freeradius.git
14 years agoMoved detail structure defs to a public header file
Alan T. DeKok [Sun, 6 Dec 2009 12:36:28 +0000 (13:36 +0100)]
Moved detail structure defs to a public header file

14 years agoFix code so that corner cases of %{%{foo}:-%{bar}} work
Alan T. DeKok [Sat, 5 Dec 2009 14:58:19 +0000 (15:58 +0100)]
Fix code so that corner cases of %{%{foo}:-%{bar}} work

The previous code was odd... this code is simpler, and works.

14 years agoRe-set FP after closing it
Alan T. DeKok [Sat, 5 Dec 2009 14:54:08 +0000 (15:54 +0100)]
Re-set FP after closing it

14 years agoBuild from v2.1.x branch instead of stable
Alan T. DeKok [Sat, 5 Dec 2009 14:57:20 +0000 (15:57 +0100)]
Build from v2.1.x branch instead of stable

14 years agoEnable new dictionaries
Alan T. DeKok [Fri, 4 Dec 2009 13:46:04 +0000 (14:46 +0100)]
Enable new dictionaries

14 years agoEnable Unix (Proftpd) dictionary
Alan T. DeKok [Fri, 4 Dec 2009 13:41:43 +0000 (14:41 +0100)]
Enable Unix (Proftpd) dictionary

14 years agoSign client certs with CA rather than server cert
Alan T. DeKok [Thu, 3 Dec 2009 09:25:33 +0000 (10:25 +0100)]
Sign client certs with CA rather than server cert

14 years agoUse intermediate buffer for error messages.
Alan T. DeKok [Wed, 2 Dec 2009 11:16:57 +0000 (12:16 +0100)]
Use intermediate buffer for error messages.

This prevents the messages from getting mangled

14 years agoNote recent changes
Alan T. DeKok [Wed, 2 Dec 2009 10:56:24 +0000 (11:56 +0100)]
Note recent changes

14 years agoRevert "Moved illegal attributes to the new dictionary"
Alan T. DeKok [Wed, 2 Dec 2009 10:55:03 +0000 (11:55 +0100)]
Revert "Moved illegal attributes to the new dictionary"

This reverts commit 4ba2853439bba32f979009fcc04b48a829bf83d4.

While these dictionaries are unfortunate, making this change
in a point release is likely a bad idea.

14 years agoRevert "Moved Ascends illegal attributes to their own file"
Alan T. DeKok [Wed, 2 Dec 2009 10:54:23 +0000 (11:54 +0100)]
Revert "Moved Ascends illegal attributes to their own file"

This reverts commit 0241615ea5e98a13c92c266daab356e057d6a27d.

While these dictionaries are unfortunate, making this change
in a point release is likely a bad idea.

14 years agoRemoved erroneous 'break'
Alan T. DeKok [Wed, 2 Dec 2009 10:37:33 +0000 (11:37 +0100)]
Removed erroneous 'break'

14 years agoFixed string copying in sub variable.
Alan T. DeKok [Wed, 2 Dec 2009 07:51:27 +0000 (08:51 +0100)]
Fixed string copying in sub variable.

Apparently the only reason this worked before is that no one used it.
The decode_attribute() function did the string copying itself,
and therefore avoided this.

14 years agoFixed typo
Alan T. DeKok [Tue, 1 Dec 2009 09:49:18 +0000 (10:49 +0100)]
Fixed typo

14 years agoFix build problem
Alan T. DeKok [Mon, 30 Nov 2009 16:07:22 +0000 (17:07 +0100)]
Fix build problem

14 years agoPrint out more useful debugging messages
Alan T. DeKok [Mon, 30 Nov 2009 12:58:49 +0000 (13:58 +0100)]
Print out more useful debugging messages

  Rather than

rlm_ldap: ...

  do

  [foo] ...

Which prints out the instance name in a slightly better format

14 years agoCleanups and simplifications.
Alan T. DeKok [Mon, 30 Nov 2009 09:05:37 +0000 (10:05 +0100)]
Cleanups and simplifications.

The FD_SET is now calculated in the event_loop() function,
making it harder to get it wrong.

fr_event_now() ALWAYS returns a time, calling gettimeofday()
if necessary

14 years agoClean up "dead" child if there's no thread associated with the request
Alan T. DeKok [Mon, 30 Nov 2009 08:14:27 +0000 (09:14 +0100)]
Clean up "dead" child if there's no thread associated with the request

14 years agoClean up state machine.
Alan T. DeKok [Sun, 29 Nov 2009 15:07:23 +0000 (16:07 +0100)]
Clean up state machine.

