingRequests(nullptr,
-                          ContextCreationResult::FAILURE_FACTORY_SHUTDOWN);
+                          GpuChannelHostResult::FAILURE_FACTORY_SHUTDOWN);
 }
 
-ContextProviderFactoryImpl::ContextProvidersRequest::ContextProvidersRequest()
-    : context_type(command_buffer_metrics::CONTEXT_TYPE_UNKNOWN),
-      surface_handle(gpu::kNullSurfaceHandle),
-      support_locking(false),
-      automatic_flushes(false),
-      shared_context_provider(nullptr) {}
-
-ContextProviderFactoryImpl::ContextProvidersRequest::ContextProvidersRequest(
-    const ContextProvidersRequest& other) = default;
-
-ContextProviderFactoryImpl::ContextProvidersRequest::
-    ~ContextProvidersRequest() = default;
-
 scoped_refptr<cc::VulkanContextProvider>
 ContextProviderFactoryImpl::GetSharedVulkanContextProvider() {
   if (!shared_vulkan_context_provider_)
     shared_vulkan_context_provider_ =
         cc::VulkanInProcessContextProvider::Create();
 
   return shared_vulkan_context_provider_.get();
 }
 
-void ContextProviderFactoryImpl::CreateDisplayContextProvider(
+void ContextProviderFactoryImpl::RequestGpuChannelHost(
+    GpuChannelHostCallback callback) {
+  DCHECK(!in_shutdown_)
+      << "The factory is shutting down, can't handle new requests";
+
+  gpu_channel_requests_.push_back(callback);
+  // If the channel is available, the factory will run the callback
+  // synchronously so we'll handle this request there.
+  EstablishGpuChannel();
+}
+
+scoped_refptr<cc::ContextProvider>
+ContextProviderFactoryImpl::CreateDisplayContextProvider(
     gpu::SurfaceHandle surface_handle,
     gpu::SharedMemoryLimits shared_memory_limits,
     gpu::gles2::ContextCreationAttribHelper attributes,
     bool support_locking,
     bool automatic_flushes,
-    ContextProviderCallback result_callback) {
-  DCHECK(!in_shutdown_)
-      << "The factory is shutting down, can't handle new requests";
+    gpu::GpuChannelHost* gpu_channel_host) {
   DCHECK(surface_handle != gpu::kNullSurfaceHandle);
-  CreateContextProviderInternal(
+  return CreateContextProviderInternal(
       command_buffer_metrics::DISPLAY_COMPOSITOR_ONSCREEN_CONTEXT,
       surface_handle, shared_memory_limits, attributes, support_locking,
-      automatic_flushes, nullptr, result_callback);
+      automatic_flushes, nullptr, gpu_channel_host);
 }
 
-void ContextProviderFactoryImpl::CreateOffscreenContextProvider(
+scoped_refptr<cc::ContextProvider>
+ContextProviderFactoryImpl::CreateOffscreenContextProvider(
     ContextType context_type,
     gpu::SharedMemoryLimits shared_memory_limits,
     gpu::gles2::ContextCreationAttribHelper attributes,
     bool support_locking,
     bool automatic_flushes,
     cc::ContextProvider* shared_context_provider,
-    ContextProviderCallback result_callback) {
-  DCHECK(!in_shutdown_)
-      << "The factory is shutting down, can't handle new requests";
-  CreateContextProviderInternal(ToCommandBufferContextType(context_type),
-                                gpu::kNullSurfaceHandle, shared_memory_limits,
-                                attributes, support_locking, automatic_flushes,
-                                shared_context_provider, result_callback);
+    gpu::GpuChannelHost* gpu_channel_host) {
+  return CreateContextProviderInternal(
+      ToCommandBufferContextType(context_type), gpu::kNullSurfaceHandle,
+      shared_memory_limits, attributes, support_locking, automatic_flushes,
+      shared_context_provider, gpu_channel_host);
 }
 
 cc::SurfaceManager* ContextProviderFactoryImpl::GetSurfaceManager() {
   if (!surface_manager_)
     surface_manager_ = base::WrapUnique(new cc::SurfaceManager);
 
   return surface_manager_.get();
 }
 
 uint32_t ContextProviderFactoryImpl::AllocateSurfaceClientId() {
   return ++surface_client_id_;
 }
 
 cc::SharedBitmapManager* ContextProviderFactoryImpl::GetSharedBitmapManager() {
   return HostSharedBitmapManager::current();
 }
 
 gpu::GpuMemoryBufferManager*
 ContextProviderFactoryImpl::GetGpuMemoryBufferManager() {
   return BrowserGpuMemoryBufferManager::current();
 }
 
