src/dawn_wire/client/ApiProcs.cpp - dawn - Git at Google

 // Copyright 2019 The Dawn Authors
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
 // You may obtain a copy of the License at
 //
 //     http://www.apache.org/licenses/LICENSE-2.0
 //
 // Unless required by applicable law or agreed to in writing, software
 // distributed under the License is distributed on an "AS IS" BASIS,
 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 // See the License for the specific language governing permissions and
 // limitations under the License.

 #include "dawn_wire/client/ApiObjects.h"
 #include "dawn_wire/client/ApiProcs_autogen.h"
 #include "dawn_wire/client/Client.h"

 namespace dawn_wire { namespace client {

     namespace {
         template <typename Handle>
         void SerializeBufferMapAsync(const Buffer* buffer, uint32_t serial, Handle* handle) {
             // TODO(enga): Remove the template when Read/Write handles are combined in a tagged
             // pointer.
             constexpr bool isWrite =
                 std::is_same<Handle, MemoryTransferService::WriteHandle>::value;

             // Get the serialization size of the handle.
             size_t handleCreateInfoLength = handle->SerializeCreateSize();

             BufferMapAsyncCmd cmd;
             cmd.bufferId = buffer->id;
             cmd.requestSerial = serial;
             cmd.isWrite = isWrite;
             cmd.handleCreateInfoLength = handleCreateInfoLength;
             cmd.handleCreateInfo = nullptr;

             size_t commandSize = cmd.GetRequiredSize();
             size_t requiredSize = commandSize + handleCreateInfoLength;
             char* allocatedBuffer =
                 static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
             cmd.Serialize(allocatedBuffer);
             // Serialize the handle into the space after the command.
             handle->SerializeCreate(allocatedBuffer + commandSize);
         }
     }  // namespace

     void ClientHandwrittenBufferMapReadAsync(WGPUBuffer cBuffer,
                                              WGPUBufferMapReadCallback callback,
                                              void* userdata) {
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         uint32_t serial = buffer->requestSerial++;
         ASSERT(buffer->requests.find(serial) == buffer->requests.end());

         // Create a ReadHandle for the map request. This is the client's intent to read GPU
         // memory.
         MemoryTransferService::ReadHandle* readHandle =
             buffer->device->GetClient()->GetMemoryTransferService()->CreateReadHandle(buffer->size);
         if (readHandle == nullptr) {
             callback(WGPUBufferMapAsyncStatus_DeviceLost, nullptr, 0, userdata);
             return;
         }

         Buffer::MapRequestData request = {};
         request.readCallback = callback;
         request.userdata = userdata;
         // The handle is owned by the MapRequest until the callback returns.
         request.readHandle = std::unique_ptr<MemoryTransferService::ReadHandle>(readHandle);

         // Store a mapping from serial -> MapRequest. The client can map/unmap before the map
         // operations are returned by the server so multiple requests may be in flight.
         buffer->requests[serial] = std::move(request);

         SerializeBufferMapAsync(buffer, serial, readHandle);
     }

     void ClientHandwrittenBufferMapWriteAsync(WGPUBuffer cBuffer,
                                               WGPUBufferMapWriteCallback callback,
                                               void* userdata) {
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         uint32_t serial = buffer->requestSerial++;
         ASSERT(buffer->requests.find(serial) == buffer->requests.end());

         // Create a WriteHandle for the map request. This is the client's intent to write GPU
         // memory.
         MemoryTransferService::WriteHandle* writeHandle =
             buffer->device->GetClient()->GetMemoryTransferService()->CreateWriteHandle(
                 buffer->size);
         if (writeHandle == nullptr) {
             callback(WGPUBufferMapAsyncStatus_DeviceLost, nullptr, 0, userdata);
             return;
         }

         Buffer::MapRequestData request = {};
         request.writeCallback = callback;
         request.userdata = userdata;
         // The handle is owned by the MapRequest until the callback returns.
         request.writeHandle = std::unique_ptr<MemoryTransferService::WriteHandle>(writeHandle);

