Re: RF transmission problem

[Tomi Manninen OH2BNS]

On Thu, 10 Aug 2000, Steve Stuart wrote:

> Yes, I did edit ax25d.conf (smile). You can be confident that I'm not
> bootlegging your callsign. ;-)
> The radio baud rate is 9600, and I'm pretty sure the serial baudrate is as
> well. This is easy to check via minicom.
> How do I reset kiss mode without rebooting my pc? Entered "kissparms -x -p
> ax0", but minicom says /dev/ttyS0 is locked?

Kissparms -x takes the TNC out of KISS mode but it's kissattach that locks
the tty. You need to kill it. Probably easiest with "killall kissattach".

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: RF transmission problem

[Tomi Manninen OH2BNS]

On Thu, 10 Aug 2000, Steve Stuart wrote:

> I've just installed the ax-25 stuff,
>node-0.3.0-1.i386.rpm,
>   libax25-0.0.7-1.i386.rpm
>  ax25-tools-0.0.5-1.i386.rpm
> ax25-apps-0.0.4-1.i386.rpm
> on RedHat 6.2 system (2.2.14-5.0) with a pentium 133 and 80M
> ram. Recompiled/installed the kernal (and modules) with ham
> radio stuff enabled. Then set up /etc/ax25/axports with the
> line: "ax0 W8AN-4 9600 255 2 144.050 MHz (9600 baud)". Modprobe

Wild guess: your TNC serial line bit rate is not 9600 but something else?
That 9600 at the speed field of axports is used for the serial bit rate,
not the radio bit rate.

> mkiss successful.
> Next "kissattach /dev/ttyS0 ax0 44.70.161.132" followed by
> "/etc/rc.d/init.d/ax25d start". Even setup a couple routes to

Just to be on the safe side: Ax25d is not needed for outgoing connections
or tcp/ip over ax25. And if you do want incoming connections, you did edit
/etc/ax25/ax25d.conf right? Otherwise you'll be using my callsign... :)

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: GW-List - Mkiss and Polled BPQ

[Tomi Manninen OH2BNS]

On Tue, 8 Aug 2000, Albert D. Lawson wrote:

> The mkiss docs say that to use it with TNC's that have the polled kiss
> proms by G8BPQ that you use the -c and -p arguements.  I have three
> TNC's on one serial port with BPQ polled kiss proms, the addresses are
> C, D, and E.  So far, only one TNC is being polled, and it's not on
> the "port" it was assigned to be on. Are there any additonal
> arguements to pass the addresses of each prom onto mkiss so it knows
> what TNC's are there...???

No. Mkiss starts the addressing from TNC 0 and goes up from there
depending on how many pty interfaces you define on the command line. I
don't remember how the BPQKISS addressing goes but if it starts from A
then you could try and define two dummy ports and then the three "real"
ones. If the address is a hexadecimal number then I'd try to find an EPROM
programmer...

> Has any one made this work before...

Run mkiss and BPQKISS with polling and checksum enabled? Yes, that is what
I wrote the support for.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: aiee, killing interrupt handler?

[Tomi Manninen OH2BNS]

On Thu, 3 Aug 2000, Stephan Loges wrote:


> > > I don't know whether it's an 2.0.x issue, but I remember a ax25-bug where
> > > the kernel oopses when you ifconfig'ed down your ax25 network interface
> > > (or was it axctl ...-kill ??) while there was an ax25 connection open.
> > > Reason was a forgotten timer that wasn't stopped when it should.
> > 
> > That was a 2.2.x bug and fixed in 2.2.1x somewhere. The exact cause was a
> > NULL pointer dereference at the time the heartbeat timer expired the last
> > time (see ax25_timer.
> 
> kernel is 2.0.38. Was the bug there too?

As far as i can remember, no.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: EPP adapter woes

[Tomi Manninen OH2BNS]

On Wed, 2 Aug 2000, Jens David wrote:

> > Once the
> > interface is up, packets never come out the h/w even
> > though listen -a shows them going.
> 
> This is not so good. Check that in BIOS setup EPP version 1.9
> is selected and/or check for a BIOS upgrade. In addition to that
> (this might sound a little but funny) change the cabling. The
> drivers have different timings (this is quiet critical).

I have no experience with DOS but under Linux I too have noticed that the
cabling is really critical. For example even a rather short length of flat
cable can totally ruin the communication. This seems to make it a bit
difficult with the older AT motherboards where the D-connector is not
directly on the board...

BTW I'm talking about the EPPFLEX hardware. Don't know if this is
applicable to the older EPP adapter.
-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: aiee, killing interrupt handler?

[Tomi Manninen OH2BNS]

On Sun, 30 Jul 2000, Thorsten Kranzkowski wrote:

> On Sun, Jul 30, 2000 at 03:13:59PM +, Stephan Loges wrote:
> > On Sun, 30 Jul 2000, Richard Adams wrote:
> > > I must say the trace does ring some bells with me, i seem to remember more
> > > about a ax25_timer problem, as to what it was or a answer to HOWTO solve
> > > it i am at a loss at the minute.
> > Maybe, would be good to have more information on it.
> 
> I don't know whether it's an 2.0.x issue, but I remember a ax25-bug where
> the kernel oopses when you ifconfig'ed down your ax25 network interface
> (or was it axctl ...-kill ??) while there was an ax25 connection open.
> Reason was a forgotten timer that wasn't stopped when it should.

That was a 2.2.x bug and fixed in 2.2.1x somewhere. The exact cause was a
NULL pointer dereference at the time the heartbeat timer expired the last
time (see ax25_timer.c).

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: KPPP problems

[Tomi Manninen OH2BNS]

On Fri, 28 Jul 2000 [EMAIL PROTECTED] wrote:

> And once upon a time ham operators were helpful people
> Could I refer you some of the Linux Hams to the Amateur Code as 
> published in the ARRL handbook for many years.
> The Radio Amateur is 
> CONSIDERATE
> LOYAL
> PROGRESSIVE
> FRIENDLY
> BALANCED
> PARRIOTIC.
> 
> A couple of replies to Marks VK3JMA plea for help certainly did not 
> meet the critera needed to be an Amateur.

Ok. Maybe the correct answer to the original plea would then be: "Mark,
this list is not meant for such questions. Please consider subscribing to
linux-newbies and asking there. Or perhaps Red Hat might be able to answer
your question; there should be pointers to the correct place to ask
somewhere in your distributions documentation." How does that sound?

> I am like Mark becoming involved with Linux simply to apply the 
> new skill to my entertainment in amateur radio. This is the only 
> group to which I subscribe and had assumed if I had a Linux 
> question my fellow amateurs may be able to assist.

Obviously you assumed wrong.

Perhaps you are also a newbie in Internet? The purpose of this mailing
list was sent to you when you subscribed. After a few years of Internet
experience you will understand why it is rather important to stick to the
meant purpose of a mailing list.

> Another point I get the feeling you guru's just what your private little 
> club, well consider the demise of the RF side of the hobby primarily 
> because of such attitudes. How many amateur controlling bodies 
> are in a growth mode.??

No clubs here. It's a question of simple practicalities. I receive ~100
emails a day from several mailing lists. Some get more. If the signal to
noise ratio drops too low on a list, I simply have to unsubscribe. Maybe I
wouldn't be such a loss to this list but it would be yet another one,
possibly having answers to relevant questions, off the list. A few
off-topic posting every now and then doesn't really hurt, but they should
be firmly discouraged.

Believe me, a lot of the real guru's have already unsubscribed. And not
because they feel they're somehow superior to you. It's because of
practicalities.

> Let remember the Amateur code maybe  even dust of the old 
> handbook and reread it might make you all think to engage brain 
> before tapping the fingers.

I remember the time when the Internet used to be a place for educated and
conciderate people. The code used to be "listen before transmitting". The
so called netiquette (which almost was required reading before getting
access to the Net) was truly understood and obeyed. Everybody would
carefully observe if a given mailing list or news group was really the one
to ask his or her question. Oh, those were the days.

Sounds familiar? It should. The only difference is that here you can not
transmit over someone else. But you can transmit in a completely wrong
place. The consequences are just as "bad".

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: NEWQPSK tests

[Tomi Manninen OH2BNS]

On Sun, 23 Jul 2000, Hamish Moffatt wrote:

> Well, I was going to arrange a time and frequency where we could
> try to make contact. If nothing else I might leave a beacon running
> one day this week, if there's anyone hear who would try to copy it.
> 
> I would be running Tomi's NEWQPSK modem for Linux, transmitting
> AX.25 packets using the 'beacon' program. Would anyone try to decode
> this if I set it up?

Unfortunately I'm now back from my vacation and HF antennas are a problem
but I might try to set up something.

By the way if anyone here in EU is trying to connect to Joni's NEWQPSK
station at 3590.0kHz LSB, it's now running 2500bps (8000sps),
interleave=8 and fec=3. After some tests we decided that the 3000bps
(9600sps) mode is too wide for most ham rigs and the high and low
carriers are unnecessarily attenuated and performance thus degraded.
2500bps is much better at least in that sense.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: KISS over psuedo TTY

[Tomi Manninen OH2BNS]

On Sun, 23 Jul 2000, Charles Brain wrote:

> > What I am getting in my own task at the slave tty is sort of KISS data in that
> 
> What I am actually receiving is 0x00 0xCC 0x07 everytime I ping.

Strange. And somewhat hard to figure out without more info about what you
are doing. If you want you can send me (off list) your config scripts and
maybe a code fragment of your modem so I can see if I can find the
problem.

> Is it me or is this listserver taking a couple of hours to reflect each message?

The listserver. Vger is under a rather heavy load I'd think.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: KISS over psuedo TTY

[Tomi Manninen OH2BNS]

On Sat, 22 Jul 2000, Charles Brain wrote:

> I am trying to set up a kiss link over a psuedo TTY without much luck.
> I can do it with slip not not it seems with KISS. I am trying to set up 
> a link to my own application.
> 
> the Kernel is 2.2.12-20 and I am using apps-0.0.4 and tools 0.0.5
> 
> Anyone have any ideas?

How exactly does it fail? It should work just fine with the versions you
mention (though upgrading to the latest wouldn't probably hurt).

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: soundmodem

[Tomi Manninen OH2BNS]

On Fri, 14 Jul 2000 [EMAIL PROTECTED] wrote:

> Before I go into a longish talk about transforms, I's, and Q's, is there
> anyone out there that regularly runs newqpsk? I'm loopback testing the
> modem on vhf fm and I get a 50% error rate. 

I run it somewhat regularly... :)

Sorry that I didn't answer to your direct email, but I'm currently on a
summer cottage 200km away from home and using GSM dada is a bit
expensive. I'll reply later.

That error rate is probably due to a somewhat ugly feature of the
original modem desing. You see, after the symbol sync is acquired and the
modem switches to data mode it just starts to feed the symbols through the
(self synchronising) de-interleaver. However there is no way of knowing
when avctually the sync sequence is over and valid data starts so there is
bound to be several symbols worth of garbage at the beginning of every
transmission. There is no way of counting the _real_ FEC errors... I've
been thinking about reimplementing part of that but it'll mean that
compatibility with the original modem is broken.

> I've used the afsk1200 baud modem in the linux kernel, I've also written a
> packet modem that works with the 'newqpsk' userspace soundmodem code.
> Both the modulator and the demodulator work on the air, but they're mainly
> a learning tool. 

Tom has already written modems for all the modes that the kernel driver
handles. But as you said, writing them is good way of learning.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

ANNOUNCE: newqpsk-0.0.4

[Tomi Manninen OH2BNS]

Hi all,

just uploaded a new version of my newqpsk package, a NEWQPSK/Q15X25
implementation for Linux and soundcard to hes.iki.fi:

  http://www.hes.iki.fi/pub/ham/unix/linux/ax25/newqpsk/

In addition to the usual tarball there are source and i386 binary RPMs as
well.

>From the ChangeLog:

2000-07-02  Tomi Manninen  <[EMAIL PROTECTED]>

* Version 0.0.4

* RPM spec file

* Changed the default channel access parameters

2000-03-11  Tomi Manninen  <[EMAIL PROTECTED]>

* Some reordering in rxdata()

* Implemented frequency and sync tracking in data phase

* Implemented data carrier S/N reporting

* Hopefully fixed the "modulate_calib error" problem


-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Converse Bridge for RH6.1 (2.2.13) ?

