在并发写入的时候，leveldb巧妙地利用一个时间窗口做batch写入，这部分代码值得一读：
Status DBImpl::Write(const WriteOptions& options, WriteBatch* my_batch) {
// A begin
Writer w(&mutex_);
w.batch = my_batch;
w.sync = options.sync;
w.done = false;
// A end
// B begin
MutexLock l(&mutex_);
writers_.push_back(&w);
while (!w.done && &w != writers_.front()) {
w.cv.Wait();
}
if (w.done) {
return w.status;
}
// B end
// May temporarily unlock and wait.
Status status = MakeRoomForWrite(my_batch == NULL);
uint64_t last_sequence = versions_->LastSequence();
Writer* last_writer = &w;
if (status.ok() && my_batch != NULL) { // NULL batch is for compactions
WriteBatch* updates = BuildBatchGroup(&last_writer);
WriteBatchInternal::SetSequence(updates, last_sequence + 1);
last_sequence += WriteBatchInternal::Count(updates);
// Add to log and apply to memtable. We can release the lock
// during this phase since &w is currently responsible for logging
// and protects against concurrent loggers and concurrent writes
// into mem_.
{
mutex_.Unlock();
status = log_->AddRecord(WriteBatchInternal::Contents(updates));
bool sync_error = false;
if (status.ok() && options.sync) {
status = logfile_->Sync();
if (!status.ok()) {
sync_error = true;
}
}
if (status.ok()) {
status = WriteBatchInternal::InsertInto(updates, mem_);
}
mutex_.Lock();
if (sync_error) {
// The state of the log file is indeterminate: the log record we
// just added may or may not show up when the DB is re-opened.
// So we force the DB into a mode where all future writes fail.
RecordBackgroundError(status);
}
}
if (updates == tmp_batch_) tmp_batch_->Clear();
versions_->SetLastSequence(last_sequence);
}
while (true) {
Writer* ready = writers_.front();
writers_.pop_front();
if (ready != &w) {
ready->status = status;
ready->done = true;
ready->cv.Signal();
}
if (ready == last_writer) break;
}
// Notify new head of write queue
if (!writers_.empty()) {
writers_.front()->cv.Signal();
}
return status;
}
假设同时有w1, w2, w3, w4, w5, w6 并发请求写入。
B部分代码让竞争到mutex资源的w1获取了锁。w1将它要写的数据添加到了writers_队列里去，此时队列只有一个w1, 从而其顺利的进行buildbatchgroup。当运行到34行时mutex_互斥锁释放，之所以这儿可以释放mutex_，是因为其它的写操作都不满足队首条件，进而不会进入log和memtable写入阶段。这时(w2, w3, w4, w5, w6)会竞争锁，由于B段代码中不满足队首条件，均等待并释放锁了。从而队列可能会如(w3, w5, w2, w4).
继而w1进行log写入和memtable写入。 当w1完成log和memtable写入后，进入46行代码，则mutex_又锁住，这时B段代码中队列因为获取不到锁则队列不会修改。
随后59行开始，w1被pop出来，由于ready==w, 并且ready==last_writer,所以直接到71行代码，唤醒了此时处于队首的w3.
w3唤醒时，发现自己是队首，可以顺利的进行进入buildbatchgroup，在该函数中，遍历了目前所有的队列元素，形成一个update的batch，即将w3, w5, w2, w4合并为一个batch. 并将last_writer置为此时处于队尾的最后一个元素w4，34行代码运行后，因为释放了锁资源，队列可能随着dbimpl::write的调用而更改，如队列状况可能为(w3, w5, w2, w4, w6, w9, w8).
35-45行的代码将w3, w5, w2, w4整个的batch写入log和memtable. 到65行，分别对w5, w2, w4进行了一次cond signal.当判断到完w4 == lastwriter时，则退出循环。72行则对队首的w6唤醒，从而按上述步骤依次进行下去。
这样就形成了多个并发write 合并为一个batch写入log和memtable的机制。