  This error happens when "max_request_time" is set VERY low.
i.e. lower than "response_window".  (12s versus 30s).

  The current logic for enforcing the various timers is pretty bad.  There
is one timer per request, and it bounces around between the different
requirements.  At the time it was written, it seemed simpler than trying
to manage 3-4 simultaneous timers per request.

  When the request is proxied, the timer being applied is for
"response_window".  BUT by the time that expires, the "max_request_time"
has expired.  The code *does* notice that it has expired.  BUT it doesn't
notice that there's no child thread processing the request.  So it waits
for the child thread to exit... forever.

  At some point, a timer overflows, and it dies.

  There are a few changes to make:

1) check for "no child" in this situation, and clean up the request rather
  than waiting forever.

2) cap the timer to 5 minutes (this can still happen, for example, when a
   bad DB locks a thread for hours at a time).

3) don't overflow when adding timer values.

14 years agoL_INFO, "PROXY:... --> L_PROXY, "...
Alan T. DeKok [Fri, 27 Nov 2009 14:58:58 +0000 (15:58 +0100)]
L_INFO, "PROXY:...   -->  L_PROXY, "...

Simplifies and regularizes the log messages

14 years agoFix typo
Alan T. DeKok [Fri, 27 Nov 2009 12:01:06 +0000 (13:01 +0100)]
Fix typo

14 years agoInitialize timers for Status-Server
Alan T. DeKok [Fri, 27 Nov 2009 11:07:25 +0000 (12:07 +0100)]
Initialize timers for Status-Server

For some weird reason they weren't initialized in debugging mode,
so we force it here.  This WAS tested to work... so I have no idea
why it stopped.

We also ignore Status-Server packets when marking home_servers as
alive.  That way, the ping_check will work properly...

14 years agoFixed log messages
Alan T. DeKok [Fri, 27 Nov 2009 07:48:16 +0000 (08:48 +0100)]
Fixed log messages

Pulled by hand

5f712ba731daa8bc7ba56ede2f12a9d4c83a90c4
71ec7a3bd21ff91bb8b4fa99e32d4ba035d54b9e

14 years agoRemove erroneous handling of option 82
Alan T. DeKok [Thu, 26 Nov 2009 18:10:44 +0000 (19:10 +0100)]
Remove erroneous handling of option 82

14 years agoPrint names for unsupported eap types
Alan T. DeKok [Sun, 22 Nov 2009 15:56:14 +0000 (16:56 +0100)]
Print names for unsupported eap types

14 years agoMore warnings
Alan T. DeKok [Sun, 22 Nov 2009 15:47:29 +0000 (16:47 +0100)]
More warnings

Modifying the values of a virtual attribute is a bad idea.

14 years agoCheck for NULL
Alan T. DeKok [Sun, 22 Nov 2009 15:35:20 +0000 (16:35 +0100)]
Check for NULL

14 years agoFixed typo
Alan T. DeKok [Sun, 22 Nov 2009 08:26:51 +0000 (09:26 +0100)]
Fixed typo

14 years agoAs posted to the list
Alan T. DeKok [Sun, 22 Nov 2009 08:18:16 +0000 (09:18 +0100)]
As posted to the list

14 years agoMove definition so that it doesn't cause issues
Alan T. DeKok [Sun, 22 Nov 2009 08:16:22 +0000 (09:16 +0100)]
Move definition so that it doesn't cause issues

Use the RFC definitions for the name

14 years agoRevert "errormsg may be NULL"
Alan T. DeKok [Sun, 22 Nov 2009 08:12:15 +0000 (09:12 +0100)]
Revert "errormsg may be NULL"

This reverts commit 45877bf44b02d418b6fb263a39e5de07ced58b6e.

It doesn't fix the problem, and it seems to cause issues for
other people

14 years agoAs posted to the list
Alan T. DeKok [Wed, 18 Nov 2009 17:34:17 +0000 (18:34 +0100)]
As posted to the list

14 years agoMove user/group/chroot/core to bootstrap config
Alan T. DeKok [Tue, 17 Nov 2009 10:19:17 +0000 (11:19 +0100)]
Move user/group/chroot/core to bootstrap config

It adds some extra stuff when starting in debug mode as root, but it
also means that "allow_core_dumps" works again.

