Fix temporary memory leak in system table index scans

Commit 811af9786b introduced palloc() calls into systable_beginscan()
and systable_beginscan_ordered().  But there was no pfree(), as is the
usual style.

It turns out that an ANALYZE of a partitioned table can invoke many
thousand system table index scans, and this memory is not cleaned up
until the end of the command, so this can temporarily leak quite a bit
of memory.  Maybe there are improvements to be made at a higher level
about this, but for now, insert a couple of corresponding pfree()
calls to fix this particular issue.

Reported-by: Justin Pryzby <pryzby@telsasoft.com>
Discussion: https://www.postgresql.org/message-id/Z0XTfIq5xUtbkiIh@pryzbyj2023

src/backend/access/index/genam.c

@@ -449,6 +449,8 @@ systable_beginscan(Relation heapRelation,
 										 snapshot, nkeys, 0);
 		index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
 		sysscan->scan = NULL;
+
+		pfree(idxkey);
 	}
 	else
 	{
@@ -713,6 +715,8 @@ systable_beginscan_ordered(Relation heapRelation,
 	index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
 	sysscan->scan = NULL;
 
+	pfree(idxkey);
+
 	/*
 	 * If CheckXidAlive is set then set a flag to indicate that system table
 	 * scan is in-progress.  See detailed comments in xact.c where these