[Tomi Manninen OH2BNS]

On Mon, 26 Jun 2000, John F. Feist wrote:

> I was wondering where I can find the source files for Converse Bridge
> (TPP, HTPP...?) that will compile under RedHat 6.1 (2.2.13 kernel).
> RPM's will be fine as well.
> Thank you in advance, John

http://www.hes.iki.fi/htpp/

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Kernel IP routing bug

[Tomi Manninen OH2BNS]

On Tue, 20 Jun 2000, Arno Verhoeven wrote:

> Robin Gilks wrote:
> > 
> > Has anyone found a fix yet for the routing bug in 2.2 kernels that I
> > highlighted about 5 months ago?
> 
> Funny... Last weekend I saw the very same thing on our packet radio
> node. I was just about to send a question about this to the mailing
> list.
> 
> I don't have a sollution for it, but I'm very interested in any ideas
> for a sollution.

The solution is for someone to dig in the kernel networking code and find
out what is going on. I once did that but as the networking code is rather
confusing and the bug doesn't currently affect me, I soon lost interest.

The other solution is to wait for SomeOne Else (tm) to do that

Use the source Luke, that's why you have it.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: kissparams & mkiss

[Tomi Manninen OH2BNS]

On Thu, 8 Jun 2000, Richard Adams wrote:

> I belive Tomi has said before, you can ignore them, but for the record my
> comments.
> BTW; just about all of my ax25 stuff is modules.
> 
> > Jun  8 19:00:25 gb7cip kernel: kissparms uses obsolete
> > (PF_INET,SOCK_PACKET)
> 
> I have serveral machines all running kernel ax25 netrom, drivers include
> mkiss scc dxlmodem, i also get the obsolite message for;
> netromd and mheardd, but not from kissparms, further more i cant say i have
> seen your problem.

The message can come from any program that uses the packet capturing
interface the old way. These include kissparms, listen, netromd, mheardd,
xfbbd... However the kernel only prints the warning once after a boot so
what you see in your system logs depends on how you set up your AX.25.

Anyway, once again, the message is completely harmless. It just warns us
that sometime in the future this backwards compatibility feature will be
removed and the program will stop working. All this has already been taken
into accound in recent versions of ax25-tools/apps (if someone sees this
message when using the very latest tools/apps please let us know).

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: SCC causes kernel crash

[Tomi Manninen OH2BNS]

On Wed, 7 Jun 2000, Paul Lewis wrote:

> gb7cip:root /var/log # cat /proc/net/ax25
> 
> Segmentation fault
> 
> gb7cip:root /var/log # tail syslog
> Jun  7 12:48:28 gb7cip kernel: Unable to handle kernel NULL pointer
> dereference
> Jun  7 12:48:28 gb7cip kernel: current->tss.cr3 = 00a06000, %cr3 =
> 00a06000
> Jun  7 12:48:28 gb7cip kernel: *pde = 
> 
> On the console screen I get the memory dump related to the segmention.
> Can not seem to save that to a file.  Other than write out the screen to
> paper for input on my internet connect PC.

You seem to have somethign odd in /etc/syslog.conf or the kernel default
message level. Try to fix that so you get the full register dump to your
log file (hopefully with symbols resolved). Otherwise writing it to paper
and running through ksymoops is the only way...

The register dump with symbols resolved is the information were interested
in.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: SCC causes kernel crash

[Tomi Manninen OH2BNS]

On Wed, 7 Jun 2000, Paul Lewis wrote:

> Like others, I have been suffering with a problem since December 1999
> when I switched from 2.0.35 kernel +ax25 to the new 2.2.13, 2.2.14 to
> the current 2.2.15.
> 
> The problem I have is with the Ax25 code. (maybe nothing wrong with the
> code - I still have a problem to solve) 
> 
> The AX25 interface(s) normally one of the very busy ax25 interfaces.
> (they all run TCP/IP,AX25,NetROm over them) stop working decoding for
> AX25 data, yet the TCP/IP stack continues to work over the interface.

Next time this happens, could you try "cat /proc/net/ax25" and if it
succeeds send the output here. Also do you see anything strange in the
system logs?

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: New install of linux

[Tomi Manninen OH2BNS]

On Fri, 2 Jun 2000, Shane Deering wrote:

> I might leave RH 6.2 for a while or make an extra partition for it and have
> it as a boot option while working with 6.1.

I wouldn't do that. I'd just go with RH6.2 and apply all the updates (that
is what you should do anyway with any distribution). Applying the updates
is very easy.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: About SQUID (dnsserver) ...

[Tomi Manninen OH2BNS]

On Thu, 18 May 2000, r00t the LiNuXeRRR wrote:

>   We resolve our problem with the cache dir. Now we have another
> problem ... 

Please solve that on some other list than linux-hams and stop crossposting
to several mailing lists...

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Question, was; [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Sun, 14 May 2000, Stewart Wilkinson wrote:

> Gerd wrote:
> > >
> > > Do you have a case where this leads to actual problems so we can see if
> > > something needs to be fixed?
> > 
> > Yes, I am aware of such a case. You run into problems when a
> > user enters a node on L2, goes out on L2 again on the same port
> > to another node where he also enters and goes out on a L2
> > connection. If everything here happens on the same port the
> > second node sends out frames that seem to come from the original
> > user (i.e., they have the same SSID as the user). If the user's TNC
> > receives such a frame by accident, it either sends out DM- or
> > reports protocol errors (FRMR).
> 
> But that applies for other NODE software too. As the SSID gets 
> inverted twice. (ie: -0  >  -15  >  -0  )

Exactly. I guess I could come up with some other scheme for the SSID
change but then it would probably cause problems elsewhere. Not worth the
trouble IMO... 

> I recall sometime ago, having another problem (perhaps in earlier
> versions), whereby Linux NODE used the wrong callsign in outgoing
> connects. 
> 
> (I seem to recall this occured when one Linux NODE linked to another
> at L3/4 and then on again at L2, but I am not currently able to 
> reproduce the problem).

Never heard of such a problem. If you can reproduce it please send me a
bug report.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Question, was; [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Tue, 2 May 2000, Kjell Jarl wrote:

> bpq can be set up to give welcome text to the node call as well, usually
> set off though. Some thenet nodes also gives welcome texts, debending on
> the set up, also over long distance (already established link level).

Ok, maybe I shouldn't have stressed this issue too much, it's not the
point. The fact is that traditionally those systems have avoided sending
welcome texts, and the reason (or one of the reasons) is what I explained.
I know some systems do send them and that those systems also do things
like check the list of known NET/ROM neighbours to see if the connectee is
a user or not. That last mentioned thing is something you will never see
on Linux, it is just too ugly and even as such it's not at all foolproof. 

Some implementations wait for the first information frame from the user
before deciding what to do (the frame will have a protocol ID so the
system knows exactly what it is for). Now this hack could be implemented
at kernel or ax25d level I think and it wouldn't be all that ugly. Maybe
some day...

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

RE: Question, was; [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Tue, 2 May 2000, Richard Bown wrote:

> While on the subject of SSID's being used, and while there is plenty of
> responses.
> Has there been a fix in this release to decrement the SSID of a either a L2
> or L3 connection to the node,   i.e. I login as g8jvm on someone else's
> system then do an outgoing L2 connection to another
> station what I see going out is g8jvm and not g8jvm-15, as I understand it
> should be.
> What it does cause problems with is , if someone logs into a neighbouring
> BBS running linux node, this appears on my mheard list as a direct connect.
> Has this been fixed?, if so can someone point me in the direction of the
> fix.

This of course has nothing to do with the kernel implementation of AX.25,
it's a LinuxNode "problem".

LinuxNode only inverts the callsign if the user is coming in via a L2
connection and going out again on a L2 connection and on the _same_port_. 
In all other cases users callsign-ssid is left untouched. When I coded
this I tried to think of all the possible cases and only saw inverting
necessary in that one particular case.

Do you have a case where this leads to actual problems so we can see if
something needs to be fixed?

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Thu, 27 Apr 2000, Jens David wrote:

> > The later will use the netlink interface, I don't know
> > how to fix net2kiss for the new DDI, and listen can easily be tweaked
> > to work with all kernels.
> 
> As I said, listen is already done, but net2kiss really puzzles me.
> Maybe AF_AX25, SOCK_DGRAM orientated?

Umm.. What is it that is difficult with net2kiss? Taking a quick look it
receives a packet from a network interface, KISS encapsulates it and
writes to a pty. Definitely a PF_PACKET thing, (PF_AX25, SOCK_DGRAM) won't
give you enough of the received packet. Am I missing something here?

Oh, BTW, rxecho also needs to be fixed.

> > Well, those are for "node" and similar programs to view the
> > connection status of other processes. What's wrong with that?
> 
> Even if they were always updated to match the current kernel
> behaviour (i.e. formatting), or through libax25 I still think
> this would be a performance issue. Doesn´t the kernel for example
> disable all interrupts while a proc file is accessed (composed) or did
> I get this wrong?

I have understood so but at least node doesn't use them in a way that
would cause any performance issues.

> I do not yet understand why an application programmer (e.g. terminal program
> writer) has to access the axports file at all. Why not use the socket interface
> and ioctl()s as AF_INET applications do? They do not require a ipports file
> either, don´t they?
> As I understood it the axports file was created to give kernel-ax25 a little
> bit the touch of a NOS-app?

The original reason for axports was that interfaces could only be
differentiated by the interface callsign. There was no way to tell the
kernel to "open a connection through device ax0". It needed to be done
by telling the kernel to "open a connection through the interface having
hardware address N0CALL". And this still is the default method.

This would obviously have been very cumbersome to the users without an
abstraction layer of some sort and as long as that was needed, symbolic
names were easy to add.

Now we have bind-to-device and settable interface names so we can start
thinking whether all this is needed. (This btw raises another question:
with the current state of things do we still need unique interface
HW-addresses? If we could get rid of that restriction, it would help a
lot.)

The question is how will the old functionality be phased out and whether
the [ax|nr|rs]ports files still have a place for some other reason. That
needs to be discussed.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Wed, 26 Apr 2000, Jens David wrote:

> please calm down and let us deal with this on the facts. It was definetely
> not my intention to start another flamewar when I did that announcement.

Neither did I. Sorry if I seemed a little too exited earlier.

> Fact is, nobody else but Jan and Joerg came up with any concrete, factual
> ideas/founded bug reports/patches. This is _very_ irritating, especially
> if you spend most of your spare time on this stuff while the download
> counter reaches the "100" mark. I really need more feedback.

Exactly. Pardon me but the way you presented things in your previous posts
on this subject didn't exactly prompt me to contribute.

Again, I buy many of the arguments you give. That channel access stuff
etc. is self-evident. I just feel that these things should be discussed
before making hasty decisions. Especially as it seems you want to drop
things that are concidered important by others.

You are of course entitled to do as you please but if the real world isn't
taken into account, I fear your valuable work ends up something only your
"group of high standards" knows about...

Anyway we now have discussion and that is good.

> I my opinion only configuration should be in /proc. That´s settings, port
> status, ax25 circuit list (for reading only of cause) etc. That´s mostly
> what it consists of now, so it should stay there. What I have a bad feeling
> about is that terminal programs begin scanning /proc/net/ax25 to find out
> about L2 state of "their" sockets just to display them. They IMHO ought to
> use the ioctl() interface.

As I said, that /proc reading stuff was written for node. Jonathan then
thought it should be moved to the lib so others could use it. In node it
is only used to give the user information about what is happening in the
system. I don't see what's wrong in that.

Oh, and it's also used to get the NET/ROM routing info from kernel in
order to do mnemonic->callsign translation. Currently there is no other
way and I see little reason to change that.

We have seen wrongful uses of the stuff but (now that the ioctl interface
is powerful enough) they have been rectified. I'm not buying it should be
removed just because someone could use it for wrong purposes.

> Another thing is trying to find ax25-Interfaces. Some apps scan the proc
> entries instead of using ioctl()s. They break when proc format changes.
> In my opinion procfs should be reserved for user interaction with standard
> tools such as "cat", "grep" etc, applications and tools should use the
> ioctl() interfaces.

What apps? That's seems just plain stupid. Anyway in my opinion "when proc
format changes" is entirely analogous to "when the kernel api changes". If
something changes in the API the apps break just as if they would when a
proc file format changes. That should just be avoided! And if it cannot be
avoided then it should be hidden behind a library anyway.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: [ANNOUNCE] new-AX.25 for 2.2.14 Rel. 5

[Tomi Manninen OH2BNS]

On Wed, 26 Apr 2000, Jens David wrote:

> My future plans are the following:
> I will mostly rework the complete packages. There is a lot of conceptionally
> wrong stuff in there like dependencies on the stupid axports file and
> /proc/*-Entries.
> 
> Also, I will remerge ax25-tools and -apps (I do not see the reason to have
> two packages here) and call the resulting package ax25-utils again. The
> library will have to be completely reworked, too. I do not like those
> proc-scanning functions and axports stuff at all. I think I am going to
> call this libax25-2, but I´m not yet completely sure about this.