         // Store a mapping from serial -> MapRequest. The client can map/unmap before the map
         // operations are returned by the server so multiple requests may be in flight.
         buffer->requests[serial] = std::move(request);

         SerializeBufferMapAsync(buffer, serial, writeHandle);
     }

     WGPUBuffer ClientHandwrittenDeviceCreateBuffer(WGPUDevice cDevice,
                                                    const WGPUBufferDescriptor* descriptor) {
         Device* device = reinterpret_cast<Device*>(cDevice);
         Client* wireClient = device->GetClient();

         auto* bufferObjectAndSerial = wireClient->BufferAllocator().New(device);
         Buffer* buffer = bufferObjectAndSerial->object.get();
         // Store the size of the buffer so that mapping operations can allocate a
         // MemoryTransfer handle of the proper size.
         buffer->size = descriptor->size;

         DeviceCreateBufferCmd cmd;
         cmd.self = cDevice;
         cmd.descriptor = descriptor;
         cmd.result = ObjectHandle{buffer->id, bufferObjectAndSerial->generation};

         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *wireClient);

         return reinterpret_cast<WGPUBuffer>(buffer);
     }

     WGPUCreateBufferMappedResult ClientHandwrittenDeviceCreateBufferMapped(
         WGPUDevice cDevice,
         const WGPUBufferDescriptor* descriptor) {
         Device* device = reinterpret_cast<Device*>(cDevice);
         Client* wireClient = device->GetClient();

         auto* bufferObjectAndSerial = wireClient->BufferAllocator().New(device);
         Buffer* buffer = bufferObjectAndSerial->object.get();
         buffer->size = descriptor->size;

         WGPUCreateBufferMappedResult result;
         result.buffer = reinterpret_cast<WGPUBuffer>(buffer);
         result.data = nullptr;
         result.dataLength = 0;

         // Create a WriteHandle for the map request. This is the client's intent to write GPU
         // memory.
         std::unique_ptr<MemoryTransferService::WriteHandle> writeHandle =
             std::unique_ptr<MemoryTransferService::WriteHandle>(
                 wireClient->GetMemoryTransferService()->CreateWriteHandle(descriptor->size));

         if (writeHandle == nullptr) {
             // TODO(enga): Support context lost generated by the client.
             return result;
         }

         // CreateBufferMapped is synchronous and the staging buffer for upload should be immediately
         // available.
         // Open the WriteHandle. This returns a pointer and size of mapped memory.
         // |result.data| may be null on error.
         std::tie(result.data, result.dataLength) = writeHandle->Open();

         if (result.data == nullptr) {
             // TODO(enga): Support context lost generated by the client.
             return result;
         }

         // Successfully created staging memory. The buffer now owns the WriteHandle.
         buffer->writeHandle = std::move(writeHandle);

         // Get the serialization size of the WriteHandle.
         size_t handleCreateInfoLength = buffer->writeHandle->SerializeCreateSize();

         DeviceCreateBufferMappedCmd cmd;
         cmd.device = cDevice;
         cmd.descriptor = descriptor;
         cmd.result = ObjectHandle{buffer->id, bufferObjectAndSerial->generation};
         cmd.handleCreateInfoLength = handleCreateInfoLength;
         cmd.handleCreateInfo = nullptr;

         size_t commandSize = cmd.GetRequiredSize();
         size_t requiredSize = commandSize + handleCreateInfoLength;
         char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *wireClient);
         // Serialize the WriteHandle into the space after the command.
         buffer->writeHandle->SerializeCreate(allocatedBuffer + commandSize);

         return result;
     }

     void ClientHandwrittenDevicePushErrorScope(WGPUDevice cDevice, WGPUErrorFilter filter) {
         Device* device = reinterpret_cast<Device*>(cDevice);
         device->PushErrorScope(filter);
     }

