- Freeradius-Devel - lists.freeradius.org

Re: Hack way to compile freeradius causing freeradius to hang under multithread mode
by Ming-Ching Tiew 14 Jun '09

14 Jun '09

OK I will try to debug and see :- As the first cut, 'netstat -naut' shows that there is no listening on port 1812, whereas when it's working there is something listening on port 1812. 'radiusd' will daemonized, but it must be blocking at thread creation or something.

2 1

Hack way to compile freeradius causing freeradius to hang under multithread mode
by Ming-Ching Tiew 14 Jun '09

14 Jun '09

When I run ./configure .... under uclibc, it can succcessful produce a Makefile. But with that makefile, compilation will fail because the lack of THREAD_TLS under uclibc. Then I went on to modify ( by hand ) src/include/autoconf.h to comment HAVE_THREAD_TLS, then compilation could be completed. However, the produced binary will not work under multithread mode, it will only work under these modes :- radiusd -X radiusd -s radiusd -t It will not work if it's run as :- radiusd Well, this behaviour is repeatable under gnu lib environment too ( tested under debian environment), ie run ./configure, edit src/include/autoconf.h by hand to comment HAVE_THREAD_TLS. Any suggest what's next best thing to do ? Regards.

2 1

Re: Hack way to compile freeradius causing freeradius to hang under multithread mode
by Ming-Ching Tiew 14 Jun '09

14 Jun '09

> > However, the produced binary will not work under > multithread mode, it will only work under these modes :- > > radiusd -X > radiusd -s > radiusd -t > > It will not work if it's run as :- > > radiusd > > Well, this behaviour is repeatable under gnu lib > environment too ( tested > under debian environment), ie run ./configure, edit > src/include/autoconf.h by hand to comment HAVE_THREAD_TLS. > Sorry to confuse you, after I posted, I found that without commenting HAVE_THREAD_TLS, it hangs too under under gnu libraries. I guess the right question to ask is whether running 'radiusd' alone without any parameters is an intended mode. Regards

1 0

GIT Log for 2009-06-11 23:33 GMT
by aland＠deployingradius.com 11 Jun '09

11 Jun '09

commit ed4770be765bd88c1dabcac7cdcb7c5194347b23 Author: Alan T. DeKok <aland(a)freeradius.org> Date: Thu Jun 11 10:38:24 2009 +0200 More debugging messages Files changed: src/modules/frs_detail/frs_detail.c | 13 ++++++++----- 1 files changed, 8 insertions(+), 5 deletions(-) ====================================================================== commit b6340d5536ca77d1169f868dd0e66c393dde2be7 Author: Alan T. DeKok <aland(a)freeradius.org> Date: Thu Jun 11 10:37:51 2009 +0200 More debugging messages Files changed: src/main/event.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) ====================================================================== commit 8c078deef7eee1d6e6475e77acf3535c058b68da Author: Alan T. DeKok <aland(a)freeradius.org> Date: Thu Jun 11 09:55:48 2009 +0200 Include chillispot dictionary Files changed: share/dictionary | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) ======================================================================

1 0

PEAP Certificate path
by Harsha gowda 10 Jun '09

10 Jun '09

Hi, I want to know where is the private key file stored for the public key which is used to decrypt Client Key exchange, the same publick key which is Sent in Server-Hello message its private patr is stored in which location. i search in {raddb}/cers/? but could not. Thanks

2 1

GIT Log for 2009-06-09 23:33 GMT
by aland＠deployingradius.com 09 Jun '09

09 Jun '09

commit 4a5b18a653c5a35786f8a9dcfab8b4ebc0671101 Author: Alan T. DeKok <aland(a)freeradius.org> Date: Tue Jun 9 09:46:29 2009 +0200 Filter Access-Challenge packets, too raddb/attrs.access_challenge | 19 +++++++++++++++++++ raddb/modules/attr_filter | 10 ++++++++++ 2 files changed, 29 insertions(+), 0 deletions(-)

1 0

Bad advices in the documantation of rlm_ldap
by Németh Tamás 09 Jun '09

09 Jun '09

Sorry for sending anothe bugreport, but I found the following in the documentation of rlm_ldap (/usr/share/doc/packages/freeradius-server- doc/rlm_ldap for me): NOTE: As LDAP is case insensitive, you should probably also set "lower_user = yes" and "lower_time = before" in main section of radiusd.conf, to get limits on simultaneous logins working correctly. Otherwise, users will be able get large number of sessions, capitalizing parts of their login names. However, these options (e.g. lower_user) don't seem to exist in freeradius 2.x.x. You should replace this section with a description of how to emulate this behaviour with unlang. Regards, Németh Tamás

