root/content/common/gpu/gpu_command_buffer_stub.cc

/* [<][>][^][v][top][bottom][index][help] */

DEFINITIONS

This source file includes following definitions.
  1. TrackMemoryAllocatedChange
  2. EnsureGPUMemoryAvailable
  3. FastSetActiveURL
  4. total_gpu_memory_
  5. GetMemoryManager
  6. OnMessageReceived
  7. Send
  8. IsScheduled
  9. HasMoreWork
  10. PollWork
  11. HasUnprocessedCommands
  12. ScheduleDelayedWork
  13. OnEcho
  14. MakeCurrent
  15. Destroy
  16. OnInitializeFailed
  17. OnInitialize
  18. OnSetLatencyInfo
  19. OnCreateStreamTexture
  20. SetLatencyInfoCallback
  21. GetRequestedAttribute
  22. OnSetGetBuffer
  23. OnProduceFrontBuffer
  24. OnGetState
  25. OnParseError
  26. OnGetStateFast
  27. OnAsyncFlush
  28. OnRescheduled
  29. OnRegisterTransferBuffer
  30. OnDestroyTransferBuffer
  31. OnCommandProcessed
  32. ReportState
  33. PutChanged
  34. OnCreateVideoDecoder
  35. OnCreateVideoEncoder
  36. OnSetSurfaceVisible
  37. AddSyncPoint
  38. OnRetireSyncPoint
  39. OnWaitSyncPoint
  40. OnSyncPointRetired
  41. OnSignalSyncPoint
  42. OnSignalSyncPointAck
  43. OnSignalQuery
  44. OnReceivedClientManagedMemoryStats
  45. OnSetClientHasMemoryAllocationChangedCallback
  46. OnRegisterGpuMemoryBuffer
  47. OnDestroyGpuMemoryBuffer
  48. SendConsoleMessage
  49. SendCachedShader
  50. AddDestructionObserver
  51. RemoveDestructionObserver
  52. SetPreemptByFlag
  53. GetTotalGpuMemory
  54. GetSurfaceSize
  55. GetMemoryTracker
  56. SetMemoryAllocation
  57. SuggestHaveFrontBuffer
  58. CheckContextLost
  59. MarkContextLost
  60. GetMemoryUsage

// Copyright (c) 2012 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 "base/bind.h"
#include "base/bind_helpers.h"
#include "base/command_line.h"
#include "base/debug/trace_event.h"
#include "base/hash.h"
#include "base/memory/shared_memory.h"
#include "base/time/time.h"
#include "build/build_config.h"
#include "content/common/gpu/devtools_gpu_instrumentation.h"
#include "content/common/gpu/gpu_channel.h"
#include "content/common/gpu/gpu_channel_manager.h"
#include "content/common/gpu/gpu_command_buffer_stub.h"
#include "content/common/gpu/gpu_memory_manager.h"
#include "content/common/gpu/gpu_memory_tracking.h"
#include "content/common/gpu/gpu_messages.h"
#include "content/common/gpu/gpu_watchdog.h"
#include "content/common/gpu/image_transport_surface.h"
#include "content/common/gpu/media/gpu_video_decode_accelerator.h"
#include "content/common/gpu/media/gpu_video_encode_accelerator.h"
#include "content/common/gpu/sync_point_manager.h"
#include "content/public/common/content_client.h"
#include "gpu/command_buffer/common/constants.h"
#include "gpu/command_buffer/common/gles2_cmd_utils.h"
#include "gpu/command_buffer/common/mailbox.h"
#include "gpu/command_buffer/service/gl_context_virtual.h"
#include "gpu/command_buffer/service/gl_state_restorer_impl.h"
#include "gpu/command_buffer/service/gpu_control_service.h"
#include "gpu/command_buffer/service/image_manager.h"
#include "gpu/command_buffer/service/logger.h"
#include "gpu/command_buffer/service/mailbox_manager.h"
#include "gpu/command_buffer/service/memory_tracking.h"
#include "gpu/command_buffer/service/query_manager.h"
#include "ui/gl/gl_bindings.h"
#include "ui/gl/gl_switches.h"

#if defined(OS_WIN)
#include "content/public/common/sandbox_init.h"
#endif

#if defined(OS_ANDROID)
#include "content/common/gpu/stream_texture_android.h"
#endif

namespace content {
namespace {

// The GpuCommandBufferMemoryTracker class provides a bridge between the
// ContextGroup's memory type managers and the GpuMemoryManager class.
class GpuCommandBufferMemoryTracker : public gpu::gles2::MemoryTracker {
 public:
  explicit GpuCommandBufferMemoryTracker(GpuChannel* channel) :
      tracking_group_(channel->gpu_channel_manager()->gpu_memory_manager()->
          CreateTrackingGroup(channel->renderer_pid(), this)) {
  }

  virtual void TrackMemoryAllocatedChange(
      size_t old_size,
      size_t new_size,
      gpu::gles2::MemoryTracker::Pool pool) OVERRIDE {
    tracking_group_->TrackMemoryAllocatedChange(
        old_size, new_size, pool);
  }

  virtual bool EnsureGPUMemoryAvailable(size_t size_needed) OVERRIDE {
    return tracking_group_->EnsureGPUMemoryAvailable(size_needed);
  };

 private:
  virtual ~GpuCommandBufferMemoryTracker() {
  }
  scoped_ptr<GpuMemoryTrackingGroup> tracking_group_;

