(no subject)

2005-12-07 Thread Landon Cox

Nicolas Baradakis wrote:


libssl.so.0.9.7 = /usr/lib/libssl.so.0.9.7 (0x40145000)


  ^^
This has nothing to do with your problem, but can you please indicate
whether you're using a binary package from Suse? If that's the case,
Suse is distributing software with incompatible licenses linked
together. I think maybe they are violating either the GPL or the
OpenSSL license.

The file was rlm_exec so file was timestamped 11-01-05 and I know I  
haven't built anything from source for a long time on that box, so  
would have to think it was a binary package unless somehow the SuSE  
update pulls and compiles source on my behalf...I don't think so.


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


OS Update broke FR - rlm_exec-1.0.0.so not found

2005-12-06 Thread Landon Cox


Hi FR community,

I've been running FR on an updated (recently fully patched apps) SuSE  
9.2 (kernel 2.6.8-24-default)   I was successfully running an earlier  
version of FR and when I decided to do an update of FR through SuSE's  
online update, FR will no longer come up and fails with a dynamic  
link error:


radiusd:  entering modules setup
Module: Library search path is /usr/lib/freeradius
radiusd.conf[1367] Failed to link to module 'rlm_exec': /usr/lib/ 
freeradius/rlm_exec-1.0.0.so: cannot open shared object file: No such  
file or directory

radius2:/home/lcox # ls /usr/lib/freeradius/rlm_exec-1*
/usr/lib/freeradius/rlm_exec-1.0.0.la  /usr/lib/freeradius/ 
rlm_exec-1.0.0.so


As you can see, my /usr/lib/freeradius directory does have the .so  
file, but I can't tell from the output if rlm_exec is dependent upon  
some other file that is not found or the dynamic linker can't find  
rlm_exec-1.0.0.so.   So, that's one question - what is not being  
found?  Sounds obvious, except that I have the exact .so filename in  
the library search path.


SuSE Yast claims it has installed FR 1.0.0-5.6.   A radiusd -v produces:
radiusd: FreeRADIUS Version 1.0.0, for host , built on May 30 2005 at  
21:02:41

Copyright (C) 2000-2003 The FreeRADIUS server project.

I've seen numerous references to this exact linkage error on various  
freeradius lists as well as have seen it on lists for various  
architectures and *nix's, not just x86/SuSE Linux.  However, there is  
typically no response given to fix the problem except to rebuild FR  
with no shared libraries and even in those cases, the build often  
seems to break later leaving the admin stuck further down the line.


Is there a known solution to what seems like a relatively common  
problem of rlm_exec dynamic linkage issues?  What am I missing and  
what needs to be done to cause this to work with the shared .so lib  
file vs having to rebuild it with static libs?


Thanks in advance for any help or direction.

Landon

(Full text of -X output follows)
radius2:/home/lcox # /usr/sbin/radiusd -X
Starting - reading configuration files ...
reread_config:  reading radiusd.conf
Config:   including file: /etc/raddb/proxy.conf
Config:   including file: /etc/raddb/clients.conf
Config:   including file: /etc/raddb/snmp.conf
Config:   including file: /etc/raddb/eap.conf
Config:   including file: /etc/raddb/sql.conf
main: prefix = /usr
main: localstatedir = /var
main: logdir = /var/log/radius
main: libdir = /usr/lib/freeradius
main: radacctdir = /var/log/radius/radacct
main: hostname_lookups = no
main: max_request_time = 30
main: cleanup_delay = 5
main: max_requests = 1024
main: delete_blocked_requests = 0
main: port = 0
main: allow_core_dumps = no
main: log_stripped_names = no
main: log_file = /var/log/radius/radius.log
main: log_auth = yes
main: log_auth_badpass = yes
main: log_auth_goodpass = yes
main: pidfile = /var/run/radiusd/radiusd.pid
main: user = radiusd
main: group = radiusd
main: usercollide = no
main: lower_user = no
main: lower_pass = no
main: nospace_user = no
main: nospace_pass = no
main: checkrad = /usr/sbin/checkrad
main: proxy_requests = yes
proxy: retry_delay = 5
proxy: retry_count = 3
proxy: synchronous = no
proxy: default_fallback = yes
proxy: dead_time = 120
proxy: post_proxy_authorize = yes
proxy: wake_all_if_all_dead = no
security: max_attributes = 200
security: reject_delay = 1
security: status_server = no
main: debug_level = 0
read_config_files:  reading dictionary
read_config_files:  reading naslist
read_config_files:  reading clients
read_config_files:  reading realms
radiusd:  entering modules setup
Module: Library search path is /usr/lib/freeradius
radiusd.conf[1367] Failed to link to module 'rlm_exec': /usr/lib/ 
freeradius/rlm_exec-1.0.0.so: cannot open shared object file: No such  
file or directory

radius2:/home/lcox # ls /usr/lib/freeradius/rlm_exec-1*
/usr/lib/freeradius/rlm_exec-1.0.0.la  /usr/lib/freeradius/ 
rlm_exec-1.0.0.so
- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html


Re: OS Update broke FR - rlm_exec-1.0.0.so not found

2005-12-06 Thread Landon Cox

Thanks, Alan.

FYI - more research on the topic, I did an ldd on the rlm_exec file:

On a freeradius box I have which is working, I did:

radius1:/usr/lib/freeradius # ldd rlm_exec-1.0.0.so
linux-gate.so.1 =  (0xe000)
libnsl.so.1 = /lib/libnsl.so.1 (0x40018000)
libresolv.so.2 = /lib/libresolv.so.2 (0x4002f000)
libpthread.so.0 = /lib/tls/libpthread.so.0 (0x40042000)
libcrypto.so.0.9.7 = /usr/lib/libcrypto.so.0.9.7 (0x40054000)
libssl.so.0.9.7 = /usr/lib/libssl.so.0.9.7 (0x40145000)
libc.so.6 = /lib/tls/libc.so.6 (0x40175000)
/lib/ld-linux.so.2 = /lib/ld-linux.so.2 (0x8000)
libdl.so.2 = /lib/libdl.so.2 (0x4028b000)


Then on the radius box I updated and broke I did:
radius2:/usr/lib/freeradius # ldd rlm_exec-1.0.0.so
linux-gate.so.1 =  (0xe000)
libnsl.so.1 = /lib/libnsl.so.1 (0x55577000)
libresolv.so.2 = /lib/libresolv.so.2 (0x5558e000)
libpthread.so.0 = /lib/tls/libpthread.so.0 (0x555a1000)
libcrypto.so.0.9.7 = /usr/lib/libcrypto.so.0.9.7 (0x555b3000)
libssl.so.0.9.7 = /usr/lib/libssl.so.0.9.7 (0x556a4000)
libc.so.6 = /lib/tls/libc.so.6 (0x556d4000)
/lib/ld-linux.so.2 = /lib/ld-linux.so.2 (0x56555000)
libdl.so.2 = /lib/libdl.so.2 (0x557ea000)

I hunted down and verified that every library that was pointed to by  
rlm_exec indeed existed on the file system.  On both systems, they  
had everything except both had no linux-gate.so.1.  But then radius1  
works fine and it shows the same paths and library existence as the  
radius2 box which fails.


Also, googling linux-gate.so.1 I saw:
What is linux-gate.so.1?:  http://www.trilithium.com/johan/2005/08/ 
linux-gate/

and
http://kerneltrap.org/node/3405

The first link explains that an ldd report of linux-gate.so.1 that  
doesn't point to a file/path is normal in recent kernels and goes  
into detail of what it is.  In any case, it's not a problem.


After verifying that every library that is pointed to by  
rlm_exec-1.0.0.so actually exists on both the machine that works fine  
and the one that doesn't, I understand a little more but don't see  
what the problem is.   There's something else bizarre going on...or  
perhaps I still need to run ldd on each of the dependent libraries -  
maybe there's one in that tree of dependencies that's missing.


I'll also try going to 1.0.5.

Thanks,

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


Re: FreeRadius EAP-TLS

2005-08-25 Thread Landon Cox


On Aug 25, 2005, at 10:34 AM, Hamid Salim wrote:


I have a 2 part question.

1.I recall reading on this forum that, Windows XP broke EAP-TLS, does
this apply to SP2 also?


I've had XP SP2 EAP-TLS clients running against FR with no problems.

Also, for what it's worth, I've built XP Embedded configurations (XP  
for embedded devices, booting compact flash), that run SP2 and work  
fine with FR as well.


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


Example of Mac OS X EAP-TLS process?

2005-08-23 Thread Landon Cox


I've got a good, working FreeRadius running EAP-TLS on a SuSE 9.2  
box.  I've had good luck with WPA supplicants for XP SP2 and several  
vendor PCMCIA card supplicants - all on XP SP2.


I've been trying to get an OS X (Tiger) machine up with the same type  
of setup, but each time I set the 802.1x TLS check box, I always  
get an error stating that there is no valid certificate available on  
the machine.  I've imported both the client and root CA certs into  
the Mac OS X Keychain (the root CA imported into X509 anchors  
category and the client cert into 'logins'.)


