From hgoldste@news1.mpcs.com  Sun Sep 29 09:31:50 1996
Received: from bbs.mpcs.com (root@bbs.mpcs.com [204.215.226.2])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA27565
          for <FreeBSD-gnats-submit@freebsd.org>; Sun, 29 Sep 1996 09:31:48 -0700 (PDT)
Received: from news1.mpcs.com (news1.mpcs.com [204.215.226.8]) by bbs.mpcs.com (8.8.0/8.7.3/MPCS) with ESMTP id MAA29732 for <FreeBSD-gnats-submit@freebsd.org>; Sun, 29 Sep 1996 12:31:24 -0400
Received: (from hgoldste@localhost) by news1.mpcs.com (8.8.Beta.5/8.8.Beta.4/MPCS) id MAA06428; Sun, 29 Sep 1996 12:31:24 -0400 (EDT)
Message-Id: <199609291631.MAA06428@news1.mpcs.com>
Date: Sun, 29 Sep 1996 12:31:24 -0400 (EDT)
From: Howard Goldstein <hgoldste@news1.mpcs.com>
Reply-To: hg@penny.n2wx.ampr.org
To: FreeBSD-gnats-submit@freebsd.org
Subject: TCP extensions throttles throughput
X-Send-Pr-Version: 3.2

>Number:         1689
>Category:       kern
>Synopsis:       TCP extensions throttles distant connections
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    jesper
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sun Sep 29 09:40:02 PDT 1996
>Closed-Date:    Thu May 31 12:26:39 PDT 2001
>Last-Modified:  Thu May 31 12:27:01 PDT 2001
>Originator:     Howard Goldstein
>Release:        FreeBSD 2.1.5-RELEASE
>Organization:
MPC Systems 
>Environment:

	

net.inet.tcp.rfc1323=1

>Description:

TCP connections between distant (155-210ms ping times)  2.1.0R, 2.1.5R or both
systems, as for example two FreeBSD systems over a dialup CSLIP line,
with net.inet.tcp.rfc1323=1 (default) results in 3 to 4 sec delays between
transmission of subsequent TCP packets even though the dialup line is solid and
tcpdump reveals correct window ACKs making a timely return.

Topographically distant remote non-dialin http users also reported poor performance.


The severity and priority of this problem would have been ranked non-critical/low
had the shipped default for TCP extensions been set to off.

	

>How-To-Repeat:

Ensuring that both systems have rfc1323=1, telnet to the echo socket of
distant system.

Distant does not mean ethernet RTT times, the RTT must be in the range
given above. Observe the first echoed packet comes back quickly but all
subsequent packets suffer 3-5 second delay.

Other TCP sockets suffer from same delay but it is most easily observed
and tcpdumped with an echo socket.

Anyone who needs to reproduce this but can't find a system topologically
distant enough should email me or talk hg@cally.south.mpcs.com and
I'll turn on a telnettable system with the deadly extension enabled.



>Fix:


Work around is inclusion of net.inet.tcp.rfc1323=0 in the rc
for all FreeBSD systems.

(NB., there continue to be RTT computation time problems, problems discussed
in a series of postings by Matthew Dillon around 8 Mar 96 on the hackers
list but not GNATed, that have not been corrected)
>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: freebsd-bugs->wollman 
Responsible-Changed-By: jkh 
Responsible-Changed-When: Sun Dec 22 09:31:09 PST 1996 
Responsible-Changed-Why:  
This is actually one for Garrett. 

From: nobody@mpcs.com
To: freebsd-gnats-submit@freebsd.org
Cc:  Subject: Re: kern/1689 TCP extensions throttles distant connections
Date: Sun, 22 Mar 1998 17:40:30 -0500

 >Greetings, :)
 
 >	I am writing to you in regards to your FreeBSD Problem
 >Report. The FreeBSD project is currently conducting a beta test on
 >version 2.2.6 and feedback as to whether you are still experiencing
 >your problem would be very valuable. 
 
 >	If you are still experiencing the problem you reported, it
 >would help the project track the problem if you could upgrade to the
 >latest snapshot of 2.2.6-Beta (located at releng22.freebsd.org) and
 >test your problem again. 
 
 >	If you have any feedback regarding this Problem Report,
 >whether you are still experiencing the problem or whether the PR can
 >be closed, please mail your response to
 >freebsd-gnats-submit@freebsd.org. Please do not respond directly to
 >me. I am merely a humble volunteer and have no official connection to
 >the FreeBSD project. Therefore I cannot make any changes to the status
 >of your Problem Report. It is also very important that you include 
 >the category and number of your Problem Report (kern/1689)
 >in the subject line of your response.
 
