Fix costing of parallel hash joins.

Commit 1804284 established that single-batch
parallel-aware hash joins could create one large shared hash table using the
combined work_mem budget of all participants.  The costing accidentally
assumed that parallel-oblivious hash joins could also do that.  The
documentation for initial_cost_hashjoin() also failed to mention the new
argument.  Repair.

Author: Thomas Munro
Reported-By: Antonin Houska
Reviewed-By: Antonin Houska
Discussion: https://postgr.es/m/12441.1513935950%40localhost

Author: petere

src/backend/optimizer/path/costsize.c

@@ -3143,6 +3143,8 @@ cached_scansel(PlannerInfo *root, RestrictInfo *rinfo, PathKey *pathkey)
  * 'outer_path' is the outer input to the join
  * 'inner_path' is the inner input to the join
  * 'extra' contains miscellaneous information about the join
+ * 'parallel_hash' indicates that inner_path is partial and that a shared
+ *		hash table will be built in parallel
  */
 void
 initial_cost_hashjoin(PlannerInfo *root, JoinCostWorkspace *workspace,

src/backend/optimizer/path/joinpath.c

@@ -814,7 +814,7 @@ try_partial_hashjoin_path(PlannerInfo *root,
 	 * cost.  Bail out right away if it looks terrible.
 	 */
 	initial_cost_hashjoin(root, &workspace, jointype, hashclauses,
-						  outer_path, inner_path, extra, true);
+						  outer_path, inner_path, extra, parallel_hash);
 	if (!add_partial_path_precheck(joinrel, workspace.total_cost, NIL))
 		return;