| // 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. |
| |
| #ifndef GPU_IPC_CLIENT_GPU_CHANNEL_HOST_H_ |
| #define GPU_IPC_CLIENT_GPU_CHANNEL_HOST_H_ |
| |
| #include <stddef.h> |
| #include <stdint.h> |
| |
| #include <memory> |
| #include <string> |
| #include <vector> |
| |
| #include "base/atomic_sequence_num.h" |
| #include "base/containers/scoped_ptr_hash_map.h" |
| #include "base/macros.h" |
| #include "base/memory/ref_counted.h" |
| #include "base/memory/weak_ptr.h" |
| #include "base/process/process.h" |
| #include "base/synchronization/lock.h" |
| #include "gpu/config/gpu_info.h" |
| #include "gpu/gpu_export.h" |
| #include "gpu/ipc/common/gpu_stream_constants.h" |
| #include "ipc/ipc_channel_handle.h" |
| #include "ipc/ipc_sync_channel.h" |
| #include "ipc/message_filter.h" |
| #include "ipc/message_router.h" |
| #include "ui/events/latency_info.h" |
| #include "ui/gfx/gpu_memory_buffer.h" |
| |
| namespace base { |
| class WaitableEvent; |
| } |
| |
| namespace IPC { |
| class SyncMessageFilter; |
| } |
| |
| namespace gpu { |
| class GpuMemoryBufferManager; |
| } |
| |
| namespace gpu { |
| |
| class GpuChannelHost; |
| using GpuChannelEstablishedCallback = |
| base::Callback<void(scoped_refptr<GpuChannelHost>)>; |
| |
| class GPU_EXPORT GpuChannelEstablishFactory { |
| public: |
| virtual ~GpuChannelEstablishFactory() {} |
| |
| virtual void EstablishGpuChannel( |
| const GpuChannelEstablishedCallback& callback) = 0; |
| virtual scoped_refptr<GpuChannelHost> EstablishGpuChannelSync() = 0; |
| virtual GpuMemoryBufferManager* GetGpuMemoryBufferManager() = 0; |
| }; |
| |
| class GPU_EXPORT GpuChannelHostFactory { |
| public: |
| virtual ~GpuChannelHostFactory() {} |
| |
| virtual bool IsMainThread() = 0; |
| virtual scoped_refptr<base::SingleThreadTaskRunner> |
| GetIOThreadTaskRunner() = 0; |
| virtual std::unique_ptr<base::SharedMemory> AllocateSharedMemory( |
| size_t size) = 0; |
| }; |
| |
| // Encapsulates an IPC channel between the client and one GPU process. |
| // On the GPU process side there's a corresponding GpuChannel. |
| // Every method can be called on any thread with a message loop, except for the |
| // IO thread. |
| class GPU_EXPORT GpuChannelHost |
| : public IPC::Sender, |
| public base::RefCountedThreadSafe<GpuChannelHost> { |
| public: |
| // Must be called on the main thread (as defined by the factory). |
| static scoped_refptr<GpuChannelHost> Create( |
| GpuChannelHostFactory* factory, |
| int channel_id, |
| const gpu::GPUInfo& gpu_info, |
| const IPC::ChannelHandle& channel_handle, |
| base::WaitableEvent* shutdown_event, |
| gpu::GpuMemoryBufferManager* gpu_memory_buffer_manager); |
| |
| bool IsLost() const { |
| DCHECK(channel_filter_.get()); |
| return channel_filter_->IsLost(); |
| } |
| |
| int channel_id() const { return channel_id_; } |
| |
| // The GPU stats reported by the GPU process. |
| const gpu::GPUInfo& gpu_info() const { return gpu_info_; } |
| |
| // IPC::Sender implementation: |
| bool Send(IPC::Message* msg) override; |
| |
| // Set an ordering barrier. AsyncFlushes any pending barriers on other |
| // routes. Combines multiple OrderingBarriers into a single AsyncFlush. |
| // Returns the flush ID for the stream or 0 if put offset was not changed. |
| // Outputs *highest_verified_flush_id. |
| uint32_t OrderingBarrier(int32_t route_id, |
| int32_t stream_id, |
| int32_t put_offset, |
| uint32_t flush_count, |
| const std::vector<ui::LatencyInfo>& latency_info, |
| bool put_offset_changed, |
| bool do_flush, |
| uint32_t* highest_verified_flush_id); |
| |
| void FlushPendingStream(int32_t stream_id); |
| |
| // Destroy this channel. Must be called on the main thread, before |
| // destruction. |
| void DestroyChannel(); |
| |
| // Add a message route for the current message loop. |
| void AddRoute(int route_id, base::WeakPtr<IPC::Listener> listener); |
| |
| // Add a message route to be handled on the provided |task_runner|. |
| void AddRouteWithTaskRunner( |
| int route_id, |
| base::WeakPtr<IPC::Listener> listener, |
| scoped_refptr<base::SingleThreadTaskRunner> task_runner); |
| |
| // Remove the message route associated with |route_id|. |
| void RemoveRoute(int route_id); |
| |
| GpuChannelHostFactory* factory() const { return factory_; } |
| |
| gpu::GpuMemoryBufferManager* gpu_memory_buffer_manager() const { |
| return gpu_memory_buffer_manager_; |
| } |
| |
| // Returns a handle to the shared memory that can be sent via IPC to the |
| // GPU process. The caller is responsible for ensuring it is closed. Returns |
| // an invalid handle on failure. |
| base::SharedMemoryHandle ShareToGpuProcess( |
| base::SharedMemoryHandle source_handle); |
| |
| // Reserve one unused transfer buffer ID. |
| int32_t ReserveTransferBufferId(); |
| |
| // Returns a GPU memory buffer handle to the buffer that can be sent via |
| // IPC to the GPU process. The caller is responsible for ensuring it is |
| // closed. Returns an invalid handle on failure. |