Oh, how nice of you to discuss all of this with the other developers and
giving constructive critisism/feedback to those of us who have written the
old code! 

The /proc reading stuff was somethign I hacked quickly together to support
node. I don't think anyone else uses it. Would you care to elaborate on
what exactly is so evil about that.

> I know that this will pose a lot of administrative problems. But I will
> not accept any performance-compatibility tradeoffs here, except perhaps
> the binary compatibility with the old socket interface.
> 
> Comments?

Good luck getting other developers interested in your (apprently personal)
crusade to save us from the evil of the old implementation.

> Remember the flexnet slogan?: "It´s like re-inventing the wheel, and doing
> it the right way this time."

Are you referring to "doing everything under secrecy and without
discussing with others" when you talk about the "right way" á la Flexnet?
Give us a break... 

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: sendmail problem

[Tomi Manninen OH2BNS]

On Mon, 17 Apr 2000, Aaron Sethman wrote:

> What I do is on my systems is allow mail to be relayed from any 44.0.0.0/8
> address. I don't know if sendmail supports this, but I do know that
> postfix does. Now all I need to do is hack LZW support into postfix, and
> I'll be happy :).

Long time ago I did that to ZMailer but only on smtp server (incoming). If
you want I could try to find the code.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Unproto Path in AX25

[Tomi Manninen OH2BNS]

On Wed, 12 Apr 2000, Hamish Moffatt wrote:

> On Mon, Apr 10, 2000 at 09:23:19AM -0500, John Gorkos wrote:
> > Unfortunately, with APRS being a UI-only protocol, my beacons don't get
> > propagated through the APRS system.  I need to be able to do something
> > like this:
> > beacon -d "APRS via WIDE,WIDE,WIDE" ...
> 
> That should work fine, use "v" instead of "via".

Both should work fine. It's the comma signs that are the problem...

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Unproto Path in AX25

[Tomi Manninen OH2BNS]

On Mon, 10 Apr 2000, jerome schatten wrote:

> John... I went through this years ago running a 1278 in kiss mode on
> an fbb bbs. As soon as it receives a string which is the same as
> the KISSOFF command, it drops out of kiss mode. I can't remember what
> the hex sequence was, but it was reproducible. No-one was able to 
> fix it, and it seemed true across all 1278's in this area.

The "kiss off" sequence in hex is 0xC0 0xFF 0xC0 and it can never appear
in a valid kiss data flow. So there must be more than that to it if it
really was happening in response to serial data to the TNC.

Long ago I had a problem with a Symek TNC2S with the TAPR firmware that
came with it. It would just reset itself after a while. But IIRC it stayed
in KISS mode, it flashed the LEDs three times and forgot the channel
access parameters... Anyway it was a bug in the firmware.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Unproto Path in AX25

[Tomi Manninen OH2BNS]

On Mon, 10 Apr 2000, John Gorkos wrote:

> Background:  I am running APRSdigi on my RedHat 6.1 box, with a MFJ-1278
> hanging on ttyS0.  I am using the modified AXtools from Alan Crosswell,
> N2YGK.  Everything is working well (aprsmon allows TCP/IP connections,
> beacons are sent, etc.).

What is "modified AXtools from Alan Crosswell" ?

> Problem 2:  The beacon command I use looks like this:
> beacon -d APRS -t 15 ax0 ''
> Unfortunately, with APRS being a UI-only protocol, my beacons don't get
> propagated through the APRS system.  I need to be able to do something
> like this:
> beacon -d "APRS via WIDE,WIDE,WIDE" ...
> Is there an axparm that I'm missing in the documentation that will allow
> me to do this?  Thanks.

Have you tried

  beacon -d "APRS v WIDE WIDE WIDE" ...

That should work.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Node query

[Tomi Manninen OH2BNS]

On Mon, 10 Apr 2000, Robin Gilks wrote:

> > > The line in my node.conf that is affected is:
> > > Alias   CONVers   "telnet %{2:gb7ipd.ampr.org} 3600 \"/n %u %{1:160}\""
> > > 
> > > and its the bit in the quotes at the end that don't work no more...

Actually you'll probably find that the first part won't work either. You
won't be able to specify another convers server at runtime...

> > What does work is;
> > 
> > 'telnet %{2:gb7ipd.ampr.org} 3600 "/n %u %{1:160}\n/w *"'
>
> Changing the double quotes to single quotes and removing the escape character
> from the double quotes fixed it!!
> 
> Not an obvious problem!!

The reason is that those %-escapes get processed twice: first when the
configuration file is read and later when the alias is actually executed.
The problem is that positional parameters (%1 and %2 in this case) don't
really have a meaning when the configuration file is read. They only have
a meaning when the alias is executed. So you need to be able to enter the
alias definition so that the %-escapes are not processed while reading the
config but only at alias execution time. That required a slight change in
the config syntax. The change is documented (see the HISTORY file,
node.conf(5) manual page and the sample files) but probably very easy to
miss...

The way it works is probably easiest explained with the example included
in the sample file:

ExtCmd TIme 1 nobody /bin/echo echo %N Node session started at %I, current time is \%I.

The %N will be expanded to the NodeId (at startup so NodeId needs to be
specified above this). %I will be the current time again at the time the
config file is read. The second %I has the percent sign escaped so it will
be expanded only when the extcmd is executed. (The same effect would have
resulted from enclosing it in single quotes: '%I')

I know this may sound confusing (I certainly was first confused when
trying to code it) but IMHO when you understand it, you'll see that it is
the only clean way of implementing it.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Attention UK netrom users

[Tomi Manninen OH2BNS]

On Mon, 10 Apr 2000, Robin Gilks wrote:

> The other problem is, using 2.0.36 kernel, how do I get the netrom broadcast to
> send the alias as well as the netrom call. I've set nrports the same as a 2.2
> kernel box which works OK. An adjacent node adds the call to its node list but
> no alias :-((

The sending node alias is part of every NODES broadcast sent by the node.
It should work the same with regardless of the kernel version. Have you
watched the 2.0.36 box sending the broadcasts? With listen -a on that box?
>From another box over the radio? Is there perhaps something strange in the
nrports file on that box?

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Kernel based AX25 on Sparc?

[Tomi Manninen OH2BNS]

On Fri, 24 Mar 2000, Geoff Blake wrote:

> I may be going off at "half cock" but during a spare moment at work today 
> I was looking at compiling an "AX25 aware" kernel on a Sparc 10. I was 
> doing about a dozen other things at the time, but it appeared that 
> Menuconfig did not offer the same options as the usual i386 config.
> 
> Should the Sparc Linux kerne source tree include the same AX25 utilities 
> or is the RedHat 6.1 (2.2.12) distribution a stripped down version? (I 
> did not have a recent kernel tree to hand nor did I want to d/load one 
> over a 28K link!) 

An excerpt from the announcement of Alan's 2.2.15pre1 kernel:

  ...
  o Use amateur radio drivers on Sparc (Tom Sailer)
  ...

Seems they were accidentally left off at some point.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: OT: T4 terminal program?

[Tomi Manninen OH2BNS]

On Sun, 26 Mar 2000, Gerd wrote:

> sorry for being offtopic, but here at the list for me there was the 
> only place the T4 packet radio terminal program was mentioned. I 
> could not find it in the Internet, however.
> Does anyone know where to get it from?

ftp://ftp.funet.fi/pub/ham/Finnish/oh2mkp/

I'm not sure if newer versions can be found anywhere on the net. The
latest is something like 1.5a24. After that Mikko lost interest in it...

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: VC and datagram mode?

[Tomi Manninen OH2BNS]

On Fri, 24 Mar 2000, Thomas Sailer wrote:

> "Wilson G. Hein" wrote:
> 
> > Is this a part of the current implementation of linux ax25 and x1j nodes?
> 
> It's not in the stock kernel, however part of the DG2FEF/DG1KJD patch.
> No idea about x1j. It works in Flexnet and Xnet. Wampes also has
> some h2h ack code.
> 
> In theory, you could also use net/rom to achieve this, but this adds
> overhead (net/rom header), a window size problem (if you don't
> use mod128 sequence numbers), and the original net/rom routing
> protocol is quite broken (INP is said to solve this though).

On the other hand both Linux (current implementation) and X1J are capable
of IP routing and support Virtual Circuit mode so if configured to do that
the whole problem disappears.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Syslog message: pg uses obsolete (PF_INET,SOCK_PACKET)

[Tomi Manninen OH2BNS]

On Mon, 20 Mar 2000, Bent Bagger wrote:

> I am hunting an error in PG (the program for upload of files to the
> Microsats) and I have turned socket debug information on by doing a
> 'setsockopt(s, SOL_SOCKET, SO_DEBUG, &opt, sizeof(opt)' call. When 
> I inspect the system log, I now find this message:
> 
> pg uses obsolete (PF_INET,SOCK_PACKET)

This shouldn't depend on having SO_DEBUG on. You probably just haven't
noticed it before. Also you only get it once after a reboot and it may be
that you didn't see it because some other program has already got the
complaint. So it depends on the order in which you run programs after a
reboot.

> Could somebody please shed some light on the significance of this
> message, more specifically: what has been changed in the ax25utilities
> since the days of kernel 2.0.35 (I'm now on kernel 2.2.13 with glibc 2.1
> and libax25 0.0 (?)) and what changes should I do in PG to make this message
> go away?

Nothing changed in the utils, it's the kernels generic packet capture
interface that was changed. The message is completely harmless (other than
a program causing it may stop working sometime in the future if not
fixed). You can get rid of the message just by changing PF_INET ->
PF_PACKET in the source. While that is actually also obsolete (see "man
packet") I believe it will continue to work for some time still. I have
not checked what it would take to properly use (PF_PACKET, SOCK_RAW) like
suggested in the packet(4) manual page.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: ARP Timeout problem

[Tomi Manninen OH2BNS]

On Sat, 18 Mar 2000, F. D. Johnson wrote:

> Hi Folks, can anyone point me in tthe right direction ?
> 
> My system (Suse6.3), when initiating any tcpip/radio session sends out 3
> very rapid ARP requests
> then, more often than not, times out before the remote end has had chance
> to reply.
> 
> Can I lengthen the ARP timeout ?

Yes, through the proc interface. See /usr/src/linux/Documentation/proc.txt
section Network Neighbor handling. The parameter you want to adjust is
retrans_time.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: DSP boards.

[Tomi Manninen OH2BNS]

On Sun, 19 Mar 2000, Geoff Bagley wrote:

> I had not realised that stand-alone DSP boards were getting scarce.

The Motorola DSP56002EVM was discontinued some time ago. Newer models are
of course available but there is no direct replacement. There was a
discussion on this topic on the TAPR DSPSIG (or was it HFSIG?) so if
anyone is interested you can check the archives on TAPR home page.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: newqpsk Activity?

[Tomi Manninen OH2BNS]

On Mon, 13 Mar 2000, Luc Langehegermann wrote:

> Thanks you for reading this. I have just compiled the newqpsk package and it
> seems to work. Are there already stations operating this? If yes, where?

.lu is Luxembourg right? You might want to try if you can hear
Joni OH2NJR's station at 3590.0 MHz LSB. It is on 24h/day, there is a
LinuxNode and a XFBB bbs listening. He uses an Alef Null DSP card 4 but
the settings are compatible with the defaults in my package.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Q15X25 - newpsk.0.0.1

[Tomi Manninen OH2BNS]

On Mon, 6 Mar 2000, Jörgen Overgaard wrote:

> Is there any one who has successfully compiled and run the Q15X25
> (newpsk-0.0.1) package on RedHat 6.1??

Well, I still run RH6.0 but I really don't think there are any 6.1
spesific problems...

> For me it all compiles well,but when i try to start it with
> ./soundmodem it stops and shouts angry at me that threre is a problem
> with TIOCSETD.

Do you have MKISS support compiled in the kernel or if it is compiled as a
module, is the module loaded?

> After some VERY rude hacking in the code it runs a bit longer,but not much.
> Now it says it cannot open PTY "".
> (I understand that you cannot open "" nothing,but why is it trying)

Better undo those hacks... :)

Anyway there is a newqpsk-0.0.2 available in the same place you downloaded
0.0.1. There is nothing really new but the nasty tx-lockup bug mentioned
in BUGS is fixed. Oh, and I should probably soon upload 0.0.3 as there is
also a problem when running newqpsk with the standard sound drivers (I use
ALSA drivers to develop). 