     bool ClientHandwrittenDevicePopErrorScope(WGPUDevice cDevice,
                                               WGPUErrorCallback callback,
                                               void* userdata) {
         Device* device = reinterpret_cast<Device*>(cDevice);
         return device->RequestPopErrorScope(callback, userdata);
     }

     uint64_t ClientHandwrittenFenceGetCompletedValue(WGPUFence cSelf) {
         auto fence = reinterpret_cast<Fence*>(cSelf);
         return fence->completedValue;
     }

     void ClientHandwrittenFenceOnCompletion(WGPUFence cFence,
                                             uint64_t value,
                                             WGPUFenceOnCompletionCallback callback,
                                             void* userdata) {
         Fence* fence = reinterpret_cast<Fence*>(cFence);
         if (value > fence->signaledValue) {
             ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
                                     WGPUErrorType_Validation,
                                     "Value greater than fence signaled value");
             callback(WGPUFenceCompletionStatus_Error, userdata);
             return;
         }

         if (value <= fence->completedValue) {
             callback(WGPUFenceCompletionStatus_Success, userdata);
             return;
         }

         Fence::OnCompletionData request;
         request.completionCallback = callback;
         request.userdata = userdata;
         fence->requests.Enqueue(std::move(request), value);
     }

     void ClientHandwrittenBufferSetSubData(WGPUBuffer cBuffer,
                                            uint64_t start,
                                            uint64_t count,
                                            const void* data) {
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         BufferSetSubDataInternalCmd cmd;
         cmd.bufferId = buffer->id;
         cmd.start = start;
         cmd.count = count;
         cmd.data = static_cast<const uint8_t*>(data);

         Client* wireClient = buffer->device->GetClient();
         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer);
     }

     void ClientHandwrittenBufferUnmap(WGPUBuffer cBuffer) {
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         // Invalidate the local pointer, and cancel all other in-flight requests that would
         // turn into errors anyway (you can't double map). This prevents race when the following
         // happens, where the application code would have unmapped a buffer but still receive a
         // callback:
         //   - Client -> Server: MapRequest1, Unmap, MapRequest2
         //   - Server -> Client: Result of MapRequest1
         //   - Unmap locally on the client
         //   - Server -> Client: Result of MapRequest2
         if (buffer->writeHandle) {
             // Writes need to be flushed before Unmap is sent. Unmap calls all associated
             // in-flight callbacks which may read the updated data.
             ASSERT(buffer->readHandle == nullptr);

             // Get the serialization size of metadata to flush writes.
             size_t writeFlushInfoLength = buffer->writeHandle->SerializeFlushSize();

             BufferUpdateMappedDataCmd cmd;
             cmd.bufferId = buffer->id;
             cmd.writeFlushInfoLength = writeFlushInfoLength;
             cmd.writeFlushInfo = nullptr;

             size_t commandSize = cmd.GetRequiredSize();
             size_t requiredSize = commandSize + writeFlushInfoLength;
             char* allocatedBuffer =
                 static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
             cmd.Serialize(allocatedBuffer);
             // Serialize flush metadata into the space after the command.
             // This closes the handle for writing.
             buffer->writeHandle->SerializeFlush(allocatedBuffer + commandSize);
             buffer->writeHandle = nullptr;

         } else if (buffer->readHandle) {
             buffer->readHandle = nullptr;
         }
         buffer->ClearMapRequests(WGPUBufferMapAsyncStatus_Unknown);

         BufferUnmapCmd cmd;
         cmd.self = cBuffer;
         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer =
             static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *buffer->device->GetClient());
     }

     void ClientHandwrittenBufferDestroy(WGPUBuffer cBuffer) {
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         // Cancel or remove all mappings
         buffer->writeHandle = nullptr;
         buffer->readHandle = nullptr;
         buffer->ClearMapRequests(WGPUBufferMapAsyncStatus_Unknown);