14 years agoNotes for v2.1.8
Alan T. DeKok [Sun, 22 Nov 2009 15:16:15 +0000 (16:16 +0100)]
Notes for v2.1.8

14 years agoMake templates work again
Alan T. DeKok [Sat, 14 Nov 2009 09:29:40 +0000 (10:29 +0100)]
Make templates work again

14 years agoAllow !* to work
Alan T. DeKok [Fri, 13 Nov 2009 13:15:51 +0000 (14:15 +0100)]
Allow !* to work

The code to delete all attributes (as documented in unlang) was
previously added in evaluate.c.  But the parser hadn't been updated,
so it was impossible to actually use that functionality.

14 years agoPrint commands from radmin when in debug mode
Alan T. DeKok [Thu, 12 Nov 2009 15:19:25 +0000 (16:19 +0100)]
Print commands from radmin when in debug mode

14 years agoFix WiMAX encoding bug introduced in 326a68b90a1a
Alan T. DeKok [Thu, 12 Nov 2009 11:25:03 +0000 (12:25 +0100)]
Fix WiMAX encoding bug introduced in 326a68b90a1a

14 years agoAs posted to the list
Alan T. DeKok [Fri, 6 Nov 2009 13:16:42 +0000 (08:16 -0500)]
As posted to the list

14 years agoFrom bug #45
Alan T. DeKok [Tue, 3 Nov 2009 21:40:03 +0000 (16:40 -0500)]
From bug #45

14 years agoRemoved recursive mutexes
Alan T. DeKok [Fri, 30 Oct 2009 00:35:53 +0000 (20:35 -0400)]
Removed recursive mutexes

14 years agoAs posted to the list
Alan T. DeKok [Fri, 23 Oct 2009 09:37:44 +0000 (11:37 +0200)]
As posted to the list

14 years agoerrormsg may be NULL
Alan T. DeKok [Wed, 28 Oct 2009 13:44:38 +0000 (09:44 -0400)]
errormsg may be NULL

14 years agoIf the port is already used, try another.
Alan T. DeKok [Fri, 30 Oct 2009 00:29:26 +0000 (20:29 -0400)]
If the port is already used, try another.

14 years agoStart simplifying the code that encodes attributes
Alan T. DeKok [Sat, 3 Oct 2009 17:58:17 +0000 (19:58 +0200)]
Start simplifying the code that encodes attributes

14 years agoBump to version 2.1.8
Alan T. DeKok [Wed, 16 Sep 2009 15:14:55 +0000 (17:14 +0200)]
Bump to version 2.1.8

14 years agoConf for debugging
Alan T. DeKok [Tue, 20 Oct 2009 14:28:58 +0000 (16:28 +0200)]
Conf for debugging

14 years agoFix openssl checks
Alan T. DeKok [Tue, 20 Oct 2009 14:07:18 +0000 (16:07 +0200)]
Fix openssl checks

14 years agoRetry if there was no response to the packet.
Alan T. DeKok [Tue, 20 Oct 2009 10:14:36 +0000 (12:14 +0200)]
Retry if there was no response to the packet.

14 years agoPrint env vars in parent, not child
Alan T. DeKok [Sun, 18 Oct 2009 15:19:22 +0000 (17:19 +0200)]
Print env vars in parent, not child

14 years agoChanged stop packet msg to debug rather than error
Alan T. DeKok [Sun, 18 Oct 2009 11:47:06 +0000 (13:47 +0200)]
Changed stop packet msg to debug rather than error

14 years agoDefine names
Alan T. DeKok [Sun, 18 Oct 2009 07:04:36 +0000 (09:04 +0200)]
Define names

14 years agoCall detach only if function exists
Alan T. DeKok [Fri, 16 Oct 2009 16:53:49 +0000 (18:53 +0200)]
Call detach only if function exists

14 years agoAs posted to the list
Alan T. DeKok [Thu, 15 Oct 2009 09:23:30 +0000 (11:23 +0200)]
As posted to the list

14 years agoFixed typo
Alan T. DeKok [Thu, 15 Oct 2009 08:08:40 +0000 (10:08 +0200)]
Fixed typo

14 years agoFix to prevent compiler optimizations
Alan T. DeKok [Wed, 14 Oct 2009 12:09:41 +0000 (14:09 +0200)]
Fix to prevent compiler optimizations

from bug #30

14 years agoWrite the PID file as late as possible
Alan T. DeKok [Tue, 13 Oct 2009 10:53:49 +0000 (12:53 +0200)]
Write the PID file as late as possible

i.e. after checking the config, and after opening any sockets

14 years agoFix typo
Alan T. DeKok [Tue, 13 Oct 2009 10:52:12 +0000 (12:52 +0200)]
Fix typo

14 years agofix debian/rules to honour CFLAGS
Alexander Clouter [Sat, 10 Oct 2009 12:25:29 +0000 (13:25 +0100)]
fix debian/rules to honour CFLAGS

Fixed up debian/rules to allow CFLAGS to be honoured.

Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
14 years agoDon't use source IP for EAP packets.
Alan T. DeKok [Mon, 12 Oct 2009 11:17:26 +0000 (13:17 +0200)]
Don't use source IP for EAP packets.

We control State, so it should be unique, too

14 years agoMake client/port/key-balance more like fail-over
Alan T. DeKok [Mon, 12 Oct 2009 11:14:09 +0000 (13:14 +0200)]
Make client/port/key-balance more like fail-over

Previously, they would default to load-balancing across all
found servers.  Now they don't.  This makes keyed-balance proxying
more robust with EAP

14 years agoReturn rather than use the same ptr twice
Alan T. DeKok [Mon, 12 Oct 2009 10:14:09 +0000 (12:14 +0200)]
Return rather than use the same ptr twice

14 years agoFix typo
Alan T. DeKok [Wed, 7 Oct 2009 13:08:55 +0000 (15:08 +0200)]
Fix typo

14 years agoEnsure that there is a cleanup event for proxied packets
Alan T. DeKok [Tue, 6 Oct 2009 08:21:45 +0000 (10:21 +0200)]
Ensure that there is a cleanup event for proxied packets

If there was no reply, clean up, reject, etc. the request.

This doesn't matter so much for normal clients, as they will retransmit
and cause the old request to be deleted from the request hash.
But detail requests have random ports (for other reasons), so
they won't be cleaned up by new packets.  Therefore, we need to clean
them up...

14 years agoAdded more debugging messages
Alan T. DeKok [Tue, 6 Oct 2009 08:21:17 +0000 (10:21 +0200)]
Added more debugging messages

14 years agoMark home server dead if it doesn't respond to pings
Alan T. DeKok [Tue, 6 Oct 2009 06:48:02 +0000 (08:48 +0200)]
Mark home server dead if it doesn't respond to pings

14 years agoCheck for undefined types, too
Alan T. DeKok [Mon, 5 Oct 2009 15:32:39 +0000 (17:32 +0200)]
Check for undefined types, too

14 years agoSimplify the code
Alan T. DeKok [Sun, 4 Oct 2009 16:12:12 +0000 (18:12 +0200)]
Simplify the code

14 years agoMore detailed debugging for detail
Alan T. DeKok [Sat, 3 Oct 2009 18:07:53 +0000 (20:07 +0200)]
More detailed debugging for detail

14 years agoBe more restrictive on bad input
Alan T. DeKok [Sat, 3 Oct 2009 18:07:22 +0000 (20:07 +0200)]
Be more restrictive on bad input

Addresses bug #27

14 years agoFix passwords to have even length
Alan T. DeKok [Fri, 2 Oct 2009 08:52:44 +0000 (10:52 +0200)]
Fix passwords to have even length

14 years agoIncrease max_sessions
Alan T. DeKok [Thu, 1 Oct 2009 13:07:51 +0000 (15:07 +0200)]
Increase max_sessions

14 years agoPrint more descriptive error message for too many EAP sessions
Alan T. DeKok [Thu, 1 Oct 2009 13:06:59 +0000 (15:06 +0200)]
Print more descriptive error message for too many EAP sessions

14 years agoMoved illegal attributes to the new dictionary
Alan T. DeKok [Thu, 1 Oct 2009 08:39:22 +0000 (10:39 +0200)]
Moved illegal attributes to the new dictionary

14 years agoAllow old-style dictionary formats, too
Alan T. DeKok [Wed, 30 Sep 2009 07:54:25 +0000 (09:54 +0200)]
Allow old-style dictionary formats, too

14 years agoUse packet codes from libradius
Alan T. DeKok [Tue, 29 Sep 2009 18:24:40 +0000 (20:24 +0200)]
Use packet codes from libradius

14 years agoMoved Ascends illegal attributes to their own file
Alan T. DeKok [Tue, 29 Sep 2009 08:10:59 +0000 (10:10 +0200)]
Moved Ascends illegal attributes to their own file

