Hold interrupts while running dsm_detach() callbacks.

macdice · macdice · commit 637668fb1d17 · 2021-02-15T14:27:33.000+13:00
While cleaning up after a parallel query or parallel index creation that created temporary files, we could be interrupted by a statement timeout. The error handling path would then fail to clean up the files when it ran dsm_detach() again, because the callback was already popped off the list. Prevent this hazard by holding interrupts while the cleanup code runs. Thanks to Heikki Linnakangas for this suggestion, and also to Kyotaro Horiguchi, Masahiko Sawada, Justin Pryzby and Tom Lane for discussion of this and earlier ideas on how to fix the problem. Back-patch to all supported releases. Reported-by: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/20191212180506.GR2082@telsasoft.com
diff --git a/src/backend/storage/ipc/dsm.c b/src/backend/storage/ipc/dsm.c
@@ -771,8 +771,12 @@ dsm_detach(dsm_segment *seg)
 	/*
 	 * Invoke registered callbacks.  Just in case one of those callbacks
 	 * throws a further error that brings us back here, pop the callback
-	 * before invoking it, to avoid infinite error recursion.
+	 * before invoking it, to avoid infinite error recursion.  Don't allow
+	 * interrupts while running the individual callbacks in non-error code
+	 * paths, to avoid leaving cleanup work unfinished if we're interrupted by
+	 * a statement timeout or similar.
 	 */
+	HOLD_INTERRUPTS();
 	while (!slist_is_empty(&seg->on_detach))
 	{
 		slist_node *node;
@@ -788,6 +792,7 @@ dsm_detach(dsm_segment *seg)
 
 		function(seg, arg);
 	}
+	RESUME_INTERRUPTS();
 
 	/*
 	 * Try to remove the mapping, if one exists.  Normally, there will be, but