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: pa4ab

[Tomi Manninen OH2BNS]

On Mon, 7 Feb 2000, Gregory Maxwell wrote:

> On Mon, 7 Feb 2000, Ing. Jose A. Amador wrote:
> 
> > On Mon, 7 Feb 2000, Tomi Manninen wrote:
> > 
> > > Why not try Q15X25 (aka NEWQPSK). Download the package from
> > >   http://www.hes.iki.fi/pub/ham/unix/linux/ax25/
> > > and give it a try!
> > 
> > I already got it, precisely from THAT URL, but have not compiled it
> > yetIs anyone using it on HF ?
> 
> Have I been out of the soundmodem loop too long or are all soundmodems now
> implimented as very nice userspace daemons like this one?

As explained in the README file, I used an early beta version of a
userspace implementation from Thomas HB9JNX as the core of my Q15 code. 
I stripped all the other modulator/demodulator libraries from my package
in order to avoid confusion.

Unfortunately it seems Thomas is so busy with more important projects
(read work) so he hasn't been able to finish the soundmodem code. The beta
I have used works, but still has issues to be solved and probably wouldn't
work too well on a busy VHF channel.

So were stuck with the kernel soundmodem implementation at least for a
while.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

RE: Routing problem

[Tomi Manninen OH2BNS]

On Fri, 4 Feb 2000, Robin Gilks wrote:

> Does anyone have a 2.0 system they can check to see if it behaves the
> same as 2.2 - be nice to know if this is a 2.2 problem only. If it is
> its move back time

2.0 lets you route back to the interface the datagram came in. I can
verify that anytime on my home system.

> (despite other threads urging us to move on the bleeding edge).

2.2 is not bleeding edge...

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Routing problem

[Tomi Manninen OH2BNS]

On Sat, 5 Feb 2000, Hamish Moffatt wrote:

> On Thu, Feb 03, 2000 at 07:36:57PM -, Robert A Jenkins wrote:
> > For machines within the same subnet to work via an intermediate system, you
> > must add specific routes to the target machines with the intermediate as a
> > gateway. I set mine up in rc.local.
> 
> Well, they are not really on the same subnet then if an active
> intermediate is required. Can you change the subnet mask to the correct
> value, which may even be 255.255.255.254? (I don't know that such
> a mask is actually supported to be honest..)

I'm not sure but I don't think the subnet / netmask issue is of relevance
here. The main point of the problem is that Linux 2.2 seems to refuse to
send out an IP datagram on the same interface it came in. At least with
AX.25 interfaces.

I tested this at home by setting up a route to certain host through my
only radio port. I tested that the route works from my machine and from
another machine on my home ethernet. I then went to another site that is
on the same frequency as mine (actually I dialed it up but anyway) and set
up the route there to point _through_my_box_. With older kernels, this
would have worked. With kernel 2.2 my box just throws the datagram away.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: no more ipc channels

[Tomi Manninen OH2BNS]

On Mon, 3 Jan 2000, Kjell Jarl wrote:

> Thanks Tomi, yes kill -9 my.script has been performed regularly. Seems
> to be a bad habit of mine, have to change it. Also, I may be doing wrong
> in the expect script, node simply is left in the dark?

Well sending SIGKILL to terminate node certainly doesn't give it any
chance to free the IPC message channel. Besides using SIGKILL should
always be the absolute last resort when trying to make a process
terminate.

If you can't get node to exit cleanly you can send a SIGTERM or a SIGQUIT
to it.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Linux FBB and Unproto list.

[Tomi Manninen OH2BNS]

On Wed, 29 Dec 1999, Dirk Koopman wrote:

> Because I wish to receive packets addressed to generalised, as it were,
> "broadcast" addresses eg "DX", "ANN", "LINK" (these are examples, not
> necessarily real) as well as to my callsign. I also may want to gateway UI
> frames to callsigns that I know (by various higher level means) to be part of
> "my" network (i.e to "digi" UI frames selectively across ports according to
> heuristics under my control).
> 
> The point of this that a reliable multicast protocol requires there to be
> some group addresses, for DX cluster work there are several logical
> groups that one could envisage including, but not limited to, "node-to-user"
> things, "network" things, "inter-user" things and so on. As well as
> multicast, there is a point to point message requirement that could be done
> by "normal" connected mode but I would prefer, for orthogonality, to also be
> done by a simple UI based, window=1, protocol.

Some kind of braodcast/multicast method should probably be added to the
AX.25 stack. But if you want that much flexibility it might be easiest to
really use PF_PACKET. Sure you need root for that but only when calling
socket().

> I was not envisaging a PF_PACKET based thing - I understand why this is a
> problem. I was under the impression that we were talking about SOCK_RAW here
> and that that required root as well.

Calling socket(PF_AX25, SOCK_RAW, ...) doesn't require root but then
binding the socket to any local address other than your own does require
it.

> Have you any sample code that does things with SOCK_RAW? I am perfectly happy
> to generate raw AX25 packets (I can do that bit), it is the ioctl and
> socket magic that I am confused (and cannot find any documentation for
> SOCK_RAW) about. 

SOCK_RAW with AX25 seems to be just about the same as SOCK_DGRAM except
that when you receive you get the PID byte as well. Thus I'm not sure if
it's really very usefull.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Error opening terminal

[Tomi Manninen OH2BNS]

On Mon, 3 Jan 2000 [EMAIL PROTECTED] wrote:

> I have two TNOS executables, the second just compiled for the y2k
> bug. When I try to run the new exe I get the following message:
> 
> Error opening terminal: linux

Probably it doesn't find the terminfo database. I guess the executable was
compiled on another machine? Running tnos with strace might reveal where
it expects to find the database and then you can add a symbolic link
there. For example on one box I had to do

 ln -s /usr/share/terminfo /usr/local/lib/terminfo

to get some old stuff running after upgrading the system. Possible places
for where the database is and/or where tnos searches for it are /usr/lib,
/usr/share, /usr/local/lib and /usr/local/share at least.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: no more ipc channels

[Tomi Manninen OH2BNS]

On Mon, 3 Jan 2000, Richard Adams wrote:

> On Mon, 03 Jan 2000,  Kjell Jarl wrote about,  no more ipc channels:
> > 
> > After using /usr/bin/node alot (I am learning script writing...) I am
> > getting:
> > 
> > [sm7gvf@pc2 test2]$ node
> > LINUX:SM7GVF-5} ipc_open: Could not get an IPC channel: No space left on
> > device
> 
> This indicates that the partition on which the software operates is full.

No it does NOT ! It means that the system is out of IPC channels.

Kjell, have you by any chance been killing node with something like
"kill -KILL" or otherwise made it terminate abnormally while testing your
stuff? Node used to have a bug that caused it to leak IPC channels but
that should be fixed (see the HISTORY file). 

Note that node not getting a free IPC channel is not a fatal bug. At least
not to node. Only the the "TAlk" command is affected (disabled). Some
other software on your system might however need a free channel too...

> > RH 6.0, node 0.3.0 and alot of other stuff (clx/postgresql). Kernel
> > 2.2.5.

I seem to have vague recollection that clx has had problems with IPC
channels as well. I'm not sure however as I don't run it my self.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Linux FBB and Unproto list.

[Tomi Manninen OH2BNS]

On Wed, 29 Dec 1999, Hamish Moffatt wrote:

> On Tue, Dec 28, 1999 at 04:39:06PM -, Dirk Koopman wrote:
> > Please would you answer the question? Why does it _actually_ need root
> > privileges for ax25 protocol traffic? We are talking here about something
> > which, whatever you do, is insecure. Anybody can program a TNC to be any
> > callsign. Anybody can monitor anything using a TNC. Why does linux have to be
> > different? 
> 
> Because Linux tries to integrate AX.25 with Unix, rather than Unix with AX.25.
> It treats AX.25 like just another protocol; it should be as secure as IP.
> Obviously it is trivially easy to imitate someone else's callsign if you
> really want to, even with linux.

But not without having root and/or hardware access to the system. Or
that's how it is supposed to be. IMHO in this case root access doesn't
have to mean a process running as root however.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Linux FBB and Unproto list.

[Tomi Manninen OH2BNS]

On Tue, 28 Dec 1999, Dirk Koopman wrote:

> On 23-Dec-1999 Tomi Manninen wrote:
> > On Thu, 23 Dec 1999, Leszek A. Szczepanowski wrote:
> > 
> >> Is ANY posibility to use unproto list (raw socket), running
> >> FBB as ordinal user, not root? I was looking in kernel
> >> sources, and found in AF_AX25 there is one 'suser' function
> >> checking if user has priviledges to open such a socket.
> >> I think it is uneseseary there, security solutions in this
> >> case aren't needed! What a stupid reason. If I'll make
> >> patch, to open raw socket for AX25 by any user, it will
> >> be placed on this list.
> > 
> > Please don't distribute such bugs on this list. There are good reasons for
> > why the unix security model is as it is. Besides, doing what you want
> > (running FBB as non-root) would be fairly simple to do the right way. If
> > you want to be usefull, just help Jean-Paul in doing that.
> 
> Erm.. a "bug"? I am not at all sure that this is a "bug". I also cannot, for
> the life of me, see why a "feature" of this sort cannot be discussed here.

Well, of course it can be discussed. Maybe my reaction was too strong.

> I would like to hear the reasons why this feature is "root only", especially
> bearing in mind that computers running ax25 should be amateur use only.

Amateur use only? Why? 

> Running ax25 stacks on "commercially sensitive" machines especially with
> ax25 programs running as root (UI generating or not) strikes me a security
> exploit waiting to happen!

There is _NO_ need to have root privileges after opening the socket! Just
drop them and suddenly you have 100% secure application. The stack it self
should be secure. Unless some bright one decides to distribute a patch
removing super user checks...

> For the record, I too would like to be able to both receive and generate
> UI frames from programs that are non-root. This is going to become an issue
> during the course of the year with my DXSpider cluster program.

Generating UI frames as non-root is easy, just use a datagram socket.
Receiving should be just as easy but I haven't tested. Of course you are
then restricted to your own source call.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Error compiling libax25

[Tomi Manninen OH2BNS]

On Fri, 17 Dec 1999, Craig Small wrote:

> Tomi Manninen OH2BNS said:
> > On Wed, 15 Dec 1999, Hamish Moffatt wrote:
> > 
> > > On Wed, Dec 15, 1999 at 02:24:43AM +, Jorge Matias wrote:
> > > > Probably you don't have symbolic links:
> > > >   /usr/include/asm to /usr/src/linux/include/asm-i386
> > > >   /usr/include/linux to /usr/src/linux/include/linux
> > > 
> > > And be glad that you don't, because you shouldn't.
> > 
> > Oh? Really? Then how come a _LOT_ of my glibc headers include 
> > and/or  ? Is glibc supposed to replace all those some day (good
> > luck!) ?
> A lot? Hmm, I have approximately 20 headers including linux/*, mostly in
> the net,netinet and sys areas and they're probably grabbing magic
> numbers and about 10 headers including asm/* also mostly in netinet and
> sys.

Ok, maybe the "a lot" part was a bit exaggerated.

> I have resisted the temptation to launch into yet another tirade at the
> totally brokeness of some distro's headers. Most people know my
> opinions already about that.

I hope you don't mean Red Hat and the netax25/*.h fiasco. Because that
wasn't at all Red Hats fault. We could try to blame the glibc folks but
that too were unfair. No, it's us, the ax.25 developers who are to blame.

> Yes, the general idea is to remove or reduce the number of linux/*.h and
> asm/*.h headers and other headers that include them. There are very good
> reasons for doing so.

Out of curiosity: How do they intend to cope with kernel api changes? Like
the one we had in 2.0.35 -> 36 and again 2.0.x -> 2.2.x (and probably
2.2.x -> 2.4.x). I hope they have a plan.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Error compiling libax25

[Tomi Manninen OH2BNS]

On Fri, 17 Dec 1999, Hamish Moffatt wrote:

> On Wed, Dec 15, 1999 at 10:25:03PM +0200, Tomi Manninen OH2BNS wrote:
> > On Wed, 15 Dec 1999, Hamish Moffatt wrote:
> > 
> > > On Wed, Dec 15, 1999 at 02:24:43AM +, Jorge Matias wrote:
> > > > Probably you don't have symbolic links:
> > > >   /usr/include/asm to /usr/src/linux/include/asm-i386
> > > >   /usr/include/linux to /usr/src/linux/include/linux
> > > 
> > > And be glad that you don't, because you shouldn't.
> > 
> > Oh? Really? Then how come a _LOT_ of my glibc headers include 
> > and/or  ? Is glibc supposed to replace all those some day (good
> > luck!) ?
> > 
> > Anyway a small correction: /usr/include/asm should point to
> > /usr/src/linux/include/asm not asm-i386.
> 
> Including  and  is fine. The glibc kernel headers
> include the files in /usr/include/linux and /usr/include/asm directly;
> they are not links into /usr/src/linux (there is no need).