  DISALLOW_COPY_AND_ASSIGN(GpuCommandBufferMemoryTracker);
};

// FastSetActiveURL will shortcut the expensive call to SetActiveURL when the
// url_hash matches.
void FastSetActiveURL(const GURL& url, size_t url_hash) {
  // Leave the previously set URL in the empty case -- empty URLs are given by
  // WebKitPlatformSupportImpl::createOffscreenGraphicsContext3D. Hopefully the
  // onscreen context URL was set previously and will show up even when a crash
  // occurs during offscreen command processing.
  if (url.is_empty())
    return;
  static size_t g_last_url_hash = 0;
  if (url_hash != g_last_url_hash) {
    g_last_url_hash = url_hash;
    GetContentClient()->SetActiveURL(url);
  }
}

// The first time polling a fence, delay some extra time to allow other
// stubs to process some work, or else the timing of the fences could
// allow a pattern of alternating fast and slow frames to occur.
const int64 kHandleMoreWorkPeriodMs = 2;
const int64 kHandleMoreWorkPeriodBusyMs = 1;

// Prevents idle work from being starved.
const int64 kMaxTimeSinceIdleMs = 10;

}  // namespace

GpuCommandBufferStub::GpuCommandBufferStub(
    GpuChannel* channel,
    GpuCommandBufferStub* share_group,
    const gfx::GLSurfaceHandle& handle,
    gpu::gles2::MailboxManager* mailbox_manager,
    gpu::gles2::ImageManager* image_manager,
    const gfx::Size& size,
    const gpu::gles2::DisallowedFeatures& disallowed_features,
    const std::vector<int32>& attribs,
    gfx::GpuPreference gpu_preference,
    bool use_virtualized_gl_context,
    int32 route_id,
    int32 surface_id,
    GpuWatchdog* watchdog,
    bool software,
    const GURL& active_url)
    : channel_(channel),
      handle_(handle),
      initial_size_(size),
      disallowed_features_(disallowed_features),
      requested_attribs_(attribs),
      gpu_preference_(gpu_preference),
      use_virtualized_gl_context_(use_virtualized_gl_context),
      route_id_(route_id),
      surface_id_(surface_id),
      software_(software),
      last_flush_count_(0),
      last_memory_allocation_valid_(false),
      watchdog_(watchdog),
      sync_point_wait_count_(0),
      delayed_work_scheduled_(false),
      previous_messages_processed_(0),
      active_url_(active_url),
      total_gpu_memory_(0) {
  active_url_hash_ = base::Hash(active_url.possibly_invalid_spec());
  FastSetActiveURL(active_url_, active_url_hash_);

  gpu::gles2::ContextCreationAttribHelper attrib_parser;
  attrib_parser.Parse(requested_attribs_);

  if (share_group) {
    context_group_ = share_group->context_group_;
    DCHECK(context_group_->bind_generates_resource() ==
           attrib_parser.bind_generates_resource_);
  } else {
    context_group_ = new gpu::gles2::ContextGroup(
        mailbox_manager,
        image_manager,
        new GpuCommandBufferMemoryTracker(channel),
        channel_->gpu_channel_manager()->shader_translator_cache(),
        NULL,
        attrib_parser.bind_generates_resource_);
  }

  use_virtualized_gl_context_ |=
      context_group_->feature_info()->workarounds().use_virtualized_gl_contexts;
}

GpuCommandBufferStub::~GpuCommandBufferStub() {
  Destroy();

  GpuChannelManager* gpu_channel_manager = channel_->gpu_channel_manager();
  gpu_channel_manager->Send(new GpuHostMsg_DestroyCommandBuffer(surface_id()));
}

GpuMemoryManager* GpuCommandBufferStub::GetMemoryManager() const {
    return channel()->gpu_channel_manager()->gpu_memory_manager();
}

bool GpuCommandBufferStub::OnMessageReceived(const IPC::Message& message) {
  devtools_gpu_instrumentation::ScopedGpuTask task(channel());
  FastSetActiveURL(active_url_, active_url_hash_);

  // Ensure the appropriate GL context is current before handling any IPC
  // messages directed at the command buffer. This ensures that the message
  // handler can assume that the context is current (not necessary for
  // Echo, RetireSyncPoint, or WaitSyncPoint).
  if (decoder_.get() &&
      message.type() != GpuCommandBufferMsg_Echo::ID &&
      message.type() != GpuCommandBufferMsg_GetStateFast::ID &&
      message.type() != GpuCommandBufferMsg_RetireSyncPoint::ID &&
      message.type() != GpuCommandBufferMsg_SetLatencyInfo::ID) {
    if (!MakeCurrent())
      return false;
  }

