Don't reset the oc->ban pointer from BAN_CheckObject

Resetting the pointer would race against the ban lurker moving the OC
in the ban list and any object event listeners wanting to see the new
ban timestamp on OEV_BANCHG.

Fixes: #1863

bin/varnishd/cache/cache_ban.c

@@ -575,16 +575,14 @@ BAN_CheckObject(struct worker *wrk, struct objcore *oc, struct req *req)
 	VSC_C_main->bans_tested++;
 	VSC_C_main->bans_tests_tested += tests;
 
-	oc->ban->refcount--;
-	VTAILQ_REMOVE(&oc->ban->objcore, oc, ban_list);
 	if (b == bn) {
 		/* not banned */
+		oc->ban->refcount--;
+		VTAILQ_REMOVE(&oc->ban->objcore, oc, ban_list);
 		VTAILQ_INSERT_TAIL(&b0->objcore, oc, ban_list);
 		b0->refcount++;
 		oc->ban = b0;
 		b = NULL;
-	} else {
-		oc->ban = NULL;
 	}
 
 	if (VTAILQ_LAST(&ban_head, banhead_s)->refcount == 0)