         BufferDestroyCmd cmd;
         cmd.self = cBuffer;
         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer =
             static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *buffer->device->GetClient());
     }

     WGPUFence ClientHandwrittenQueueCreateFence(WGPUQueue cSelf,
                                                 WGPUFenceDescriptor const* descriptor) {
         Queue* queue = reinterpret_cast<Queue*>(cSelf);
         Device* device = queue->device;

         QueueCreateFenceCmd cmd;
         cmd.self = cSelf;
         auto* allocation = device->GetClient()->FenceAllocator().New(device);
         cmd.result = ObjectHandle{allocation->object->id, allocation->generation};
         cmd.descriptor = descriptor;

         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer = static_cast<char*>(device->GetClient()->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *device->GetClient());

         WGPUFence cFence = reinterpret_cast<WGPUFence>(allocation->object.get());

         Fence* fence = reinterpret_cast<Fence*>(cFence);
         fence->queue = queue;

         uint64_t initialValue = descriptor != nullptr ? descriptor->initialValue : 0u;
         fence->signaledValue = initialValue;
         fence->completedValue = initialValue;
         return cFence;
     }

     void ClientHandwrittenQueueSignal(WGPUQueue cQueue, WGPUFence cFence, uint64_t signalValue) {
         Fence* fence = reinterpret_cast<Fence*>(cFence);
         Queue* queue = reinterpret_cast<Queue*>(cQueue);
         if (fence->queue != queue) {
             ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
                                     WGPUErrorType_Validation,
                                     "Fence must be signaled on the queue on which it was created.");
             return;
         }
         if (signalValue <= fence->signaledValue) {
             ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
                                     WGPUErrorType_Validation,
                                     "Fence value less than or equal to signaled value");
             return;
         }
         fence->signaledValue = signalValue;

         QueueSignalCmd cmd;
         cmd.self = cQueue;
         cmd.fence = cFence;
         cmd.signalValue = signalValue;

         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer =
             static_cast<char*>(fence->device->GetClient()->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer, *fence->device->GetClient());
     }

     void ClientHandwrittenQueueWriteBuffer(WGPUQueue cQueue,
                                            WGPUBuffer cBuffer,
                                            uint64_t bufferOffset,
                                            const void* data,
                                            size_t size) {
         Queue* queue = reinterpret_cast<Queue*>(cQueue);
         Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

         QueueWriteBufferInternalCmd cmd;
         cmd.queueId = queue->id;
         cmd.bufferId = buffer->id;
         cmd.bufferOffset = bufferOffset;
         cmd.data = static_cast<const uint8_t*>(data);
         cmd.size = size;

         Client* wireClient = buffer->device->GetClient();
         size_t requiredSize = cmd.GetRequiredSize();
         char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
         cmd.Serialize(allocatedBuffer);
     }

     void ClientDeviceReference(WGPUDevice) {
     }

     void ClientDeviceRelease(WGPUDevice) {
     }

     WGPUQueue ClientHandwrittenDeviceGetDefaultQueue(WGPUDevice cSelf) {
         Device* device = reinterpret_cast<Device*>(cSelf);
         return device->GetDefaultQueue();
     }

     void ClientHandwrittenDeviceSetUncapturedErrorCallback(WGPUDevice cSelf,
                                                            WGPUErrorCallback callback,
                                                            void* userdata) {
         Device* device = reinterpret_cast<Device*>(cSelf);
         device->SetUncapturedErrorCallback(callback, userdata);
     }
     void ClientHandwrittenDeviceSetDeviceLostCallback(WGPUDevice cSelf,
                                                       WGPUDeviceLostCallback callback,
                                                       void* userdata) {
         Device* device = reinterpret_cast<Device*>(cSelf);
         device->SetDeviceLostCallback(callback, userdata);
     }

 }}  // namespace dawn_wire::client
	// Copyright 2019 The Dawn Authors
	//
	// Licensed under the Apache License, Version 2.0 (the "License");
	// you may not use this file except in compliance with the License.
	// You may obtain a copy of the License at
	//
	// http://www.apache.org/licenses/LICENSE-2.0
	//
	// Unless required by applicable law or agreed to in writing, software
	// distributed under the License is distributed on an "AS IS" BASIS,
	// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
	// See the License for the specific language governing permissions and
	// limitations under the License.

