jcarneal [Sun, 1 Oct 2000 21:59:14 +0000 (21:59 +0000)]
Updated xxx_authenticate() to accept check/reply pairs
jcarneal [Sun, 1 Oct 2000 21:58:34 +0000 (21:58 +0000)]
Changed module_authenicate function to accept check/reply pairs
jcarneal [Sun, 1 Oct 2000 21:57:59 +0000 (21:57 +0000)]
Changed rad_check_password() to pass user_reply to authenticate funciton
jcarneal [Sun, 1 Oct 2000 21:56:58 +0000 (21:56 +0000)]
Changed module_authenticate function parameters
jcarneal [Sat, 30 Sep 2000 21:50:58 +0000 (21:50 +0000)]
Added user collision code for users file only
jcarneal [Sat, 30 Sep 2000 21:50:07 +0000 (21:50 +0000)]
Added DEBUG for user collide code
jcarneal [Sat, 30 Sep 2000 21:49:28 +0000 (21:49 +0000)]
Added --with-user-collide
jcarneal [Sat, 30 Sep 2000 21:48:46 +0000 (21:48 +0000)]
Added #undef WITH_USERCOLLIDE
jcarneal [Sat, 30 Sep 2000 21:47:14 +0000 (21:47 +0000)]
Added --with-usercollide option (hopefully correctly)
aland [Fri, 29 Sep 2000 18:52:09 +0000 (18:52 +0000)]
Gnome MIB for FreeRADIUS, as submitted by
Jochen Friedrich <jochen+freeradius-devel@scram.de>
For others, see:
http://www.gxsnmp.org/nic/
jcarneal [Fri, 29 Sep 2000 17:21:19 +0000 (17:21 +0000)]
Added trivial patch to show which list a module is added to during debug
aland [Fri, 29 Sep 2000 14:43:37 +0000 (14:43 +0000)]
added 'static const char rcsid[]', so we can tell which version
of what is being built
aland [Fri, 29 Sep 2000 14:41:06 +0000 (14:41 +0000)]
minor bug fix by Jochen Friedrich <jochen+freeradius-devel@scram.de>
aland [Thu, 28 Sep 2000 19:36:24 +0000 (19:36 +0000)]
new attributes from Chris Adams <cmadams@hiwaay.net>
aland [Thu, 28 Sep 2000 19:27:43 +0000 (19:27 +0000)]
redback dictionary supplied by Chris Adams <cmadams@hiwaay.net>
From http://hiwaay.net/~cmadams/files/cistron/
jcarneal [Thu, 28 Sep 2000 00:19:44 +0000 (00:19 +0000)]
Added ability to send multiple, distinct packets.
aland [Tue, 26 Sep 2000 20:45:47 +0000 (20:45 +0000)]
let's see people get credit for their work...
aland [Tue, 26 Sep 2000 20:39:24 +0000 (20:39 +0000)]
snmp / SMUX changes to make the server listen on the smux FD,
and handle snmp / smux requests.
Not tested.
aland [Tue, 26 Sep 2000 20:38:26 +0000 (20:38 +0000)]
minor cleanups and code changes
aland [Tue, 26 Sep 2000 20:06:26 +0000 (20:06 +0000)]
new header file, for simplicity
aland [Tue, 26 Sep 2000 19:55:07 +0000 (19:55 +0000)]
added WITH_SNMP to output
aland [Tue, 26 Sep 2000 17:16:48 +0000 (17:16 +0000)]
smux header file, from patch by
Jochen Friedrich <jochen+freeradius-devel@scram.de>
aland [Tue, 26 Sep 2000 17:14:41 +0000 (17:14 +0000)]
adding SNMP / SMUX support to the server, from patches by
Jochen Friedrich <jochen+freeradius-devel@scram.de>
- adding sources to the tree
- updating make file
aland [Tue, 26 Sep 2000 17:02:09 +0000 (17:02 +0000)]
first step of snmp/smux patches from
Jochen Friedrich <jochen+freeradius-devel@scram.de>
- configuration settings, no source code changes
aland [Tue, 26 Sep 2000 16:18:44 +0000 (16:18 +0000)]
grr... corrected dumb typo
aland [Tue, 26 Sep 2000 15:57:30 +0000 (15:57 +0000)]
print pid's as (int) to radiusd.pid
if proxy_retry_delay == 0, then don't bother doing proxy retries,
as we assume that proxy+synchronous == TRUE.
