* Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
@ 2009-01-18 8:22 Curt, WE7U
2009-01-18 21:08 ` Ken Koster
2009-01-19 21:49 ` Curt, WE7U
0 siblings, 2 replies; 11+ messages in thread
From: Curt, WE7U @ 2009-01-18 8:22 UTC (permalink / raw)
To: linux-hams; +Cc: archer
I saw several bugs in the AX.25 stuff yesterday/today. First let me
say that I've configured AX.25 quite a few times on several systems
but haven't done played with "ax25d" or "axspawn" in more than ten
years. I've done ax.25 KISS configurations and ax.25 soundmodem and
serial KISS soundmodem configs to test/use Xastir much more recently
and without a hitch.
So... I get AX.25 kernel networking configured and working. I use
my old 44.x IP address. No problems so far. "listen" is decoding
packets, I can "ping" and/or use "call" and I see packets go out.
So far so good. I also set up Xastir on that port and it is
sending/receiving just fine over the AX.25 kernel networking port.
It's attached to a TNC-X TNC via a USB->Serial cable.
Now I start playing with "/etc/ax25/ax25d.conf" and
"/etc/ax25/axspawn.conf". I refer to the AX.25 HOWTO (yea, it was
never up-to-date back then either, so I don't expect an easy go of
it). I also consult two friends locally who did TCP/IP with me
years back. They have great bits of knowledge and one helped with
on-air testing tonight (Hi Bob/Ken!).
To the meat of the matter:
--------------------------
1) First scenario: Specifying "create_with_useradd yes" in
"axspawn.conf" to auto-create users when they first log in doesn't
work. The account gets created just fine with LOTS of dot files in
it (apparently all the dot files "useradd" would normally stick in
there for a regular user). There's an entry created in
"/etc/passwd" and "/etc/shadow" but the "useradd" executable stays
running.
The xterm where I earlier started "ax25d" starts spewing "waitpid"
errors and goes to 75% CPU usage ("waitpid (15185) failed: No child
process"). I kill the xterm and "useradd" goes to 92% CPU usage.
I kill "useradd" and things go to normal except for an empty
password for the new user, so he can't log in. The user gets
disconnected by the Linux box. After root changes his password he's
able to log in at the point.
2) Second scenario: Specifying "create_with_useradd no" in
"axspawn.conf". In this case there are fewer files created in the
new account (two dot files) but the new account creation step goes
without a hitch.
Except: There's still no password for the user so the root user has
to set it, and GET THIS: The encrypted password string ends up in
"/etc/passwd" instead of "/etc/shadow"! Do the AX.25 programs know
about and check "/etc/shadow" for the encrypted password strings?
3) The next problem: Information that the user has requested seems
to only send the last 256 bytes or less across RF. The earlier
stuff just disappears. For instance he did this:
cd /usr/bin
ls
and got one line sent to him containing two filenames starting with
'z'. I have 7275 files in that directory. I watched the packet go
out with "listen" and nothing was sent before that.
4) Oh yea, when the remote user types "exit" I see a stack trace and
memory dump from "axspawn" on the xterm I started "ax25d" from.
Doesn't sound like a nice exit.
5) There were also cases, I believe right after the account was
created but before the password was set by root, that the remote
user got odd messages from "login" about not specifying the correct
number of parameters. After the password was set by root these
messages went away. Yea, documented below.
6) The remote user got messages about ".bcpasswd" not being present.
I don't know what that was about, but I see that a
"/etc/ax25/bcpasswd" file was touched by the system and I didn't
edit that one by hand. It mentions "$HOME/.bcpasswd".
OpenSuSE 11.0 32-bit, all the updates applied. 2.6.26.18 kernel.
AX.25 RPM's installed from OpenSuSE-11.0 ham repository. AX.25
stuff loaded as modules:
Module Size Used by
af_packet 38656 4
mkiss 28292 1
ax25 74728 7 mkiss
crc16 18432 1 mkiss
Configs available on request.
Here's a few notes from the latest attempts:
---------------------------------------------------------------------
Useradd YES:
------------
xterm at 75% CPU printing:
waitpid (15185) failed: No child process
Killed xterm
useradd at 92% CPU.
kill 15184
kill 15184
kill -9 15184
Bob got this:
"Notice: No .bcpasswd file found in your homedirectory (for more
secure password authentication than plaintext). Generating example
file, with unique passwords (which may be changed). Please edit
~/.bcpasswd, and enable your prefered authentication type; MD5 is
recommended."
Passwd line:
kd7nm:x:2001:1000:kd7nm:/home/ax25/kd7.../kd7nm:/bin/bash
Shadow line:
kd7nm:+:14262:0:99999:7:::
Blank login didn't work.
Login incorrect
wapiti2 login:
Root changed his password to "kd7nm".