 >	Another option if you need a refresher on the details of your
 >problem or would like to submit a followup is to use the web page
 >interface and look up your PR by number.
 >http://www.freebsd.org/cgi/query-pr-summary.cgi
 
 >	Thank you for helping to make this the greatest release of
 >FreeBSD ever.
 
 >Doug
 
 
 >-- 
 >***         Chief Operations Officer, DALnet IRC network       ***
 >*** Proud operator, designer and maintainer of the world's largest
 >*** Internet Relay Chat server.  5,328 clients and still growing.
 >*** Try spider.dal.net on ports 6662-4    (Powered by FreeBSD)
 
 **NOTICE**
 Hello, please do not panic.  The
 above message was received but my automated script cannot
 tell if this is unwanted junk mail.
 
 If you need or expect a rapid reply, please resend
 with words 'this is spam' somewhere in the Subject: header and I'll see it
 sooner.  I will see it even if you dont resent it so dont fret.
 
 
 P.S. Dont reply to this message without fixing the To header
 probably to hgoldste at bbs.mpcs.com
 
 hg
 -----
State-Changed-From-To: open->closed 
State-Changed-By: sheldonh 
State-Changed-When: Tue Dec 21 04:49:44 PST 1999 
State-Changed-Why:  
RFC1323 extensions have been turned off by default since  
FreeBSD 3.0-RELEASE. 

From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
To: sheldonh@FreeBSD.org
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Tue, 21 Dec 1999 13:04:49 -0500 (EST)

 <<On Tue, 21 Dec 1999 04:51:20 -0800 (PST), sheldonh@FreeBSD.org said:
 
 > RFC1323 extensions have been turned off by default since 
 > FreeBSD 3.0-RELEASE.
 
 This is a bug.
 
 -GAWollman
 
 --
 Garrett A. Wollman   | O Siem / We are all family / O Siem / We're all the same
 wollman@lcs.mit.edu  | O Siem / The fires of freedom 
 Opinions not those of| Dance in the burning flame
 MIT, LCS, CRS, or NSA|                     - Susan Aglukark and Chad Irschick
 

From: Sheldon Hearn <sheldonh@uunet.co.za>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections 
Date: Wed, 22 Dec 1999 11:52:58 +0200

 On Tue, 21 Dec 1999 13:04:49 EST, Garrett Wollman wrote:
 
 > > RFC1323 extensions have been turned off by default since 
 > > FreeBSD 3.0-RELEASE.
 > 
 > This is a bug.
 
 Que?  It's a bug that tcp_extensions is turned off by default?  Or the
 problems caused by the extensions are the manifestation of a bug?  If
 so, any idea who's looking into a fix?
 
 Ciao,
 Sheldon.
 

From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
To: Sheldon Hearn <sheldonh@uunet.co.za>
Cc: jlemon@freebsd.org, freebsd-gnats-submit@freebsd.org
Subject: Re: kern/1689: TCP extensions throttles distant connections 
Date: Wed, 22 Dec 1999 11:45:38 -0500 (EST)

 <<On Wed, 22 Dec 1999 11:52:58 +0200, Sheldon Hearn <sheldonh@uunet.co.za> said:
 
 > Que?  It's a bug that tcp_extensions is turned off by default?  
 
 Yes.  There should be some form of ``black hole detection'' which,
 after some number of SYNs, turns off TCP options and tries again.
 (I believe 1323 is still on by default in the kernel, but the default
 rc.conf turns it off again.  1323 is vital for TCP performance over
 fast wide-area networks.)
 
 -GAWollman
 
 --
 Garrett A. Wollman   | O Siem / We are all family / O Siem / We're all the same
 wollman@lcs.mit.edu  | O Siem / The fires of freedom 
 Opinions not those of| Dance in the burning flame
 MIT, LCS, CRS, or NSA|                     - Susan Aglukark and Chad Irschick
 
 --LAA97904.945881074/khavrinen.lcs.mit.edu--
 
State-Changed-From-To: closed->open 
State-Changed-By: sheldonh 
State-Changed-When: Wed Dec 22 08:51:39 PST 1999 
State-Changed-Why:  
wollman insisted on it, after explaining that this is a real 
bug. 


Responsible-Changed-From-To: wollman->freebsd-bugs 
Responsible-Changed-By: sheldonh 
Responsible-Changed-When: Wed Dec 22 08:51:39 PST 1999 
Responsible-Changed-Why:  
wollman isn't looking at it. :-) 
I'm going to ask luigi, fenner and des to take a look. 

