Fix the initialization of atomic variable introduced by the

group clearing mechanism.

Commit 0e141c0fbb introduced initialization of atomic variable in
InitProcess which means that it's not safe to look at it for backends that
aren't currently in use.  Fix that by initializing the same during postmaster
startup.

Reported-by: Andres Freund
Author: Amit Kapila
Backpatch-through: 9.6
Discussion: https://postgr.es/m/20181027104138.qmbbelopvy7cw2qv@alap3.anarazel.de
This commit is contained in:
Amit Kapila 2018-11-13 10:43:35 +05:30
parent fe66fc6f94
commit 1a8bde490c

View File

@ -268,6 +268,12 @@ InitProcGlobal(void)
/* Initialize lockGroupMembers list. */
dlist_init(&procs[i].lockGroupMembers);
/*
* Initialize the atomic variable, otherwise, it won't be safe to
* access it for backends that aren't currently in use.
*/
pg_atomic_init_u32(&(procs[i].procArrayGroupNext), INVALID_PGPROCNO);
}
/*
@ -401,7 +407,7 @@ InitProcess(void)
/* Initialize fields for group XID clearing. */
MyProc->procArrayGroupMember = false;
MyProc->procArrayGroupMemberXid = InvalidTransactionId;
pg_atomic_init_u32(&MyProc->procArrayGroupNext, INVALID_PGPROCNO);
Assert(pg_atomic_read_u32(&MyProc->procArrayGroupNext) == INVALID_PGPROCNO);
/* Check that group locking fields are in a proper initial state. */
Assert(MyProc->lockGroupLeader == NULL);