x
Yes
No
Do you want to visit DriveHQ English website?
首页
产品服务
价格
免费试用
下载客户端
关于我们
云文件服务
|
云备份服务
|
FTP服务
|
企业邮箱服务
|
网站托管
|
客户端软件
云文件服务
云备份服务
FTP服务
企业级邮箱服务
网站托管
客户端软件
OgreHardwareBufferManager.h - Hosted on DriveHQ Cloud IT Platform
返回上层目录
上传
下载
共享
发布
新建文件夹
新建文件
复制
剪切
删除
粘贴
评论
升级服务
路径: \\game3dprogramming\materials\GameFactory\GameFactoryDemo\references\ogre\include\OgreHardwareBufferManager.h
旋转
特效
属性
历史版本
/* ----------------------------------------------------------------------------- This source file is part of OGRE (Object-oriented Graphics Rendering Engine) For the latest info, see http://www.ogre3d.org/ Copyright (c) 2000-2006 Torus Knot Software Ltd Also see acknowledgements in Readme.html This program is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more details. You should have received a copy of the GNU Lesser General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA, or go to http://www.gnu.org/copyleft/lesser.txt. You may alternatively use this source under the terms of a specific version of the OGRE Unrestricted License provided you have obtained such a license from Torus Knot Software Ltd. ----------------------------------------------------------------------------- */ #ifndef __HardwareBufferManager__ #define __HardwareBufferManager__ // Precompiler options #include "OgrePrerequisites.h" #include "OgreSingleton.h" #include "OgreHardwareVertexBuffer.h" #include "OgreHardwareIndexBuffer.h" namespace Ogre { /** Abtract interface representing a 'licensee' of a hardware buffer copy. remarks Often it's useful to have temporary buffers which are used for working but are not necessarily needed permanently. However, creating and destroying buffers is expensive, so we need a way to share these working areas, especially those based on existing fixed buffers. This class represents a licensee of one of those temporary buffers, and must be implemented by any user of a temporary buffer if they wish to be notified when the license is expired. */ class _OgreExport HardwareBufferLicensee { public: virtual ~HardwareBufferLicensee() { } /** This method is called when the buffer license is expired and is about to be returned to the shared pool. */ virtual void licenseExpired(HardwareBuffer* buffer) = 0; }; /** Structure for recording the use of temporary blend buffers */ class _OgreExport TempBlendedBufferInfo : public HardwareBufferLicensee { private: // Pre-blended HardwareVertexBufferSharedPtr srcPositionBuffer; HardwareVertexBufferSharedPtr srcNormalBuffer; // Post-blended HardwareVertexBufferSharedPtr destPositionBuffer; HardwareVertexBufferSharedPtr destNormalBuffer; /// Both positions and normals are contained in the same buffer bool posNormalShareBuffer; unsigned short posBindIndex; unsigned short normBindIndex; bool bindPositions; bool bindNormals; public: ~TempBlendedBufferInfo(void); /// Utility method, extract info from the given VertexData void extractFrom(const VertexData* sourceData); /// Utility method, checks out temporary copies of src into dest void checkoutTempCopies(bool positions = true, bool normals = true); /// Utility method, binds dest copies into a given VertexData struct void bindTempCopies(VertexData* targetData, bool suppressHardwareUpload); /** Overridden member from HardwareBufferLicensee. */ void licenseExpired(HardwareBuffer* buffer); /** Detect currently have buffer copies checked out and touch it */ bool buffersCheckedOut(bool positions = true, bool normals = true) const; }; /** Abstract singleton class for managing hardware buffers, a concrete instance of this will be created by the RenderSystem. */ class _OgreExport HardwareBufferManager : public Singleton
{ friend class HardwareVertexBufferSharedPtr; friend class HardwareIndexBufferSharedPtr; protected: /** WARNING: The following two members should place before all other members. Members destruct order is very important here, because destructing other members will cause notify back to this class, and then will access to this two members. */ typedef std::set
VertexBufferList; typedef std::set
IndexBufferList; VertexBufferList mVertexBuffers; IndexBufferList mIndexBuffers; typedef std::set
