Date: Thu, 09 Aug 2001 05:47:09 +0900 From: Jun-ichiro itojun Hagino <itojun@iijlab.net> To: Mike Tancsa <mike@sentex.net> Cc: freebsd-net@freebsd.org Subject: Re: gif MTU of 1280 ? Message-ID: <20010808204709.308547BA@starfruit.itojun.org> In-Reply-To: mike's message of Wed, 08 Aug 2001 16:32:01 -0400. <5.1.0.14.0.20010808162550.0433abb0@marble.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
>172.16.1.1/24 <--> 10.0.0.1/30 - - - 10.0.0.2/30<--->192.168.1.1 > >With 192.168.1.1 and 172.16.1.1 being non RFC 1918 addresses in this >example and the 10.0.0.1 IPs being the internal addresses. > >Lets say a machine (216.136.204.21 wants to get to 192.168.1.1, and it >comes in via the public gateway 172.16.1.1. To get to 192.168.1.1, it must >be fragmented. The question is, when the ICMP message is sent out, with >what source address does the packet leave ? e.g. if 216.136.204.21 is >configured to ignore all traffic from the RFC 1918 space, and the ICMP >message goes out with a source address of 10.0.0.1 or 10.0.0.2, then >216.136.204.21 will never hear it. i don't understand what is 10.0.0.1 and 10.0.0.2. is it a tunnel outer address like configured with gifconfig(8), or tunnel inner address like configured with ifconfig(8)? i assume it as the latter. when an ICMP need fragmented message is emitted from 172.16.1.1 device to 216.136.204.21, normal routing decisions take place for source address determination. that is, 172.16.1.1 in the normal case. original packet: IP[216.136.204.21 -> 192.168.1.1] payload ICMP message: IP[172.16.1.1 -> 216.136.204.21] ICMP [original packet] successfully-encapsulated packet itojun To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010808204709.308547BA>