leveldb分析——Arena内存管理

leveldb中实现了一个简单的内存管理工具Arena,其基本思想为：先预先向系统申请一块内存，此后需要申请内存时，直接到预先分配的内存中申请。

那么这样做的目的是什么呢？

(1)避免了频率地进行malloc/new和free/delete操作，同时对于内存管理变得简单，对于内存的释放工作交给Arena。

(2)避免造成大量的内存碎片。(还需去了解一下)

下面看具体的源码分析：

Arena定义：

class Arena {
 public:
  Arena();
  ~Arena();

  // Return a pointer to a newly allocated memory block of "bytes" bytes.
  char* Allocate(size_t bytes);

  // Allocate memory with the normal alignment guarantees provided by malloc
  char* AllocateAligned(size_t bytes);

  // Returns an estimate of the total memory usage of data allocated
  // by the arena (including space allocated but not yet used for user
  // allocations).
  size_t MemoryUsage() const {
    return blocks_memory_ + blocks_.capacity() * sizeof(char*);
  }

 private:
  char* AllocateFallback(size_t bytes);
  char* AllocateNewBlock(size_t block_bytes);

  // Allocation state
  char* alloc_ptr_;
  size_t alloc_bytes_remaining_;

  // Array of new[] allocated memory blocks
  std::vector<char*> blocks_;

  // Bytes of memory in blocks allocated so far
  size_t blocks_memory_;

  // No copying allowed
  Arena(const Arena&);
  void operator=(const Arena&);
};

Arena提供两种分配方式：所分配的内存严格对齐、不一定严格对齐的分配方式。每次预先分配的4K（为什么是4K？）保存到blocks_ vector中，最后统一释放。这种内存管理方式是具有一定的适用范围，如需不断分配小内存，最终一并全释放的场景。对于leveldb来说，memtable恰好就是这样的，每次向memtable中insert一条k/v时，就申请一块内存，当memtable被flush到磁盘且不再使用时，将整个memtable释放掉。

inline char* Arena::Allocate(size_t bytes) {
  // The semantics of what to return are a bit messy if we allow
  // 0-byte allocations, so we disallow them here (we don't need
  // them for our internal use).
  assert(bytes > 0);
  if (bytes <= alloc_bytes_remaining_) {
    char* result = alloc_ptr_;
    alloc_ptr_ += bytes;
    alloc_bytes_remaining_ -= bytes;
    return result;
  }
  return AllocateFallback(bytes);   //预先分配的不足
}

char* Arena::AllocateFallback(size_t bytes) {
  if (bytes > kBlockSize / 4) {
    // Object is more than a quarter of our block size.  Allocate it separately
    // to avoid wasting too much space in leftover bytes.
    char* result = AllocateNewBlock(bytes);  ///对于大内存，直接单独给分配一块，原先预分配的内存还能使用
    return result;
  }

  // We waste the remaining space in the current block.  预分配的内存剩下的已很少，所以直接重新分配一块，也就是说浪费了一点内存
  alloc_ptr_ = AllocateNewBlock(kBlockSize);
  alloc_bytes_remaining_ = kBlockSize;

  char* result = alloc_ptr_;
  alloc_ptr_ += bytes;
  alloc_bytes_remaining_ -= bytes;
  return result;
}

下面来看下严格对齐的分配方式

char* Arena::AllocateAligned(size_t bytes) {
  const int align = sizeof(void*);    // We'll align to pointer size
  assert((align & (align-1)) == 0);   // Pointer size should be a power of 2
  size_t current_mod = reinterpret_cast<uintptr_t>(alloc_ptr_) & (align-1);  ///计算出alloc_ptr_ % align 
  size_t slop = (current_mod == 0 ? 0 : align - current_mod);                ///对齐还需要向前移动的大小
  size_t needed = bytes + slop; 
  char* result;
  if (needed <= alloc_bytes_remaining_) {
    result = alloc_ptr_ + slop;
    alloc_ptr_ += needed;
    alloc_bytes_remaining_ -= needed;
  } else {
    // AllocateFallback always returned aligned memory
    result = AllocateFallback(bytes);
  }
  assert((reinterpret_cast<uintptr_t>(result) & (align-1)) == 0);
  return result;
}

char* Arena::AllocateNewBlock(size_t block_bytes) {
  char* result = new char[block_bytes];
  blocks_memory_ += block_bytes;
  blocks_.push_back(result);
  return result;
}

leveldb分析——Arena内存管理

推荐阅读