Change the routing entry of the pool addresses from blackhole to

reject.  This is better because the routing cache can be purged by the
reject.

diff --git a/usr.sbin/npppd/npppd/npppd.c b/usr.sbin/npppd/npppd/npppd.c
index 295cb9e..b9f68b3 100644 (file)
--- a/usr.sbin/npppd/npppd/npppd.c
+++ b/usr.sbin/npppd/npppd/npppd.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: npppd.c,v 1.54 2024/07/11 14:05:59 yasuoka Exp $ */
+/*     $OpenBSD: npppd.c,v 1.55 2024/08/01 00:05:16 yasuoka Exp $ */
 
 /*-
  * Copyright (c) 2005-2008,2009 Internet Initiative Japan Inc.
@@ -29,7 +29,7 @@
  * Next pppd(nppd). This file provides a npppd daemon process and operations
  * for npppd instance.
  * @author     Yasuoka Masahiko
- * $Id: npppd.c,v 1.54 2024/07/11 14:05:59 yasuoka Exp $
+ * $Id: npppd.c,v 1.55 2024/08/01 00:05:16 yasuoka Exp $
  */
 #include "version.h"
 #include <sys/param.h> /* ALIGNED_POINTER */
@@ -608,7 +608,7 @@ npppd_reset_routing_table(npppd *_this, int pool_only)
                                if (is_first)
                                        in_route_add(&snp->snp_addr,
                                            &snp->snp_mask, &loop,
-                                           LOOPBACK_IFNAME, RTF_BLACKHOLE, 0);
+                                           LOOPBACK_IFNAME, RTF_REJECT, 0);
                                break;
 
                        case SNP_PPP: