aland [Mon, 12 Feb 2007 15:07:37 +0000 (15:07 +0000)]
Pass the correct arguments to packet list walk in rl_deinit()
aland [Mon, 12 Feb 2007 14:51:52 +0000 (14:51 +0000)]
call paircompare_unregister
aland [Mon, 12 Feb 2007 14:47:07 +0000 (14:47 +0000)]
Unregister PW_TIME_OF_DAY, too
aland [Mon, 12 Feb 2007 14:45:40 +0000 (14:45 +0000)]
We register Expiration, not Current-Time.
aland [Mon, 12 Feb 2007 14:24:20 +0000 (14:24 +0000)]
Don't free strings allocated by cf_section_parse()
aland [Mon, 12 Feb 2007 14:07:20 +0000 (14:07 +0000)]
Free the parsed data prior to calling the modules detach()
routine. This means that any string that gets parsed by a
module automatically gets free'd by the configuration file code,
so that the modules don't have to do it.
Next step is to remove most of the calls to free() in the
modules. After that, if there's nothing in the modules detach
routine, the configuration file code can clean up the module
instance, too
aland [Mon, 12 Feb 2007 14:01:41 +0000 (14:01 +0000)]
Don't free formatstring on instantiate, free it on detach
aland [Mon, 12 Feb 2007 13:20:39 +0000 (13:20 +0000)]
Added code to free module configuration automagically.
It's not currently used because of conflicts with the client
code, which calls cf_section_parse(), and then client_free()
aland [Mon, 12 Feb 2007 13:07:02 +0000 (13:07 +0000)]
Free allocated memory
aland [Mon, 12 Feb 2007 12:49:15 +0000 (12:49 +0000)]
Corrected typo
aland [Mon, 12 Feb 2007 12:37:43 +0000 (12:37 +0000)]
Free allocated memory
aland [Mon, 12 Feb 2007 12:37:27 +0000 (12:37 +0000)]
Added '-m' option from 1.1
aland [Mon, 12 Feb 2007 10:44:52 +0000 (10:44 +0000)]
Don't initialize the wait hash if we're not threaded
aland [Mon, 12 Feb 2007 10:43:35 +0000 (10:43 +0000)]
Correctly parse Base64 encoded strings
aland [Mon, 12 Feb 2007 10:39:28 +0000 (10:39 +0000)]
Include filenames, too
aland [Mon, 12 Feb 2007 10:38:19 +0000 (10:38 +0000)]
Added 'base' and 'variables' to CONF_SECTION, so that the
allocated memory can get automatically freed.
Ported from 1.1.x
aland [Mon, 12 Feb 2007 10:36:21 +0000 (10:36 +0000)]
Free allocated memory
aland [Mon, 12 Feb 2007 10:01:19 +0000 (10:01 +0000)]
use chap->foo, not request->password->foo
aland [Fri, 9 Feb 2007 15:40:30 +0000 (15:40 +0000)]
Free memory on detach
aland [Fri, 9 Feb 2007 15:32:13 +0000 (15:32 +0000)]
When a duplicate attribute is found, delete the old one
aland [Fri, 9 Feb 2007 15:05:37 +0000 (15:05 +0000)]
Fix lrad_hash_table_free()
aland [Fri, 9 Feb 2007 13:09:35 +0000 (13:09 +0000)]
request->password is User-Password or CHAP-Password.
That's probably a bad idea
aland [Fri, 9 Feb 2007 13:09:10 +0000 (13:09 +0000)]
Caching request->password is probably a bad idea
aland [Fri, 9 Feb 2007 13:05:25 +0000 (13:05 +0000)]
Use pairfind() to get CHAP-Password, not request->password
aland [Fri, 9 Feb 2007 13:03:18 +0000 (13:03 +0000)]
Don't start if radius_dir is globally writable
Relax the checks on readability
aland [Fri, 9 Feb 2007 10:52:32 +0000 (10:52 +0000)]
Updated with new attributes from bug #417
aland [Fri, 9 Feb 2007 10:38:25 +0000 (10:38 +0000)]
use correct lineo.
