Avoid harmless Valgrind no-buffer-pin errors.

petergeoghegan · petergeoghegan · commit a766d6ca22ac · 2020-07-19T16:12:51.000-07:00
Valgrind builds with assertions enabled sometimes perform a theoretically unsafe page access inside an assertion in heapam_tuple_lock(). This happened when the eval-plan-qual isolation test ran one of the permutations added by commit a2418f9. Avoid complaints from Valgrind by moving the assertion ever so slightly. This is minor cleanup for commit 1e0dfd1, which added Valgrind buffer access instrumentation. No backpatch, since this only happens within an assertion, and seems very unlikely to cause any real problems even with assert-enabled builds.
diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
@@ -368,10 +368,11 @@ heapam_tuple_lock(Relation relation, ItemPointer tid, Snapshot snapshot,
 	if (result == TM_Updated &&
 		(flags & TUPLE_LOCK_FLAG_FIND_LAST_VERSION))
 	{
-		ReleaseBuffer(buffer);
 		/* Should not encounter speculative tuple on recheck */
 		Assert(!HeapTupleHeaderIsSpeculative(tuple->t_data));
 
+		ReleaseBuffer(buffer);
+
 		if (!ItemPointerEquals(&tmfd->ctid, &tuple->t_self))
 		{
 			SnapshotData SnapshotDirty;

Original file line number	Diff line number	Diff line change
`@@ -368,10 +368,11 @@ heapam_tuple_lock(Relation relation, ItemPointer tid, Snapshot snapshot,`
`368`	`368`	`if (result == TM_Updated &&`
`369`	`369`	`(flags & TUPLE_LOCK_FLAG_FIND_LAST_VERSION))`
`370`	`370`	`{`
`371`		`- ReleaseBuffer(buffer);`
`372`	`371`	`/* Should not encounter speculative tuple on recheck */`
`373`	`372`	`Assert(!HeapTupleHeaderIsSpeculative(tuple->t_data));`
`374`	`373`
	`374`	`+ ReleaseBuffer(buffer);`
	`375`	`+`
`375`	`376`	`if (!ItemPointerEquals(&tmfd->ctid, &tuple->t_self))`
`376`	`377`	`{`
`377`	`378`	`SnapshotData SnapshotDirty;`