*BSD News Article 7611


Return to BSD News archive

Newsgroups: comp.unix.bsd,ics.general
Path: sserve!manuel.anu.edu.au!munnari.oz.au!network.ucsd.edu!usc!zaphod.mps.ohio-state.edu!cs.utexas.edu!newsfeed.rice.edu!rice!news.Rice.edu!rich
From: rich@Rice.edu (Richard Murphey)
Subject: Re: [386bsd]: SLIP woes (packetsize > 876)
In-Reply-To: bvickers@valentine.ics.uci.edu's message of 11 Nov 92 01:00:42 GMT
Message-ID: <RICH.92Nov10220937@omicron.Rice.edu>
Sender: news@rice.edu (News)
Reply-To: Rich@rice.edu
Organization: Department of Electrical and Computer Engineering, Rice
	University
References: <2B005B3A.17158@ics.uci.edu>
Date: Wed, 11 Nov 1992 04:09:37 GMT
Lines: 23

In article <2B005B3A.17158@ics.uci.edu> bvickers@valentine.ics.uci.edu (Brett J. Vickers) writes:
   I just got SLIP up and running on my computer.  However, whenever I
   try to send a packet of size greater than or equal to 877 bytes,
   nothing happens.  The process trying to send such a packet just sits
   there and hangs.  I found this out using "ping -s869 wherever" (ping
   adds 8 bytes to the packet).

   I'm using the standard 386BSD if_sl.[ch] files.  Anyone know why
   this is happening?


Look at /sys/net/if_sl.c.  The output packet size is limited to 296
(the value of SLMTU).

The comments there contain a careful discussion on the trade offs
between fractional overhead of ip headers at lower values v.s. latency
in interactive TCP connections at higher values.

Maybe ICMP packets are not fragmented.  Rich

   --
   Brett J. Vickers
   bvickers@ics.uci.edu