On Sat, Jul 12, 2014 at 7:38 AM, Camilo Aguilar <camilo.agui...@gmail.com> wrote: >>Hm, if I read the DHCP spec correctly it requires the networks to deal >>with broadcast packets, as NAK is always sent as broadcast, so if this >>is the case we have a bigger problem. > > My interpretation is that a DHCPNAK is usually sent through unicast unless > the broadcast bit flag is ON in the DHCPDISCOVERY or DHCPREQUEST packet. > (Page 25)
This is what I had in mind (section 4.1)"In all cases, when 'giaddr' is zero, the server broadcasts any DHCPNAK messages to 0xffffffff." Anyway, see below. > I was also cc'ed recently to a new case in Linode: > https://bugs.freedesktop.org/show_bug.cgi?id=81225 > If what Linode says in this blog post still holds true, then it seems Linode > is filtering ARP broadcasts. Seems like you are right, and that there are cases where we must use unicast, and other cases where we must use broadcast (regardless of what the RFC says). Appears to be confirmed by the pcaps on fdo. I put your suggestion (unicast by default, but broadcast opt-in) in the TODO, I'd be happy to take a patch, otherwise I'll get to it soon. Cheers, Tom _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
