Looking for dictionary files for mobile VSAs

[John Padula]

>   Is there a RADIUS client which uses 3GPP VSA's?

The devices that handle wireless access to the 
IP world (SGSNs and GGSNs) advertise these VSAs.
The 'vendor' is actually 3GPP (id=10415) or
cmda2000 (id=5535).

For 3GPP, see Appendicies in:

ftp://ftp.3gpp.org/specs/2002-06/R1999/29_series/29061-3a0.zip.


Regards,
JP



- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Looking for dictionary files for mobile VSAs

[John Padula]

Hi:

I've been looking for dictionary files for 3GPP and CDMA-related VSAs.

I've searched over the web and at their sites and have been unable to find them.

Just wondering if these exist or if I need to piece them
together from their specs and submit a home-grown version.

Thanks for any info,
JohnP


- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Questions on IP address allocation

[John Padula]

I'm still looking into IP address allocation via the "+" option.

I wanted to check if there's any sense of a lease or expiration on the
IP address
given out.  I've checked the RFC and didn't see any mention of this, but
wanted to 
make sure.  Coming from the DHCP world, where there's a large blob of
processing
devoted to lease negotiation, it seems strange that (a) we just give out
an IP 'forever'
and (b) we just keep incrementing based on nas-port-id.  There's no way
to define 
a range of IPs.

Is there any way to establish a hook (in the 'module' model) where
a user could call some external function that would return an IP and
possibly manage
that IP?

Thanks,
JohnP

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

RE: Re: Configure script problems and compiler pickiness

[John Padula]

Alan:

Thanks for the speedy response.

> Of course, if you use 'gcc' when running 'configure', 
> and 'cc' for the compile, then that would explain some
inconsistencies. 

Maybe I wasn't clear, but the configure-make sequence on the new machine
was
done with a fresh 0.2 TAR file, so I tried to keep things consistent and
only used CC throughout.

I've since loaded GCC on the target machine and (1) the horrible
compiler warnings
were cut down significantly (not suprising - Solaris' CC is usually
pickier) and (2) the 
gdbm errors continued.  So, at least gcc did not fix the path problem.

I examined the machine that all this initially worked on and tracked
down this
include file (and the associated library) as belonging to the "KDE - X11
Windowing Environment".
The second machine did not have this installed - though it did have a
Solaris equivalent
(called Gnome).  I pointed my paths to the Gnome directories, did a
"make clean"
and tried again.  It still failed!  Ultimately, I added the paths to the
gnome area by hand
to the libtool and it worked!!  Not sure why these weren't picked up
from my environment.
I even fixed LD_LIBRARY_PATH to point to these...?!

Don't know how much of this is useful information that might prevent
others from
spinning their wheels for 3+ hours trying to get this resolved, but
there you have it.

By the way - regarding the $PGROOT problem - to get around this, I
copied the test 
from the 0.2 "configure" script.  Basically, the line is:

if test $PGROOT ; then
IS_LOCAL=1

versus

if test -d $PGROOT ; then
IS_LOCAL=1

and later:

if test $PGROOT && test IS_LOCAL ; then

versus

if test -d $PGROOT && test "x$IS_LOCAL" != "x" ; then


Not sure why this was changed from what was in 0.2, but maybe it can be
changed back.

Hope all that helps.  I'll finally test the dynamic IP allocation fix
you made a few weeks ago
and let you know how it goes.

Thanks,
JohnP

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Configure script problems and compiler pickiness

[John Padula]

Hello:

I saw a posting a few weeks back about a problem with the configure
script
and $PGROOT not being properly defined.

I just grabbed the latest (09/10) TAR file and configure seems to still
be dying 
right around that spot (where it says "checking for PostgreSQL
includes").

Was a fix ever installed?

Also, are there known issues with the Solaris CC compiler?  I used gcc
on 
one machine and the 0.2 release compiled fine.  I then used CC on
another machine
(same release) and I got a lot of complaints (e.g.:

cc  -g -mt -D_REENTRANT -DNDEBUG -D_LIBRADIUS -I../include -c radius.c
"radius.c", line 341: warning: argument #5 is incompatible with
prototype:
prototype: pointer to const struct sockaddr {unsigned short
sa_family, array[14] of char sa_data} : "/usr/include/sys/socket.h",
line 454
argument : pointer to struct sockaddr_in {unsigned short
sin_family, unsigned short sin_port, struct in_addr {..} sin_addr,
array[8] of char sin_zero}
 
The compile ultimately failed with 

   cc  -g -mt -D_REENTRANT -DNDEBUG -I../../include  -c rlm_counter.c -o
rlm_counter.o
   "rlm_counter.c", line 37: cannot find include file: 
   "rlm_counter.c", line 71: syntax error before or at: GDBM_FILE
   "rlm_counter.c", line 71: cannot recover from previous errors
   cc: acomp failed for rlm_counter.c

so I wasn't sure how much of this was complier (versus config) related..


Thanks again,
JohnP

-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Is Dynamic IP address assignment working??

[John Padula]

Hi there:

I've been interested in the dynamic IP address capabilities of
FreeRadius.

I've also searched the mailing lists for the last 6 months and have seen

repeated postings that talk about the infamous "+" option and how it
allows incrementing IP addresses to be assigned.  It seems, though, 
that most of the advice on how to get this going had to do with passing
back some pool name id to the NAS and let *it* manage the IP addresses.

It didn't seem to complicated, so I gave it a try, but my users keeps
getting
assigned the same IP address..?!  Now, I should qualify that and say
that 
my "users" are really via radtest, but I don't think that should matter.

I decided to check the code and I discovered this comment in
rad_authenticate():

/*
 *  Add the port number to the Framed-IP-Address if
 *  vp->addport is set, or if the Add-Port-To-IP-Address
 *  pair is present.
 *
 *  FIXME:  This doesn't work because
PW_ADD_PORT_TO_IP_ADDRESS
 *  is never added to the request pairs!
 */

So, since I don't see any bug on the recent bug list that describes a
problem with 
dynamic IP addresses, I have to ask: is it broken and/or has it ever
worked??

Thanks,
John Padula 



-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html