From owner-freebsd-bugs Sat Jun 17 04:49:33 1995 Return-Path: bugs-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id EAA21809 for bugs-outgoing; Sat, 17 Jun 1995 04:49:33 -0700 Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id EAA21802 ; Sat, 17 Jun 1995 04:49:19 -0700 Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id VAA25496; Sat, 17 Jun 1995 21:47:20 +1000 Date: Sat, 17 Jun 1995 21:47:20 +1000 From: Bruce Evans Message-Id: <199506171147.VAA25496@godzilla.zeta.org.au> To: davidg@freefall.cdrom.com, freebsd-bugs@freefall.cdrom.com, gibbs@estienne.cs.berkeley.edu Subject: Re: Changed information for PR misc/292 Sender: bugs-owner@FreeBSD.org Precedence: bulk >Synopsis: rfc1323 and rfc1644 support can confuse terminal servers w/SLIP >State-Changed-From-To: open-closed >State-Changed-By: davidg >State-Changed-When: Fri Jun 16 22:58:46 PDT 1995 >State-Changed-Why: >As mentioned in the PR, this isn't a bug in FreeBSD and the decision on >this is that TCP options negotiation should remain enabled. It does seem rather wasteful for slip. The overhead is about 4% for the default mtu. At 115200 bps, as measured by ttcp -r for 80 packets of size 8K: protocol mtu throughput -------------- ---- ----------- cslip 552 9.98 K/sec cslip/no rfc 552 10.39 cslip 1500 10.76 cslip/no rfc 1500 10.92 For catting and `rz -q'ing /kernel, and for larger `ttcp -r's. -------- overheads (%) ------- 86DX2/66-16550 486DX/33-16450 --------------- -------------- protocol mtu throughput r w r w -------------- ---- ----------- ---- ---- ---- ---- cat - 11.25 6.7 2.9 20.3 15.6 cslip 1500 10.76 6.7 3.6 21.5 16.7 pppd 1500 10.74 9.1 4.3 26.2 18.2 ppp (no pred1) 1500 10.80 11.3 6.6 28.6 21.7 zmodem - 10.73 11.4 5.8 27.6 19.2 The overheads were measured by counting to 1 or 2 billion in the background (this takes about 150 seconds on an idle machine) and measuring the relative slowdown. Bruce