Fix mishandling of resjunk columns in ON CONFLICT ... UPDATE tlists.

tglsfdc · tglsfdc · commit b7d1f32ff658 · 2021-05-10T11:02:29.000-04:00
It's unusual to have any resjunk columns in an ON CONFLICT ... UPDATE list, but it can happen when MULTIEXPR_SUBLINK SubPlans are present. If it happens, the ON CONFLICT UPDATE code path would end up storing tuples that include the values of the extra resjunk columns. That's fairly harmless in the short run, but if new columns are added to the table then the values would become accessible, possibly leading to malfunctions if they don't match the datatypes of the new columns. This had escaped notice through a confluence of missing sanity checks, including * There's no cross-check that a tuple presented to heap_insert or heap_update matches the table rowtype. While it's difficult to check that fully at reasonable cost, we can easily add assertions that there aren't too many columns. * The output-column-assignment cases in execExprInterp.c lacked any sanity checks on the output column numbers, which seems like an oversight considering there are plenty of assertion checks on input column numbers. Add assertions there too. * We failed to apply nodeModifyTable's ExecCheckPlanOutput() to the ON CONFLICT UPDATE tlist. That wouldn't have caught this specific error, since that function is chartered to ignore resjunk columns; but it sure seems like a bad omission now that we've seen this bug. In HEAD, the right way to fix this is to make the processing of ON CONFLICT UPDATE tlists work the same as regular UPDATE tlists now do, that is don't add "SET x = x" entries, and use ExecBuildUpdateProjection to evaluate the tlist and combine it with old values of the not-set columns. This adds a little complication to ExecBuildUpdateProjection, but allows removal of a comparable amount of now-dead code from the planner. In the back branches, the most expedient solution seems to be to (a) use an output slot for the ON CONFLICT UPDATE projection that actually matches the target table, and then (b) invent a variant of ExecBuildProjectionInfo that can be told to not store values resulting from resjunk columns, so it doesn't try to store into nonexistent columns of the output slot. (We can't simply ignore the resjunk columns altogether; they have to be evaluated for MULTIEXPR_SUBLINK to work.) This works back to v10. In 9.6, projections work much differently and we can't cheaply give them such an option. The 9.6 version of this patch works by inserting a JunkFilter when it's necessary to get rid of resjunk columns. In addition, v11 and up have the reverse problem when trying to perform ON CONFLICT UPDATE on a partitioned table. Through a further oversight, adjust_partition_tlist() discarded resjunk columns when re-ordering the ON CONFLICT UPDATE tlist to match a partition. This accidentally prevented the storing-bogus-tuples problem, but at the cost that MULTIEXPR_SUBLINK cases didn't work, typically crashing if more than one row has to be updated. Fix by preserving resjunk columns in that routine. (I failed to resist the temptation to add more assertions there too, and to do some minor code beautification.) Per report from Andres Freund. Back-patch to all supported branches. Security: CVE-2021-32028
diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c
@@ -2459,6 +2459,10 @@ heap_insert(Relation relation, HeapTuple tup, CommandId cid,
 	Buffer		vmbuffer = InvalidBuffer;
 	bool		all_visible_cleared = false;
 
+	/* Cheap, simplistic check that the tuple matches the rel's rowtype. */
+	Assert(HeapTupleHeaderGetNatts(tup->t_data) <=
+		   RelationGetNumberOfAttributes(relation));
+
 	/*
 	 * Fill in tuple header fields, assign an OID, and toast the tuple if
 	 * necessary.
@@ -3573,6 +3577,10 @@ heap_update(Relation relation, ItemPointer otid, HeapTuple newtup,
 
 	Assert(ItemPointerIsValid(otid));
 