  // Always use IPC_MESSAGE_HANDLER_DELAY_REPLY for synchronous message handlers
  // here. This is so the reply can be delayed if the scheduler is unscheduled.
  bool handled = true;
  IPC_BEGIN_MESSAGE_MAP(GpuCommandBufferStub, message)
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_Initialize,
                                    OnInitialize);
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_SetGetBuffer,
                                    OnSetGetBuffer);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_ProduceFrontBuffer,
                        OnProduceFrontBuffer);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_Echo, OnEcho);
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_GetState, OnGetState);
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_GetStateFast,
                                    OnGetStateFast);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_AsyncFlush, OnAsyncFlush);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_SetLatencyInfo, OnSetLatencyInfo);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_Rescheduled, OnRescheduled);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_RegisterTransferBuffer,
                        OnRegisterTransferBuffer);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_DestroyTransferBuffer,
                        OnDestroyTransferBuffer);
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_CreateVideoDecoder,
                                    OnCreateVideoDecoder)
    IPC_MESSAGE_HANDLER_DELAY_REPLY(GpuCommandBufferMsg_CreateVideoEncoder,
                                    OnCreateVideoEncoder)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_SetSurfaceVisible,
                        OnSetSurfaceVisible)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_RetireSyncPoint,
                        OnRetireSyncPoint)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_SignalSyncPoint,
                        OnSignalSyncPoint)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_SignalQuery,
                        OnSignalQuery)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_SendClientManagedMemoryStats,
                        OnReceivedClientManagedMemoryStats)
    IPC_MESSAGE_HANDLER(
        GpuCommandBufferMsg_SetClientHasMemoryAllocationChangedCallback,
        OnSetClientHasMemoryAllocationChangedCallback)
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_RegisterGpuMemoryBuffer,
                        OnRegisterGpuMemoryBuffer);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_DestroyGpuMemoryBuffer,
                        OnDestroyGpuMemoryBuffer);
    IPC_MESSAGE_HANDLER(GpuCommandBufferMsg_CreateStreamTexture,
                        OnCreateStreamTexture)
    IPC_MESSAGE_UNHANDLED(handled = false)
  IPC_END_MESSAGE_MAP()

  // Ensure that any delayed work that was created will be handled.
  ScheduleDelayedWork(kHandleMoreWorkPeriodMs);

  DCHECK(handled);
  return handled;
}

bool GpuCommandBufferStub::Send(IPC::Message* message) {
  return channel_->Send(message);
}

bool GpuCommandBufferStub::IsScheduled() {
  return (!scheduler_.get() || scheduler_->IsScheduled());
}

bool GpuCommandBufferStub::HasMoreWork() {
  return scheduler_.get() && scheduler_->HasMoreWork();
}

void GpuCommandBufferStub::PollWork() {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::PollWork");
  delayed_work_scheduled_ = false;
  FastSetActiveURL(active_url_, active_url_hash_);
  if (decoder_.get() && !MakeCurrent())
    return;

  if (scheduler_) {
    bool fences_complete = scheduler_->PollUnscheduleFences();
    // Perform idle work if all fences are complete.
    if (fences_complete) {
      uint64 current_messages_processed =
          channel()->gpu_channel_manager()->MessagesProcessed();
      // We're idle when no messages were processed or scheduled.
      bool is_idle =
          (previous_messages_processed_ == current_messages_processed) &&
          !channel()->gpu_channel_manager()->HandleMessagesScheduled();
      if (!is_idle && !last_idle_time_.is_null()) {
        base::TimeDelta time_since_idle = base::TimeTicks::Now() -
            last_idle_time_;
        base::TimeDelta max_time_since_idle =
            base::TimeDelta::FromMilliseconds(kMaxTimeSinceIdleMs);

        // Force idle when it's been too long since last time we were idle.
        if (time_since_idle > max_time_since_idle)
          is_idle = true;
      }

      if (is_idle) {
        last_idle_time_ = base::TimeTicks::Now();
        scheduler_->PerformIdleWork();
      }
    }
  }
  ScheduleDelayedWork(kHandleMoreWorkPeriodBusyMs);
}

bool GpuCommandBufferStub::HasUnprocessedCommands() {
  if (command_buffer_) {
    gpu::CommandBuffer::State state = command_buffer_->GetLastState();
    return state.put_offset != state.get_offset &&
        !gpu::error::IsError(state.error);
  }
  return false;
}

void GpuCommandBufferStub::ScheduleDelayedWork(int64 delay) {
  if (!HasMoreWork()) {
    last_idle_time_ = base::TimeTicks();
    return;
  }

  if (delayed_work_scheduled_)
    return;
  delayed_work_scheduled_ = true;

  // Idle when no messages are processed between now and when
  // PollWork is called.
  previous_messages_processed_ =
      channel()->gpu_channel_manager()->MessagesProcessed();
  if (last_idle_time_.is_null())
    last_idle_time_ = base::TimeTicks::Now();