Well, this simply is not true with Red Hat, at least not on 6.0 that the
original question was about.

> If you need the exact .h files for the current kernel, you should
> get them out of /usr/src/linux/include directly using -I. On Debian
> at least, /usr/include contains the header files provided by glibc,
> and not the current kernel versions. Craig's ax25-tools sources cope
> just fine with this.

Yes, you told us that could compile the tools with 2.0.35 kernel in
your /usr/src/linux. Well, do those tools work if you really run that
kernel? ... I didn't think so. I rest my case.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Error compiling libax25

[Tomi Manninen OH2BNS]

On Wed, 15 Dec 1999, Hamish Moffatt wrote:

> On Wed, Dec 15, 1999 at 02:24:43AM +, Jorge Matias wrote:
> > Probably you don't have symbolic links:
> >   /usr/include/asm to /usr/src/linux/include/asm-i386
> >   /usr/include/linux to /usr/src/linux/include/linux
> 
> And be glad that you don't, because you shouldn't.

Oh? Really? Then how come a _LOT_ of my glibc headers include 
and/or  ? Is glibc supposed to replace all those some day (good
luck!) ?

Anyway a small correction: /usr/include/asm should point to
/usr/src/linux/include/asm not asm-i386.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Config help required..

[Tomi Manninen OH2BNS]

On Fri, 10 Dec 1999, Rob Compton wrote:

> >  Hamish, after loading soundmodem.o/baycom.o/yam.o module and
> > correctly configure this driver we got network device, such
> > as sm0, bc0 etc. This device can be used for IP over AX25, but,
> > we can't make simple 'call' for connecting to other station and
> > use this device in LinuxNode/xfbbd etc ...
> >
> >  For this utils we need to make KISS stream w/net2kiss (man net2kiss),
> > and w/kissattach create standart AX.25 port.
> 
> Now I'm confused more.
> 
> In my YAM docs, it says nothing about kissattach to use the port for xfbb.
> 
> Don't you just love incomplete documentation.
> 
> Looks like I'll have to go all the way through my config and do some more
> work!

No, no! Don't panic. What Yuri says above isn't true.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Route

[Tomi Manninen OH2BNS]

On Mon, 29 Nov 1999 [EMAIL PROTECTED] wrote:

> I need help to lock a route in net/rom. Sk7hw is heard on my port1 but i
> cant connect this way. My connectway is on port2 via sk7bk-5. How can i lock
> the port so that the connection goes the right way? I have tryed nrparms
> -nodes sk7hw-5 - vxo 50 5 port1 to remove it from the list. When sk7hw sends
> out nodeinfo it comes back. How?

You can not lock a node in Linux. What you can do is lock a route (a
neighbour). If you lock the quality of that particular neighbour (sk7hw on
port1) to a low value, the node routes it advertises won't be used except
as a last resort. Locking a route is of course done with "nrparms -route".

However if there are also no other "good" neighbours on your port1
frequency it might be a better idea to lower the default quality for that
port in /etc/ax25/nrbroadcast. The effect will be the same but it will by
default affect all neighbours on that port.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: ax25 mod being opened and not closed!

[Tomi Manninen OH2BNS]

On Mon, 22 Nov 1999, Richard Adams wrote:

> No it certanly does not look normal, just how o you attaqch the
> interfaces,? i hope not with kissattach, ifconfig is the way to go.

What???

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Interesting High Speed Engineering Project from WSU

[Tomi Manninen OH2BNS]

On Sun, 7 Nov 1999, Cathryn Mataga wrote:

> >> Yeah, I wonder if it would work just sending raw TCP/IP without any ax25 headers
> >> at all.  Aren't the crc's and packet lengths and all that stuff just transmitted 
>twice?
> >
> >I'm not sure. Need to include amateur callsigns in there somewhere, though,
> >since IP addresses aren't official identification.
> 
> Or do we need the callsigns in every single packet?  US-wise, I thought we only
> needed to ID every 10 minuts.  

I guess we still need some sort of MAC level addresses anyway. Ethernet
has them too. IP over AX.25 in datagram mode actually isn't much more than
AX.25 addresses in front of an IP frame.

> >Is Van Jacobsen compression used? I haven't used TCP/IP over AX.25 yet,
> >still waiting on an IP allocation. VJ header compression is routinely
> >used on dialup PPP links, and if a link that fast (comparatively) can
> >benefit from it, I suspect our links could too.
> 
> My understanding is that tcp/ip headers are not compressed.  I believe every packet
> goes out with an ax25 header and a TCP/ip header. Isn't
> some kind of header compression going to be part of the 2.3/2.4 kernels?  Does
> *NOS understand any kind of header compression?

Matthias' new AX.25 stuff supports VJ compression over connected mode
AX.25. VJ won't work with datagram mode as it depends on the incoming
frames being ordered.

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: Interesting High Speed Engineering Project from WSU

[Tomi Manninen OH2BNS]

On Tue, 9 Nov 1999, Riley Williams wrote:

> Perhaps the following URL will be of interest to you...
> 
>  Q> http://www.tapr.org/tapr/pdf/AX25.2.2.pdf
> 
> ...it is the current 2.2 specification of the AX.25 protocol in Adobe
> Acrobat format. Be warned in advance that it's around 1.6M in size, so
> don't try downloading it over a 1200 Baud link...

Just remember that document specifies the version 2.2 of the protocol.
Version 2.2 hasn't been implemented anywhere AFAIK and judging from some
comments (for example on this list) it may never become a standard...

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Virtual console corrupted with xfbbC

[Tomi Manninen OH2BNS]

On Tue, 2 Nov 1999, Alexandre Fornieles wrote:

> When i start a chat with a user connected to FBB, with '=F1CAY', the
> concole gets completely corrupted. Characters become unreadable and it
> remains in this state even when leaving xfbbC and logging out from that
> console. I mean, there are characters, but they're not displayed as
> normal letters but as strange symbols that are mixed with a few readable
> normal letters.
> 
> It just remains unusable and haven't found yet how to restore it until i
> reboot the machine which does not happen so often.

Try typing "reset" when you get back to the shell prompt and see if that
helps.

The problem is possibly that you receive the "switch to alternate
character set" -character. That happens often here when one reads a packet
bbs message written with the DOS character set (IBM codepage 437) and
someone or something has stripped the 8th bit. The "ö" (o-diaresis)
character then becomes the switch character... I wish people would start
to use ISO character sets on packet too. 

-- 
Tomi Manninen   Internet:  [EMAIL PROTECTED]
OH2BNS  AX.25: [EMAIL PROTECTED]
KP20ME04Amprnet:   [EMAIL PROTECTED]

Re: mkiss.c: uninitialized function (kernel)

[Tomi Manninen OH2BNS]

On Wed, 27 Oct 1999, Jorge Matias wrote:

> >gcc -D__KERNEL__ -I/usr/src/linux-2.2.13/include -Wall -Wstrict-prototypes
> >-O2 -fomit-frame-pointer -fno-strict-aliasing -pipe -fno-strength-reduce
> >-m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=586
> >-DMODULE   -c -o mkiss.o mkiss.c
> >mkiss.c: In function `kiss_esc_crc':
> >mkiss.c:805: warning: `c' might be used uninitialized in this function
> 
>   Anyone knows if this is a possible problem?

http://www.hes.iki.fi/archive/linux-hams/199910/0139.html

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: 1 mswin, 1 linux, 1 TNC

[Tomi Manninen OH2BNS]

On Wed, 27 Oct 1999, Chris Ruffin wrote:

> -Connect the TNC to a serial port on the linux machine, run
> kissattach.
> -Connect the windows and linux machines together using the
> serial ports.  Run kiss mode software on the windows station,
> run kissattach on the linux machine on that serial interface.
> -Route the packets to/from the windows station with 
> axparms -forward.  If you set up all the ax ports right, it
> should forward the packets to and from the windows station.

This is _exactly_ what I wrote rxecho for. So instead of axparms
-forward (which I think wouldn't work, not sure though) use that.

Our club BBS used to be a DOS FBB (Linux FBB wasn't quite mature at the
time) and I wanted to connect all the radios to a Linux machine without
the BBS users even noticing anything. So the DOS box was connected to
the Linux box with a serial cable running multi-kiss (one channel for
each radio port and one channel for the traffic between the two machines) 
and the radios to the Linux box.

> Doing things this way will allow arbitrary software to be able to
> run on the windows station, since it will think that it is
> connected directly to the TNC.

In our case the DOS box thought it was talking to several TNCs
multidropped on one serial line.

> The big concern I have is the kiss connection from the linux
> station to the windows station.  I've read over the kiss spec,
> and it would *seem* that it would work.  But if it does seem that
> way, then it would also seem that you could connect two TNCs
> directly together and have them complement each other, i.e. all
> packets received on one TNC will be transmitted by the other, and
> vice versa.  That doesn't seem right to me for some reason.

That's the way it works...

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: QRZ! + node

[Tomi Manninen OH2BNS]

On Thu, 28 Oct 1999, James S. Kaplan KG7FU wrote:

> Actually, I was looking to usse the cdrom and perhaps the unix/linux
> files supplied on it.

If there is a command line tool to query a callsign then interfacing that
to LinuxNode should be simple. Just use the extcmd facility, probably you
want to use the "pipe" flag there. See the node.conf manual page.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

/proc/.../neigh interface (Was: Re: TCP/IP config/connect problem)

[Tomi Manninen OH2BNS]

I wrote:

> > PS. Does anyone have a pointer to a document describing the
> > /proc/sys/net/ipv4/neigh/ interface?

/usr/src/linux/Documentation/proc.txt

(searching for a brown paper bag...)

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: TCP/IP config/connect problem

[Tomi Manninen OH2BNS]

On Tue, 26 Oct 1999, Ron Jochems wrote:

> Can you help me find the archive .?.
> Is it online, or also downloadable.?..



-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Killing KIssattach Kills Linux!

[Tomi Manninen OH2BNS]

On Mon, 25 Oct 1999, Joachim Holst wrote:

> > What makes you (two) think the crashing is caused by anything AX.25
> > related? Do you see something AX.25 related in the Oops dump when it
> > happens? If so, what? Please give us something usable to work with. Find
> > the Oops in your system log and send it here for us to see or if your
> > machine freezes before anything gets written in the log then write the
> > Oops down and run it through ksymoops.
> 
> From my side, it's a very loose guess. Since I didn't have this
> problem with older kernels 2.0.x and the old AX25 utils but I do now
> with new kernel and old AX25 utils, I found it quite natural to assume
> that it was related with incompatible AX25 utils.

Well, I don't find that very natural at all. There are million other
things that changed in the kernel between 2.0.x and 2.2.x. It's actually
much more probable that the problem is elsewhere. BUT, if you can show us
an Oops trace (symbols resolved) with something AX.25 related then we can
start debugging...

> The old ones don't
> compile on a 2.2.x kernel (at least not the last time I tried) and
> neither did the new AX25 libs. I know that newer versions have been
> released that do compile on a RedHat system (the reason why they
> didn't compile earlier, weas RedHat related).

Installing the new lib/tools/apps is very simple and there are now even
RPMs available. Anyway if the problem turns out to be this (old utils)
then it still is a kernel bug. No usermode application should be able to
crash the kernel, even a faulty one!

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: nrparms command error with new ax25 tools

[Tomi Manninen OH2BNS]

On Fri, 22 Oct 1999, Kai Altenfelder wrote:

> > It's now available at ftp.hes.iki.fi aka ftp.oh7lzb.ampr.org:
> > 
> >   ftp://ftp.hes.iki.fi/pub/ham/linux/ax25/ax25-tools-0.0.5-oh2bns.patch.gz
> 
> Is this patch needed for every distribution or for RH only?

The patch fixes a bug in nrparms and in the user_call set of utils. It's
not Red Hat specific.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Baycom installation fails

[Tomi Manninen OH2BNS]

On Fri, 22 Oct 1999 [EMAIL PROTECTED] wrote:

> No error messages whatsoever. Still the device /dev/bc0 (51,0) remains
> dead. I also read about the new devices bcsf0 etc., but those do not exist on
> my system. I'd like to create them using mknod, however I do not know which
> major number to use.