+	/* Cheap, simplistic check that the tuple matches the rel's rowtype. */
+	Assert(HeapTupleHeaderGetNatts(newtup->t_data) <=
+		   RelationGetNumberOfAttributes(relation));
+
 	/*
 	 * Forbid this during a parallel operation, lest it allocate a combocid.
 	 * Other workers might need that combocid for visibility checks, and we
diff --git a/src/backend/executor/execExpr.c b/src/backend/executor/execExpr.c
@@ -351,6 +351,32 @@ ExecBuildProjectionInfo(List *targetList,
 						TupleTableSlot *slot,
 						PlanState *parent,
 						TupleDesc inputDesc)
+{
+	return ExecBuildProjectionInfoExt(targetList,
+									  econtext,
+									  slot,
+									  true,
+									  parent,
+									  inputDesc);
+}
+
+/*
+ *		ExecBuildProjectionInfoExt
+ *
+ * As above, with one additional option.
+ *
+ * If assignJunkEntries is true (the usual case), resjunk entries in the tlist
+ * are not handled specially: they are evaluated and assigned to the proper
+ * column of the result slot.  If assignJunkEntries is false, resjunk entries
+ * are evaluated, but their result is discarded without assignment.
+ */
+ProjectionInfo *
+ExecBuildProjectionInfoExt(List *targetList,
+						   ExprContext *econtext,
+						   TupleTableSlot *slot,
+						   bool assignJunkEntries,
+						   PlanState *parent,
+						   TupleDesc inputDesc)
 {
 	ProjectionInfo *projInfo = makeNode(ProjectionInfo);
 	ExprState  *state;
@@ -388,7 +414,8 @@ ExecBuildProjectionInfo(List *targetList,
 		 */
 		if (tle->expr != NULL &&
 			IsA(tle->expr, Var) &&
-			((Var *) tle->expr)->varattno > 0)
+			((Var *) tle->expr)->varattno > 0 &&
+			(assignJunkEntries || !tle->resjunk))
 		{
 			/* Non-system Var, but how safe is it? */
 			variable = (Var *) tle->expr;
@@ -452,6 +479,10 @@ ExecBuildProjectionInfo(List *targetList,
 			ExecInitExprRec(tle->expr, state,
 							&state->resvalue, &state->resnull);
 
+			/* This makes it easy to discard resjunk results when told to. */
+			if (!assignJunkEntries && tle->resjunk)
+				continue;
+
 			/*
 			 * Column might be referenced multiple times in upper nodes, so
 			 * force value to R/O - but only if it could be an expanded datum.
diff --git a/src/backend/executor/execExprInterp.c b/src/backend/executor/execExprInterp.c
@@ -560,6 +560,7 @@ ExecInterpExpr(ExprState *state, ExprContext *econtext, bool *isnull)
 			 * care of at compilation time.  But see EEOP_INNER_VAR comments.
 			 */
 			Assert(attnum >= 0 && attnum < innerslot->tts_nvalid);
+			Assert(resultnum >= 0 && resultnum < resultslot->tts_tupleDescriptor->natts);
 			resultslot->tts_values[resultnum] = innerslot->tts_values[attnum];
 			resultslot->tts_isnull[resultnum] = innerslot->tts_isnull[attnum];
 
@@ -576,6 +577,7 @@ ExecInterpExpr(ExprState *state, ExprContext *econtext, bool *isnull)
 			 * care of at compilation time.  But see EEOP_INNER_VAR comments.
 			 */
 			Assert(attnum >= 0 && attnum < outerslot->tts_nvalid);
+			Assert(resultnum >= 0 && resultnum < resultslot->tts_tupleDescriptor->natts);
 			resultslot->tts_values[resultnum] = outerslot->tts_values[attnum];
 			resultslot->tts_isnull[resultnum] = outerslot->tts_isnull[attnum];
 
@@ -592,6 +594,7 @@ ExecInterpExpr(ExprState *state, ExprContext *econtext, bool *isnull)
 			 * care of at compilation time.  But see EEOP_INNER_VAR comments.
 			 */
 			Assert(attnum >= 0 && attnum < scanslot->tts_nvalid);
+			Assert(resultnum >= 0 && resultnum < resultslot->tts_tupleDescriptor->natts);
 			resultslot->tts_values[resultnum] = scanslot->tts_values[attnum];
 			resultslot->tts_isnull[resultnum] = scanslot->tts_isnull[attnum];
 
@@ -602,6 +605,7 @@ ExecInterpExpr(ExprState *state, ExprContext *econtext, bool *isnull)
 		{
 			int			resultnum = op->d.assign_tmp.resultnum;
 
+			Assert(resultnum >= 0 && resultnum < resultslot->tts_tupleDescriptor->natts);
 			resultslot->tts_values[resultnum] = state->resvalue;
 			resultslot->tts_isnull[resultnum] = state->resnull;
 
@@ -612,6 +616,7 @@ ExecInterpExpr(ExprState *state, ExprContext *econtext, bool *isnull)
 		{
 			int			resultnum = op->d.assign_tmp.resultnum;
 