  // IsScheduled() returns true after passing all unschedule fences
  // and this is when we can start performing idle work. Idle work
  // is done synchronously so we can set delay to 0 and instead poll
  // for more work at the rate idle work is performed. This also ensures
  // that idle work is done as efficiently as possible without any
  // unnecessary delays.
  if (scheduler_.get() &&
      scheduler_->IsScheduled() &&
      scheduler_->HasMoreIdleWork()) {
    delay = 0;
  }

  base::MessageLoop::current()->PostDelayedTask(
      FROM_HERE,
      base::Bind(&GpuCommandBufferStub::PollWork, AsWeakPtr()),
      base::TimeDelta::FromMilliseconds(delay));
}

void GpuCommandBufferStub::OnEcho(const IPC::Message& message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnEcho");
  Send(new IPC::Message(message));
}

bool GpuCommandBufferStub::MakeCurrent() {
  if (decoder_->MakeCurrent())
    return true;
  DLOG(ERROR) << "Context lost because MakeCurrent failed.";
  command_buffer_->SetContextLostReason(decoder_->GetContextLostReason());
  command_buffer_->SetParseError(gpu::error::kLostContext);
  CheckContextLost();
  return false;
}

void GpuCommandBufferStub::Destroy() {
  if (handle_.is_null() && !active_url_.is_empty()) {
    GpuChannelManager* gpu_channel_manager = channel_->gpu_channel_manager();
    gpu_channel_manager->Send(new GpuHostMsg_DidDestroyOffscreenContext(
        active_url_));
  }

  memory_manager_client_state_.reset();

  while (!sync_points_.empty())
    OnRetireSyncPoint(sync_points_.front());

  if (decoder_)
    decoder_->set_engine(NULL);

  // The scheduler has raw references to the decoder and the command buffer so
  // destroy it before those.
  scheduler_.reset();

  bool have_context = false;
  if (decoder_ && command_buffer_ &&
      command_buffer_->GetState().error != gpu::error::kLostContext)
    have_context = decoder_->MakeCurrent();
  FOR_EACH_OBSERVER(DestructionObserver,
                    destruction_observers_,
                    OnWillDestroyStub());

  if (decoder_) {
    decoder_->Destroy(have_context);
    decoder_.reset();
  }

  command_buffer_.reset();

  // Remove this after crbug.com/248395 is sorted out.
  surface_ = NULL;
}

void GpuCommandBufferStub::OnInitializeFailed(IPC::Message* reply_message) {
  Destroy();
  GpuCommandBufferMsg_Initialize::WriteReplyParams(
      reply_message, false, gpu::Capabilities());
  Send(reply_message);
}

void GpuCommandBufferStub::OnInitialize(
    base::SharedMemoryHandle shared_state_handle,
    IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnInitialize");
  DCHECK(!command_buffer_.get());

  scoped_ptr<base::SharedMemory> shared_state_shm(
      new base::SharedMemory(shared_state_handle, false));

  command_buffer_.reset(new gpu::CommandBufferService(
      context_group_->transfer_buffer_manager()));

  bool result = command_buffer_->Initialize();
  DCHECK(result);

  decoder_.reset(::gpu::gles2::GLES2Decoder::Create(context_group_.get()));

  scheduler_.reset(new gpu::GpuScheduler(command_buffer_.get(),
                                         decoder_.get(),
                                         decoder_.get()));
  if (preemption_flag_.get())
    scheduler_->SetPreemptByFlag(preemption_flag_);

  decoder_->set_engine(scheduler_.get());

  if (!handle_.is_null()) {
#if defined(OS_MACOSX) || defined(UI_COMPOSITOR_IMAGE_TRANSPORT)
    if (software_) {
      LOG(ERROR) << "No software support.\n";
      OnInitializeFailed(reply_message);
      return;
    }
#endif

    surface_ = ImageTransportSurface::CreateSurface(
        channel_->gpu_channel_manager(),
        this,
        handle_);
  } else {
    GpuChannelManager* manager = channel_->gpu_channel_manager();
    surface_ = manager->GetDefaultOffscreenSurface();
  }

  if (!surface_.get()) {
    DLOG(ERROR) << "Failed to create surface.\n";
    OnInitializeFailed(reply_message);
    return;
  }

  scoped_refptr<gfx::GLContext> context;
  if (use_virtualized_gl_context_ && channel_->share_group()) {
    context = channel_->share_group()->GetSharedContext();
    if (!context.get()) {
      context = gfx::GLContext::CreateGLContext(
          channel_->share_group(),
          channel_->gpu_channel_manager()->GetDefaultOffscreenSurface(),
          gpu_preference_);
      channel_->share_group()->SetSharedContext(context.get());
    }
    // This should be a non-virtual GL context.
    DCHECK(context->GetHandle());
    context = new gpu::GLContextVirtual(
        channel_->share_group(), context.get(), decoder_->AsWeakPtr());
    if (!context->Initialize(surface_.get(), gpu_preference_)) {
      // TODO(sievers): The real context created above for the default
      // offscreen surface might not be compatible with this surface.
      // Need to adjust at least GLX to be able to create the initial context
      // with a config that is compatible with onscreen and offscreen surfaces.
      context = NULL;