I've generated the Mac client cert in the same way as I do the XP  
client cert except without the xpextensions ASN.1 options on openSSL.


I realize this isn't a FreeRadius question per-se but was hoping that  
someone else in FR land has done an 802.1x EAP-TLS setup on OS X w/FR  
and had success.


Thank you,

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


Re: XP won't authenticate with EAP TLS - log shows unknown_ca fatal error

2005-08-10 Thread Landon Cox


On Aug 8, 2005, at 9:39 AM, Landon Cox wrote:


I'm going to do some experiments later tonight and see if I can  
isolate the success factor.


Back on this topic for a moment...some things I tried to see if I  
could break the configuration were:
1) remove the certs from the /etc/ssl/certs directory, restart  
FR, no difference - still hooked up fine since the certs are also in  
raddb/certs.


I decided to generate a client cert for a Mac box and when I imported  
it into the Keychain of OS X, I noticed This certificate is not yet  
valid.


I went back and looked at the output of the certificate generation  
and the validity Not Before gave a date/time stamp that was 1 hour  
future (my timezone setting was off by one hour.)


But this made me wonderwas the unknown_ca problem caused by the  
CA cert having a Not Valid Before validity that was in the future  
from the real time when it was generated and then initially tested?


Is this a possible cause for an unknown_ca error?

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


Re: XP won't authenticate with EAP TLS - log shows unknown_ca fatal error

2005-08-08 Thread Landon Cox


On Aug 8, 2005, at 9:18 AM, Kris Benson wrote:


Did you do anything differently with your 'random' file and your  
'dh' file?


Creating those properly (as opposed to the idiotic directions of  
date 

dh; date  random) seemed to solve my dilemma when I was getting a
similar issue to what you were getting.


Hi Kris,

No, both dh and random stayed as initially generated.  I used the  
Bauer Linux journal article to do that step which used the /dev/ 
urandom method vs date.


I'm going to do some experiments later tonight and see if I can  
isolate the success factor.


Thanks,

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


Re: XP won't authenticate with EAP TLS - log shows unknown_ca fatal error

2005-08-06 Thread Landon Cox


Thanks for looking at this, Michael.

I decided to restart the certificate generation process and did it  
again from scratch following the article.  Same results.


I did it a 3rd time and but this time copied the certs to /etc/ssl/ 
certs and insured all CNs were unique (not being completely up on  
what is right or wrong w/r to input values for the cert process or  
what directories the new certs needed to live in, I wanted to make  
sure that wasn't an issue.)  So, one of those actions did the trick  
and I haven't gone back to isolate which one.


After that I was able to login - authenticated in both directions,  
too.  I did go ahead and do the pkcs export password and that worked  
fine.  I'm not sure what Bauer's comment was referring to in the  
article about XP supplicants only working with non-pw protected certs  
in the store.  Oh well, it's up and working and I'm grateful.


Thank you,

Landon

On Aug 5, 2005, at 4:30 PM, Michael Wang wrote:


Hi Landon,

I think this piece from the log is suspicious:



rlm_eap_tls:  Length Included
 eaptls_verify returned 11
 rlm_eap_tls:  TLS 1.0 Handshake [length 02ab], Certificate
-- verify error:num=18:self signed certificate
chain-depth=0,
error=18
-- User-Name = 360VL
-- BUF-Name = 360VL
-- subject = /C=US/ST=Colorado/L=Colorado Springs/O=360VL
Incorporated/CN=360VL/emailAddress=emailwithheld
-- issuer  = /C=US/ST=Colorado/L=Colorado Springs/O=360VL
Incorporated/CN=360VL/emailAddress=emailwithheld
-- verify return:0
 rlm_eap_tls:  TLS 1.0 Alert [length 0002], fatal unknown_ca
TLS Alert write:fatal:unknown CA
  TLS_accept:error in SSLv3 read client certificate B
rlm_eap_tls: SSL_read failed in a system call (-1), TLS session  
fails.




I think the problem is the user certificate that you imported into XP
is self-signed.  What you need to do is use openssl to create a
certificate request (using openssl req ...) and then sign that request
using the CA (using openssl ca).  Then package up the user key and
signed user cert into the pkcs#12 envelope (using openssl pkcs12).
Finally import into XP.  I looked at the instructions for certificate
generation in the linux format article and they look OK.  Make sure
you did not miss a step or use the wrong command somewhere.

As to using a password for the pkcs#12 envelope, go ahead and use it.
When you import the pkcs#12 file into XP, it will just ask for it, and
you enter it, and that should be it.

Hope that helps.

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