The packets that are generated before the link is up and the ip addresses
are negotiated are doomed.  There is no point in buffering them.  Any
subsequent packets that are generated have the source ip rewritten.

Buffering works for static ip-addresses, if your not dialing into a Cisco
product, which throws your packets away for the first couple seconds after
PPP has completed its mating ritual.

----- Original Message -----
From: Stan Brown <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Saturday, September 04, 1999 8:15 PM
Subject: Why does ip_dynaddr _not_ work, if I am running named?


| Thanks to help from many people on this list I have finally managed to
| arive at a workaround for the first connection hanging, when using
| diald.
|
| However I cleary don not understand what's going on here.
|
| If I set /proc/sys/net/ipv4/ip_dynaddr to 1, and put -buffer-packets in
| the diald.conf file, then my understanding of what happens is thus.
|
| A packet goes to diald, it imedialty sends it on to the kernel. The
| kernel recognizing that the source address is not valid yet, waits for
| the link to come up and rewrites the source address to the dynamicly
| assigned one.
|
| If however it works this way, then it should not matter of the packet
| is a DNS query, a ping, or any other TCP/IP packet, right, but yet if I
| run a loacal cacheing named, and if the destination that I am sending
| to is in the chache, then the conection will hanf forever.
|
| Why is this? And how can it be fixed, either long term as a design
| iissue, or short term as a workaround?
|
| I have been suing FreeBSD machines with userland ppp, doing
| masqureading for over 4 years nowm and had never seen this type of
| situation.
|
| Thanks for any insight on this.
|
| --
| Stan Brown     [EMAIL PROTECTED]
843-745-3154
| Westvaco
| Charleston SC.
| --
| Windows 98: n.
| useless extension to a minor patch release for 32-bit extensions and
| a graphical shell for a 16-bit patch to an 8-bit operating system
| originally coded for a 4-bit microprocessor, written by a 2-bit
| company that can't stand for 1 bit of competition.
| -
| (c) 1999 Stan Brown.  Redistribution via the Microsoft Network is
prohibited.
|

Reply via email to