aland [Mon, 25 Sep 2000 20:11:22 +0000 (20:11 +0000)]
added targets for radzap.o and radwho.o
aland [Mon, 25 Sep 2000 20:07:59 +0000 (20:07 +0000)]
re-arranged #include lines
aland [Mon, 25 Sep 2000 20:02:54 +0000 (20:02 +0000)]
added explicit targets for conffile.c, log.c, and timestr.c
aland [Mon, 25 Sep 2000 19:56:23 +0000 (19:56 +0000)]
cleaning up #include files
aland [Mon, 25 Sep 2000 19:55:57 +0000 (19:55 +0000)]
updated rcsid, instead of sccs_id
aland [Mon, 25 Sep 2000 19:50:52 +0000 (19:50 +0000)]
cleaning up #include files
aland [Mon, 25 Sep 2000 19:45:28 +0000 (19:45 +0000)]
cleaning up #includes
aland [Mon, 25 Sep 2000 19:44:12 +0000 (19:44 +0000)]
re-arranging header files
aland [Mon, 25 Sep 2000 19:40:09 +0000 (19:40 +0000)]
cleaning up #include files, moving them around, deleting not-needed
ones, added #if wrappers
aland [Mon, 25 Sep 2000 19:39:34 +0000 (19:39 +0000)]
added #if wrappers for #include files
aland [Mon, 25 Sep 2000 19:39:19 +0000 (19:39 +0000)]
added checks for netinet/in.h, and sys/types.h
aland [Mon, 25 Sep 2000 17:48:23 +0000 (17:48 +0000)]
fixed typos
aland [Mon, 25 Sep 2000 15:44:35 +0000 (15:44 +0000)]
changed hard-coded 0x7f00000001 to INADDR_LOOPBACK, and
ensured that INADDR_{ANY,LOOPBACK,NONE} are defined.
aland [Mon, 25 Sep 2000 15:24:05 +0000 (15:24 +0000)]
if the realm is a LOCAL one, then still add the 'Realm' attribute,
but don't actually proxy the packet, as we'll end up handling
it anyhow.
This allows the administrator to use Proxy-To-Realm := "localrealm",
and NOT have the server send itself packets in an infinite loop.
aland [Fri, 22 Sep 2000 18:11:27 +0000 (18:11 +0000)]
partial update of patch from "Bruno L. F. Cabral" <bruno@openline.com.br>
Untested...
aland [Fri, 22 Sep 2000 18:06:32 +0000 (18:06 +0000)]
removed Cistron spec file.
Added FreeRADIUS spec file, as given by
"Bruno L. F. Cabral" <bruno@openline.com.br>
aland [Thu, 21 Sep 2000 14:50:13 +0000 (14:50 +0000)]
update mkinstalldirs location to coddle dumb OS's
aland [Thu, 14 Sep 2000 18:41:51 +0000 (18:41 +0000)]
ALWAYS look up host names to IP addresses, otherwise stuff really
fails
aland [Thu, 14 Sep 2000 18:14:01 +0000 (18:14 +0000)]
updated 'compat' mode logic to hopefully work a little better.
Bugs noted by "Bruno L. F. Cabral" <bruno@openline.com.br>
aland [Thu, 14 Sep 2000 18:04:35 +0000 (18:04 +0000)]
last change to update new libltdl location
aland [Thu, 14 Sep 2000 15:44:59 +0000 (15:44 +0000)]
updated 'distclean' for new location of libltdl
aland [Thu, 14 Sep 2000 15:28:38 +0000 (15:28 +0000)]
use variables LIBLTDL and INCLTDL, instead of hard-coded paths.
These variables can get updated by the 'configure' scripts,
if the user already has libltdl on their system.
aland [Thu, 14 Sep 2000 15:26:18 +0000 (15:26 +0000)]
moved libltdl up a directory, to keep the GNU libtool
scripts happy
aland [Thu, 14 Sep 2000 15:25:48 +0000 (15:25 +0000)]
find and substitute the variable top_builddir
aland [Thu, 14 Sep 2000 15:19:53 +0000 (15:19 +0000)]
check if libltdl is to be installable, and set LIBLTDL and INCLTDL
to the appropriate paths
aland [Wed, 13 Sep 2000 18:07:19 +0000 (18:07 +0000)]
don't run dynamic log code, as it should be re-written
aland [Wed, 13 Sep 2000 16:27:19 +0000 (16:27 +0000)]
updated to use radlog() instead of log(), to avoid any possible
conflicts with the math library log() function.
aland [Wed, 13 Sep 2000 16:18:35 +0000 (16:18 +0000)]
changed log() to radlog(). Over-loading function names
is probably a bad thing...
aland [Wed, 13 Sep 2000 16:13:33 +0000 (16:13 +0000)]
actually print out the error, if we have a problem in select()
aland [Thu, 7 Sep 2000 20:17:33 +0000 (20:17 +0000)]
a number of more updates to clean up the man pages
aland [Thu, 7 Sep 2000 19:57:57 +0000 (19:57 +0000)]
formatting changes
aland [Thu, 7 Sep 2000 19:45:12 +0000 (19:45 +0000)]
set Auth-Type := again...
aland [Wed, 6 Sep 2000 16:25:47 +0000 (16:25 +0000)]
Cistron 'users' file compatibility mode is off, by default
aland [Wed, 6 Sep 2000 16:25:14 +0000 (16:25 +0000)]
added support for a Cistron 'compat' mode, so that people can
copy over their old 'users' file, and have them mostly work
with FreeRADIUS
aland [Wed, 6 Sep 2000 15:47:37 +0000 (15:47 +0000)]
more documentation, more examples
aland [Tue, 5 Sep 2000 18:00:42 +0000 (18:00 +0000)]
Place-holder file for accounting users file
aland [Tue, 5 Sep 2000 17:56:46 +0000 (17:56 +0000)]
added even more text. It might actually make sense now.
aland [Tue, 5 Sep 2000 17:07:14 +0000 (17:07 +0000)]
more minor updates
aland [Tue, 5 Sep 2000 17:04:53 +0000 (17:04 +0000)]
New documentation describing the format of the 'users' file.
