Fix ALTER COLUMN TYPE to not open a relation without any lock.

If the column being modified is referenced by a foreign key constraint
of another table, ALTER TABLE would open the other table (to re-parse
the constraint's definition) without having first obtained a lock on it.
This was evidently intentional, but that doesn't mean it's really safe.
It's especially not safe in 9.3, which pre-dates use of MVCC scans for
catalog reads, but even in current releases it doesn't seem like a good
idea.

We know we'll need AccessExclusiveLock shortly to drop the obsoleted
constraint, so just get that a little sooner to close the hole.

Per testing with a patch that complains if we open a relation without
holding any lock on it.  I don't plan to back-patch that patch, but we
should close the holes it identifies in all supported branches.

Discussion: https://postgr.es/m/2038.1538335244@sss.pgh.pa.us

1 files changed, 14 insertions, 4 deletions

diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c
index 9e60bb37a79..c145385f843 100644
--- a/src/backend/commands/tablecmds.c
+++ b/src/backend/commands/tablecmds.c
@@ -9888,8 +9888,7 @@ ATPostAlterTypeCleanup(List **wqueue, AlteredTableInfo *tab, LOCKMODE lockmode)
 	 * appropriate work queue entries.  We do this before dropping because in
 	 * the case of a FOREIGN KEY constraint, we might not yet have exclusive
 	 * lock on the table the constraint is attached to, and we need to get
-	 * that before dropping.  It's safe because the parser won't actually look
-	 * at the catalogs to detect the existing entry.
+	 * that before reparsing/dropping.
 	 *
 	 * We can't rely on the output of deparsing to tell us which relation to
 	 * operate on, because concurrent activity might have made the name
@@ -9905,6 +9904,7 @@ ATPostAlterTypeCleanup(List **wqueue, AlteredTableInfo *tab, LOCKMODE lockmode)
 		Form_pg_constraint con;
 		Oid			relid;
 		Oid			confrelid;
+		char		contype;
 		bool		conislocal;
 
 		tup = SearchSysCache1(CONSTROID, ObjectIdGetDatum(oldId));
@@ -9921,10 +9921,11 @@ ATPostAlterTypeCleanup(List **wqueue, AlteredTableInfo *tab, LOCKMODE lockmode)
 				elog(ERROR, "could not identify relation associated with constraint %u", oldId);
 		}
 		confrelid = con->confrelid;
+		contype = con->contype;
 		conislocal = con->conislocal;
 		ReleaseSysCache(tup);
 
-		ObjectAddressSet(obj, ConstraintRelationId, lfirst_oid(oid_item));
+		ObjectAddressSet(obj, ConstraintRelationId, oldId);
 		add_exact_object_address(&obj, objects);
 
 		/*
@@ -9936,6 +9937,15 @@ ATPostAlterTypeCleanup(List **wqueue, AlteredTableInfo *tab, LOCKMODE lockmode)
 		if (!conislocal)
 			continue;
 
+		/*
+		 * When rebuilding an FK constraint that references the table we're
+		 * modifying, we might not yet have any lock on the FK's table, so get
+		 * one now.  We'll need AccessExclusiveLock for the DROP CONSTRAINT
+		 * step, so there's no value in asking for anything weaker.
+		 */
+		if (relid != tab->relid && contype == CONSTRAINT_FOREIGN)
+			LockRelationOid(relid, AccessExclusiveLock);
+
 		ATPostAlterTypeParse(oldId, relid, confrelid,
 							 (char *) lfirst(def_item),
 							 wqueue, lockmode, tab->rewrite);
@@ -9951,7 +9961,7 @@ ATPostAlterTypeCleanup(List **wqueue, AlteredTableInfo *tab, LOCKMODE lockmode)
 							 (char *) lfirst(def_item),
 							 wqueue, lockmode, tab->rewrite);
 
-		ObjectAddressSet(obj, RelationRelationId, lfirst_oid(oid_item));
+		ObjectAddressSet(obj, RelationRelationId, oldId);
 		add_exact_object_address(&obj, objects);
 	}