	#include "dawn_wire/client/ApiObjects.h"
	#include "dawn_wire/client/ApiProcs_autogen.h"
	#include "dawn_wire/client/Client.h"

	namespace dawn_wire { namespace client {

	namespace {
	template <typename Handle>
	void SerializeBufferMapAsync(const Buffer* buffer, uint32_t serial, Handle* handle) {
	// TODO(enga): Remove the template when Read/Write handles are combined in a tagged
	// pointer.
	constexpr bool isWrite =
	std::is_same<Handle, MemoryTransferService::WriteHandle>::value;

	// Get the serialization size of the handle.
	size_t handleCreateInfoLength = handle->SerializeCreateSize();

	BufferMapAsyncCmd cmd;
	cmd.bufferId = buffer->id;
	cmd.requestSerial = serial;
	cmd.isWrite = isWrite;
	cmd.handleCreateInfoLength = handleCreateInfoLength;
	cmd.handleCreateInfo = nullptr;

	size_t commandSize = cmd.GetRequiredSize();
	size_t requiredSize = commandSize + handleCreateInfoLength;
	char* allocatedBuffer =
	static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer);
	// Serialize the handle into the space after the command.
	handle->SerializeCreate(allocatedBuffer + commandSize);
	}
	} // namespace

	void ClientHandwrittenBufferMapReadAsync(WGPUBuffer cBuffer,
	WGPUBufferMapReadCallback callback,
	void* userdata) {
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	uint32_t serial = buffer->requestSerial++;
	ASSERT(buffer->requests.find(serial) == buffer->requests.end());

	// Create a ReadHandle for the map request. This is the client's intent to read GPU
	// memory.
	MemoryTransferService::ReadHandle* readHandle =
	buffer->device->GetClient()->GetMemoryTransferService()->CreateReadHandle(buffer->size);
	if (readHandle == nullptr) {
	callback(WGPUBufferMapAsyncStatus_DeviceLost, nullptr, 0, userdata);
	return;
	}

	Buffer::MapRequestData request = {};
	request.readCallback = callback;
	request.userdata = userdata;
	// The handle is owned by the MapRequest until the callback returns.
	request.readHandle = std::unique_ptr<MemoryTransferService::ReadHandle>(readHandle);

	// Store a mapping from serial -> MapRequest. The client can map/unmap before the map
	// operations are returned by the server so multiple requests may be in flight.
	buffer->requests[serial] = std::move(request);

	SerializeBufferMapAsync(buffer, serial, readHandle);
	}

	void ClientHandwrittenBufferMapWriteAsync(WGPUBuffer cBuffer,
	WGPUBufferMapWriteCallback callback,
	void* userdata) {
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	uint32_t serial = buffer->requestSerial++;
	ASSERT(buffer->requests.find(serial) == buffer->requests.end());

	// Create a WriteHandle for the map request. This is the client's intent to write GPU
	// memory.
	MemoryTransferService::WriteHandle* writeHandle =
	buffer->device->GetClient()->GetMemoryTransferService()->CreateWriteHandle(
	buffer->size);
	if (writeHandle == nullptr) {
	callback(WGPUBufferMapAsyncStatus_DeviceLost, nullptr, 0, userdata);
	return;
	}

	Buffer::MapRequestData request = {};
	request.writeCallback = callback;
	request.userdata = userdata;
	// The handle is owned by the MapRequest until the callback returns.
	request.writeHandle = std::unique_ptr<MemoryTransferService::WriteHandle>(writeHandle);