+			Assert(resultnum >= 0 && resultnum < resultslot->tts_tupleDescriptor->natts);
 			resultslot->tts_isnull[resultnum] = state->resnull;
 			if (!resultslot->tts_isnull[resultnum])
 				resultslot->tts_values[resultnum] =
@@ -2046,8 +2051,10 @@ ExecJustAssignInnerVar(ExprState *state, ExprContext *econtext, bool *isnull)
 	 *
 	 * Since we use slot_getattr(), we don't need to implement the FETCHSOME
 	 * step explicitly, and we also needn't Assert that the attnum is in range
-	 * --- slot_getattr() will take care of any problems.
+	 * --- slot_getattr() will take care of any problems.  Nonetheless, check
+	 * that resultnum is in range.
 	 */
+	Assert(resultnum >= 0 && resultnum < outslot->tts_tupleDescriptor->natts);
 	outslot->tts_values[resultnum] =
 		slot_getattr(inslot, attnum, &outslot->tts_isnull[resultnum]);
 	return 0;
@@ -2064,6 +2071,7 @@ ExecJustAssignOuterVar(ExprState *state, ExprContext *econtext, bool *isnull)
 	TupleTableSlot *outslot = state->resultslot;
 
 	/* See comments in ExecJustAssignInnerVar */
+	Assert(resultnum >= 0 && resultnum < outslot->tts_tupleDescriptor->natts);
 	outslot->tts_values[resultnum] =
 		slot_getattr(inslot, attnum, &outslot->tts_isnull[resultnum]);
 	return 0;
@@ -2080,6 +2088,7 @@ ExecJustAssignScanVar(ExprState *state, ExprContext *econtext, bool *isnull)
 	TupleTableSlot *outslot = state->resultslot;
 
 	/* See comments in ExecJustAssignInnerVar */
+	Assert(resultnum >= 0 && resultnum < outslot->tts_tupleDescriptor->natts);
 	outslot->tts_values[resultnum] =
 		slot_getattr(inslot, attnum, &outslot->tts_isnull[resultnum]);
 	return 0;
diff --git a/src/backend/executor/execPartition.c b/src/backend/executor/execPartition.c
@@ -629,11 +629,11 @@ ExecInitPartitionInfo(ModifyTableState *mtstate,
 				leaf_part_rri->ri_onConflict = rootResultRelInfo->ri_onConflict;
 			else
 			{
+				OnConflictSetState *onconfl = makeNode(OnConflictSetState);
 				List	   *onconflset;
-				TupleDesc	tupDesc;
 				bool		found_whole_row;
 
-				leaf_part_rri->ri_onConflict = makeNode(OnConflictSetState);
+				leaf_part_rri->ri_onConflict = onconfl;
 
 				/*
 				 * Translate expressions in onConflictSet to account for
@@ -642,7 +642,7 @@ ExecInitPartitionInfo(ModifyTableState *mtstate,
 				 * pseudo-relation (INNER_VAR), and second to handle the main
 				 * target relation (firstVarno).
 				 */
-				onconflset = (List *) copyObject((Node *) node->onConflictSet);
+				onconflset = copyObject(node->onConflictSet);
 				if (part_attnos == NULL)
 					part_attnos =
 						convert_tuples_by_name_map(RelationGetDescr(partrel),
@@ -665,7 +665,7 @@ ExecInitPartitionInfo(ModifyTableState *mtstate,
 										&found_whole_row);
 				/* We ignore the value of found_whole_row. */
 
-				/* Finally, adjust this tlist to match the partition. */
+				/* Finally, reorder the tlist to match the partition. */
 				onconflset = adjust_partition_tlist(onconflset, map);
 
 				/*
@@ -675,13 +675,12 @@ ExecInitPartitionInfo(ModifyTableState *mtstate,
 				 * partition that's tupdesc-equal to the partitioned table;
 				 * partitions of different tupdescs must generate their own.
 				 */
-				tupDesc = ExecTypeFromTL(onconflset, partrelDesc->tdhasoid);
-				ExecSetSlotDescriptor(mtstate->mt_conflproj, tupDesc);
-				leaf_part_rri->ri_onConflict->oc_ProjInfo =
-					ExecBuildProjectionInfo(onconflset, econtext,
-											mtstate->mt_conflproj,
-											&mtstate->ps, partrelDesc);
-				leaf_part_rri->ri_onConflict->oc_ProjTupdesc = tupDesc;
+				ExecSetSlotDescriptor(mtstate->mt_conflproj, partrelDesc);
+				onconfl->oc_ProjInfo =
+					ExecBuildProjectionInfoExt(onconflset, econtext,
+											   mtstate->mt_conflproj, false,
+											   &mtstate->ps, partrelDesc);
+				onconfl->oc_ProjTupdesc = partrelDesc;
 
 				/*
 				 * If there is a WHERE clause, initialize state where it will
@@ -710,7 +709,7 @@ ExecInitPartitionInfo(ModifyTableState *mtstate,
 											RelationGetForm(partrel)->reltype,
 											&found_whole_row);
 					/* We ignore the value of found_whole_row. */
-					leaf_part_rri->ri_onConflict->oc_WhereClause =
+					onconfl->oc_WhereClause =
 						ExecInitQual((List *) clause, &mtstate->ps);
 				}
 			}
@@ -1332,18 +1331,15 @@ ExecBuildSlotPartitionKeyDescription(Relation rel,
 
 /*
  * adjust_partition_tlist
- *		Adjust the targetlist entries for a given partition to account for
- *		attribute differences between parent and the partition
+ *		Re-order the targetlist entries for a given partition to account for
+ *		column position differences between the parent and the partition.
  *
- * The expressions have already been fixed, but here we fix the list to make
- * target resnos match the partition's attribute numbers.  This results in a
- * copy of the original target list in which the entries appear in resno
- * order, including both the existing entries (that may have their resno
- * changed in-place) and the newly added entries for columns that don't exist
- * in the parent.
+ * The expressions have already been fixed, but we must now re-order the
+ * entries in case the partition has different column order, and possibly
+ * add or remove dummy entries for dropped columns.
  *
- * Scribbles on the input tlist, so callers must make sure to make a copy
- * before passing it to us.
+ * Although a new List is returned, this feels free to scribble on resno
+ * fields of the given tlist, so that should be a working copy.
  */
 static List *
 adjust_partition_tlist(List *tlist, TupleConversionMap *map)
@@ -1352,31 +1348,35 @@ adjust_partition_tlist(List *tlist, TupleConversionMap *map)
 	TupleDesc	tupdesc = map->outdesc;
 	AttrNumber *attrMap = map->attrMap;
 	AttrNumber	attrno;
+	ListCell   *lc;
 
 	for (attrno = 1; attrno <= tupdesc->natts; attrno++)
 	{
 		Form_pg_attribute att_tup = TupleDescAttr(tupdesc, attrno - 1);
+		AttrNumber	parentattrno = attrMap[attrno - 1];
 		TargetEntry *tle;
 
-		if (attrMap[attrno - 1] != InvalidAttrNumber)
+		if (parentattrno != InvalidAttrNumber)
 		{
-			Assert(!att_tup->attisdropped);
-
 			/*
 			 * Use the corresponding entry from the parent's tlist, adjusting
-			 * the resno the match the partition's attno.
+			 * the resno to match the partition's attno.
 			 */
-			tle = (TargetEntry *) list_nth(tlist, attrMap[attrno - 1] - 1);
+			Assert(!att_tup->attisdropped);
+			tle = (TargetEntry *) list_nth(tlist, parentattrno - 1);
+			Assert(!tle->resjunk);
+			Assert(tle->resno == parentattrno);
 			tle->resno = attrno;
 		}
 		else
 		{
-			Const	   *expr;
-
 			/*
 			 * For a dropped attribute in the partition, generate a dummy
-			 * entry with resno matching the partition's attno.
+			 * entry with resno matching the partition's attno.  This should
+			 * match what expand_targetlist() does.
 			 */
+			Const	   *expr;
+
 			Assert(att_tup->attisdropped);
 			expr = makeConst(INT4OID,
 							 -1,
@@ -1394,6 +1394,18 @@ adjust_partition_tlist(List *tlist, TupleConversionMap *map)
 		new_tlist = lappend(new_tlist, tle);
 	}
 