We could really us a cf_item_lineno()
aland [Fri, 9 Feb 2007 10:29:21 +0000 (10:29 +0000)]
cleaned up the exec stuff again, and moved the "wait for
10 seconds or give up" code into rad_waitpid, where it's easier
to manage
aland [Thu, 8 Feb 2007 14:06:28 +0000 (14:06 +0000)]
If a packet is too long, truncate at the last good attribute
aland [Thu, 8 Feb 2007 10:25:09 +0000 (10:25 +0000)]
Don't leave 'vp' pointing to a deleted attribute
aland [Thu, 8 Feb 2007 09:38:27 +0000 (09:38 +0000)]
Apply patch from 1.67 to accounting requests, too
pnixon [Tue, 6 Feb 2007 20:06:08 +0000 (20:06 +0000)]
Change BEGIN to SQL99 conformant START TRANSACTION
pnixon [Tue, 6 Feb 2007 16:37:12 +0000 (16:37 +0000)]
Add non-mandatory pool-check statement to allow detection of whether the Pool does not exist or rather is simply full
aland [Tue, 6 Feb 2007 14:06:55 +0000 (14:06 +0000)]
Support Status-Server on accounting ports, too
aland [Tue, 6 Feb 2007 14:06:32 +0000 (14:06 +0000)]
Don't copy too much data && crash the client
aland [Tue, 6 Feb 2007 13:57:09 +0000 (13:57 +0000)]
Return -1 on error, not NULL
The file permission checks here are probably too strong..
aland [Tue, 6 Feb 2007 13:46:37 +0000 (13:46 +0000)]
One more check
aland [Tue, 6 Feb 2007 13:02:10 +0000 (13:02 +0000)]
Build on non-IPv6 aware systems
aland [Tue, 6 Feb 2007 12:50:14 +0000 (12:50 +0000)]
Build on systems without IPv6 addresses
aland [Tue, 6 Feb 2007 10:12:20 +0000 (10:12 +0000)]
Check for existence of c->shortname before dereferencing it
pnixon [Mon, 5 Feb 2007 17:57:27 +0000 (17:57 +0000)]
Added new option sqlippool_log_clear to log when an IP address is deallocated
pnixon [Mon, 5 Feb 2007 12:17:13 +0000 (12:17 +0000)]
On query error, emit the failed query to the log so we can track it down.
pnixon [Mon, 5 Feb 2007 08:43:32 +0000 (08:43 +0000)]
Add configure to the tree
pnixon [Sun, 4 Feb 2007 23:04:30 +0000 (23:04 +0000)]
Change the order back. Turns out that we get unwanted No Pool-Name defined logs even with static ips. oops
pnixon [Sun, 4 Feb 2007 17:12:06 +0000 (17:12 +0000)]
Add sqlippool and acctlog to the list of stable modules
pnixon [Sun, 4 Feb 2007 15:28:56 +0000 (15:28 +0000)]
Apply SUSE's edir.patch to explicitly set edir_account_policy_check=no. Wont affect no edir users and saves a patch for SUSE
pnixon [Sun, 4 Feb 2007 15:10:08 +0000 (15:10 +0000)]
Applied radiusd-pamd.diff. thanks to SUSE
pnixon [Sun, 4 Feb 2007 14:39:46 +0000 (14:39 +0000)]
Add -pie patch thanks to SUSE
pnixon [Sat, 3 Feb 2007 23:09:52 +0000 (23:09 +0000)]
Add new log type (L_ACCT) and use it in rlm_acctlog
pnixon [Sat, 3 Feb 2007 23:08:44 +0000 (23:08 +0000)]
Add Acct-Session-Time to Stop logs and adjust other queries slightly
pnixon [Sat, 3 Feb 2007 22:13:43 +0000 (22:13 +0000)]
Fix spelling in comments
pnixon [Sat, 3 Feb 2007 19:19:04 +0000 (19:19 +0000)]
Return RLM_MODULE_FAIL instead of RLM_MODULE_NOOP if there is no SQL socket
pnixon [Sat, 3 Feb 2007 19:03:52 +0000 (19:03 +0000)]
Switch the order of Framed-IP-Address and Pool-Name check. May make a small speed increase for those with a mix of static and dynamic addresses
aland [Tue, 30 Jan 2007 15:46:57 +0000 (15:46 +0000)]
List include files as being OK to install
aland [Tue, 30 Jan 2007 13:10:02 +0000 (13:10 +0000)]
Don't allow globally readable or writeable configuration files.