      DLOG(ERROR) << "Failed to initialize virtual GL context.";
      OnInitializeFailed(reply_message);
      return;
    }
  }
  if (!context.get()) {
    context = gfx::GLContext::CreateGLContext(
        channel_->share_group(), surface_.get(), gpu_preference_);
  }
  if (!context.get()) {
    DLOG(ERROR) << "Failed to create context.\n";
    OnInitializeFailed(reply_message);
    return;
  }

  if (!context->MakeCurrent(surface_.get())) {
    LOG(ERROR) << "Failed to make context current.";
    OnInitializeFailed(reply_message);
    return;
  }

  if (!context->GetGLStateRestorer()) {
    context->SetGLStateRestorer(
        new gpu::GLStateRestorerImpl(decoder_->AsWeakPtr()));
  }

  if (!context->GetTotalGpuMemory(&total_gpu_memory_))
    total_gpu_memory_ = 0;

  if (!context_group_->has_program_cache()) {
    context_group_->set_program_cache(
        channel_->gpu_channel_manager()->program_cache());
  }

  // Initialize the decoder with either the view or pbuffer GLContext.
  if (!decoder_->Initialize(surface_,
                            context,
                            !surface_id(),
                            initial_size_,
                            disallowed_features_,
                            requested_attribs_)) {
    DLOG(ERROR) << "Failed to initialize decoder.";
    OnInitializeFailed(reply_message);
    return;
  }

  gpu_control_.reset(
      new gpu::GpuControlService(context_group_->image_manager(),
                                 NULL,
                                 context_group_->mailbox_manager(),
                                 NULL,
                                 decoder_->GetCapabilities()));

  if (CommandLine::ForCurrentProcess()->HasSwitch(
      switches::kEnableGPUServiceLogging)) {
    decoder_->set_log_commands(true);
  }

  decoder_->GetLogger()->SetMsgCallback(
      base::Bind(&GpuCommandBufferStub::SendConsoleMessage,
                 base::Unretained(this)));
  decoder_->SetShaderCacheCallback(
      base::Bind(&GpuCommandBufferStub::SendCachedShader,
                 base::Unretained(this)));
  decoder_->SetWaitSyncPointCallback(
      base::Bind(&GpuCommandBufferStub::OnWaitSyncPoint,
                 base::Unretained(this)));

  command_buffer_->SetPutOffsetChangeCallback(
      base::Bind(&GpuCommandBufferStub::PutChanged, base::Unretained(this)));
  command_buffer_->SetGetBufferChangeCallback(
      base::Bind(&gpu::GpuScheduler::SetGetBuffer,
                 base::Unretained(scheduler_.get())));
  command_buffer_->SetParseErrorCallback(
      base::Bind(&GpuCommandBufferStub::OnParseError, base::Unretained(this)));
  scheduler_->SetSchedulingChangedCallback(
      base::Bind(&GpuChannel::StubSchedulingChanged,
                 base::Unretained(channel_)));

  if (watchdog_) {
    scheduler_->SetCommandProcessedCallback(
        base::Bind(&GpuCommandBufferStub::OnCommandProcessed,
                   base::Unretained(this)));
  }

  const size_t kSharedStateSize = sizeof(gpu::CommandBufferSharedState);
  if (!shared_state_shm->Map(kSharedStateSize)) {
    DLOG(ERROR) << "Failed to map shared state buffer.";
    OnInitializeFailed(reply_message);
    return;
  }
  command_buffer_->SetSharedStateBuffer(gpu::MakeBackingFromSharedMemory(
      shared_state_shm.Pass(), kSharedStateSize));

  GpuCommandBufferMsg_Initialize::WriteReplyParams(
      reply_message, true, gpu_control_->GetCapabilities());
  Send(reply_message);

  if (handle_.is_null() && !active_url_.is_empty()) {
    GpuChannelManager* gpu_channel_manager = channel_->gpu_channel_manager();
    gpu_channel_manager->Send(new GpuHostMsg_DidCreateOffscreenContext(
        active_url_));
  }
}

void GpuCommandBufferStub::OnSetLatencyInfo(
    const std::vector<ui::LatencyInfo>& latency_info) {
  if (!ui::LatencyInfo::Verify(latency_info,
                               "GpuCommandBufferStub::OnSetLatencyInfo"))
    return;
  if (!latency_info_callback_.is_null())
    latency_info_callback_.Run(latency_info);
}

void GpuCommandBufferStub::OnCreateStreamTexture(uint32 texture_id,
                                                 int32* stream_id) {
#if defined(OS_ANDROID)
  *stream_id = StreamTexture::Create(this, texture_id);
#else
  *stream_id = 0;
#endif
}

void GpuCommandBufferStub::SetLatencyInfoCallback(
    const LatencyInfoCallback& callback) {
  latency_info_callback_ = callback;
}