From: Sheldon Hearn <sheldonh@uunet.co.za>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections 
Date: Wed, 22 Dec 1999 18:57:44 +0200

 On Wed, 22 Dec 1999 11:44:18 EST, Garrett Wollman wrote:
 
 > Yes.  There should be some form of ``black hole detection'' which,
 > after some number of SYNs, turns off TCP options and tries again.
 > (I believe 1323 is still on by default in the kernel, but the default
 > rc.conf turns it off again.  1323 is vital for TCP performance over
 > fast wide-area networks.)
 
 Okay, I've reopened the PR.
 
 Thanks,
 Sheldon.
 

From: Jonathan Lemon <jlemon@americantv.com>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: Sheldon Hearn <sheldonh@uunet.co.za>,
	freebsd-gnats-submit@freebsd.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Wed, 22 Dec 1999 17:30:59 -0600

 I haven't been able to reproduce this here.  Between two -current boxes:
 
     net.inet.tcp.rfc1323 = 1 	  on both machines,
     net.inet.tcp.delayed_ack = 1  on both machines,
 
 On one machine:
 
     ipfw pipe 1 config bw 128kbit/s delay 200ms
     ipfw add pipe 1 ip from hostA to hostB out
     ipfw add pipe 1 ip from hostB to hostA in
 
 This gives:
 
  recvpipe  sendpipe  ssthresh  rtt,msec    rttvar  hopcount      mtu     expire
        0         0         0       400         3         0      1500      1012
 
 E.g.: a 400ms round trip time.
 
 I don't see any abnormal delays in connections between the two
 machines.  Is there something else I'm missing here, or is the 
 problem gone in -current?
 --
 Jonathan
 

From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
To: Jonathan Lemon <jlemon@americantv.com>
Cc: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>,
	Sheldon Hearn <sheldonh@uunet.co.za>,
	freebsd-gnats-submit@freebsd.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Wed, 22 Dec 1999 20:05:10 -0500 (EST)

 <<On Wed, 22 Dec 1999 17:30:59 -0600, Jonathan Lemon <jlemon@americantv.com> said:
 
 > I don't see any abnormal delays in connections between the two
 > machines.  Is there something else I'm missing here, or is the 
 > problem gone in -current?
 
 There are two things you're missing (but neither is your fault):
 
 1) Some broken terminal servers (most of which have hopefully been
 retired) have bad VJ header compression code which trashes TCP
 segments containing unknown-to-them TCP options.
 
 2) RFC1323 timestamps effectively disable VJ header compression.
 Arguably, this is not a real-life problem, but for the lack of a
 dynamic mechanism to control use of RFC 1323 (on a per-destination
 basis).
 
 (1), and not this PR, is the actual reason why RFC 1323 was turned off
 by default.  I'm no longer concerned about it myself and have turned
 it back on on all of my machines.
 
 -GAWollman
 
 --
 Garrett A. Wollman   | O Siem / We are all family / O Siem / We're all the same
 wollman@lcs.mit.edu  | O Siem / The fires of freedom 
 Opinions not those of| Dance in the burning flame
 MIT, LCS, CRS, or NSA|                     - Susan Aglukark and Chad Irschick
 

From: Sheldon Hearn <sheldonh@uunet.co.za>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: Jonathan Lemon <jlemon@americantv.com>,
	freebsd-gnats-submit@freebsd.org
Subject: Re: kern/1689: TCP extensions throttles distant connections 
Date: Thu, 23 Dec 1999 11:14:15 +0200

 On Wed, 22 Dec 1999 20:05:10 EST, Garrett Wollman wrote:
 
 > I'm no longer concerned about it myself and have turned
 > it back on on all of my machines.
 
 But you still want this PR open?  I'm getting confused. :-)
 
 Ciao,
 Sheldon.
 
Responsible-Changed-From-To: freebsd-bugs->jesper 
Responsible-Changed-By: jesper 
Responsible-Changed-When: Sun May 27 16:06:24 PDT 2001 
Responsible-Changed-Why:  
Put it on the radar screen while waiting for feedback from sheldonh, jlemon and wollman 

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=1689 

From: Jesper Skriver <jesper@FreeBSD.org>
To: Sheldon Hearn <sheldonh@FreeBSD.org>,
	Garrett Wollman <wollman@FreeBSD.org>,
	Jonathan Lemon <jlemon@FreeBSD.org>, freebsd-gnats-submit@freebsd.org
Cc:  
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Mon, 28 May 2001 01:08:04 +0200

 Hi,
 
 I too have been trying to reproduce this problem in -current, and
 cannot.
 
 I fail to understand Garrett's comments, but I'd like to know if there
 is any objections to closing this PR ?
 