| gfx::GpuMemoryBufferHandle ShareGpuMemoryBufferToGpuProcess( |
| const gfx::GpuMemoryBufferHandle& source_handle, |
| bool* requires_sync_point); |
| |
| // Reserve one unused image ID. |
| int32_t ReserveImageId(); |
| |
| // Generate a route ID guaranteed to be unique for this channel. |
| int32_t GenerateRouteID(); |
| |
| // Generate a stream ID guaranteed to be unique for this channel. |
| int32_t GenerateStreamID(); |
| |
| // Sends a synchronous nop to the server which validate that all previous IPC |
| // messages have been received. Once the synchronous nop has been sent to the |
| // server all previous flushes will all be marked as validated, including |
| // flushes for other streams on the same channel. Once a validation has been |
| // sent, it will return the highest validated flush id for the stream. |
| // If the validation fails (which can only happen upon context lost), the |
| // highest validated flush id will not change. If no flush ID were ever |
| // validated then it will return 0 (Note the lowest valid flush ID is 1). |
| uint32_t ValidateFlushIDReachedServer(int32_t stream_id, bool force_validate); |
| |
| // Returns the highest validated flush ID for a given stream. |
| uint32_t GetHighestValidatedFlushID(int32_t stream_id); |
| |
| private: |
| friend class base::RefCountedThreadSafe<GpuChannelHost>; |
| |
| // A filter used internally to route incoming messages from the IO thread |
| // to the correct message loop. It also maintains some shared state between |
| // all the contexts. |
| class MessageFilter : public IPC::MessageFilter { |
| public: |
| MessageFilter(); |
| |
| // Called on the IO thread. |
| void AddRoute(int32_t route_id, |
| base::WeakPtr<IPC::Listener> listener, |
| scoped_refptr<base::SingleThreadTaskRunner> task_runner); |
| // Called on the IO thread. |
| void RemoveRoute(int32_t route_id); |
| |
| // IPC::MessageFilter implementation |
| // (called on the IO thread): |
| bool OnMessageReceived(const IPC::Message& msg) override; |
| void OnChannelError() override; |
| |
| // The following methods can be called on any thread. |
| |
| // Whether the channel is lost. |
| bool IsLost() const; |
| |
| private: |
| struct ListenerInfo { |
| ListenerInfo(); |
| ListenerInfo(const ListenerInfo& other); |
| ~ListenerInfo(); |
| |
| base::WeakPtr<IPC::Listener> listener; |
| scoped_refptr<base::SingleThreadTaskRunner> task_runner; |
| }; |
| |
| ~MessageFilter() override; |
| |
| // Threading notes: |listeners_| is only accessed on the IO thread. Every |
| // other field is protected by |lock_|. |
| base::hash_map<int32_t, ListenerInfo> listeners_; |
| |
| // Protects all fields below this one. |
| mutable base::Lock lock_; |
| |
| // Whether the channel has been lost. |
| bool lost_; |
| }; |
| |
| struct StreamFlushInfo { |
| StreamFlushInfo(); |
| StreamFlushInfo(const StreamFlushInfo& other); |
| ~StreamFlushInfo(); |
| |
| // These are global per stream. |
| uint32_t next_stream_flush_id; |
| uint32_t flushed_stream_flush_id; |
| uint32_t verified_stream_flush_id; |
| |
| // These are local per context. |
| bool flush_pending; |
| int32_t route_id; |
| int32_t put_offset; |
| uint32_t flush_count; |
| uint32_t flush_id; |
| std::vector<ui::LatencyInfo> latency_info; |
| }; |
| |
| GpuChannelHost(GpuChannelHostFactory* factory, |
| int channel_id, |
| const gpu::GPUInfo& gpu_info, |
| gpu::GpuMemoryBufferManager* gpu_memory_buffer_manager); |
| ~GpuChannelHost() override; |
| void Connect(const IPC::ChannelHandle& channel_handle, |
| base::WaitableEvent* shutdown_event); |
| bool InternalSend(IPC::Message* msg); |
| void InternalFlush(StreamFlushInfo* flush_info); |
| |
| // Threading notes: all fields are constant during the lifetime of |this| |
| // except: |
| // - |next_image_id_|, atomic type |
| // - |next_route_id_|, atomic type |
| // - |next_stream_id_|, atomic type |
| // - |channel_| and |stream_flush_info_|, protected by |context_lock_| |
| GpuChannelHostFactory* const factory_; |
| |
| const int channel_id_; |
| const gpu::GPUInfo gpu_info_; |
| |
| scoped_refptr<MessageFilter> channel_filter_; |
| |
| gpu::GpuMemoryBufferManager* gpu_memory_buffer_manager_; |
| |
| // A filter for sending messages from thread other than the main thread. |
| scoped_refptr<IPC::SyncMessageFilter> sync_filter_; |
| |
| // Image IDs are allocated in sequence. |
| base::AtomicSequenceNumber next_image_id_; |
| |
| // Route IDs are allocated in sequence. |
| base::AtomicSequenceNumber next_route_id_; |
| |
| // Stream IDs are allocated in sequence. |
| base::AtomicSequenceNumber next_stream_id_; |
| |
| // Protects channel_ and stream_flush_info_. |
| mutable base::Lock context_lock_; |
| std::unique_ptr<IPC::SyncChannel> channel_; |
| base::hash_map<int32_t, StreamFlushInfo> stream_flush_info_; |
| |
| DISALLOW_COPY_AND_ASSIGN(GpuChannelHost); |
| }; |
| |
| } // namespace gpu |
| |
| #endif // GPU_IPC_CLIENT_GPU_CHANNEL_HOST_H_ |