int32 GpuCommandBufferStub::GetRequestedAttribute(int attr) const {
  // The command buffer is pairs of enum, value
  // search for the requested attribute, return the value.
  for (std::vector<int32>::const_iterator it = requested_attribs_.begin();
       it != requested_attribs_.end(); ++it) {
    if (*it++ == attr) {
      return *it;
    }
  }
  return -1;
}

void GpuCommandBufferStub::OnSetGetBuffer(int32 shm_id,
                                          IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnSetGetBuffer");
  if (command_buffer_)
    command_buffer_->SetGetBuffer(shm_id);
  Send(reply_message);
}

void GpuCommandBufferStub::OnProduceFrontBuffer(const gpu::Mailbox& mailbox) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnProduceFrontBuffer");
  if (!decoder_) {
    LOG(ERROR) << "Can't produce front buffer before initialization.";
    return;
  }

  decoder_->ProduceFrontBuffer(mailbox);
}

void GpuCommandBufferStub::OnGetState(IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnGetState");
  if (command_buffer_) {
    gpu::CommandBuffer::State state = command_buffer_->GetState();
    CheckContextLost();
    GpuCommandBufferMsg_GetState::WriteReplyParams(reply_message, state);
  } else {
    DLOG(ERROR) << "no command_buffer.";
    reply_message->set_reply_error();
  }
  Send(reply_message);
}

void GpuCommandBufferStub::OnParseError() {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnParseError");
  DCHECK(command_buffer_.get());
  gpu::CommandBuffer::State state = command_buffer_->GetState();
  IPC::Message* msg = new GpuCommandBufferMsg_Destroyed(
      route_id_, state.context_lost_reason);
  msg->set_unblock(true);
  Send(msg);

  // Tell the browser about this context loss as well, so it can
  // determine whether client APIs like WebGL need to be immediately
  // blocked from automatically running.
  GpuChannelManager* gpu_channel_manager = channel_->gpu_channel_manager();
  gpu_channel_manager->Send(new GpuHostMsg_DidLoseContext(
      handle_.is_null(), state.context_lost_reason, active_url_));

  CheckContextLost();
}

void GpuCommandBufferStub::OnGetStateFast(IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnGetStateFast");
  DCHECK(command_buffer_.get());
  CheckContextLost();
  gpu::CommandBuffer::State state = command_buffer_->GetState();
  GpuCommandBufferMsg_GetStateFast::WriteReplyParams(reply_message, state);
  Send(reply_message);
}

void GpuCommandBufferStub::OnAsyncFlush(int32 put_offset,
                                        uint32 flush_count) {
  TRACE_EVENT1("gpu", "GpuCommandBufferStub::OnAsyncFlush",
               "put_offset", put_offset);
  DCHECK(command_buffer_.get());
  if (flush_count - last_flush_count_ < 0x8000000U) {
    last_flush_count_ = flush_count;
    command_buffer_->Flush(put_offset);
  } else {
    // We received this message out-of-order. This should not happen but is here
    // to catch regressions. Ignore the message.
    NOTREACHED() << "Received a Flush message out-of-order";
  }

  ReportState();
}

void GpuCommandBufferStub::OnRescheduled() {
  gpu::CommandBuffer::State pre_state = command_buffer_->GetLastState();
  command_buffer_->Flush(pre_state.put_offset);
  gpu::CommandBuffer::State post_state = command_buffer_->GetLastState();

  if (pre_state.get_offset != post_state.get_offset)
    ReportState();
}

void GpuCommandBufferStub::OnRegisterTransferBuffer(
    int32 id,
    base::SharedMemoryHandle transfer_buffer,
    uint32 size) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnRegisterTransferBuffer");

  // Take ownership of the memory and map it into this process.
  // This validates the size.
  scoped_ptr<base::SharedMemory> shared_memory(
      new base::SharedMemory(transfer_buffer, false));
  if (!shared_memory->Map(size)) {
    DVLOG(0) << "Failed to map shared memory.";
    return;
  }

  if (command_buffer_) {
    command_buffer_->RegisterTransferBuffer(
        id, gpu::MakeBackingFromSharedMemory(shared_memory.Pass(), size));
  }
}

void GpuCommandBufferStub::OnDestroyTransferBuffer(int32 id) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnDestroyTransferBuffer");

  if (command_buffer_)
    command_buffer_->DestroyTransferBuffer(id);
}

void GpuCommandBufferStub::OnCommandProcessed() {
  if (watchdog_)
    watchdog_->CheckArmed();
}

void GpuCommandBufferStub::ReportState() {
  if (!CheckContextLost())
    command_buffer_->UpdateState();
}

void GpuCommandBufferStub::PutChanged() {
  FastSetActiveURL(active_url_, active_url_hash_);
  scheduler_->PutChanged();
}