The new baycom devices are _network_ devices and they don't have an entry
in /dev/. Try "ifconfig bcsf0" to see if the network device is there.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: nrparms command error with new ax25 tools

[Tomi Manninen OH2BNS]

On Thu, 21 Oct 1999, Arno Verhoeven wrote:

> > > Everything is looking pretty good except the nrparms portion of my startup
> > > script. When it executes an nrparms command directly in the script ... and
> > > when it runs the embedded nrparms command in the nodes.save file saved by
> > > the nodesave command, it reports the following error:
> > >
> > > nrparms: SIOCADDRT - invalid operation
> > 
> > A bug in nrparms. I'd recommend you get the RPMs from the Red Hat
> > Powertools 6.1 collection and try those. That particular bug is fixed
> > there.
> 
> Is there a diff available that fixes his problem?

It's now available at ftp.hes.iki.fi aka ftp.oh7lzb.ampr.org:

  ftp://ftp.hes.iki.fi/pub/ham/linux/ax25/ax25-tools-0.0.5-oh2bns.patch.gz

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Notebook and Linux/AX25.....

[Tomi Manninen OH2BNS]

On Mon, 11 Oct 1999 [EMAIL PROTECTED] wrote:

> On 10 Oct, richadr wrote:
> 
> > I want to buy a notebook, but it must be able to run Linux
> > (preferably SuSE 6.2). Anyone out there that can tell me
> > something about the possibillities and impossibillities 
> > or perhaps redirect me to the right news-group?
> 
> Richard,
> Just about any notebook should work fine.
> There is a Linux LapTop-HOWTO (http://www.linuxdoc.org/) that should
> provide hints and pointers to what to buy and what to watch out for.
> 
> The usual precautions apply: treat it as a PC with hardware that you
> can't easily change, make sure the hardware is supported.

And since we are talking about possible ham use, you might want to make
sure the soundcard is well supported in general and preferably by the
soundmodem too. I believe quite a few laptops have cs423x based soundcards
so one of them might be a good choise.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: TCP/IP config/connect problem

[Tomi Manninen OH2BNS]

On Sun, 3 Oct 1999, Elan Portnoy wrote:

> I have kernel based ax25 connects working ok. The trouble I am experiencing
> is with my TCP/IP connections. My IP is 44.68.48.216, the IP of the nearest
> node is 44.68.48.1. I used ifconfig to set ax0 with my IP and added the
> route to the 44.68.48.0 network. 
> 
> When I try a telnet, the transmitter quickly kicks on and off several times
> and I get "No route to host" error message. 

Most probably your ARP query times out. For some reason the new kernels
have very short timeouts. One solution is to enter a manual ARP entry for
the gw and then route everything through it. Another solution (in some
cases a better one) would be to fix the timeouts. Someone posted howto do
that on this list a few weeks ago. Maybe you can find it in the archive.

PS. Does anyone have a pointer to a document describing the
/proc/sys/net/ipv4/neigh/ interface?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Problems compiling ax25-utils-2.1.42a

[Tomi Manninen OH2BNS]

On Wed, 29 Sep 1999, Dirk Koopman wrote:

> I am confused. Are the new ax25 tools a complete replacement for 2.1.42a on
> all kernels (ie 2.0.35 upwards) or just 2.2.x kernels or 2.3.x kernels?

The lib/tools/apps are an almost complete replacement for the utils for
2.2.x kernels. Almost complete means that some packages that previously
were in utils are now distributed separately. Like node for instance.

Anyway mixing ax25-utils and libax25/ax25-tools/ax25-apps is asking for
trouble.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: AX.25 Layer 1

[Tomi Manninen OH2BNS]

On Tue, 28 Sep 1999, Dirk Koopman wrote:

> On 28-Sep-99 Samuel A. Falvo II wrote:
> > I'm looking for the physical characteristics for AX.25, as used on the VHF
> > bands.  Any ideas?  Would they be on tapr.org?  (Of course, I'll look there
> > anyway, but it doesn't hurt to ask.)
> 
> layer 1 = bell 202 tones.

Just to be sure we don't confuse anyone here: The AX.25 specification
makes no attempt at defining layer 1 issues. It is however true that on
VHF the most popular signaling method is AFSK with bell 202 tones (1200Hz
and 2200Hz) with signalling rate of 1200 bps. Also HDLC bit stuffing
and NRZI is used.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Problems compiling ax25-utils-2.1.42a

[Tomi Manninen OH2BNS]

On Tue, 28 Sep 1999, Jaime Robles wrote:

> Hello all!
> I am having BIG problems with the last ax25-utils...
> Since i update to linux 2.2.12 (from 2.0.36) my ax25 connection stopt
> working...
> I only used "call" to connect the DX-cluster, but for a week or less i
> have been "fighting" the PC and i have been unable to make the call work!
> 
> The problem i got when i try "make" in the ax-util directory is:
> 
> 
> gcc -Wall -Wstrict-prototypes -O2 -c -o procutils.o procutils.c
> procutils.c: In function `read_proc_ax25':
> procutils.c:36: `errno' undeclared (first use this function)
> procutils.c:36: (Each undeclared identifier is reported only once
> procutils.c:36: for each function it appears in.)

Umm... Are you trying to compile ax25-utils-2.1.42a? Forget that. Remove
everything that was installed by that in the past. Then get the latest
ax25-lib/tools/apps from ftp.hes.iki.fi, configure and install them as
described in the INSTALL file that comes with each package. (I recommend
using the "--prefix=/usr ..." stuff when configuring.)

> I start the ax25 with:
> 
> kissattach -m 512 -v /dev/tnc radio 44.133.228.22
^^^
> Then i get:
> 
> ea4abw:~/.pgp# tnc.bat
> kissattach: 0.0.5

That's the problem. The "-v" option causes kissattach to display the
version number and _exit_immediately_. So your kissattach line doesn't do
anything. Remove the "-v" and try again.

BTW. Judging from the kissattach version number you obviously have already
installed ax25-lib/tools/apps. Why are you trying to compile ax25-utils?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Antwort: Re: new kernels

[Tomi Manninen OH2BNS]

On Wed, 22 Sep 1999, Richard Adams wrote:

> Well the way in which i would do it, a way which has been flamed many
> times is the same way G4KLX once sent out a patch for, i have los the
> patch, but the idea is just change the default max backoff time from
> 120 seconds to what ever you want, i still use 60 for a link which
> plays up every now and again. I use 2.0.36, but looking at 2.2.x code
> its the same so i see no reason why it should nopt work.

I seem to remember Jonathan had a patch that really made the timers linear
(as I said in the other email). Anyway I have also used the method you
describe above a few times.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Patch for kernel 2.2.12

[Tomi Manninen OH2BNS]

Hello all!

I meant to send this last weekend but...

Anyway according to Alan kernel version 2.2.13 is still at least a couple
of weeks away, so I uploaded my patch to hes.iki.fi for people to test.
The patch can be downloaded at:

  ftp://ftp.hes.iki.fi/pub/ham/linux/ax25/linux-2.2.12-bns.diff.gz

or

  ftp://ftp.oh7lzb.ampr.org/pub/ham/linux/ax25/linux-2.2.12-bns.diff.gz

The patch is also already included in Alans linux-2.2.13preXX kernels (at
least it is in pre10).

This patch fixes the "unknown address family" bug that people have seen
with node and possibly other programs using the getpeername() and
getsockname() system calls. The patch should also finally once and for all
fix the "crash on ifconfig ax0 down" bug. If and when people are testing
this, please make sure you don't have any of the other "fixes" that have
been proposed to fix this one. I'd like people to verify that this really
is The Correct Fix (tm).

Finally there is a small change to the netrom loopback mechanism. If you
have ever experianced an oops or crash relating to this (if you've seen an
oops dump in your system log with nr_loopback_timer() mentioned) I'd like
to know about it.

PS. Note that much of the credit for finding the getpeername bug goes to
Cathryn Mataga. Thanks Cathryn!
-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Antwort: Re: new kernels

[Tomi Manninen OH2BNS]

On Wed, 22 Sep 1999 [EMAIL PROTECTED] wrote:

> Hi Tom , tnx for the reply, now I'm getting really confused, as I have
> already set the retry timers to linear, but they still behave as
> exponential timers. I was told that the kernel timers in 2.0.36 are
> exponential, Can this be clarrified ?

The AX.25 T1 (retry) timer has three modes in the linux implementation.
First is no backoff, second is linear backoff and third exponential.
(The numbers for the sysctl interface are 0, 1 and 2 respectively.)
AFAIK the timers work just fine both in 2.0.x and 2.2.x.

> But if you either telnet or ftp kernel to kernel over a not to good
> path, once a couple of packets have been missed the backoff time just
> get worse. I cant do much about the path , it got a big lump of iron
> ore in the way, it a great microwave reflector !!! but if I can
> improve the rate of retry attemps it would help..

You are talking about the TCP timers here. Nothing to do with the kernel
AX.25 implementation. That code is done by completely different people and
I doubt we have much chances of getting that changed. Everything except
exponential TCP timers is simply concidered evil in the real world.
(However Jonathan did a patch long time ago that set the TCP timers to be
linear. Jonathan called the patch something like "the evil sick TCP timer
patch...")

> the retry timers in tnos can be set to linear and when set are linear,
> the Tampa convers server, and the modified and cleaned up variant HTPP
> also behave linearly.

This however confuses me. Are you saying that you are using HTPP, a normal
Linux application, and TNOS timers have an effect on convers connections?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Kernel 2.2.12 select()/accept() broken ?

[Tomi Manninen OH2BNS]

On Tue, 14 Sep 1999, Cathryn Mataga wrote:

> >I had that problem on my system (RedHat 5.1, kernel 2.2.9). Because I
> >couldn't solve the problem, I decided to reinstall on a new harddisk the
> >latest RedHat with the 3 ax25 packages. Guess what? LinuxNode started to
> >work again perfectly fine.
> 
> Yup, the nature of this thing is that it's likely completely random in how
> or when the bug strikes.  I think 'node' was unusually succeptible, for
> reasons which are no fault of node at all.  Is someone going to upload
> a kernel patch on ftp.hes.iki.fi or radio.linux.org.au?  Err how is this
> kind of bug normally handled.

I have prepared a patch that fixes this (also for ROSE, NET/ROM was ok all
along) and some other stuff. I'll submit it to Alan today and ask him
about the release schedule for 2.2.13. If it takes very long I'll put the
patch available somewhere.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Kernel 2.2.12 select()/accept() broken ?

[Tomi Manninen OH2BNS]

On Tue, 14 Sep 1999, Jorge Matias wrote:

> On Mon, 6 Sep 1999, Tomi Manninen OH2BNS wrote:
> 
> > This sounds very similar to the problem _some_ people have had with
> > LinuxNode on 2.2.x kernels. The most visible symptom of the bug is that an
> > incoming AX.25 connection to the node fails immediately and a message
> > about unsupported address family is logged in the syslogs.
> 
> I had that problem on my system (RedHat 5.1, kernel 2.2.9). Because I
> couldn't solve the problem, I decided to reinstall on a new harddisk the
> latest RedHat with the 3 ax25 packages. Guess what? LinuxNode started to
> work again perfectly fine.

This bug has now been fixed. The bug is in the kernel and hopefully
2.2.13 will contain the fix.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: kissattach : cannot open axports file

[Tomi Manninen OH2BNS]

On Mon, 13 Sep 1999, Jose Angel Amador Fundora wrote:

> Weeellsince I jumped to RedHat, I have had to make a symlink from 
> a fake /axports to the real /etc/ax25/axports for it to work.
> 
> Somebody explained that it is a bug somewhere in RedHat or the ax25-
> utils that the proper directory is not properly referred to during 
> compile time, and so, it goes to / looking for axports. My memory 
> fails to remember the details.
> 
> The symlink was the easy way out for me, before I knew the above 
> explanation. I was in a hurry and didn't mend the cause when it 
> was posted to this list.
> 
> Could we get a refresh of what is wrong with the sources by whom 
> knows it better ?

What utils are you talking about? Version info tnx.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Kernel 2.2.12 select()/accept() broken ?