He logged in.
Passwd line is the same, Shadow line is now:
kd7nm:$2a$10$KNDr8D1fMk92vie7/T6wd.g35q4yhp4aBA4bojFJ8plAtw60Zstdu:14262:0:99999:7:::
---------------------------------------------------------------------
Useradd NO:
-----------
Account created, user disconnected right away. Two files in user
account: .bcpasswd and .profile
Passwd line:
kd7nm:+:2001:1000:kd7nm:/home/ax25/kd7.../kd7nm:/bin/bash
No Shadow line.
User cannot connect. Says:
"login: Too many arguments
Try `login --help' or `login --usage' for more information."
and then immediate disconnect.
Root changed his password to "kd7nm".
He logged in.
Passwd line:
kd7nm:$2a$10$djAnco8ODzBCug0lWYVareg1IjVgf3SzrFhltBeN9daYnIwYaLqEC:2001:1000:kd7nm:/home/ax25/kd7.../kd7nm:/bin/bash
No Shadow line.
"Last login: Sat Jan 17 23:53:34 PST 2009 from AX.25 on pts/20
Have a lot of fun...
kd7nm@wapiti2:~>"
---------------------------------------------------------------------
Comments?
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-18 8:22 Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit Curt, WE7U
@ 2009-01-18 21:08 ` Ken Koster
2009-01-19 21:49 ` Curt, WE7U
1 sibling, 0 replies; 11+ messages in thread
From: Ken Koster @ 2009-01-18 21:08 UTC (permalink / raw)
To: Curt, WE7U; +Cc: linux-hams
On Sunday 18 January 2009, Curt, WE7U wrote:
...
> Except: There's still no password for the user so the root user has
> to set it, and GET THIS: The encrypted password string ends up in
> "/etc/passwd" instead of "/etc/shadow"! Do the AX.25 programs know
> about and check "/etc/shadow" for the encrypted password strings?
Why I didn't remember this as we played last night I don't know, but it
never did support shadow passwords.
I remember adding in support for the WETNET network but I can't recall
if I ever submitted the changes. (hey it was over ten years ago after all:-)
Obviously I must not have.
I'll dig thru my archives and see if I can remember how I did it. If I recall
correctly it wasn't all that hard.
Ken
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-18 8:22 Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit Curt, WE7U
2009-01-18 21:08 ` Ken Koster
@ 2009-01-19 21:49 ` Curt, WE7U
2009-01-19 22:41 ` Ray Wells
2009-01-23 18:19 ` Curt, WE7U
1 sibling, 2 replies; 11+ messages in thread
From: Curt, WE7U @ 2009-01-19 21:49 UTC (permalink / raw)
To: linux-hams
On Sun, 18 Jan 2009, Curt, WE7U wrote:
> I saw several bugs in the AX.25 stuff yesterday/today.
An update: I figured newer versions of the AX.25 bits in the SuSE
ham repository but my update methods were not snagging them. The
SuSE update methods need a bit of revision as neither "Online
Update", nor "zypper ref; zypper up --type package" gets everything.
The only method that does (thanks Ken!) is Software Management ->
Package -> All Packages -> Update if newer version available.
So... I need to repeat my on-air tests with Bob using the newer
AX.25 modules and tools to see if any or all of the bugs reported
are still evident. Stay tuned.
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-19 21:49 ` Curt, WE7U
@ 2009-01-19 22:41 ` Ray Wells
2009-01-21 17:35 ` Curt, WE7U
2009-01-23 18:19 ` Curt, WE7U
1 sibling, 1 reply; 11+ messages in thread
From: Ray Wells @ 2009-01-19 22:41 UTC (permalink / raw)
To: Curt, WE7U; +Cc: linux-hams
Curt, WE7U wrote:
> On Sun, 18 Jan 2009, Curt, WE7U wrote:
>
>> I saw several bugs in the AX.25 stuff yesterday/today.
>
> An update: I figured newer versions of the AX.25 bits in the SuSE
> ham repository but my update methods were not snagging them. The
> SuSE update methods need a bit of revision as neither "Online
> Update", nor "zypper ref; zypper up --type package" gets everything.
> The only method that does (thanks Ken!) is Software Management ->
> Package -> All Packages -> Update if newer version available.
>
> So... I need to repeat my on-air tests with Bob using the newer
> AX.25 modules and tools to see if any or all of the bugs reported
> are still evident. Stay tuned.
>
Are you guys familiar with the work being done by Bernard, F6BVP, on
kernel ax25 related stuff? Some of his patches have been included in
later kernels.