-void ContextProviderFactoryImpl::CreateContextProviderInternal(
+scoped_refptr<cc::ContextProvider>
+ContextProviderFactoryImpl::CreateContextProviderInternal(
     command_buffer_metrics::ContextType context_type,
     gpu::SurfaceHandle surface_handle,
     gpu::SharedMemoryLimits shared_memory_limits,
     gpu::gles2::ContextCreationAttribHelper attributes,
     bool support_locking,
     bool automatic_flushes,
     cc::ContextProvider* shared_context_provider,
-    ContextProviderCallback result_callback) {
-  DCHECK(!result_callback.is_null());
-
-  ContextProvidersRequest context_request;
-  context_request.context_type = context_type;
-  context_request.surface_handle = surface_handle;
-  context_request.shared_memory_limits = shared_memory_limits;
-  context_request.attributes = attributes;
-  context_request.support_locking = support_locking;
-  context_request.automatic_flushes = automatic_flushes;
-  context_request.shared_context_provider = shared_context_provider;
-  context_request.result_callback = result_callback;
-
-  context_provider_requests_.push_back(context_request);
-
-  // If the channel is available, the factory will run the callback
-  // synchronously so we'll handle this request there.
-  EstablishGpuChannel();
+    gpu::GpuChannelHost* gpu_channel_host) {
+  return make_scoped_refptr(new ContextProviderCommandBuffer(
+      gpu_channel_host, gpu::GPU_STREAM_DEFAULT, gpu::GpuStreamPriority::NORMAL,

[danakj, 2016/09/01 23:35:19]

this gpu_channel_host is an implicit conversion to refptr. make all the
functions here receive this by value, and std::move() it around instead.

[Khushal, 2016/09/02 00:24:33]

Didn't see that it was a refptr. Done.

+      surface_handle,
+      GURL(std::string("chrome://gpu/ContextProviderFactoryImpl::") +
+           std::string("CompositorContextProvider")),
+      automatic_flushes, support_locking, shared_memory_limits, attributes,
+      static_cast<ContextProviderCommandBuffer*>(shared_context_provider),
+      context_type));
 }
 
 void ContextProviderFactoryImpl::HandlePendingRequests(
     scoped_refptr<gpu::GpuChannelHost> gpu_channel_host,
-    ContextCreationResult result) {
-  DCHECK(!context_provider_requests_.empty())
+    GpuChannelHostResult result) {
+  DCHECK(!gpu_channel_requests_.empty())
       << "We don't have any pending requests?";
 
-  // Failure to initialize the context could result in new requests. Handle
+  // Failure to initialize the channel could result in new requests. Handle
   // them after going through the current list.
   if (in_handle_pending_requests_)
     return;
 
   {
     base::AutoReset<bool> auto_reset_in_handle_requests(
         &in_handle_pending_requests_, true);
 
-    std::list<ContextProvidersRequest> context_requests =
-        context_provider_requests_;
-    context_provider_requests_.clear();
+    std::list<GpuChannelHostCallback> gpu_channel_requests =

[danakj, 2016/09/01 23:35:19]

Normally I'd see this written as

std::list<T> requests;
requests.swap(gpu_channel_requests_);

(ps why are we using std::list :( oh noes)

[Khushal, 2016/09/02 00:24:33]

Much better. Done.

