On Jun 13 20:45:07, Eric Faurot wrote:
> On Wed, Jun 13, 2012 at 07:39:46PM +0200, Jan Stary wrote:
> > Through my smtpd, I am sending an email to an address
> > in the fjfi.cvut.cz domain. The email get delivered alright,
> > but the maillog message is strange:
> >
> > Jun 13 19:32:08 mini smtpd[7138]: 93fe4a4f: from=<r...@mini.stare.cz>,
> > size=427, nrcpts=1, proto=ESMTP, relay=0@localhost [IPv6:::1]
> > Jun 13 19:32:10 mini smtpd[6586]: 93fe4a4f700df332:
> > to=<jan.st...@fjfi.cvut.cz>, delay=2, relay=n003-000-000-000.static.ge.com
> > [147.32.9.3], stat=Sent (2.0.0 Ok: queued as CECD3C0083)
> >
> > $ host -t mx fjfi.cvut.cz.
> > fjfi.cvut.cz mail is handled by 70 brk1.fjfi.cvut.cz.
> > fjfi.cvut.cz mail is handled by 30 mailgw1.fjfi.cvut.cz.
> > fjfi.cvut.cz mail is handled by 50 mailgw2.fjfi.cvut.cz.
> > fjfi.cvut.cz mail is handled by 60 brk2.fjfi.cvut.cz.
> >
> > So smtpd correctly chooses the first priority MX, which is
> >
> > $ host mailgw1.fjfi.cvut.cz.
> > mailgw1.fjfi.cvut.cz has address 147.32.9.3
> >
> > $ host 147.32.9.3
> > 3.9.32.147.in-addr.arpa domain name pointer mailgw1.fjfi.cvut.cz.
> >
> > All the resolvers in my /etc/resolv.conf resolve
> > mailgw1.fjfi.cvut.cz. to 147.32.9.3 and back, as above.
> >
> > So how did "relay=n003-000-000-000.static.ge.com [147.32.9.3]"
> > get into the log? Am I missing something?
> >
> > Jan
>
> That's an endianness issue. I thought I fixed that already...
> Can you try this?
>
> Eric.
>
> Index: sockaddr.c
> ===================================================================
> RCS file: /cvs/src/usr.sbin/smtpd/sockaddr.c,v
> retrieving revision 1.4
> diff -u -p -u -r1.4 sockaddr.c
> --- sockaddr.c 16 May 2011 10:57:41 -0000 1.4
> +++ sockaddr.c 13 Jun 2012 18:40:52 -0000
> @@ -114,7 +114,7 @@ sockaddr_as_fqdn(const struct sockaddr *
>
> switch (sa->sa_family) {
> case AF_INET:
> - addr = ((const struct sockaddr_in *)sa)->sin_addr.s_addr;
> + addr = ntohl(((const struct sockaddr_in *)sa)->sin_addr.s_addr);
> snprintf(dst, max,
> "%d.%d.%d.%d.in-addr.arpa.",
> (addr >> 24) & 0xff,
It is still happening with this patch.