void GpuCommandBufferStub::OnCreateVideoDecoder(
    media::VideoCodecProfile profile,
    IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnCreateVideoDecoder");
  int32 decoder_route_id = channel_->GenerateRouteID();
  GpuVideoDecodeAccelerator* decoder = new GpuVideoDecodeAccelerator(
      decoder_route_id, this, channel_->io_message_loop());
  decoder->Initialize(profile, reply_message);
  // decoder is registered as a DestructionObserver of this stub and will
  // self-delete during destruction of this stub.
}

void GpuCommandBufferStub::OnCreateVideoEncoder(
    media::VideoFrame::Format input_format,
    const gfx::Size& input_visible_size,
    media::VideoCodecProfile output_profile,
    uint32 initial_bitrate,
    IPC::Message* reply_message) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnCreateVideoEncoder");
  int32 encoder_route_id = channel_->GenerateRouteID();
  GpuVideoEncodeAccelerator* encoder =
      new GpuVideoEncodeAccelerator(encoder_route_id, this);
  encoder->Initialize(input_format,
                      input_visible_size,
                      output_profile,
                      initial_bitrate,
                      reply_message);
  // encoder is registered as a DestructionObserver of this stub and will
  // self-delete during destruction of this stub.
}

void GpuCommandBufferStub::OnSetSurfaceVisible(bool visible) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnSetSurfaceVisible");
  if (memory_manager_client_state_)
    memory_manager_client_state_->SetVisible(visible);
}

void GpuCommandBufferStub::AddSyncPoint(uint32 sync_point) {
  sync_points_.push_back(sync_point);
}

void GpuCommandBufferStub::OnRetireSyncPoint(uint32 sync_point) {
  DCHECK(!sync_points_.empty() && sync_points_.front() == sync_point);
  sync_points_.pop_front();
  if (context_group_->mailbox_manager()->UsesSync() && MakeCurrent())
    context_group_->mailbox_manager()->PushTextureUpdates();
  GpuChannelManager* manager = channel_->gpu_channel_manager();
  manager->sync_point_manager()->RetireSyncPoint(sync_point);
}

bool GpuCommandBufferStub::OnWaitSyncPoint(uint32 sync_point) {
  if (!sync_point)
    return true;
  GpuChannelManager* manager = channel_->gpu_channel_manager();
  if (manager->sync_point_manager()->IsSyncPointRetired(sync_point))
    return true;

  if (sync_point_wait_count_ == 0) {
    TRACE_EVENT_ASYNC_BEGIN1("gpu", "WaitSyncPoint", this,
                             "GpuCommandBufferStub", this);
  }
  scheduler_->SetScheduled(false);
  ++sync_point_wait_count_;
  manager->sync_point_manager()->AddSyncPointCallback(
      sync_point,
      base::Bind(&GpuCommandBufferStub::OnSyncPointRetired,
                 this->AsWeakPtr()));
  return scheduler_->IsScheduled();
}

void GpuCommandBufferStub::OnSyncPointRetired() {
  --sync_point_wait_count_;
  if (sync_point_wait_count_ == 0) {
    TRACE_EVENT_ASYNC_END1("gpu", "WaitSyncPoint", this,
                           "GpuCommandBufferStub", this);
  }
  scheduler_->SetScheduled(true);
}

void GpuCommandBufferStub::OnSignalSyncPoint(uint32 sync_point, uint32 id) {
  GpuChannelManager* manager = channel_->gpu_channel_manager();
  manager->sync_point_manager()->AddSyncPointCallback(
      sync_point,
      base::Bind(&GpuCommandBufferStub::OnSignalSyncPointAck,
                 this->AsWeakPtr(),
                 id));
}

void GpuCommandBufferStub::OnSignalSyncPointAck(uint32 id) {
  Send(new GpuCommandBufferMsg_SignalSyncPointAck(route_id_, id));
}

void GpuCommandBufferStub::OnSignalQuery(uint32 query_id, uint32 id) {
  if (decoder_) {
    gpu::gles2::QueryManager* query_manager = decoder_->GetQueryManager();
    if (query_manager) {
      gpu::gles2::QueryManager::Query* query =
          query_manager->GetQuery(query_id);
      if (query) {
        query->AddCallback(
          base::Bind(&GpuCommandBufferStub::OnSignalSyncPointAck,
                     this->AsWeakPtr(),
                     id));
        return;
      }
    }
  }
  // Something went wrong, run callback immediately.
  OnSignalSyncPointAck(id);
}


void GpuCommandBufferStub::OnReceivedClientManagedMemoryStats(
    const gpu::ManagedMemoryStats& stats) {
  TRACE_EVENT0(
      "gpu",
      "GpuCommandBufferStub::OnReceivedClientManagedMemoryStats");
  if (memory_manager_client_state_)
    memory_manager_client_state_->SetManagedMemoryStats(stats);
}

void GpuCommandBufferStub::OnSetClientHasMemoryAllocationChangedCallback(
    bool has_callback) {
  TRACE_EVENT0(
      "gpu",
      "GpuCommandBufferStub::OnSetClientHasMemoryAllocationChangedCallback");
  if (has_callback) {
    if (!memory_manager_client_state_) {
      memory_manager_client_state_.reset(GetMemoryManager()->CreateClientState(
          this, surface_id_ != 0, true));
    }
  } else {
    memory_manager_client_state_.reset();
  }
}