VertexDeclarationList; typedef std::set
VertexBufferBindingList; VertexDeclarationList mVertexDeclarations; VertexBufferBindingList mVertexBufferBindings; // Mutexes OGRE_MUTEX(mVertexBuffersMutex) OGRE_MUTEX(mIndexBuffersMutex) OGRE_MUTEX(mVertexDeclarationsMutex) OGRE_MUTEX(mVertexBufferBindingsMutex) /// Internal method for destroys all vertex declarations virtual void destroyAllDeclarations(void); /// Internal method for destroys all vertex buffer bindings virtual void destroyAllBindings(void); /// Internal method for creates a new vertex declaration, may be overridden by certain rendering APIs virtual VertexDeclaration* createVertexDeclarationImpl(void); /// Internal method for destroys a vertex declaration, may be overridden by certain rendering APIs virtual void destroyVertexDeclarationImpl(VertexDeclaration* decl); /// Internal method for creates a new VertexBufferBinding, may be overridden by certain rendering APIs virtual VertexBufferBinding* createVertexBufferBindingImpl(void); /// Internal method for destroys a VertexBufferBinding, may be overridden by certain rendering APIs virtual void destroyVertexBufferBindingImpl(VertexBufferBinding* binding); public: enum BufferLicenseType { /// Licensee will only release buffer when it says so BLT_MANUAL_RELEASE, /// Licensee can have license revoked BLT_AUTOMATIC_RELEASE }; protected: /** Struct holding details of a license to use a temporary shared buffer. */ class _OgrePrivate VertexBufferLicense { public: HardwareVertexBuffer* originalBufferPtr; BufferLicenseType licenseType; size_t expiredDelay; HardwareVertexBufferSharedPtr buffer; HardwareBufferLicensee* licensee; VertexBufferLicense( HardwareVertexBuffer* orig, BufferLicenseType ltype, size_t delay, HardwareVertexBufferSharedPtr buf, HardwareBufferLicensee* lic) : originalBufferPtr(orig) , licenseType(ltype) , expiredDelay(delay) , buffer(buf) , licensee(lic) {} }; /// Map from original buffer to temporary buffers typedef std::multimap
FreeTemporaryVertexBufferMap; /// Map of current available temp buffers FreeTemporaryVertexBufferMap mFreeTempVertexBufferMap; /// Map from temporary buffer to details of a license typedef std::map
TemporaryVertexBufferLicenseMap; /// Map of currently licensed temporary buffers TemporaryVertexBufferLicenseMap mTempVertexBufferLicenses; /// Number of frames elapsed since temporary buffers utilization was above half the available size_t mUnderUsedFrameCount; /// Number of frames to wait before free unused temporary buffers static const size_t UNDER_USED_FRAME_THRESHOLD; /// Frame delay for BLT_AUTOMATIC_RELEASE temporary buffers static const size_t EXPIRED_DELAY_FRAME_THRESHOLD; // Mutexes OGRE_MUTEX(mTempBuffersMutex) /// Creates a new buffer as a copy of the source, does not copy data HardwareVertexBufferSharedPtr makeBufferCopy( const HardwareVertexBufferSharedPtr& source, HardwareBuffer::Usage usage, bool useShadowBuffer); public: HardwareBufferManager(); virtual ~HardwareBufferManager(); /** Create a hardware vertex buffer. @remarks This method creates a new vertex buffer; this will act as a source of geometry data for rendering objects. Note that because the meaning of the contents of the vertex buffer depends on the usage, this method does not specify a vertex format; the user of this buffer can actually insert whatever data they wish, in any format. However, in order to use this with a RenderOperation, the data in this vertex buffer will have to be associated with a semantic element of the rendering pipeline, e.g. a position, or texture coordinates. This is done using the VertexDeclaration class, which itself contains VertexElement structures referring to the source data. @remarks Note that because vertex buffers can be shared, they are reference counted so you do not need to worry about destroying themm this will be done automatically. @param vertexSize The size in bytes of each vertex in this buffer; you must calculate this based on the kind of data you expect to populate this buffer with. @param numVerts The number of vertices in this buffer. @param usage One or more members of the HardwareBuffer::Usage enumeration; you are strongly advised to use HBU_STATIC_WRITE_ONLY