+        gpu_channel_requests_;
+    gpu_channel_requests_.clear();
 
-    for (ContextProvidersRequest& context_request : context_requests) {
-      scoped_refptr<cc::ContextProvider> context_provider;
-      ContextCreationResult result_to_report = result;
-
-      const bool create_onscreen_context =
-          context_request.surface_handle != gpu::kNullSurfaceHandle;
-
-      // Is the request for an onscreen context? Make sure the surface is
-      // still valid in that case.
-      if (create_onscreen_context &&
-          !GpuSurfaceTracker::GetInstance()->IsValidSurfaceHandle(
-              context_request.surface_handle)) {
-        // Choose what to report based on severity, factory shutdown trumps
-        // everything, otherwise report GpuSurfaceHandle loss.
-        result_to_report =
-            result_to_report == ContextCreationResult::FAILURE_FACTORY_SHUTDOWN
-                ? result_to_report
-                : ContextCreationResult::FAILURE_GPU_SURFACE_HANDLE_LOST;
-      } else if (gpu_channel_host) {
-        DCHECK_EQ(ContextCreationResult::SUCCESS, result);
-
-        context_provider = new ContextProviderCommandBuffer(
-            gpu_channel_host, gpu::GPU_STREAM_DEFAULT,
-            gpu::GpuStreamPriority::NORMAL, context_request.surface_handle,
-            GURL(std::string("chrome://gpu/ContextProviderFactoryImpl::") +
-                 std::string("CompositorContextProvider")),
-            context_request.automatic_flushes, context_request.support_locking,
-            context_request.shared_memory_limits, context_request.attributes,
-            static_cast<ContextProviderCommandBuffer*>(
-                context_request.shared_context_provider),
-            context_request.context_type);
-      }
-
-      context_request.result_callback.Run(context_provider, result_to_report);
+    for (GpuChannelHostCallback& gpu_channel_request : gpu_channel_requests) {
+      DCHECK(gpu_channel_host || result != GpuChannelHostResult::SUCCESS);
+      gpu_channel_request.Run(gpu_channel_host, result);
     }
   }
 
-  if (!context_provider_requests_.empty())
+  if (!gpu_channel_requests_.empty())
     EstablishGpuChannel();
 }
 
 void ContextProviderFactoryImpl::EstablishGpuChannel() {
 #if defined(ADDRESS_SANITIZER) || defined(THREAD_SANITIZER) || \
     defined(SYZYASAN) || defined(CYGPROFILE_INSTRUMENTATION)
   const int64_t kGpuChannelTimeoutInSeconds = 40;
 #else
   const int64_t kGpuChannelTimeoutInSeconds = 10;
 #endif
 
   // Start the timer first, if the result comes synchronously, we want it to
   // stop in the callback.
   establish_gpu_channel_timeout_.Start(
       FROM_HERE, base::TimeDelta::FromSeconds(kGpuChannelTimeoutInSeconds),
       this, &ContextProviderFactoryImpl::OnGpuChannelTimeout);
 
   gpu_channel_factory_->EstablishGpuChannel(
       base::Bind(&ContextProviderFactoryImpl::OnGpuChannelEstablished,
                  weak_factory_.GetWeakPtr()));
 }
 
 void ContextProviderFactoryImpl::OnGpuChannelEstablished(
     scoped_refptr<gpu::GpuChannelHost> gpu_channel) {
   establish_gpu_channel_timeout_.Stop();
 
   // We can queue the Gpu Channel initialization requests multiple times as
   // we get context requests. So we might have already handled any pending
   // requests when this callback runs.
-  if (context_provider_requests_.empty())
+  if (gpu_channel_requests_.empty())
     return;
 
   if (gpu_channel) {
     HandlePendingRequests(std::move(gpu_channel),
-                          ContextCreationResult::SUCCESS);
+                          GpuChannelHostResult::SUCCESS);
   } else {
     HandlePendingRequests(
         nullptr,
-        ContextCreationResult::FAILURE_GPU_PROCESS_INITIALIZATION_FAILED);
+        GpuChannelHostResult::FAILURE_GPU_PROCESS_INITIALIZATION_FAILED);
   }
 }
 
 void ContextProviderFactoryImpl::OnGpuChannelTimeout() {
   LOG(FATAL) << "Timed out waiting for GPU channel.";
 }
 
 }  // namespace content