Scott Long <scott4l...@yahoo.com> wrote in <2ae597f6-02a5-405f-b1d8-34054a0da...@yahoo.com>:
sc> sc> On Jul 17, 2013, at 7:00 PM, Glen Barber <g...@freebsd.org> wrote: sc> sc> > On Wed, Jul 17, 2013 at 02:05:20PM +0000, Hiroki Sato wrote: sc> >> Author: hrs sc> >> Date: Wed Jul 17 14:05:20 2013 sc> >> New Revision: 253427 sc> >> URL: http://svnweb.freebsd.org/changeset/base/253427 sc> >> sc> >> Log: sc> >> - Add support of MK_INET_SUPPORT=no. sc> >> - Fix a bug in sodump() which prevented struct sockaddr_in6 from sc> >> - displaying. sc> >> - Fix a bug in in fiboptlist_csv() which could cause free() of sc> >> - uninitialized sc> >> pointer. sc> >> - Style cleanups: sc> >> . Add missing "static" keywords. sc> >> . Use an array of struct sockaddr_storage instead of sockunion for sc> >> rtmsg. sc> >> . Use err() and errx() instead of pair of fprintf(stderr, "...") + sc> >> exit(1). sc> >> . Use nitems() macro. sc> >> . Various style(9) fixes. sc> >> sc> >> Modified: sc> >> head/sbin/route/Makefile sc> >> head/sbin/route/route.c sc> >> sc> > sc> > [...] sc> > sc> >> Modified: head/sbin/route/route.c sc> >> ============================================================================== sc> >> --- head/sbin/route/route.c Wed Jul 17 14:04:18 2013 (r253426) sc> >> +++ head/sbin/route/route.c Wed Jul 17 14:05:20 2013 (r253427) sc> > [...] sc> >> @@ -1470,14 +1498,12 @@ rtmsg(int cmd, int flags, int fib) sc> >> rtm.rtm_rmx = rt_metrics; sc> >> rtm.rtm_inits = rtm_inits; sc> >> sc> >> - if (rtm_addrs & RTA_NETMASK) sc> >> - mask_addr(); sc> >> - NEXTADDR(RTA_DST, so_dst); sc> >> - NEXTADDR(RTA_GATEWAY, so_gate); sc> >> - NEXTADDR(RTA_NETMASK, so_mask); sc> >> - NEXTADDR(RTA_GENMASK, so_genmask); sc> >> - NEXTADDR(RTA_IFP, so_ifp); sc> >> - NEXTADDR(RTA_IFA, so_ifa); sc> >> + NEXTADDR(RTA_DST, so[RTAX_DST]); sc> >> + NEXTADDR(RTA_GATEWAY, so[RTAX_GATEWAY]); sc> >> + NEXTADDR(RTA_NETMASK, so[RTAX_NETMASK]); sc> >> + NEXTADDR(RTA_GENMASK, so[RTAX_GENMASK]); sc> >> + NEXTADDR(RTA_IFP, so[RTAX_IFP]); sc> >> + NEXTADDR(RTA_IFA, so[RTAX_IFA]); sc> >> rtm.rtm_msglen = l = cp - (char *)&m_rtmsg; sc> >> if (verbose) sc> >> print_rtmsg(&rtm, l); sc> > sc> > I think this breaks head/, at least for ia64. sc> > sc> sc> Yes, as well as for amd64. This breakage occurred at the exact moment sc> that we tried to cut over to HEAD at Netflix. It's indeed a Brave New sc> World. Should we back it out? Sorry, should be fixed just now. It seems I need more sleep. -- Hiroki
pgpqAhC3a8Hha.pgp
Description: PGP signature