	// Store a mapping from serial -> MapRequest. The client can map/unmap before the map
	// operations are returned by the server so multiple requests may be in flight.
	buffer->requests[serial] = std::move(request);

	SerializeBufferMapAsync(buffer, serial, writeHandle);
	}

	WGPUBuffer ClientHandwrittenDeviceCreateBuffer(WGPUDevice cDevice,
	const WGPUBufferDescriptor* descriptor) {
	Device* device = reinterpret_cast<Device*>(cDevice);
	Client* wireClient = device->GetClient();

	auto* bufferObjectAndSerial = wireClient->BufferAllocator().New(device);
	Buffer* buffer = bufferObjectAndSerial->object.get();
	// Store the size of the buffer so that mapping operations can allocate a
	// MemoryTransfer handle of the proper size.
	buffer->size = descriptor->size;

	DeviceCreateBufferCmd cmd;
	cmd.self = cDevice;
	cmd.descriptor = descriptor;
	cmd.result = ObjectHandle{buffer->id, bufferObjectAndSerial->generation};

	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *wireClient);

	return reinterpret_cast<WGPUBuffer>(buffer);
	}

	WGPUCreateBufferMappedResult ClientHandwrittenDeviceCreateBufferMapped(
	WGPUDevice cDevice,
	const WGPUBufferDescriptor* descriptor) {
	Device* device = reinterpret_cast<Device*>(cDevice);
	Client* wireClient = device->GetClient();

	auto* bufferObjectAndSerial = wireClient->BufferAllocator().New(device);
	Buffer* buffer = bufferObjectAndSerial->object.get();
	buffer->size = descriptor->size;

	WGPUCreateBufferMappedResult result;
	result.buffer = reinterpret_cast<WGPUBuffer>(buffer);
	result.data = nullptr;
	result.dataLength = 0;

	// Create a WriteHandle for the map request. This is the client's intent to write GPU
	// memory.
	std::unique_ptr<MemoryTransferService::WriteHandle> writeHandle =
	std::unique_ptr<MemoryTransferService::WriteHandle>(
	wireClient->GetMemoryTransferService()->CreateWriteHandle(descriptor->size));

	if (writeHandle == nullptr) {
	// TODO(enga): Support context lost generated by the client.
	return result;
	}

	// CreateBufferMapped is synchronous and the staging buffer for upload should be immediately
	// available.
	// Open the WriteHandle. This returns a pointer and size of mapped memory.
	// \|result.data\| may be null on error.
	std::tie(result.data, result.dataLength) = writeHandle->Open();

	if (result.data == nullptr) {
	// TODO(enga): Support context lost generated by the client.
	return result;
	}

	// Successfully created staging memory. The buffer now owns the WriteHandle.
	buffer->writeHandle = std::move(writeHandle);

	// Get the serialization size of the WriteHandle.
	size_t handleCreateInfoLength = buffer->writeHandle->SerializeCreateSize();

	DeviceCreateBufferMappedCmd cmd;
	cmd.device = cDevice;
	cmd.descriptor = descriptor;
	cmd.result = ObjectHandle{buffer->id, bufferObjectAndSerial->generation};
	cmd.handleCreateInfoLength = handleCreateInfoLength;
	cmd.handleCreateInfo = nullptr;

	size_t commandSize = cmd.GetRequiredSize();
	size_t requiredSize = commandSize + handleCreateInfoLength;
	char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *wireClient);
	// Serialize the WriteHandle into the space after the command.
	buffer->writeHandle->SerializeCreate(allocatedBuffer + commandSize);

	return result;
	}

	void ClientHandwrittenDevicePushErrorScope(WGPUDevice cDevice, WGPUErrorFilter filter) {
	Device* device = reinterpret_cast<Device*>(cDevice);
	device->PushErrorScope(filter);
	}