wherever possible, if you need to update regularly, consider HBU_DYNAMIC_WRITE_ONLY and useShadowBuffer=true. @param useShadowBuffer If set to true, this buffer will be 'shadowed' by one stored in system memory rather than GPU or AGP memory. You should set this flag if you intend to read data back from the vertex buffer, because reading data from a buffer in the GPU or AGP memory is very expensive, and is in fact impossible if you specify HBU_WRITE_ONLY for the main buffer. If you use this option, all reads and writes will be done to the shadow buffer, and the shadow buffer will be synchronised with the real buffer at an appropriate time. */ virtual HardwareVertexBufferSharedPtr createVertexBuffer(size_t vertexSize, size_t numVerts, HardwareBuffer::Usage usage, bool useShadowBuffer = false) = 0; /** Create a hardware index buffer. @remarks Note that because buffers can be shared, they are reference counted so you do not need to worry about destroying themm this will be done automatically. @param itype The type in index, either 16- or 32-bit, depending on how many vertices you need to be able to address @param numIndexes The number of indexes in the buffer @param usage One or more members of the HardwareBuffer::Usage enumeration. @param useShadowBuffer If set to true, this buffer will be 'shadowed' by one stored in system memory rather than GPU or AGP memory. You should set this flag if you intend to read data back from the index buffer, because reading data from a buffer in the GPU or AGP memory is very expensive, and is in fact impossible if you specify HBU_WRITE_ONLY for the main buffer. If you use this option, all reads and writes will be done to the shadow buffer, and the shadow buffer will be synchronised with the real buffer at an appropriate time. */ virtual HardwareIndexBufferSharedPtr createIndexBuffer(HardwareIndexBuffer::IndexType itype, size_t numIndexes, HardwareBuffer::Usage usage, bool useShadowBuffer = false) = 0; /** Creates a new vertex declaration. */ virtual VertexDeclaration* createVertexDeclaration(void); /** Destroys a vertex declaration. */ virtual void destroyVertexDeclaration(VertexDeclaration* decl); /** Creates a new VertexBufferBinding. */ virtual VertexBufferBinding* createVertexBufferBinding(void); /** Destroys a VertexBufferBinding. */ virtual void destroyVertexBufferBinding(VertexBufferBinding* binding); /** Registers a vertex buffer as a copy of another. @remarks This is useful for registering an existing buffer as a temporary buffer which can be allocated just like a copy. */ virtual void registerVertexBufferSourceAndCopy( const HardwareVertexBufferSharedPtr& sourceBuffer, const HardwareVertexBufferSharedPtr& copy); /** Allocates a copy of a given vertex buffer. @remarks This method allocates a temporary copy of an existing vertex buffer. This buffer is subsequently stored and can be made available for other purposes later without incurring the cost of construction / destruction. @param sourceBuffer The source buffer to use as a copy @param licenseType The type of license required on this buffer - automatic release causes this class to release licenses every frame so that they can be reallocated anew. @param licensee Pointer back to the class requesting the copy, which must implement HardwareBufferLicense in order to be notified when the license expires. @param copyData If true, the current data is copied as well as the structure of the buffer */ virtual HardwareVertexBufferSharedPtr allocateVertexBufferCopy( const HardwareVertexBufferSharedPtr& sourceBuffer, BufferLicenseType licenseType, HardwareBufferLicensee* licensee, bool copyData = false); /** Manually release a vertex buffer copy for others to subsequently use. @remarks Only required if the original call to allocateVertexBufferCopy included a licenseType of BLT_MANUAL_RELEASE. @param bufferCopy The buffer copy. The caller is expected to delete or at least no longer use this reference, since another user may well begin to modify the contents of the buffer. */ virtual void releaseVertexBufferCopy( const HardwareVertexBufferSharedPtr& bufferCopy); /** Tell engine that the vertex buffer copy intent