osd/OSD: wakeup all threads of shard.

In our test(4NVME), we found for 4K randread(8/16 mean shard, 2/1 mean
thread_per_shard):
QD	8_2(IOPS(k))	16_1(IOPS(K))	8_2(apply patch)
32	 191		 263		 263.5

Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>
(cherry picked from commit d1cf3fb80bce50a13da0fd45dba65248cbe38383)

diff --git a/src/osd/OSD.cc b/src/osd/OSD.cc
index e72f9f6f650f451cff898293f12390de85547849..9fe1653489a8140704d8f86085e468acb47dd6bc 100644 (file)
--- a/src/osd/OSD.cc
+++ b/src/osd/OSD.cc
@@ -10711,7 +10711,7 @@ void OSD::ShardedOpWQ::_enqueue(OpSchedulerItem&& item) {
 
   if (empty) {
     std::lock_guard l{sdata->sdata_wait_lock};
-    sdata->sdata_cond.notify_one();
+    sdata->sdata_cond.notify_all();
   }
 }