	bool ClientHandwrittenDevicePopErrorScope(WGPUDevice cDevice,
	WGPUErrorCallback callback,
	void* userdata) {
	Device* device = reinterpret_cast<Device*>(cDevice);
	return device->RequestPopErrorScope(callback, userdata);
	}

	uint64_t ClientHandwrittenFenceGetCompletedValue(WGPUFence cSelf) {
	auto fence = reinterpret_cast<Fence*>(cSelf);
	return fence->completedValue;
	}

	void ClientHandwrittenFenceOnCompletion(WGPUFence cFence,
	uint64_t value,
	WGPUFenceOnCompletionCallback callback,
	void* userdata) {
	Fence* fence = reinterpret_cast<Fence*>(cFence);
	if (value > fence->signaledValue) {
	ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
	WGPUErrorType_Validation,
	"Value greater than fence signaled value");
	callback(WGPUFenceCompletionStatus_Error, userdata);
	return;
	}

	if (value <= fence->completedValue) {
	callback(WGPUFenceCompletionStatus_Success, userdata);
	return;
	}

	Fence::OnCompletionData request;
	request.completionCallback = callback;
	request.userdata = userdata;
	fence->requests.Enqueue(std::move(request), value);
	}

	void ClientHandwrittenBufferSetSubData(WGPUBuffer cBuffer,
	uint64_t start,
	uint64_t count,
	const void* data) {
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	BufferSetSubDataInternalCmd cmd;
	cmd.bufferId = buffer->id;
	cmd.start = start;
	cmd.count = count;
	cmd.data = static_cast<const uint8_t*>(data);

	Client* wireClient = buffer->device->GetClient();
	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer);
	}

	void ClientHandwrittenBufferUnmap(WGPUBuffer cBuffer) {
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	// Invalidate the local pointer, and cancel all other in-flight requests that would
	// turn into errors anyway (you can't double map). This prevents race when the following
	// happens, where the application code would have unmapped a buffer but still receive a
	// callback:
	// - Client -> Server: MapRequest1, Unmap, MapRequest2
	// - Server -> Client: Result of MapRequest1
	// - Unmap locally on the client
	// - Server -> Client: Result of MapRequest2
	if (buffer->writeHandle) {
	// Writes need to be flushed before Unmap is sent. Unmap calls all associated
	// in-flight callbacks which may read the updated data.
	ASSERT(buffer->readHandle == nullptr);

	// Get the serialization size of metadata to flush writes.
	size_t writeFlushInfoLength = buffer->writeHandle->SerializeFlushSize();

	BufferUpdateMappedDataCmd cmd;
	cmd.bufferId = buffer->id;
	cmd.writeFlushInfoLength = writeFlushInfoLength;
	cmd.writeFlushInfo = nullptr;

	size_t commandSize = cmd.GetRequiredSize();
	size_t requiredSize = commandSize + writeFlushInfoLength;
	char* allocatedBuffer =
	static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer);
	// Serialize flush metadata into the space after the command.
	// This closes the handle for writing.
	buffer->writeHandle->SerializeFlush(allocatedBuffer + commandSize);
	buffer->writeHandle = nullptr;

	} else if (buffer->readHandle) {
	buffer->readHandle = nullptr;
	}
	buffer->ClearMapRequests(WGPUBufferMapAsyncStatus_Unknown);

	BufferUnmapCmd cmd;
	cmd.self = cBuffer;
	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer =
	static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *buffer->device->GetClient());
	}

	void ClientHandwrittenBufferDestroy(WGPUBuffer cBuffer) {
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	// Cancel or remove all mappings
	buffer->writeHandle = nullptr;
	buffer->readHandle = nullptr;
	buffer->ClearMapRequests(WGPUBufferMapAsyncStatus_Unknown);

	BufferDestroyCmd cmd;
	cmd.self = cBuffer;
	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer =
	static_cast<char*>(buffer->device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *buffer->device->GetClient());
	}

