在Ogre 1.7中，对多线程的支持有三个选择：
要在Ogre 1.7中使用多线程：
1 在OgreBuildSettings.h改写2行
1 OGRE_THREAD_SUPPORT 2 //
通过ResourceBackgroundQueue实现。
1 OGRE_THREAD_PROVIDER 3 // TBB library provides threadingfunctionality
我采用C++的并行编程模板库Threading
Building Blocks。采用CMake构造的需要把TBB的文件拷贝到Ogre的Dependence的相关目录下。
Ogre 1.7
的 Threading changes
· WorkQueue added to accept generalised work items to beexecuted in background worker threads
· WorkQueue starts the number of workers based on hardware, orcan be told to start a different number
· Main Ogre WorkQueue is in Root::getWorkQueue. You can alsosubclass WorkQueue and provide your own if you want
· ResourceBackgroundQueuenow uses WorkQueue instead of using its own queueand can have multipletasks running at
once
· New focus on data-driven, task-based parallel execution withseparation of GPU and CPU activities
· Boost, POCO and Thread Building Blocks supported as threadingback-ends (Boost preferred)
· All Ogre thread support will use WorkQueue in future
备注：1.7添加了一个Character Sample，其中SinbadCharacterController有参考价值。
WorkQueue
使用方法
1、继承WorkQueue::RequestHandler、WorkQueue::ResponseHandler，准备一个uint16mWorkQueueChannel，在类的构造函数中通过以下代码初始化：
WorkQueue* wq =Root::getSingleton().getWorkQueue();
mWorkQueueChannel = wq->getChannel("Ogre/Terrain"); // 这个名字自己取好
wq->addRequestHandler(mWorkQueueChannel, this);
wq->addResponseHandler(mWorkQueueChannel, this);
析构时移除：
waitForDerivedProcesses(); //
这个最后说明
WorkQueue* wq =Root::getSingleton().getWorkQueue();
wq->removeRequestHandler(mWorkQueueChannel, this);
wq->removeResponseHandler(mWorkQueueChannel, this);
_____________________________________________________________________________________________________________________________________________________
2、准备两个结构体，用于记录Request/Response的Contex，Request与Response容易看错，注意颜色区分
/// A data holder for
communicating with thebackground derived data update
struct DerivedDataRequest
{
Terrain* terrain;
_OgreTerrainExport friendstd::ostream& operator<<(std::ostream& o, constDerivedDataRequest& r)
{ return o; } // 这个函数要导出
};
/// A data holder forcommunicating with the background derived data update
struct DerivedDataResponse
{
Terrain* terrain;
PixelBox* lightMapBox;
_OgreTerrainExport friendstd::ostream& operator<<(std::ostream& o, constDerivedDataResponse& r)
{ return o; } // 这个函数要导出
};
_____________________________________________________________________________________________________________________________________________________
3、重载以下4个函数
///WorkQueue::RequestHandler override
bool canHandleRequest(const WorkQueue::Request* req, const WorkQueue* srcQ);
/// WorkQueue::RequestHandler override
WorkQueue::Response* handleRequest(const WorkQueue::Request* req, constWorkQueue* srcQ);
/// WorkQueue::ResponseHandler override
bool canHandleResponse(const WorkQueue::Response* res, const WorkQueue* srcQ);
/// WorkQueue::ResponseHandler override
void handleResponse(const WorkQueue::Response* res, const WorkQueue* srcQ);
其中handleRequest是在多线程中处理实际加载数据的函数，具体的读取、创建之类的操作都在此完成。
对于 canHandleRequest、canHandleResponse基本上就是判定是不是自己的处理对象，一般可以不管，但为了谨慎起见还是需要对请求任务判定，以免多个同类任务执行时会有冲突：
boolTerrain::canHandleRequest(const WorkQueue::Request* req, const WorkQueue* srcQ)
{
DerivedDataRequestddr =any_cast<DerivedDataRequest>(req->getData());
// only deal with own requests
// we do this because if we deletea terrain we want any pending tasks to be discarded
if (ddr.terrain != this)
return false;
else
returnRequestHandler::canHandleRequest(req, srcQ);
}
//---------------------------------------------------------------------
bool Terrain::canHandleResponse(const WorkQueue::Response* res, constWorkQueue* srcQ)
{
DerivedDataRequestddreq =any_cast<DerivedDataRequest>(res->getRequest()->getData());
// only deal with own requests
// we do this because if we deletea terrain we want any pending tasks to be discarded
if (ddreq.terrain != this)
return false;
else
return true;
}
上面就是要注意用any转型时取得的源对象是什么。（res->getRequest()->getData() 与req->getData()）
对于handleRequest最后要做的就是返回一个新的处理结果：
WorkQueue::Response*Terrain::handleRequest(const WorkQueue::Request* req, const WorkQueue* srcQ)
{
DerivedDataResponse ddres;
把操作的结果都放到ddres中返回给主线程中去处理
WorkQueue::Response* response = OGRE_NEWWorkQueue::Response(req, true, Any(ddres));
return response;
}
_____________________________________________________________________________________________________________________________________________________
4、添加具体的任务请求（部分代码省略）
DerivedDataRequest req;
req.terrain = this;
Root::getSingleton().getWorkQueue()->addRequest(
mWorkQueueChannel,WORKQUEUE_DERIVED_DATA_REQUEST,
Any(req), 0, synchronous);
其中 ：const uint16 Terrain::WORKQUEUE_DERIVED_DATA_REQUEST = 1，对于同个对象，如果有不同的数据操作类型，可以用多个不同的ID区别。
_____________________________________________________________________________________________________________________________________________________
注意：在析构时可能还有请求的任务没完成，因此要等待任务完成后才能移除，要不会CRASH。这个需要自己维护一个标记，并让线程队列处理。基本上是在addRequest的时候设置为true，在handleResponse时设置为false：
boolmDerivedDataUpdateInProgress;
voidTerrain::waitForDerivedProcesses()
{
while(mDerivedDataUpdateInProgress)
{
// we need to wait forthis to finish
OGRE_THREAD_SLEEP(50);
Root::getSingleton().getWorkQueue()->processResponses();
}
}
voidTerrain::handleResponse(const WorkQueue::Response* res, const WorkQueue* srcQ)
｛
mDerivedDataUpdateInProgress = false;
｝
void Terrain::updateDerivedDataImpl(constRect& rect, const Rect& lightmapExtraRect,
bool synchronous, uint8 typeMask)
{
mDerivedDataUpdateInProgress =true;
Root::getSingleton().getWorkQueue()->addRequest(
mWorkQueueChannel,WORKQUEUE_DERIVED_DATA_REQUEST,
Any(req), 0, synchronous);
}