[ntp:hackers] Why are we going down this road? Question on listen-on, query-on, -I

Poul-Henning Kamp phk at phk.freebsd.dk
Sun May 31 17:53:31 UTC 2009

In message <4A201556.5010100 at sun.com>, Brian Utterback writes:

>I have heard the argument over and over that we have to bind all the 
>addresses to ensure that the src address of outgoing packets matches 
>the dst packets of the incoming packets.

Can I point out that POSIX adopted and standardized the
recvmsg(2)/sendmsg(2) API, for this very purpose ?

These API calls are widely unknown, but exactly there to not burden
applications with topology information, while giving them the
ability to reply with a consistent source address on multihomed

Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the hackers mailing list