Bernard's source files for libax25, ax25-apps & ax25-tools are available
here
http://f6bvp.free.fr/logiciels/ax25/
Ray vk2tv
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-19 22:41 ` Ray Wells
@ 2009-01-21 17:35 ` Curt, WE7U
2009-01-22 9:11 ` John Ronan
0 siblings, 1 reply; 11+ messages in thread
From: Curt, WE7U @ 2009-01-21 17:35 UTC (permalink / raw)
To: Ray Wells; +Cc: linux-hams
On Tue, 20 Jan 2009, Ray Wells wrote:
> Are you guys familiar with the work being done by Bernard, F6BVP, on kernel
> ax25 related stuff? Some of his patches have been included in later kernels.
>
> Bernard's source files for libax25, ax25-apps & ax25-tools are available here
>
> http://f6bvp.free.fr/logiciels/ax25/
Now that you mention it, yea, I've seen stuff come across this list.
At the time I wasn't interested in playing with AX.25 right then, so
I skipped over most of it. Now that I'm playing with it in order to
develop good config files to hand to someone else, I'm interested
again. Funny how that works.
Thanks for the info!
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-21 17:35 ` Curt, WE7U
@ 2009-01-22 9:11 ` John Ronan
2009-01-22 16:51 ` Curt, WE7U
0 siblings, 1 reply; 11+ messages in thread
From: John Ronan @ 2009-01-22 9:11 UTC (permalink / raw)
To: Linux-Hams
On 21 Jan 2009, at 17:35, Curt, WE7U wrote:
> On Tue, 20 Jan 2009, Ray Wells wrote:
>
>> Are you guys familiar with the work being done by Bernard, F6BVP,
>> on kernel ax25 related stuff? Some of his patches have been
>> included in later kernels.
>>
>> Bernard's source files for libax25, ax25-apps & ax25-tools are
>> available here
>>
>> http://f6bvp.free.fr/logiciels/ax25/
>
> Now that you mention it, yea, I've seen stuff come across this list.
> At the time I wasn't interested in playing with AX.25 right then, so
> I skipped over most of it. Now that I'm playing with it in order to
> develop good config files to hand to someone else, I'm interested
> again. Funny how that works.
>
I guess I could say "Snap"... I've not been really too interested, as
what is bundled with ubuntu seems to be ok/enough for my purposes.
Should be be trying to test it on different Distro's or, what exactly
'should' we be doing? From seeing posts from Bernard go by, it seems
to me that he's being doing a solid enough job in keeping his sources
current.
Cheers
John
--
John Ronan <jronan@tssg.org>, +353-51-302938
Telecommunications Software & Systems Group, http://www.tssg.org
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-22 9:11 ` John Ronan
@ 2009-01-22 16:51 ` Curt, WE7U
0 siblings, 0 replies; 11+ messages in thread
From: Curt, WE7U @ 2009-01-22 16:51 UTC (permalink / raw)
To: John Ronan; +Cc: Linux-Hams
On Thu, 22 Jan 2009, John Ronan wrote:
> I guess I could say "Snap"... I've not been really too interested, as what is
> bundled with ubuntu seems to be ok/enough for my purposes.
>
> Should be be trying to test it on different Distro's or, what exactly
> 'should' we be doing? From seeing posts from Bernard go by, it seems to me
> that he's being doing a solid enough job in keeping his sources current.
For my purposes on OpenSuSE it wasn't working. As I discover
OS-specific tweaks and/or what modules work with what, it'd be nice
to update a Wiki page to alert others that may be using the same OS.
Similar to the OS-specific instructions we have on the Xastir Wiki.
Ok, had... The site has been down for a bit.
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-19 21:49 ` Curt, WE7U
2009-01-19 22:41 ` Ray Wells
@ 2009-01-23 18:19 ` Curt, WE7U
2009-01-25 6:02 ` Curt, WE7U
1 sibling, 1 reply; 11+ messages in thread
From: Curt, WE7U @ 2009-01-23 18:19 UTC (permalink / raw)
To: linux-hams
On Mon, 19 Jan 2009, Curt, WE7U wrote:
> So... I need to repeat my on-air tests with Bob using the newer
> AX.25 modules and tools to see if any or all of the bugs reported
> are still evident. Stay tuned.
The RF tests were repeated with ALL of the SuSE-11.0 updates
applied, no differences were seen. Each of the bugs described in my
first post on this thread were experienced again.
I've upgraded all machines but this one to OpenSuSE 11.1. After I
upgrade it too I'll probably ask Bob to help me repeat the tests.
As things stand now:
*) Unproto packets to/from Xastir over the port are working fine.
*) Ax25d and/or axspawn have problems with respect to the automatic
account creation.
*) Connected mode using axspawn loses LOTS of data. It might be
useful to test another connected-mode AX.25 app to see where the
problem lies, in the AX.25 modules or in axspawn.
