Ensure first ModifyTable rel initialized if all are pruned

Commit cbc127917e introduced tracking of unpruned relids to avoid
processing pruned relations, and changed ExecInitModifyTable() to
initialize only unpruned result relations. As a result, MERGE
statements that prune all target partitions can now lead to crashes
or incorrect behavior during execution.

The crash occurs because some executor code paths rely on
ModifyTableState.resultRelInfo[0] being present and initialized,
even when no result relations remain after pruning. For example,
ExecMerge() and ExecMergeNotMatched() use the first resultRelInfo
to determine the appropriate action. Similarly,
ExecInitPartitionInfo() assumes that at least one result relation
exists.

To preserve these assumptions, ExecInitModifyTable() now includes the
first result relation in the initialized result relation list if all
result relations for that ModifyTable were pruned. To enable that,
ExecDoInitialPruning() ensures the first relation is locked if it was
pruned and locking is necessary.

To support this exception to the pruning logic, PlannedStmt now
includes a list of RT indexes identifying the first result relation
of each ModifyTable node in the plan. This allows
ExecDoInitialPruning() to check whether each such relation was
pruned and, if so, lock it if necessary.

Bug: #18830
Reported-by: Robins Tharakan <tharakan@gmail.com>
Diagnozed-by: Tender Wang <tndrwang@gmail.com>
Diagnozed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Co-authored-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Reviewed-by: Tender Wang <tndrwang@gmail.com>
Reviewed-by: Dean Rasheed <dean.a.rasheed@gmail.com>
Discussion: https://postgr.es/m/18830-1f31ea1dc930d444%40postgresql.org

1 files changed, 30 insertions, 1 deletions

diff --git a/src/backend/executor/execPartition.c b/src/backend/executor/execPartition.c
index 5cd5e2eeb80..84ccd7d457d 100644
--- a/src/backend/executor/execPartition.c
+++ b/src/backend/executor/execPartition.c
@@ -1819,6 +1819,7 @@ adjust_partition_colnos_using_map(List *colnos, AttrMap *attrMap)
 void
 ExecDoInitialPruning(EState *estate)
 {
+	PlannedStmt *stmt = estate->es_plannedstmt;
 	ListCell   *lc;
 	List	   *locked_relids = NIL;
 
@@ -1868,6 +1869,34 @@ ExecDoInitialPruning(EState *estate)
 	}
 
 	/*
+	 * Lock the first result relation of each ModifyTable node, even if it was
+	 * pruned.  This is required for ExecInitModifyTable(), which keeps its
+	 * first result relation if all other result relations have been pruned,
+	 * because some executor paths (e.g., in nodeModifyTable.c and
+	 * execPartition.c) rely on there being at least one result relation.
+	 *
+	 * There's room for improvement here --- we actually only need to do this
+	 * if all other result relations of the ModifyTable node were pruned, but
+	 * we don't have an easy way to tell that here.
+	 */
+	if (stmt->resultRelations && ExecShouldLockRelations(estate))
+	{
+		foreach(lc, stmt->firstResultRels)
+		{
+			Index		firstResultRel = lfirst_int(lc);
+
+			if (!bms_is_member(firstResultRel, estate->es_unpruned_relids))
+			{
+				RangeTblEntry *rte = exec_rt_fetch(firstResultRel, estate);
+
+				Assert(rte->rtekind == RTE_RELATION && rte->rellockmode != NoLock);
+				LockRelationOid(rte->relid, rte->rellockmode);
+				locked_relids = lappend_int(locked_relids, firstResultRel);
+			}
+		}
+	}
+
+	/*
 	 * Release the useless locks if the plan won't be executed.  This is the
 	 * same as what CheckCachedPlan() in plancache.c does.
 	 */
@@ -2076,7 +2105,7 @@ CreatePartitionPruneState(EState *estate, PartitionPruneInfo *pruneinfo,
 			 * because that entry will be held open and locked for the
 			 * duration of this executor run.
 			 */
-			partrel = ExecGetRangeTableRelation(estate, pinfo->rtindex);
+			partrel = ExecGetRangeTableRelation(estate, pinfo->rtindex, false);
 
 			/* Remember for InitExecPartitionPruneContext(). */
 			pprune->partrel = partrel;