Buffer info logs when picking compactions and write them out after releasing the mutex

Summary: Now while the background thread is picking compactions, it writes out multiple info_logs, especially for universal compaction, which introduces a chance of waiting log writing in mutex, which is bad. To remove this risk, write all those info logs to a buffer and flush it after releasing the mutex.

Test Plan:
make all check
check the log lines while running some tests that trigger compactions.

Reviewers: haobo, igor, dhruba

Reviewed By: dhruba

CC: i.am.jin.lei, dhruba, yhchiang, leveldb, nkg-

Differential Revision: https://reviews.facebook.net/D16515

include/rocksdb/env.h

@@ -575,6 +575,26 @@ class Logger {
   InfoLogLevel log_level_;
 };
 
+// A class to buffer info log entries and flush them in the end.
+class LogBuffer {
+ public:
+  // log_level: the log level for all the logs
+  // info_log:  logger to write the logs to
+  LogBuffer(const InfoLogLevel log_level, const shared_ptr<Logger>& info_log);
+  ~LogBuffer();
+
+  // Add a log entry to the buffer.
+  void AddLogToBuffer(const char* format, ...);
+
+  // Flush all buffered log to the info log.
+  void FlushBufferToLog() const;
+
+ private:
+  struct Rep;
+  Rep* rep_;
+  const InfoLogLevel log_level_;
+  const shared_ptr<Logger>& info_log_;
+};
 
 // Identifies a locked file.
 class FileLock {
@@ -587,6 +607,9 @@ class FileLock {
   void operator=(const FileLock&);
 };
 
+// Add log to the LogBuffer for a delayed info logging. It can be used when
+// we want to add some logs inside a mutex.
+extern void LogToBuffer(LogBuffer* log_buffer, const char* format, ...);
 
 extern void LogFlush(const shared_ptr<Logger>& info_log);