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-23 18:19 ` Curt, WE7U
@ 2009-01-25 6:02 ` Curt, WE7U
2009-01-25 22:33 ` Ken Koster
0 siblings, 1 reply; 11+ messages in thread
From: Curt, WE7U @ 2009-01-25 6:02 UTC (permalink / raw)
To: linux-hams
On Fri, 23 Jan 2009, Curt, WE7U wrote:
> As things stand now:
>
> *) Unproto packets to/from Xastir over the port are working fine.
>
> *) Ax25d and/or axspawn have problems with respect to the automatic
> account creation.
>
> *) Connected mode using axspawn loses LOTS of data. It might be
> useful to test another connected-mode AX.25 app to see where the
> problem lies, in the AX.25 modules or in axspawn.
Today at the ham brunch n7ipb and kd7nm were present among others.
N7ipb and myself connected our laptops with a null-modem cable and
had KISS ports configured on each. I ran ax25d and had axspawn
configured to launch from it.
Ken had no problems running SSH over that connection after he
created his initial account via connected-mode and I manually
changed his password as root.
Ax25d and axspawn demonstrated the same bugs I reported before.
--
Curt, WE7U. archer at eskimo dot com
http://www.eskimo.com/~archer
Lotto: A tax on people who are bad at math. - unknown
Windows: Microsoft's tax on computer illiterates. - WE7U.
The world DOES revolve around me: I picked the coordinate system!"
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-25 6:02 ` Curt, WE7U
@ 2009-01-25 22:33 ` Ken Koster
2009-01-26 16:51 ` Thomas Osterried
0 siblings, 1 reply; 11+ messages in thread
From: Ken Koster @ 2009-01-25 22:33 UTC (permalink / raw)
To: linux-hams
On Saturday 24 January 2009, Curt, WE7U wrote:
> On Fri, 23 Jan 2009, Curt, WE7U wrote:
>
> > As things stand now:
> >
> > *) Unproto packets to/from Xastir over the port are working fine.
> >
> > *) Ax25d and/or axspawn have problems with respect to the automatic
> > account creation.
The account creation issues should be relatively easy to troubleshoot but I
was more intrigued by the data loss issues so I'm not looking at account
creation.
> > *) Connected mode using axspawn loses LOTS of data. It might be
> > useful to test another connected-mode AX.25 app to see where the
> > problem lies, in the AX.25 modules or in axspawn.
To follow up on Curt's comment, it's probably not really axspawn or ax25d
since ultimately they just use execve to execute 'login'.
...
I spent several hours last night with a bench test setup verifying what we saw,
and verifying that the old releases worked.
I'm going to try again with the latest CVS now that I know where it is.
Ken, N7IPB
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit
2009-01-25 22:33 ` Ken Koster
@ 2009-01-26 16:51 ` Thomas Osterried
0 siblings, 0 replies; 11+ messages in thread
From: Thomas Osterried @ 2009-01-26 16:51 UTC (permalink / raw)
To: Ken Koster; +Cc: linux-hams
On 2009-01-25 14:33:24 -0800, Ken Koster <n7ipb@wetnet.net>
wrote in <200901251433.26055.n7ipb@wetnet.net>:
> > > *) Connected mode using axspawn loses LOTS of data. It might be
> > > useful to test another connected-mode AX.25 app to see where the
> > > problem lies, in the AX.25 modules or in axspawn.
>
> To follow up on Curt's comment, it's probably not really axspawn or ax25d
> since ultimately they just use execve to execute 'login'.
the old axspawn had the issue that write()'s return value was not used
correctly.
The new axspawn is even compatible with the #BIN# protocol.
I have written the tools bput and bget, which enables you to up/download
any file with your favourite packet program to the shell session at the
remote site.
An issue concerning the kernel-ax25 we found a few days past:
we observed, that it could happen that the rf transmission did not has
incremented the session number of the I-frame. This means, that there
may be two or more different data-packets with the same sequence number -
with the result that the receiving site misses that data.
We think it's a smp-problem: non-smp kernels did not show this behaviour
(or at least it is rare that we still have not seen it when testing).
We testet this least with kernel 2.6.24 to 2.6.28.1, but also heard that
the problem was seen many years past with kernel 2.4.x.
73,
- Thomas dl9sau
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2009-01-26 16:51 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-01-18 8:22 Problems with ax25d/axspawn on OpenSuSE 11.0 32-bit Curt, WE7U
2009-01-18 21:08 ` Ken Koster
2009-01-19 21:49 ` Curt, WE7U
2009-01-19 22:41 ` Ray Wells
2009-01-21 17:35 ` Curt, WE7U
2009-01-22 9:11 ` John Ronan
2009-01-22 16:51 ` Curt, WE7U
2009-01-23 18:19 ` Curt, WE7U
2009-01-25 6:02 ` Curt, WE7U
2009-01-25 22:33 ` Ken Koster
2009-01-26 16:51 ` Thomas Osterried
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.