 /Jesper
 
 -- 
 Jesper Skriver, jesper(at)skriver(dot)dk  -  CCIE #5456
 Work:    Network manager   @ AS3292 (Tele Danmark DataNetworks)
 Private: FreeBSD committer @ AS2109 (A much smaller network ;-)
 
 One Unix to rule them all, One Resolver to find them,
 One IP to bring them all and in the zone to bind them.

From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
To: Jesper Skriver <jesper@FreeBSD.org>
Cc: Sheldon Hearn <sheldonh@FreeBSD.org>,
	Jonathan Lemon <jlemon@FreeBSD.org>, freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Tue, 29 May 2001 18:19:48 -0400 (EDT)

 <<On Mon, 28 May 2001 01:08:04 +0200, Jesper Skriver <jesper@FreeBSD.org> said:
 
 > I too have been trying to reproduce this problem in -current, and
 > cannot.
 
 You can't reproduce the problem because you don't have one of the old
 terminal servers which chokes.  Like I said, I hope most if not all of
 them are out of service by now.
 
 I still think it's a bug that we can't deal with the chokage, but
 finding a work-around is getting less and less important all the time.
 
 -GAWollman
 

From: Jesper Skriver <jesper@skriver.dk>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: Sheldon Hearn <sheldonh@FreeBSD.org>,
	Jonathan Lemon <jlemon@FreeBSD.org>, freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Wed, 30 May 2001 10:35:33 +0200

 On Tue, May 29, 2001 at 06:19:48PM -0400, Garrett Wollman wrote:
 > <<On Mon, 28 May 2001 01:08:04 +0200, Jesper Skriver <jesper@FreeBSD.org> said:
 > 
 > > I too have been trying to reproduce this problem in -current, and
 > > cannot.
 > 
 > You can't reproduce the problem because you don't have one of the old
 > terminal servers which chokes.  Like I said, I hope most if not all of
 > them are out of service by now.
 >
 > I still think it's a bug that we can't deal with the chokage, but
 > finding a work-around is getting less and less important all the time.
 
 The work-around you describe about disabling rfc1323 extensions
 (and probably rfc1644 too) when multiple SYN's are recieved, would
 it be sufficient to disable it when the second SYN is recieved ?
 
 If so, I have a patch that I need to test a bit more, I'll send it
 to you for review later.
 
 /Jesper
 
 -- 
 Jesper Skriver, jesper(at)skriver(dot)dk  -  CCIE #5456
 Work:    Network manager   @ AS3292 (Tele Danmark DataNetworks)
 Private: FreeBSD committer @ AS2109 (A much smaller network ;-)
 
 One Unix to rule them all, One Resolver to find them,
 One IP to bring them all and in the zone to bind them.

From: Jesper Skriver <jesper@skriver.dk>
To: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
Cc: Sheldon Hearn <sheldonh@FreeBSD.org>,
	Jonathan Lemon <jlemon@FreeBSD.org>, freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/1689: TCP extensions throttles distant connections
Date: Thu, 31 May 2001 20:07:03 +0200

 On Wed, May 30, 2001 at 10:35:32AM +0200, Jesper Skriver wrote:
 > On Tue, May 29, 2001 at 06:19:48PM -0400, Garrett Wollman wrote:
 > > <<On Mon, 28 May 2001 01:08:04 +0200, Jesper Skriver <jesper@FreeBSD.org> said:
 > > 
 > > > I too have been trying to reproduce this problem in -current, and
 > > > cannot.
 > > 
 > > You can't reproduce the problem because you don't have one of the old
 > > terminal servers which chokes.  Like I said, I hope most if not all of
 > > them are out of service by now.
 > >
 > > I still think it's a bug that we can't deal with the chokage, but
 > > finding a work-around is getting less and less important all the time.
 > 
 > I have a patch that I need to test a bit more, I'll send it
 > to you for review later.
 
 Please review http://people.FreeBSD.org/~jesper/pr1689.diff
 
 Which will disable rfc1323 and rfc1644, if we havn't got a response
 to our third SYN.
 
 /Jesper
 
 -- 
 Jesper Skriver, jesper(at)skriver(dot)dk  -  CCIE #5456
 Work:    Network manager   @ AS3292 (Tele Danmark DataNetworks)
 Private: FreeBSD committer @ AS2109 (A much smaller network ;-)
 
 One Unix to rule them all, One Resolver to find them,
 One IP to bring them all and in the zone to bind them.
State-Changed-From-To: open->closed 
State-Changed-By: jesper 
State-Changed-When: Thu May 31 12:26:39 PDT 2001 
State-Changed-Why:  
Fix committed, will MFC in about 2 weeks. 

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=1689 
>Unformatted:
