JournalingObjectStore: lock journal_lock at top of journal_replay

CID 716968: Data race condition (MISSING_LOCK)At (6): Accessing "this->op_seq"
("JournalingObjectStore.op_seq") requires the "Mutex._m" lock.

Doesn't really matter, but this should satisfy coverity.

Signed-off-by: Samuel Just <sam.just@inktank.com>

diff --git a/src/os/JournalingObjectStore.cc b/src/os/JournalingObjectStore.cc
index 2119f9538cf80c30331019637528513860f0c5c2..83864fc00cc525719638f8027c7b5baf06cf17c9 100644 (file)
--- a/src/os/JournalingObjectStore.cc
+++ b/src/os/JournalingObjectStore.cc
@@ -37,10 +37,12 @@ int JournalingObjectStore::journal_replay(uint64_t fs_op_seq)
     fs_op_seq = g_conf->journal_replay_from - 1;
   }
 
+  journal_lock.Lock();
   op_seq = fs_op_seq;
   committed_seq = fs_op_seq;
   committing_seq = fs_op_seq;
   applied_seq = fs_op_seq;
+  journal_lock.Unlock();
 
   if (!journal)
     return 0;