TNL::DataChunker class Reference

TNL::DataChunker Class Reference

#include <tnlDataChunker.h>

Inheritance diagram for TNL::DataChunker:

Detailed Description

Implements a chunked data allocater.

Calling new/malloc all the time is a time consuming operation. Therefore, we provide the DataChunker, which allocates memory in blocks of chunkSize (by default 16k, see ChunkSize, though it can be set in the constructor), then doles it out as requested, in chunks of up to chunkSize in size.

It will assert if you try to get more than ChunkSize bytes at a time, and it deals with the logic of allocating new blocks and giving out word-aligned chunks.

Note that new/free/realloc WILL NOT WORK on memory gotten from the DataChunker. This also only grows (you can call freeBlocks to deallocate and reset things).

Public Types

enum { ChunkSize = 16376 }

Public Member Functions

void * alloc (S32 size)

allocate a pointer to memory of size bytes from the DataChunker

void freeBlocks ()

free all pages currently allocated in the DataChunker

DataChunker (S32 size=ChunkSize)

Construct a DataChunker with a page size of size bytes.

~DataChunker ()

void swap (DataChunker &d)

Swaps the memory allocated in one data chunker for another.

Private Attributes

DataBlock * curBlock

current page we're allocating data from

S32 chunkSize

The size allocated for each page in the DataChunker.

Member Enumeration Documentation

anonymous enum

Enumeration values:

ChunkSize Default size of each DataBlock page in the DataChunker.

Constructor & Destructor Documentation

TNL::DataChunker::DataChunker ( S32 size = ChunkSize )

Construct a DataChunker with a page size of size bytes.

TNL::DataChunker::~DataChunker ( )

Member Function Documentation

void* TNL::DataChunker::alloc ( S32 size )

allocate a pointer to memory of size bytes from the DataChunker

void TNL::DataChunker::freeBlocks ( )

free all pages currently allocated in the DataChunker

Reimplemented in TNL::ClassChunker< T >, TNL::ClassChunker< BucketEntry >, and TNL::ClassChunker< EventNote >.

void TNL::DataChunker::swap ( DataChunker & d ) [inline]

Swaps the memory allocated in one data chunker for another.
This can be used to implement packing of memory stored in a DataChunker.

Field Documentation

DataBlock* TNL::DataChunker::curBlock [private]

current page we're allocating data from
If the data size request is greater than the memory space currently available in the current page, a new page will be allocated.

S32 TNL::DataChunker::chunkSize [private]

The size allocated for each page in the DataChunker.