void GpuCommandBufferStub::OnRegisterGpuMemoryBuffer(
    int32 id,
    gfx::GpuMemoryBufferHandle gpu_memory_buffer,
    uint32 width,
    uint32 height,
    uint32 internalformat) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnRegisterGpuMemoryBuffer");
#if defined(OS_ANDROID)
  // Verify that renderer is not trying to use a surface texture it doesn't own.
  if (gpu_memory_buffer.type == gfx::SURFACE_TEXTURE_BUFFER &&
      gpu_memory_buffer.surface_texture_id.secondary_id !=
          channel()->client_id()) {
    LOG(ERROR) << "Illegal surface texture ID for renderer.";
    return;
  }
#endif
  if (gpu_control_) {
    gpu_control_->RegisterGpuMemoryBuffer(id,
                                          gpu_memory_buffer,
                                          width,
                                          height,
                                          internalformat);
  }
}

void GpuCommandBufferStub::OnDestroyGpuMemoryBuffer(int32 id) {
  TRACE_EVENT0("gpu", "GpuCommandBufferStub::OnDestroyGpuMemoryBuffer");
  if (gpu_control_)
    gpu_control_->DestroyGpuMemoryBuffer(id);
}

void GpuCommandBufferStub::SendConsoleMessage(
    int32 id,
    const std::string& message) {
  GPUCommandBufferConsoleMessage console_message;
  console_message.id = id;
  console_message.message = message;
  IPC::Message* msg = new GpuCommandBufferMsg_ConsoleMsg(
      route_id_, console_message);
  msg->set_unblock(true);
  Send(msg);
}

void GpuCommandBufferStub::SendCachedShader(
    const std::string& key, const std::string& shader) {
  channel_->CacheShader(key, shader);
}

void GpuCommandBufferStub::AddDestructionObserver(
    DestructionObserver* observer) {
  destruction_observers_.AddObserver(observer);
}

void GpuCommandBufferStub::RemoveDestructionObserver(
    DestructionObserver* observer) {
  destruction_observers_.RemoveObserver(observer);
}

void GpuCommandBufferStub::SetPreemptByFlag(
    scoped_refptr<gpu::PreemptionFlag> flag) {
  preemption_flag_ = flag;
  if (scheduler_)
    scheduler_->SetPreemptByFlag(preemption_flag_);
}

bool GpuCommandBufferStub::GetTotalGpuMemory(uint64* bytes) {
  *bytes = total_gpu_memory_;
  return !!total_gpu_memory_;
}

gfx::Size GpuCommandBufferStub::GetSurfaceSize() const {
  if (!surface_.get())
    return gfx::Size();
  return surface_->GetSize();
}

gpu::gles2::MemoryTracker* GpuCommandBufferStub::GetMemoryTracker() const {
  return context_group_->memory_tracker();
}

void GpuCommandBufferStub::SetMemoryAllocation(
    const gpu::MemoryAllocation& allocation) {
  if (!last_memory_allocation_valid_ ||
      !allocation.Equals(last_memory_allocation_)) {
    Send(new GpuCommandBufferMsg_SetMemoryAllocation(
        route_id_, allocation));
  }

  last_memory_allocation_valid_ = true;
  last_memory_allocation_ = allocation;
}

void GpuCommandBufferStub::SuggestHaveFrontBuffer(
    bool suggest_have_frontbuffer) {
  // This can be called outside of OnMessageReceived, so the context needs
  // to be made current before calling methods on the surface.
  if (surface_.get() && MakeCurrent())
    surface_->SetFrontbufferAllocation(suggest_have_frontbuffer);
}

bool GpuCommandBufferStub::CheckContextLost() {
  DCHECK(command_buffer_);
  gpu::CommandBuffer::State state = command_buffer_->GetState();
  bool was_lost = state.error == gpu::error::kLostContext;
  // Lose all other contexts if the reset was triggered by the robustness
  // extension instead of being synthetic.
  if (was_lost && decoder_ && decoder_->WasContextLostByRobustnessExtension() &&
      (gfx::GLContext::LosesAllContextsOnContextLost() ||
       use_virtualized_gl_context_))
    channel_->LoseAllContexts();
  return was_lost;
}

void GpuCommandBufferStub::MarkContextLost() {
  if (!command_buffer_ ||
      command_buffer_->GetState().error == gpu::error::kLostContext)
    return;

  command_buffer_->SetContextLostReason(gpu::error::kUnknown);
  if (decoder_)
    decoder_->LoseContext(GL_UNKNOWN_CONTEXT_RESET_ARB);
  command_buffer_->SetParseError(gpu::error::kLostContext);
}

uint64 GpuCommandBufferStub::GetMemoryUsage() const {
  return GetMemoryManager()->GetClientMemoryUsage(this);
}

}  // namespace content

/* [<][>][^][v][top][bottom][index][help] */