1 0

conditional variable syntax doesn't work in LDAP xlat strings
by Németh Tamás 09 Jun '09

09 Jun '09

Dear developers, Sorry for posting a a bugreport here, but http://bugs.freeradius.org is not working for several days (or is it intentionally and utterly down?). I wanted to evaluate this expression (in /etc/raddb/sites-enabled/default in the authorize section): if ( "%{%{Stripped-User-Name}:-%{User-Name}}" == "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=*))}" && "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?radiusCallingStationId?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=%{request:Calling-Station-Id}))}" != "%{request:Calling-Station-Id}" ) { but freeradius 2.1.6 behaved like this: +++? if ("%{%{Stripped-User-Name}:-%{User-Name}}" == "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=*))}" && "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?radiusCallingStationId?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=%{request:Calling-Station-Id}))}" != "%{request:Calling-Station-Id}" ) expand: %{Stripped-User-Name} -> expand: %{User-Name} -> nice expand: %{%{Stripped-User-Name}:-%{User-Name}} -> nice rlm_ldap: - ldap_xlat No trailing :- after variable at %{Stripped-User-Name} expand: ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%{%{Stripped-User-Name} -> ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn= rlm_ldap: ldap_get_conn: Checking Id: 0 rlm_ldap: ldap_get_conn: Got Id: 0 rlm_ldap: performing search in ou=People,ou=sopron,o=nyme,dc=hu, with filter (&(cn= rlm_ldap: ldap_search() failed: Bad search filter: (&(cn= rlm_ldap: Search returned error rlm_ldap: ldap_release_conn: Release Id: 0 expand: %{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=*))} -> nice)(radiusCallingStationId=*)) That is, it expanded "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%{%{Stripped-User-Name}:-%{User-Name}})(radiusCallingStationId=*))}" as a mere "ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub? (&(cn=%{%{Stripped-User-Name}" concatenated with "%{User-Name}})(radiusCallingStationId=*))", and thus it evaluated it as "nice)(radiusCallingStationId=*))" instead of "nice". However, it works if I'm using "%u" (or eventually %{control:Tmp-String-0}) instead of "%{%{Stripped-User-Name}:-%{User-Name}}" : +++? if ("%u" == "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%u)(radiusCallingStationId=*))}" && "%{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?radiusCallingStationId?sub?(&(cn=%u) (radiusCallingStationId=%{request:Calling-Station-Id}))}" != "%{request:Calling-Station-Id}" ) expand: %u -> nice rlm_ldap: - ldap_xlat expand: ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%u)(radiusCallingStationId=*)) -> ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=nice)(radiusCallingStationId=*)) rlm_ldap: ldap_get_conn: Checking Id: 0 rlm_ldap: ldap_get_conn: Got Id: 0 rlm_ldap: performing search in ou=People,ou=sopron,o=nyme,dc=hu, with filter (&(cn=nice)(radiusCallingStationId=*)) rlm_ldap: Adding attribute cn, value: nice rlm_ldap: ldap_release_conn: Release Id: 0 rlm_ldap: - ldap_xlat end expand: %{ldap:ldap:///ou=People,ou=sopron,o=nyme,dc=hu?cn?sub?(&(cn=%u)(radiusCallingStationId=*))} -> nice

1 0

GIT Log for 2009-06-06 23:33 GMT
by aland＠deployingradius.com 06 Jun '09

06 Jun '09

commit 632daac8d9969a5c32242db08123190adb58d043 Author: Alan T. DeKok <aland(a)freeradius.org> Date: Sat Jun 6 09:42:40 2009 +0200 Some clients end option 53 buried inside of the packet. Pointed out on the list by Martin Lorentz src/modules/frs_dhcp/dhcp.c | 59 ++++++++++++++++++++++++++++++++++++++---- 1 files changed, 53 insertions(+), 6 deletions(-)

1 0

GIT Log for 2009-06-02 23:33 GMT
by aland＠deployingradius.com 02 Jun '09

02 Jun '09

commit c9d65dc24fcb353f5300299aa71de846dc0a5d71 Author: Alan T. DeKok <aland(a)freeradius.org> Date: Tue Jun 2 10:35:38 2009 +0200 Mark the proxy mutex as being recursive Some systems needs this to avoid deadlocks. Others (Linux) don't src/main/event.c | 14 ++++++++++++++ 1 files changed, 14 insertions(+), 0 deletions(-)

1 0