| |
PA2AGA > TCPDIG 31.08.96 04:05l 188 Lines 6700 Bytes #-10902 (0) @ EU
BID : TCP_96_175B
Read: DG2CV GUEST
Subj: TCP-Group Digest 96/175B
Path: DB0AAB<DB0KCP<DB0ULM<DB0LX<DB0RBS<DB0SEL<DB0ZDF<DB0SRS<DB0SIF<DB0GV<
DB0EAM<DK0MWX<DB0IZ<DB0RWI<DB0END<DB0ACC<PI8DRS<PI8DAZ<PI8GCB<PI8MBQ<
PI8VNW
Sent: 960830/2233Z @:PI8VNW.#ZH2.NLD.EU #:58551 [Hoek v Holland] FBB5.15c
From: PA2AGA@PI8VNW.#ZH2.NLD.EU
To : TCPDIG@EU
Received: from pa2aga by pi1hvh with SMTP
id AA16538 ; Fri, 30 Aug 96 21:50:59 UTC
Received: from pa2aga by pa2aga (NET/Mac 2.3.61/7.1) with SMTP
id AA00003722 ; Fri, 30 Aug 96 22:49:30 MET
Received: from pa2aga-1 by pa2aga with SMTP
id AA00003688 ; Fri, 30 Aug 96 22:44:30 MET
Received: from pa2aga-1 by pa2aga-1 (NET/Mac 2.3.61/7.5.4) with SMTP
id AA00008004 ; Fri, 30 Aug 96 22:44:29 MET
Date: Fri, 30 Aug 96 22:38:37 MET
Message-Id: <tcp_96_175B>
From: pa2aga
To: tcp_broadcast@pa2aga-1
Subject: TCP-Group Digest 96/175B
X-BBS-Msg-Type: B
actual AX.25 layer is sufficiently different it doesnt quite come off
(as in been there, done that, looked an idiot ;))
> AX.25 connections will be managed autonomously by the AX.25 driver, and
> the AX.25 driver will generate virtual connection requests based on
> the PID of incoming I frames.
That prevents a user process that knows the protocol will be the one it
desires from issuing an initial greeting.
> Any thoughts? Anyone else interested in STREAMS? They're cool, you
> know.
Mutter..performance..mutter..latency..mutter... No Im not convinced,
even the Sun stuff which is much faster than normal streams. One thing
however - there is a kind of defacto unix socket api standard for AX.25
now, it would be nice to support that as well. In theory you can then
get stuff like the node software for free and might even get the FBB
folks to compile it under Solaris as well as Linux.
Alan
------------------------------
Date: Wed, 28 Aug 1996 08:42:56 -0400 (EDT)
From: "Brian A. Lantz" <brian@lantz.com>
Subject: AX.25 - that darn PID
On Wed, 28 Aug 1996, Carl Makin wrote:
> It's how the "jumpstart" problem was solved in jnos.
Correction! Solved in TNOS, copied in JNOS ;-)
-----------------------------------------------------------
Brian A. Lantz http://www.lantz.com brian@lantz.com
REAL PORTION of Microsoft Windows code:
while (memory_available) {
eat_major_portion_of_memory (no_real_reason);
if (feel_like_it)
make_user_THINK (this_is_an_OS);
gates_bank_balance++;
}
------------------------------
Date: Wed, 28 Aug 1996 10:16:53 -0700 (PDT)
From: jmorriso@bogomips.com (John Paul Morrison)
Subject: AX.25 - that darn PID
I'm interested in the AX.25 streams implementation. I'm trying
to work up the courage to do a Windows 95 or Windows NT port to
run 56kbps packet (hey, I was cornered in a dark parking lot by a
couple of ornery hams!)
Windows NT has a streams driver you can load. I don't know anything
about it, but maybe it can be used to port existing streams modules.