to reuse. @remarks Ogre internal keep an expired delay counter of BLT_AUTOMATIC_RELEASE buffers, when the counter count down to zero, it'll release for other purposes later. But you can use this function to reset the counter to the internal configured value, keep the buffer not get released for some frames. @param bufferCopy The buffer copy. The caller is expected to keep this buffer copy for use. */ virtual void touchVertexBufferCopy( const HardwareVertexBufferSharedPtr& bufferCopy); /** Free all unused vertex buffer copies. @remarks This method free all temporary vertex buffers that not in used. In normally, temporary vertex buffers are subsequently stored and can be made available for other purposes later without incurring the cost of construction / destruction. But in some cases you want to free them to save hardware memory (e.g. application was runs in a long time, you might free temporary buffers periodically to avoid memory overload). */ virtual void _freeUnusedBufferCopies(void); /** Internal method for releasing all temporary buffers which have been allocated using BLT_AUTOMATIC_RELEASE; is called by OGRE. @param forceFreeUnused If true, free all unused temporary buffers. If false, auto detect and free all unused temporary buffers based on temporary buffers utilization. */ virtual void _releaseBufferCopies(bool forceFreeUnused = false); /** Internal method that forces the release of copies of a given buffer. @remarks This usually means that the buffer which the copies are based on has been changed in some fundamental way, and the owner of the original wishes to make that known so that new copies will reflect the changes. @param sourceBuffer the source buffer as a shared pointer. Any buffer copies created from the source buffer are deleted. */ virtual void _forceReleaseBufferCopies( const HardwareVertexBufferSharedPtr& sourceBuffer); /** Internal method that forces the release of copies of a given buffer. @remarks This usually means that the buffer which the copies are based on has been changed in some fundamental way, and the owner of the original wishes to make that known so that new copies will reflect the changes. @param sourceBuffer the source buffer as a pointer. Any buffer copies created from the source buffer are deleted. */ virtual void _forceReleaseBufferCopies(HardwareVertexBuffer* sourceBuffer); /// Notification that a hardware vertex buffer has been destroyed void _notifyVertexBufferDestroyed(HardwareVertexBuffer* buf); /// Notification that a hardware index buffer has been destroyed void _notifyIndexBufferDestroyed(HardwareIndexBuffer* buf); /** Override standard Singleton retrieval. @remarks Why do we do this? Well, it's because the Singleton implementation is in a .h file, which means it gets compiled into anybody who includes it. This is needed for the Singleton template to work, but we actually only want it compiled into the implementation of the class based on the Singleton, not all of them. If we don't change this, we get link errors when trying to use the Singleton-based class from an outside dll. @par This method just delegates to the template version anyway, but the implementation stays in this single compilation unit, preventing link errors. */ static HardwareBufferManager& getSingleton(void); /** Override standard Singleton retrieval. @remarks Why do we do this? Well, it's because the Singleton implementation is in a .h file, which means it gets compiled into anybody who includes it. This is needed for the Singleton template to work, but we actually only want it compiled into the implementation of the class based on the Singleton, not all of them. If we don't change this, we get link errors when trying to use the Singleton-based class from an outside dll. @par This method just delegates to the template version anyway, but the implementation stays in this single compilation unit, preventing link errors. */ static HardwareBufferManager* getSingletonPtr(void); }; } #endif
OgreHardwareBufferManager.h
网页地址
文件地址
上一页
67/217
下一页
下载
( 19 KB )
Comments
Total ratings:
0
Average rating:
无评论
of 10
Would you like to comment?
Join now
, or
Logon
if you are already a member.