	WGPUFence ClientHandwrittenQueueCreateFence(WGPUQueue cSelf,
	WGPUFenceDescriptor const* descriptor) {
	Queue* queue = reinterpret_cast<Queue*>(cSelf);
	Device* device = queue->device;

	QueueCreateFenceCmd cmd;
	cmd.self = cSelf;
	auto* allocation = device->GetClient()->FenceAllocator().New(device);
	cmd.result = ObjectHandle{allocation->object->id, allocation->generation};
	cmd.descriptor = descriptor;

	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer = static_cast<char*>(device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *device->GetClient());

	WGPUFence cFence = reinterpret_cast<WGPUFence>(allocation->object.get());

	Fence* fence = reinterpret_cast<Fence*>(cFence);
	fence->queue = queue;

	uint64_t initialValue = descriptor != nullptr ? descriptor->initialValue : 0u;
	fence->signaledValue = initialValue;
	fence->completedValue = initialValue;
	return cFence;
	}

	void ClientHandwrittenQueueSignal(WGPUQueue cQueue, WGPUFence cFence, uint64_t signalValue) {
	Fence* fence = reinterpret_cast<Fence*>(cFence);
	Queue* queue = reinterpret_cast<Queue*>(cQueue);
	if (fence->queue != queue) {
	ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
	WGPUErrorType_Validation,
	"Fence must be signaled on the queue on which it was created.");
	return;
	}
	if (signalValue <= fence->signaledValue) {
	ClientDeviceInjectError(reinterpret_cast<WGPUDevice>(fence->device),
	WGPUErrorType_Validation,
	"Fence value less than or equal to signaled value");
	return;
	}
	fence->signaledValue = signalValue;

	QueueSignalCmd cmd;
	cmd.self = cQueue;
	cmd.fence = cFence;
	cmd.signalValue = signalValue;

	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer =
	static_cast<char*>(fence->device->GetClient()->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer, *fence->device->GetClient());
	}

	void ClientHandwrittenQueueWriteBuffer(WGPUQueue cQueue,
	WGPUBuffer cBuffer,
	uint64_t bufferOffset,
	const void* data,
	size_t size) {
	Queue* queue = reinterpret_cast<Queue*>(cQueue);
	Buffer* buffer = reinterpret_cast<Buffer*>(cBuffer);

	QueueWriteBufferInternalCmd cmd;
	cmd.queueId = queue->id;
	cmd.bufferId = buffer->id;
	cmd.bufferOffset = bufferOffset;
	cmd.data = static_cast<const uint8_t*>(data);
	cmd.size = size;

	Client* wireClient = buffer->device->GetClient();
	size_t requiredSize = cmd.GetRequiredSize();
	char* allocatedBuffer = static_cast<char*>(wireClient->GetCmdSpace(requiredSize));
	cmd.Serialize(allocatedBuffer);
	}

	void ClientDeviceReference(WGPUDevice) {
	}

	void ClientDeviceRelease(WGPUDevice) {
	}

	WGPUQueue ClientHandwrittenDeviceGetDefaultQueue(WGPUDevice cSelf) {
	Device* device = reinterpret_cast<Device*>(cSelf);
	return device->GetDefaultQueue();
	}

	void ClientHandwrittenDeviceSetUncapturedErrorCallback(WGPUDevice cSelf,
	WGPUErrorCallback callback,
	void* userdata) {
	Device* device = reinterpret_cast<Device*>(cSelf);
	device->SetUncapturedErrorCallback(callback, userdata);
	}
	void ClientHandwrittenDeviceSetDeviceLostCallback(WGPUDevice cSelf,
	WGPUDeviceLostCallback callback,
	void* userdata) {
	Device* device = reinterpret_cast<Device*>(cSelf);
	device->SetDeviceLostCallback(callback, userdata);
	}

	}} // namespace dawn_wire::client