---------------------------------------------------------------------------
BogoMIPS Research Labs -- bogosity research & simulation -- VE7JPM --
jmorriso@bogomips.com ve7jpm@ve7jpm.ampr.org jmorriso@ve7ubc.ampr.org
---------------------------------------------------------------------------
------------------------------
Date: Wed, 28 Aug 1996 21:00:15 -0500 (CDT)
From: David Kelly <dkelly@hiwaay.net>
Subject: AX.25 - that darn PID
On 20:27:06 Carl Makin wrote:
>>
>On Tue, 27 Aug 1996, Dana H. Myers wrote:
>
>> So, I've prototyped a DLPI AX.25 driver (under SunOS 5.x) and it works as w
ell
>> as I expected. I'm currently only using datagram mode for IP frames, and
>> don't support AX.25 connections at all.
>
>Hmmm.. I run FreeBSD. I wonder how portable your implementation is?
FreeBSD doesn't have streams for much the same sentiments as Alan voiced,
namely "lack of respect for performance."
--
David Kelly N4HHE, dkelly@tomcat1.tbe.com (wk), dkelly@hiwaay.net (hm)
=====================================================================
The human mind ordinarily operates at only ten percent of its
capacity -- the rest is overhead for the operating system.
------------------------------
Date: Wed, 28 Aug 1996 08:32:17 -0700 (PDT)
From: "Dana H. Myers" <myers@bigboy.West.Sun.COM>
Subject: AX.25 - that darn PID
> > Ether type in the PID octet. The problem with this is that the PID
> > octet is only present in I frames; when a connection is established,
there's
> > no way to know what SAP the connection is going to end up using.
>
> Yes. We have this problem with the Linux AX.25, at the moment an incoming
> netrom connect will cause an AX.25 connect which will wait in the user
> process not receiving the netrom frames then time out. Its a pig because
> you can't send the user an initial greeting if you don't know what
> protocol they are.
This is the problem that 'jumpstart' tries to answer in JNOS.
> > The reason I care about all this is I'd like to complete the project
> > I envisioned - a STREAMS implementation of the AX.25 protocols (LAPB,
> > net/rom, etc.) using standard interfaces (DLPI).
>
> Don't use a standard LAPB streams module for the AX.25 layer - the
> actual AX.25 layer is sufficiently different it doesnt quite come off
> (as in been there, done that, looked an idiot ;))
Of course, I don't plan to use a standard LAPB layer from the outset.
A look at the AX.25 spec pretty clearly indicates that a custom built
protocol driver is a good idea.
> > AX.25 connections will be managed autonomously by the AX.25 driver, and
> > the AX.25 driver will generate virtual connection requests based on
> > the PID of incoming I frames.
>
> That prevents a user process that knows the protocol will be the one it
> desires from issuing an initial greeting.
Yup - I'll have to implement some kind of jumpstart mechanism.
> > Any thoughts? Anyone else interested in STREAMS? They're cool, you
> > know.
>
> Mutter..performance..mutter..latency..mutter... No Im not convinced,
> even the Sun stuff which is much faster than normal streams. One thing
Where exactly is the performance problem in streams? How exactly
does streams cause significant latency? (hint: it depends on how a
streams module operates). People grumble about the overhead of
streams but almost never explain where the overhead is.
> however - there is a kind of defacto unix socket api standard for AX.25
> now, it would be nice to support that as well. In theory you can then
> get stuff like the node software for free and might even get the FBB
> folks to compile it under Solaris as well as Linux.
What exactly is this node software? FBB? I've never really looked
at it.
Dana KK6JQ
Dana@Source.Net
------------------------------
Date: Wed, 28 Aug 1996 19:55:06 -0700 (PDT)
From: "Dana H. Myers" <myers@bigboy.West.Sun.COM>
Subject: AX.25 - that darn PID (Fwd)
[forgot to cc the group]
To be continued in digest: tcp_96_175C
Read previous mail | Read next mail
| |