aland [Tue, 30 Jan 2007 13:02:05 +0000 (13:02 +0000)]
Don't allow globally writable dictionaries. This allows any
user to control the behavior of the server.
aland [Sun, 28 Jan 2007 08:09:18 +0000 (08:09 +0000)]
As posted to the list
pnixon [Sat, 27 Jan 2007 15:09:58 +0000 (15:09 +0000)]
Add version line and allow code to be executed from the command line
aland [Fri, 26 Jan 2007 09:57:14 +0000 (09:57 +0000)]
use fork() where we don't need to wait.
We don't need the argument to rad_fork() any more.
Always call waitpid(), even if no one called rad_fork(), because
someone might have called fork()
aland [Fri, 26 Jan 2007 09:39:19 +0000 (09:39 +0000)]
Use strlcpy rather than strncpy, and use the size of the output
buffer rather than the size of the input buffer.
aland [Wed, 24 Jan 2007 12:38:25 +0000 (12:38 +0000)]
Pull fix from 1.1 branch
pnixon [Tue, 23 Jan 2007 16:26:06 +0000 (16:26 +0000)]
Clean up the comments a little
pnixon [Tue, 23 Jan 2007 02:00:06 +0000 (02:00 +0000)]
Update comments to correctly reflect which detail files contain User Passwords and additionally add an example "suppress" statement to each.
pnixon [Mon, 22 Jan 2007 16:51:27 +0000 (16:51 +0000)]
fix 2 minor typos in the comments
pnixon [Wed, 10 Jan 2007 16:06:42 +0000 (16:06 +0000)]
teach sqlippool to report if it allocated and IP or not via return code. Closed bug 414
pnixon [Tue, 9 Jan 2007 01:58:18 +0000 (01:58 +0000)]
Fix Bug 413 sqlippool_query1 freeing results too early
Cleanup a bunch of other rubbish including a memory leak in sql_postgresql.c
pnixon [Mon, 8 Jan 2007 19:19:49 +0000 (19:19 +0000)]
Someone added attrs.access_reject and attrs.accounting_response to radiusd.conf but forgot to add them to the Makefile creating a broken install.
pnixon [Mon, 8 Jan 2007 16:14:53 +0000 (16:14 +0000)]
in case of a non-SQL_DOWN fatal error that occurs on the postgresql database server the function incorrectly returns 0 instead of -1
pnixon [Fri, 5 Jan 2007 15:27:21 +0000 (15:27 +0000)]
Add X-Ascend-Session-Svr-Key to table to allow Packet of Disconnects
pnixon [Tue, 2 Jan 2007 12:34:30 +0000 (12:34 +0000)]
Fixed incorrect spelling of "occurred" in 2 places
nbk [Thu, 14 Dec 2006 14:20:36 +0000 (14:20 +0000)]
Include "autoconf.h" before testing the macro WITH_SNMP.
nbk [Thu, 14 Dec 2006 14:20:01 +0000 (14:20 +0000)]
Include "autoconf.h" before testing the macro WITH_UDPFROMTO.
pnixon [Wed, 13 Dec 2006 13:33:34 +0000 (13:33 +0000)]
Fix list address
aland [Tue, 12 Dec 2006 19:56:48 +0000 (19:56 +0000)]
Fix stupid bug. rad_send() takes secret as second argument,
listener->send takes REQUEST*
pnixon [Tue, 12 Dec 2006 16:15:06 +0000 (16:15 +0000)]
Minor formatting cleanups.
pnixon [Tue, 12 Dec 2006 16:03:38 +0000 (16:03 +0000)]
Code is a bit clearer this way. Functionality should not change.
pnixon [Tue, 12 Dec 2006 12:08:48 +0000 (12:08 +0000)]
Fix syntax typo
pnixon [Tue, 12 Dec 2006 11:51:44 +0000 (11:51 +0000)]
Fix a STUPID memory leak introduced by yours truely..
bjordanov [Mon, 4 Dec 2006 08:49:38 +0000 (08:49 +0000)]
Added new hash RAD_CONFIG
aland [Wed, 29 Nov 2006 23:01:11 +0000 (23:01 +0000)]
Workarounds for Vista
aland [Tue, 28 Nov 2006 19:08:41 +0000 (19:08 +0000)]
If less than 4 bytes are read, tell the caller that 0 bytes were read
If the MSG_PEEK'd length says that the packet is less than
the RADIUS header length, tell the caller that 0 bytes were read.
