[Sis-SCPS-INTEREST] Latest Version of the SCPS RI

[Feighery, Patrick D.]

Absolutely, if you forward your TUN/TAP changes for FreeBSD/NetBSD to
me I would be happy to add them to the SCPS RI.  My first kernel
programming was on a BSD 4.2 based OS on a DEC Microvaxen, so I have
too have an affinity to BSD based systems.  Many thanks for the
changes.

	Pat

>>-----Original Message-----
>>From: Marcin Jessa [mailto:yazzy at yazzy.org] 
>>Sent: Friday, September 23, 2005 2:50 PM
>>To: Feighery, Patrick D.
>>Cc: sis-scps-interest at mailman.ccsds.org
>>Subject: Re: [Sis-SCPS-INTEREST] Latest Version of the SCPS RI
>>
>>Hi Pat.
>>
>>Could you please send me the latest release?
>>I have also managed to port SCPS to both FreeBSD and NetBSD 
>>and run it with tun and tap devices.
>>Both FreeBSD and NetBSD(from 3.0) have tap and tun in the base system

>>so it was a child play to make both work after I patched the 
>>SCPS sources.
>>No need for 3.rd party patches like it's necessarly dealing 
>>with Linux :)
>>I can send you all my patches if you are interested.
>>Would be great to have them included in the official source.
>>To be honest it really surprises me BSD's are so forgotten 
>>being older and more mature compared to Linux...
>>
>>Cheers,
>>Marcin.
>>
>>
>>On Fri, 23 Sep 2005 12:53:42 -0400
>>"Feighery, Patrick D." <feighery at mitre.org> wrote:
>>
>>> For those of you that have a copy of the SCPS RI, please read on.
>>> 
>>> During the last couple of days, we have found and corrected a bug
in
>>> the SCPS RI gateway which you should be aware of.  Under certain
>>> conditions it will stop traffic from traversing the gateway.
>>> 
>>> Here is a quick synopsis
>>> 
>>> The problem only occurs when you are running the gateway under
linux
>>> via the TAP interface and you configure the gateway to not 
>>pass non-IP
>>> based traffic (e.g., IPX, appletalk, etc.)  In our case the problem
>>> reared its head when the gateway received CISCO's Cisco Discovery
>>> Protocol (CDP) which operates directly over ethernet. If the
gateway
>>> receives a packet/frame that is not IP based and the gateway is
>>> configured to not pass it we simply drop it on the floor.  So far
so
>>> good.  Unfortunately we do not free up the memory by 
>>placing it back in
>>> the memory pool.  Bad programmer.  In our case we see a CDP packet
>>> every 30 seconds or so.  So after about 8 hours or so we 
>>have no more
>>> buffers to receive the packets.  Thus the packets will 
>>starting filling
>>> up the receive queue for the tap interface. After the tap 
>>interface is
>>> full, the ethernet frames can no longer be bridged.  A 
>>simple two line
>>> fix corrects this problem.  For those of you that may be 
>>unaware, the
>>> gateway does have the ability to either allow or deny non-IP based
>>> traffic from traversing the RF side and thus consuming its 
>>resources.
>>>  
>>> The newest release of the SCPS RI is version 1.1.119.
>>> 
>>> Besides the change listed above, following are some even shorter
>>> synopses of some changes from the 1.1.113 release - which 
>>we have been
>>> sending out for about a year now.
>>> 
>>>   o  Updated some of the existing documentation which was pointed
>>>      out by some of you to be well.. Antiquated to put it nicely.
>>> 
>>>   o  Added some additional documents
>>> 	o  The SCPS gateway user's guide - which I have been including
>>> 	   in recent requests.
>>>       o  A Frequently Asked Questions (FAQ) about the SCPS gateway.
>>>          If there are additional questions/answers please send them
>>>          to me and I will include them in future versions of this
>>>          document.
>>>       o  Some of you have brought up performance issues with SCPS
>>>          and Microsoft's file sharing.  This mini-white paper
>>> 	   discusses the interaction issues with Microsoft's
>>>          Server Message Block (SMB) protocol used to implement
>>> 	   their File Sharing over the net, how it performs over
>>>          large bandwidth*delay paths, and its interaction 
>>with SCPS. 
>>> 
>>>   O  Fixed some compilation issues with newer versions of gcc.
>>>      Unfortunately, I am an old K&R C programmer and well..
ensuring
>>>      correct typing of variables has never been a strong point.
>>>      Thanks to those who have pointed this out and sent in diffs.
>>> 
>>>   O  Fixed minor rate control bug when emitting non TCP based IP
>>>      traffic.
>>> 
>>>   O  Corrected some bookkeeping issues concerning lower level
buffer
>>>      utilization and added some stats which can be log to
>>>      /var/log/messages. This is how I tracked down the bug
mentioned
>>>      in the first part of the e-mail.
>>>  
>>>   O  Changed the default behavior to keep the transport layer
source
>>>      port constant as the gateway creates the two peer transport
>>>      layer connections.  I originally kept the transport 
>>layer source
>>>      ports different to assist in the debugging of various things.
>>>      However it was brought to my attention that some testing tools
>>>      assume the source port received by the test sink apps must be
>>>      the same as the source port sent the test source app.  To
>>>      revert to the previous behaving you need to
>>>      #define CHANGE_SRC_PORT.  This may be significant to some...
>>> 
>>> If you already have a copy of the SCPS RI, send me an 
>>e-mail and I will
>>> be happy to reply with the latest release.
>>> 
>>> Best Regards
>>> 
>>> 	Pat
>>> 
>>> _______________________________________________
>>> Sis-SCPS-INTEREST mailing list
>>> Sis-SCPS-INTEREST at mailman.ccsds.org
>>> http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-scps-interest
>>> 
>>