14 years agoBe more flexible about parsing detail files
Alan T. DeKok [Fri, 25 Sep 2009 09:21:50 +0000 (11:21 +0200)]
Be more flexible about parsing detail files

Allow any operator, not just '='

14 years agoFixed typo && include attrs.access_challenge in build
Alan T. DeKok [Thu, 24 Sep 2009 20:01:54 +0000 (22:01 +0200)]
Fixed typo && include attrs.access_challenge in build

14 years agoFixed compile error
Alan T. DeKok [Thu, 24 Sep 2009 15:37:07 +0000 (17:37 +0200)]
Fixed compile error

14 years ago64-bit fixes.
bduncan [Tue, 22 Sep 2009 16:24:57 +0000 (09:24 -0700)]
64-bit fixes.

14 years agoCheck for NOOP from opendir.c
bduncan [Tue, 22 Sep 2009 16:23:47 +0000 (09:23 -0700)]
Check for NOOP from opendir.c

14 years ago64-bit fixes and return NOOP for AD users.
bduncan [Tue, 22 Sep 2009 16:22:49 +0000 (09:22 -0700)]
64-bit fixes and return NOOP for AD users.

14 years agoGet private key passphrase from keychain using certadmin command.
bduncan [Tue, 22 Sep 2009 16:20:02 +0000 (09:20 -0700)]
Get private key passphrase from keychain using certadmin command.

14 years agoChanged order of code to avoid race conditions
Alan T. DeKok [Thu, 24 Sep 2009 15:12:53 +0000 (17:12 +0200)]
Changed order of code to avoid race conditions

14 years agoEnsure that cached SSL sessions have data
Alan T. DeKok [Thu, 24 Sep 2009 15:12:33 +0000 (17:12 +0200)]
Ensure that cached SSL sessions have data

14 years agoUpdate change logs before 2.1.7 release_2_1_7
Alan T. DeKok [Mon, 14 Sep 2009 14:43:29 +0000 (16:43 +0200)]
Update change logs before 2.1.7

14 years agoPrint out a little more information
Alan T. DeKok [Mon, 14 Sep 2009 12:29:23 +0000 (14:29 +0200)]
Print out a little more information

14 years agoMove "set state" to before log message
Alan T. DeKok [Mon, 14 Sep 2009 12:26:15 +0000 (14:26 +0200)]
Move "set state" to before log message

To avoid race conditions... though I'm not sure why this would
happen

14 years agoAs posted to the list.
Alan T. DeKok [Sun, 13 Sep 2009 15:07:22 +0000 (17:07 +0200)]
As posted to the list.

It isn't made live yet, because we have no idea if it works or not

14 years agoRemoved redundant code
Alan T. DeKok [Sun, 13 Sep 2009 14:01:51 +0000 (16:01 +0200)]
Removed redundant code

14 years agoMake a better name
Alan T. DeKok [Fri, 11 Sep 2009 15:14:09 +0000 (17:14 +0200)]
Make a better name

14 years agoWrap crap with more crap
Alan T. DeKok [Fri, 11 Sep 2009 15:12:06 +0000 (17:12 +0200)]
Wrap crap with more crap

14 years agoMore updates
Alan T. DeKok [Thu, 10 Sep 2009 13:16:40 +0000 (15:16 +0200)]
More updates

14 years agoAdded sample configs for MySQL cluster
Alan T. DeKok [Thu, 10 Sep 2009 13:02:39 +0000 (15:02 +0200)]
Added sample configs for MySQL cluster

14 years agoUpdate to new syntax
Alan T. DeKok [Wed, 9 Sep 2009 14:27:16 +0000 (16:27 +0200)]
Update to new syntax

14 years agoCheck value of Fall-Through, too
Alan T. DeKok [Wed, 9 Sep 2009 12:12:35 +0000 (14:12 +0200)]
Check value of Fall-Through, too

14 years agoBe less forgiving about the allowed operators.
Alan T. DeKok [Tue, 8 Sep 2009 11:56:54 +0000 (13:56 +0200)]
Be less forgiving about the allowed operators.

If they're not a "foo = bar" type of operator, then return
an error

14 years agoWrap header file in protective #ifdef
Alan T. DeKok [Mon, 7 Sep 2009 05:49:34 +0000 (07:49 +0200)]
Wrap header file in protective #ifdef

14 years agoNo need to include modules.h twice
Alan T. DeKok [Mon, 7 Sep 2009 05:48:57 +0000 (07:48 +0200)]
No need to include modules.h twice