+	/* Finally, attach any resjunk entries to the end of the new tlist */
+	foreach(lc, tlist)
+	{
+		TargetEntry *tle = (TargetEntry *) lfirst(lc);
+
+		if (tle->resjunk)
+		{
+			tle->resno = list_length(new_tlist) + 1;
+			new_tlist = lappend(new_tlist, tle);
+		}
+	}
+
 	return new_tlist;
 }
 
diff --git a/src/backend/executor/nodeModifyTable.c b/src/backend/executor/nodeModifyTable.c
@@ -2533,9 +2533,9 @@ ExecInitModifyTable(ModifyTable *node, EState *estate, int eflags)
 	 */
 	if (node->onConflictAction == ONCONFLICT_UPDATE)
 	{
+		OnConflictSetState *onconfl = makeNode(OnConflictSetState);
 		ExprContext *econtext;
 		TupleDesc	relationDesc;
-		TupleDesc	tupDesc;
 
 		/* insert may only have one plan, inheritance is not expanded */
 		Assert(nplans == 1);
@@ -2562,7 +2562,7 @@ ExecInitModifyTable(ModifyTable *node, EState *estate, int eflags)
 		mtstate->mt_excludedtlist = node->exclRelTlist;
 
 		/* create state for DO UPDATE SET operation */
-		resultRelInfo->ri_onConflict = makeNode(OnConflictSetState);
+		resultRelInfo->ri_onConflict = onconfl;
 
 		/*
 		 * Create the tuple slot for the UPDATE SET projection.
@@ -2573,19 +2573,27 @@ ExecInitModifyTable(ModifyTable *node, EState *estate, int eflags)
 		 * the tupdesc in the parent's state: it can be reused by partitions
 		 * with an identical descriptor to the parent.
 		 */
-		tupDesc = ExecTypeFromTL((List *) node->onConflictSet,
-								 relationDesc->tdhasoid);
 		mtstate->mt_conflproj =
 			ExecInitExtraTupleSlot(mtstate->ps.state,
 								   mtstate->mt_partition_tuple_routing ?
-								   NULL : tupDesc);
-		resultRelInfo->ri_onConflict->oc_ProjTupdesc = tupDesc;
+								   NULL : relationDesc);
+		onconfl->oc_ProjTupdesc = relationDesc;
+
+		/*
+		 * The onConflictSet tlist should already have been adjusted to emit
+		 * the table's exact column list.  It could also contain resjunk
+		 * columns, which should be evaluated but not included in the
+		 * projection result.
+		 */
+		ExecCheckPlanOutput(resultRelInfo->ri_RelationDesc,
+							node->onConflictSet);
 
 		/* build UPDATE SET projection state */
-		resultRelInfo->ri_onConflict->oc_ProjInfo =
-			ExecBuildProjectionInfo(node->onConflictSet, econtext,
-									mtstate->mt_conflproj, &mtstate->ps,
-									relationDesc);
+		onconfl->oc_ProjInfo =
+			ExecBuildProjectionInfoExt(node->onConflictSet, econtext,
+									   mtstate->mt_conflproj, false,
+									   &mtstate->ps,
+									   relationDesc);
 
 		/* initialize state to evaluate the WHERE clause, if any */
 		if (node->onConflictWhere)
@@ -2594,7 +2602,7 @@ ExecInitModifyTable(ModifyTable *node, EState *estate, int eflags)
 
 			qualexpr = ExecInitQual((List *) node->onConflictWhere,
 									&mtstate->ps);
-			resultRelInfo->ri_onConflict->oc_WhereClause = qualexpr;
+			onconfl->oc_WhereClause = qualexpr;
 		}
 	}
 
diff --git a/src/include/executor/executor.h b/src/include/executor/executor.h
@@ -269,6 +269,12 @@ extern ProjectionInfo *ExecBuildProjectionInfo(List *targetList,
 						TupleTableSlot *slot,
 						PlanState *parent,
 						TupleDesc inputDesc);
+extern ProjectionInfo *ExecBuildProjectionInfoExt(List *targetList,
+												  ExprContext *econtext,
+												  TupleTableSlot *slot,
+												  bool assignJunkEntries,
+												  PlanState *parent,
+												  TupleDesc inputDesc);
 extern ExprState *ExecPrepareExpr(Expr *node, EState *estate);
 extern ExprState *ExecPrepareQual(List *qual, EState *estate);
 extern ExprState *ExecPrepareCheck(List *qual, EState *estate);
diff --git a/src/test/regress/expected/update.out b/src/test/regress/expected/update.out
diff --git a/src/test/regress/sql/update.sql b/src/test/regress/sql/update.sql