[Tomi Manninen OH2BNS]

On Wed, 8 Sep 1999, Craig Small wrote:

> Tomi Manninen said:
> > It seems the interface to getname has changed a bit. From that code I
> > deduce that &len is now supposed be _filled_in_ by sock->ops->getname(). 
> > At least inet_getname() does that. And it's now move_addr_to_user() that
> > handles truncating the address struct if needed. 
> > 
> > Now it should only be a matter of a small adjustment to ax25_getname(),
> > nr_getname() and rose_getname() and we are back to business...
> 
> Did this change from 2.2.11 to 2.2.12?  What the hell were they thinking
> changing this in a stable kernel?  Isn't this the whole point of 2.3.x
> kernels.

No, I think it was changed quite some time ago, in the 2.1.x series. The
reason we haven't noticed it is that as len is uninitialized, everything
works if it just happens to be greater than the size of the struct and it
seems that happens surprisingly often...

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: ax25_get_info() - TNT&BayCom modem

[Tomi Manninen OH2BNS]

On Mon, 6 Sep 1999, Jan Wasserbauer wrote:

> Is there any reason why linux/net/ax25/af_ax25.c - ax25_get_info()
> disables interrupts ?

Probably because ax25_get_info reads kernel internal tables that could be
changed from an interrupt. You would crash the kernel nicely if you
happened to be just reading an ax25 control block that is suddenly
removed from an interrupt...

> This causes problems with BayCom modem (and possibly others) since
> its driver cannot receive interrupts from serial port while
> something is reading /proc/net/ax25. TNT reads it every 300ms
> which makes it unusable with BayCom modem (255B packets just can't
> get into 300ms on 1200bd)

IMHO TNT is broken in this case. Why does it need to read it so often? Ah,
don't tell me, the DOS version shows that info too...? ;)

> PS: strerror() returns inet-style link state messages which are
> sometimes quite confusing to PR users (like "Connection timed out").

Why is that confusing? Timed out _is_ timed out. 

> What about some ax25_strerror() (in libax25 probably) which
> would return PR-style warnings (like "Link failure") ?

This on the other hand is under concideration already. It is possible
Craig has already written some code, I don't know. Anyway it will be only
error messages that are specific to libax25, eg. "Invalid callsign", "Port
xxx not active" etc.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Kernel 2.2.12 select()/accept() broken ?

[Tomi Manninen OH2BNS]

On Mon, 6 Sep 1999, Jan Wasserbauer wrote:

> Hi everyone on the list ..
> 
> Part of checklogin() function:
> ...
>   tv.tv_sec = 0;
>   tv.tv_usec = 0;
> 
>   if (select (maxsock + 1, &rfds, NULL, NULL, &tv) <= 0)
> return;
> 
>   for (i = 0; i < listen_sock_num; i++)
>   {
>   if (!FD_ISSET (listen_sock[i], &rfds))
>  continue;
> 
>   addrlen = sizeof (struct full_sockaddr_ax25);
>   ret = accept (listen_sock[i], (struct sockaddr *) &newconn, &addrlen);
> ...
> 
> This does work OK with any kernel below 2.2.12.
> 
> With 2.2.12 accept() does not correctly fill sockaddr struct
> when tv.tv_sec=tv.tv_usec=0 (some binary mess instead of callsigns)
> It works when tv is nonzero.
> 
> Reffering to select() man page tv=0 should cause select
> to check descriptors and exit immediately (and not just
> break accept() :) )
> To me it seems like kernel bug ..

This sounds very similar to the problem _some_ people have had with
LinuxNode on 2.2.x kernels. The most visible symptom of the bug is that an
incoming AX.25 connection to the node fails immediately and a message
about unsupported address family is logged in the syslogs.

Also seems that in these cases ax25d logs the connection as coming from
port (null).

At least in node the problem seems to be that a
getpeername(s, &addr, &len) system call returns with len==0 and bogus data
in the addr struct. The ax25d problem could be similarly failing
getsockname() but I haven't been able to verify that (see below).

It's noteworthy that this has happened to different people with different
2.2.x kernels, with different ax25d and node versions, different
distributions with different libc... For one guy this all suddenly went
away as he upgraded to RH6 but I've had reports on that platform too...

It's also noteworthy that I haven't been able to duplicate this error EVER
on ANY kernel version, including 2.2.12. This makes debugging very
difficult... :(

> My system - Slackware 4.0 (gcc2.7.2.3, libc5.4.46)
> 
> If this is known bug I hope this at least saves some time
> to people wondering why their software doesn't work with 2.2.12.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: ax25ipd errors out

[Tomi Manninen OH2BNS]

On Thu, 2 Sep 1999, Peter Grace wrote:

> device /dev/ttyS0  (error was 'fetching tty device parameters: Invalid
> argument')

Don't know whats wrong here. Oh, except in the script below you have
already used ttyS0 for kissattach. You can't of course have two processes
use the same serial port at once.

> device /dev/ttyp0 (I/o error)
> and also /dev/ttyq0 (i/o error)

These are both slave ends of a pty/tty pair. I guess you don't have
anything in the master end (I don't see anything in the script below).
That would cause exactly the error message you get.

In case you didn't know the pty/tty pair (or pipe) has two ends that are
not quite equal. The pty one (eg. /dev/ptyp0) is the master and it needs
to always be opened first. Also if the process that opened the master end
dies or otherwise closes the master end, the slave end will instantly get
an I/O error.

So what you seem to miss is a kissattach process opening and doing the
kiss stuff at the master end. After that you can start ax25ipd.

> /usr/sbin/kissattach -i 44.56.20.25 -m 512 /dev/ttyS0 radio
> /usr/sbin/axparms -assoc kb1cvh pgrace
> /sbin/route add -host 44.56.20.0 ax0
> /sbin/route add ax0
> /usr/sbin/kissparms -p radio -t 300 -s 300 -r 25
> /usr/sbin/nrattach -i 44.56.20.25 -m 512 netrom
> route add 44.56.20.25 nr0
> /usr/sbin/ax25ipd
> /usr/sbin/ax25d
> /usr/sbin/netromd -i -t 120
> /usr/sbin/mheardd
> /usr/sbin/nrparms -routes radio K1CF-1 + 120
> /sbin/route add 44.56.14.11 nr0
> /sbin/arp -t netrom -s 44.56.14.11 K1CF-1
> /usr/sbin/nrparms -nodes KA1TUZ + BBSTUZ 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes KA1TUZ-11 + "*" 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes WZ1L + BBSWZL 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes WB1DSW-2 + EKINGS 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes W1UU + MASS 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes WZ1L-5 + POWOW 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes K1TR-10 + WHDHM 120 5 radio K1CF-1
> /usr/sbin/nrparms -nodes K1UGM + BBSUGM 120 5 radio K1CF-1

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: New lib,tools and apps

[Tomi Manninen OH2BNS]

On Thu, 26 Aug 1999, Craig Small wrote:

> Bob Meyer said:
> > Craig Small wrote:
> > >   The latest lib, tools and apps for ax25 are out.  The main difference
> > > between this version and the last is I finally got my act together and
> > > put Tomi's patches in.
> > >
> > 
> > Call is broken here, on Redhat6.  Executing any of the menu commands causes
> > a "Segmentation fault".

> This is very strange because the patch changes one value to be static
> and the rest is just cleaning up of the configure script so it detects
> broken redhat includes properly. Don't tell me those library/includes
> are broken in yet another way.

Actually not even that. The patch removes a compiler warning by explicitly
declaring the variable "int". Before it was implicitly declared "int" by
the compiler. So that can not be the problem.

My guess is there might a problem with ncurses. Red Hat 6 might have a
version on ncurses that call isn't quite compatible with. If I understood
Richard correctly this segfault has been there in previous versions too?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Is Netrom/node ok in 2.2.5 ?

[Tomi Manninen OH2BNS]

On Tue, 24 Aug 1999, pa3gcu wrote:

> I have Redht 6.0 libax25-0.0.5 ax25-apps-0.0.3 ax25-tools-0.0.4
> installed following the INSTALL and README, however NETROMD gives;
> Aug 19 20:16:55 pa3gcu netromd[]: netromr: SIOCADDRT: Invalid argument
> I have a bpq interface between two computers.
> 
> Installing the netromd from ax25-utils-2.1.42a solves this problem.

Try installing the new libax25, ax25-tools and ax25-apps that Craig
announced just today and tell us if that fixed the problem. It should.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: listen utility and permissions

[Tomi Manninen OH2BNS]

On Thu, 19 Aug 1999, Ronnie Hale wrote:

> I'm with you on this too, Gerd.   Fortunately, there exists such a program,
> TNT.  TNT is similar to SP (Eskay Packet), and while it works as a front-end
> to DPBOX, it is also an excellent stand-alone packet terminal program.  It
> doesn't have the million or so remote features as SP, but it isn't as
> complicated to set up.  It will work with Linux kernel/ax25 ports, hostmode
> tnc, kiss tnc (via tfkiss driver). It has //mheard (selectable) for remote
> users, real-time monitor window in terminal (like SP 3-way split screen),
> F11 monitor channel, huffman encode/decode in terminal qso and monitor,
> Extended monitor (to monitor a specific callsign qso).  Lots of goodies us
> old-timer packet operators enjoyed when using Stupid programs like SP, GP,
> etc.

Did I say these programs are stupid? No, I didn't. I said the platform
they run on sucks and thus required them to be designed as they are. That
doesn't automatically make the desing good or bad. I'm just slightly
annoyed everytime someone goes "it works this way on DOS, why the heck
does it work differently on Linux."

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: listen utility and permissions

[Tomi Manninen OH2BNS]

On Fri, 20 Aug 1999, Gerd wrote:

> Hmm *thinking*...
> I thought you need root permissions in order to open the socket? 
> Did you have a version of listen that used to have only that part for 
> opening the socket set to UID root? 
> How about two different versions of a listen utility? So it would 
> depend on the remote server's admin whether stupid users are able 
> to trash the links or not :)

I think you are missing the point. In the new ax25-tools nothing has been
changed in listen (well some formatting changes have been done but
anyway). It's just that "make install" doesn't set the setuid bit anymore.
Craig gave the explanation why and I think it makes perfect sense.

BUT. You have the root password so you can do whatever you want. If you
want everyone to be able to "listen" then just set the setuid bit. By
playing with the file ownerships and permissions you can even make it so
that only certain regular users can "listen". This is Linux, all that is
possible.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: listen utility and permissions

[Tomi Manninen OH2BNS]

On Wed, 18 Aug 1999, Gerd wrote:

> Hello Tomi, hello all,

> > I don't see the need for such an elaborate arrangement, at least if the
> > only reason was to replace the current "listen". Listen has no user
> > interaction and it should be pretty easy to make sure it has no buffer
> > overflows or other security risks. So there really isn't any reason a
> > sysop couldn't make it setuid root if he/she wants to. It's just that now
> > it isn't setuid root by default.
> 
> A thing that should be mentioned in the HOWTO, though.

Of course.

> But I was thinking a little bit further. Packet Radio terminal 
> programs that want to be able to monitor and show the traffic on 
> the QRG (similar to let's say Eskay Packet under DOS) have to 
> open and establish interaction with the same socket as the listen 
> program.

Ahem, well, the reason Eskay Packet under DOS or similar systems under
DOS/Windows have a band monitoring tool built in is that on those
(stupid) platforms something as simple as "listen" is not possible. IMNSHO
there is _no_ reason to duplicate that kind of design in Linux. Linux is a
good multitasking operating system and the networking model nicely permits
running these things in separate programs that are well designed for their
purpose.

You won't run out of virtual consoles on Linux unless you are doing
something really weird to occupy all of them. With X there is of course no
limit in that sense.

> In another mail Heikki Hannikainen <[EMAIL PROTECTED]> mentioned
> setting only certain program parts UID root since this is only
> required to _open_ the socket. Is this at all possible? Until now, I
> thought that only whole programs could be set to run under UID root.

Yes, it is possible. I recommend reading the man pages Hessu mentioned and
the books too.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Problems with ax25-tools-0.0.4 and Kernel 2.2.11

[Tomi Manninen OH2BNS]

On Wed, 18 Aug 1999, Robert Schelander wrote:

> I've tried to compile ax25-tools-0.0.4
> './configure' worked without any problems, but
> when I try to 'make install' I get the following message:
> 
> axctl.c:93: structure has no member named 'digi_count'
> 
> what could be the reason?

This has been answered quite a few times here already...

You probably use Red Hat 6.0. Currently the tools don't compile on that
platform. A new version of the libs/tools/apps is on it's way.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: listen utility and permissions

[Tomi Manninen OH2BNS]

On Wed, 18 Aug 1999, Gerd wrote:

> So it seems there must be a solution like the following available to 
> avoid setting the whole terminal program UID root:
> 
> On the socket listens a daemon with UID root that could be started 
> upon system boot or with sudo or something else to assure its root 
> permissions. The terminal then must be able to talk to this daemon 
> to get information about the actual traffic on the QRG.
> 
> Is this a suitable approach or do I miss something here?