If the MSG_PEEK'd length says that the packet is more than the
allowed RFC maximum of 4k, tell the caller that more than 4K was
read, but do NOT allocate memory for the packet.
rad_recv() takes care of catching these cases, and returns a
descriptive error to the caller.
In all 3 cases, recvmsg() is called to read a few bytes of data
from the packet. The kernel takes care of deciding that the packet
was read, and discards the rest of the data.
aland [Tue, 28 Nov 2006 18:24:02 +0000 (18:24 +0000)]
Updated names so that upper/lower case matches. Patch from
Daniel Larsson
aland [Mon, 27 Nov 2006 18:52:51 +0000 (18:52 +0000)]
VALUEs can only be defined for BYTE, SHORT, and INTEGER
aland [Mon, 27 Nov 2006 18:43:06 +0000 (18:43 +0000)]
Simplify parsing of byte/short/string, based on a patch by
Krzysztof Oledzki
aland [Sun, 26 Nov 2006 19:48:57 +0000 (19:48 +0000)]
pull fixes from branch_1_1
aland [Sun, 26 Nov 2006 19:48:10 +0000 (19:48 +0000)]
ATTRIBUTEs of type STRING shouldn't have VALUEs
aland [Sun, 26 Nov 2006 19:44:55 +0000 (19:44 +0000)]
Correct definitions for stats, from
http://www.juniper.net/techpubs/software/erx/junose60/unisphere6-0.dct
aland [Sun, 26 Nov 2006 19:29:19 +0000 (19:29 +0000)]
Include dictionary.sofaware
aland [Sun, 26 Nov 2006 18:21:42 +0000 (18:21 +0000)]
From branch_1_1
aland [Sun, 26 Nov 2006 18:21:02 +0000 (18:21 +0000)]
As posted to the list by Jacques Bourdeau
aland [Fri, 24 Nov 2006 14:54:39 +0000 (14:54 +0000)]
Add MS-MPPE-*, as suggested by Stefan Winter
aland [Wed, 22 Nov 2006 22:48:43 +0000 (22:48 +0000)]
Remove rfc_clean function
aland [Wed, 22 Nov 2006 21:48:35 +0000 (21:48 +0000)]
Filter the contents of Access-Reject and Accounting-Response packets,
to enforce RFC compliance.
aland [Wed, 22 Nov 2006 21:44:19 +0000 (21:44 +0000)]
Added 'key' field, so that we can later get rid of the rfc_clean()
function.
Added a bit of a wildcard for to handle 'Vendor-Specific =* ANY'
aland [Wed, 22 Nov 2006 21:37:12 +0000 (21:37 +0000)]
Move buffer to enclosing block, so we don't point to a buffer
on the stack which may be invalid
aland [Wed, 22 Nov 2006 17:00:34 +0000 (17:00 +0000)]
Pull fix from branch_1_1, so proxied EAP replies work
aland [Wed, 22 Nov 2006 17:00:01 +0000 (17:00 +0000)]
Permit EAP-Message and State from the home server, so that
EAP works. Bug noted by Stefan Winter
nbk [Mon, 20 Nov 2006 13:45:25 +0000 (13:45 +0000)]
Update the types that were mistakenly inversed. (closes: #401)
Thanks to Andrea Scharfe <schmedi@gmail.com>
nbk [Mon, 20 Nov 2006 13:44:46 +0000 (13:44 +0000)]
Update the types that were mistakenly inversed. (closes: #401)
Thanks to Andrea Scharfe <schmedi@gmail.com>
nbk [Sun, 19 Nov 2006 18:46:31 +0000 (18:46 +0000)]
Pull from CVS head:
Patch to allow /var/run to be a tmpfs.