I don't see the need for such an elaborate arrangement, at least if the
only reason was to replace the current "listen". Listen has no user
interaction and it should be pretty easy to make sure it has no buffer
overflows or other security risks. So there really isn't any reason a
sysop couldn't make it setuid root if he/she wants to. It's just that now
it isn't setuid root by default.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: soundmodem tcp/ip point2point

[Tomi Manninen OH2BNS]

On Sun, 15 Aug 1999, Bernd Thallner wrote:

> i want to make a point to point connection via soundmodem.
> i can ping the points/computers but no one is answering.
> in smdiag i can see the signal.
> 
> my configuration:
> 
> -- computer a kernel 2.2.10
> insmod ax25
> insmod hdlcdrv
> setcrystal -s 0x220 -i 5 -d 1
> insmod soundmodem
> sethdlc -p -i sm0 mode sbc:afsk1200 io 0x220 irq 5 dma 1 pario 0x378
> ifconfig sm0 hw ax25 OH2BNS-1 up
> ifconfig sm0 192.42.172.11
> smmixer -i sm0 i=0 o=0 s=line
> route add -net 192.42.172.0 sm0
> sethdlc -i sm0 -a txd 500 slot 50 ppersist 128
> axparms -route add 1 OH2BNS-2
> arp -H ax25 -i sm0 -s 192.42.172.12 OH2BNS-2
> 
> /etc/ax25/axports:
> 1   OH2BNS-11200255 2   144.675 MHz (1200  bps)
> 
> --
> what could be the problem?
> did i forget something?

The problem is you are using my callsign in your experiments. Stop that
IMMEDIATELY !

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: sucess story! upgrade from RH4.2 -> RH6.0

[Tomi Manninen OH2BNS]

On Mon, 16 Aug 1999, Thomas M. wrote:

>  To get this far, I've had to patch a couple of files :
> 
>  in libax25 kernel_netrom.h is missing the definition of SOL_NETROM, after
> browsing some sources i found out that this should be 259, so I added a
> #define SOL_NETROM 259 to the file...

This one is ok. Though there are other errors in that file so you will
find that many netrom tools won't work.

>  in ax25_tools, the axctl program has an error, i removed the line near
> the bottom of the file, saying ax25_ctl.digi_count, as digi_count doesn't
> appear in the ax25_ctl struct...

There is no error in axctl. Zeroing ax25_ctl.digi_count is crucial to
axctl's function, in fact with kernels older than 2.2.11 you can crash the
kernel big time if it's not properly initialized...

The problem is that the kernel_*.h "compatibility headers" are broken.
Patch sent. Waiting for something to happen.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: ANNOUNCE: dev-hams mailing list

[Tomi Manninen OH2BNS]

On Thu, 12 Aug 1999, Terry Dawson wrote:

> On Wed, Aug 11, 1999 at 08:22:35PM +, Julian Munoz Dominguez wrote:

> > So it's the freedom of the subscribers of this mail-list to keep on
> > beeing a closed group (less closed that before, but still closed)...
> >  Personally I am sure that I would have learn a lot of interesting 
> > things reading the content of this list. Don't forget the eductional
> > point of view.
> 
> It is for this reason that we discussed the notion of archiving it and
> making it publicly readable. I'm not sure what happened to that idea.

Quoting Hessu's original announcement of the dev-hams list:

  Following recent discussion it was decided to publish the existence of
  the list, and make the mailing list archive available for anyone to
  read.

  
  
  The mailing list archive can be found at:
  http://hes.iki.fi/archive/dev-hams/

I really wish the people complaining about this evil secret society of
ours would read the announcement a bit more carefully.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Compiling new AX25 stuff on Slackware 3.6

[Tomi Manninen OH2BNS]

On Tue, 10 Aug 1999, Gareth Rowlands wrote:

> The slackware 3.6 has been augmented with Kernel 2.2.10 and net-tools 1.52 
> The glibc already supplied with the distribution is 2.0.7

> ax25-tools-0.0.4
> 
> 
> Stops compiling with an error relating to a structure member in axctl.c
> line 93
> 
> ax25_ctl.digi_count = 0;
> 
> Within the file ax25-tools-0.0.4/ax25/axctl.c  I just commented out this
> particular line for now, as I couldn't work out the specifics of the
> problem (digi_count belongs to the structure ax25_routes ?)
> 
> /* ax25_ctl.digi_count = 0; */
> 
> Time will tell as to whether I've made a duff move.

Unfortunately with that "fix" axctl won't work and with kernels earlier
than 2.2.11 one can produce a major kernel screw-up with it... :(
(The AX.25 related security update mentioned at
http://www.uk.linux.org/VERSION/relnotes.2211.html fixes the latter.)

The problem is a little more complicated and the real fix needs to be done
in the "compatibility headers" that libax25 installs. I have sent patches
for libax25 to Craig so we'll just have to wait...

> ax25-apps-0.0.3
> 
> 
> Compiles and installs OK.  If you have a 'heritage' installation like
> mine, you may find that 'call' and 'listen' etc. complain of being
> unable to open the axports file
> 
> Within the directory  /usr/local/etc  set a link to  /etc/ax25
> 
> ln -s /etc/ax25 ax25

I'd suggest you instead do the following:

  1. Run "make uninstall"
  2. Run "./configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var"
  3. Run "make" and "make install".

You do this to all the three packages starting with libax25 and suddenly
you have all the stuff in the "right" places... The "make uninstall" is of
course needed only if you have already installed the packages to
/usr/local.

Hope this helps.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Compiling new AX25 stuff on Slackware 3.6

[Tomi Manninen OH2BNS]

On Thu, 12 Aug 1999, Craig Small wrote:

> Kai Altenfelder said:
> > open("/usr/etc/ax25/axports", O_RDONLY) = -1 ENOENT (No such file or
> >   ^ 
> [...] 
> > Where the hell does this prefix "/usr/etc" come from? An error of the
> > libax25 function "ax25_config_load_ports()"?

> Yep, so you can put whatever you like for the ax25-tools compile and it is 
> not going to make one bit of difference. 
> Recompile libax25.

Kissattach is different in that it doesn't use libax25 to read the config
but has its own readconfig() routine. I really don't know why but it does.
Kissattach tries to fopen(CONF_AXPORTS_FILE, "r")) and CONF_AXPORTS_FILE
is defined in "../pathnames.h":

  #define CONF_AXPORTS_FILE AX25_SYSCONFDIR"axports"

AX25_SYSCONFDIR on the other hand is defined from gcc command line so you
should see what it is when you compile the stuff.

My guess is something went wrong when you ran ./configure. Have you tried
starting over with freshly unpacked source code? Autoconf can be confusing
sometimes as it caches all sorts of information.

Just for reference, kissattach works for me and reads the correct config
file. I just checked yesterday evening.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: 6pack: spattach won't die !

[Tomi Manninen OH2BNS]

On Mon, 9 Aug 1999, Andreas Koensgen wrote:

> I've tested the problem on the weekend and could reconstruct it.
> I've discovered even another error -- when killing spattach
> after a TCP connection has been opened the kernel panics with
> the messages "killing interrupt handler" and "unable to kill the idle
> task". I think the problem is that the driver code which interfaces the
> kernel is very old -- it has been taken from the SLIP driver for 2.0.x
> kernels, later minor changes were done to get it compiled on 2.1.x
> kernels. But meanwhile so many things in the networking code have been
> changed that it is probably easier if I take the kernel interface code
> from the latest 2.2.x mkiss driver and replace/add the 6pack-specific
> routines. This will also be a good opportunity to fix the old
> documentation attached with the driver ...

I think the mkiss driver may suffer from the same problem. I once had
similar behaviour with kissattach but didn't investigate it further...

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Virtual Netrom circiut/Ip over Netrom

[Tomi Manninen OH2BNS]

On Sun, 8 Aug 1999, Glasscock Family wrote:

> I haven't seen anything on this so here is my question: I have the
> standard ax25 tools, version 2.1.42 or whatever that is, and I am
> trying to set up a link to someone using TCP/IP carried over netrom,
> for the netrom man page: IP traffic may be stacked on top of NET/ROM
> frames using a non-standard extension to the NET/ROM protocol. I am
> wondering how do I set it up.

Ok. This is from memory so some details might be wrong. But I think it
should get you going. After this script your end should be ok. Then you
have to get the #ALW router to do the same so it can route any reply
IP datagrams to you. A route and arp needs to be set up there.

# Set up the netrom interface. (Assuming you haven't already done that)
#
nrattach -i 44.12.0.144 -m 236 netrom_port_name_as_in_nrports
#
# Route the gw through the nr0 interface.
#
route add -host 44.12.3.131 window 392 dev nr0
#
# NET/ROM doesn't support automatic ARP so this has to be done manually.
# You will have to replace "#ALW" with the real callsign, alias won't work.
#
arp -H netrom -s 44.12.3.131 #ALW
#
# The rest of the world is behind the router.
#
route add -net 44.0.0.0 netmask 255.0.0.0 gw 44.12.3.131
#
# All done!

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: Redhat aout -> tpp

[Tomi Manninen OH2BNS]

On Thu, 5 Aug 1999, Kjell Jarl wrote:

> RH 6.0 and tpp (convers server).
> I have no lib /usr/lib/gcc-lib/i486-linuxaout/egc-2.91.66/
> thus, tpp will not compile.
> Any sugestion to this?
> Would any one's else binary run here?

Try HTPP at .

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

RE: radio.linux.org.au

[Tomi Manninen OH2BNS]

On Thu, 5 Aug 1999 [EMAIL PROTECTED] wrote:

> > is § a defined SGML/HTML entity? I'm betting if we renamed the
> > variable to something else it wouldn't happen.
> > 
> > I'm betting also it's a browser bug.

That would make sense. My dictionary describes § as section mark...
Clearly a browser bug.

> Try it now, I reordered the variables so § doesn't appear.
> If my hunch is correct it will work ok now.

Works fine now. Thanks!

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

RE: radio.linux.org.au

[Tomi Manninen OH2BNS]

On Thu, 5 Aug 1999, BBAGGER wrote:

> I didn't get it when I looked at PB/PG for Linux.

Hmm... It seems to be dependent on he browser version. Just tested with
Nescape 4.61 (Linux) and viewing the document source (the first item on
the first page):

  7plus/FBB

This link works without any problems. But with Netscape 3.04 Gold (HP-UX):

  7plus/FBB

This gives the error I mentioned before. I don't know what's going on but
maybe this can be fixed somewhere...?

Other than that radio.linux.org.au looks very nice! Thanks Joop and Terry.

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

radio.linux.org.au

[Tomi Manninen OH2BNS]

Hi!

Am I the only one that gets an error message when trying to look at the
package description:

  1064: You have an error in your SQL syntax near '§pat=packet' at line 1
  Warning: 0 is not a MySQL result index in /home/pa4tu/radio/pkgdetail.phtml on line 
22

Manually changing the '§' in the URL to a '&' fixes th problem. A typo
somewhere?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---

Re: soundmodem on SBPro

[Tomi Manninen OH2BNS]

On Wed, 4 Aug 1999, Thomas Sailer wrote:

> Tomi Manninen OH2BNS wrote:
> > 
> > On Wed, 4 Aug 1999, Hamish Moffatt wrote:
> 
> > > Do you think it is a problem to operate with the squelch open always?
> > > The soundmodem seems to do a pretty good carrier detect anyway.
> > 
> > As far as I can see, using squelch with soundmodem can only make things
> > worse. It _will_ make carrier detection slower and possibly also receiver
> > recovery (from tx). Both are very important factors for efficient channel
> > usage. False carrier detection is much less dangerous. CPU usage might be
> > slightly lower with squelch but I doubt you can even notice that.
> 
> On the other hand, if a hardware squelch decides wrongly, the packet is
> trashed. If the software squelch (or rather carrier detect) errs,
> the packet reception is not affected.

Umm... Isn't that just another reason not to use squelch (or rather run
squelch open). Did I miss something?

> So it depends. If you have a very fast hardware squelch (preferably
> that operates on the RF power level), then use it, but I doubt that
> most FM rigs have anything but an extremely slow squelch.

But in the case of soundmodem, the carrier detection is done in software
anyway and any kind of squelch, no matter how fast, can only make the DCD
slower. Right?

-- 
--- Tomi Manninen / [EMAIL PROTECTED] / OH2BNS @ OH2RBI.FIN.EU ---