This should help to make some things clearer...
aland [Tue, 5 Sep 2000 16:37:00 +0000 (16:37 +0000)]
added a lot of text describing the format of the configuration
file.
aland [Tue, 5 Sep 2000 16:13:37 +0000 (16:13 +0000)]
updated the Auth-Type attributes to use ':=' and '+=', where
appropriate
aland [Tue, 5 Sep 2000 16:11:17 +0000 (16:11 +0000)]
*add* in the auth-tyep attribute, and don't *over-ride* it.
aland [Tue, 5 Sep 2000 14:40:22 +0000 (14:40 +0000)]
update the configuration file reader to ignore semi-colons
pam [Tue, 5 Sep 2000 07:12:13 +0000 (07:12 +0000)]
Fixed compilation for OpenLDAP _without_ Cyrus SASL libraries
Fixed module operation if only authentication is done via LDAP (no authorization)
Fixed module to support new instance framework
aland [Fri, 1 Sep 2000 19:36:22 +0000 (19:36 +0000)]
configuration option to turn DNS lookups on/off.
Removed the now-redundant command-line option to do the same
thing.
aland [Fri, 1 Sep 2000 19:33:24 +0000 (19:33 +0000)]
configuration option to turn DNS lookups on or off.
aland [Fri, 1 Sep 2000 18:21:44 +0000 (18:21 +0000)]
add port numbers to log messages about requests from/to clients.
e.g. "127.0.0.1:1038", instead of "127.0.0.1"
If we see a duplicate request, then send the duplicate reply
to the port of the *new* request, not the port of the *old* request
aland [Fri, 1 Sep 2000 17:58:25 +0000 (17:58 +0000)]
the setuptimeout() difference MAY be negative, so we short-circuit
if it's <=0, not if it's ==0.
Also, if the smallest is < 0, then we set it to zero.
aland [Fri, 1 Sep 2000 15:38:50 +0000 (15:38 +0000)]
only kill all of the processes in the process group if we're
the process group leader.
aland [Fri, 1 Sep 2000 14:52:14 +0000 (14:52 +0000)]
add destination port to the debugging output, so we can
see where the packet is going.
aland [Fri, 1 Sep 2000 14:43:46 +0000 (14:43 +0000)]
made pairlist_read take a PAIR_LIST **, and return 0/-1 for OK/ERROR
This allows the functions calling it to tell the difference between
an empty PAIR_LIST, and an erroneous one.
aland [Fri, 1 Sep 2000 13:56:25 +0000 (13:56 +0000)]
corrected behaviour of setuptimeout() when the curreq->timestamp
is zero. Bug noted by Alan Curry.
aland [Thu, 31 Aug 2000 21:07:27 +0000 (21:07 +0000)]
Patches from "Alan Curry" <pacman-radius@cqc.com>
I invented a new structure called CONF_ITEM, and made both CONF_SECTION
and CONF_PAIR subclasses (in the C-only libXt/libgtk style) of it. I also
made all the CONF_* structs opaque to everything outside conffile.c, which I
gather was the original intent since everything is using those *_next()
accessor functions instead of walking the ->next list directly.
The semicolon stuff goes away, and now you can mix pairs (including
word-on-a-line-by-itself) with sections, and order is preserved:
aland [Thu, 31 Aug 2000 20:52:36 +0000 (20:52 +0000)]
cleaned up option handling to make somewhat more sense
aland [Thu, 31 Aug 2000 20:40:36 +0000 (20:40 +0000)]
made the dictionary case insensitive, by default. The user
can still over-ride this on the configure line.
Added -D_REENTRANT to CFLAGS, if building with threads. This
tells systems like Solaris to be more careful...
aland [Thu, 31 Aug 2000 20:38:48 +0000 (20:38 +0000)]
made internal version of strncasecmp actually case insensitive
aland [Thu, 31 Aug 2000 20:32:27 +0000 (20:32 +0000)]
made all of the 'strcmp' operations use DICT_STRCMP, which
is case (in)sensitive, depending on the system configuration
aland [Fri, 25 Aug 2000 19:49:19 +0000 (19:49 +0000)]
minor cleanups on code, with more comments, less global variables,
but no functionality changes
aland [Fri, 25 Aug 2000 18:23:59 +0000 (18:23 +0000)]
print out thread numbers, not pthread_id. pthread_id MAY be
a structure/pointer/integer on different platforms, so it's not
safe to print it out like an 'int'.
Each thread now has a unique number assigned to it, and these
numbers are printed out during debugging.
aland [Fri, 25 Aug 2000 15:09:14 +0000 (15:09 +0000)]
renamed a few variables for RFC complicance.
Patch from Alan Curry
aland [Fri, 25 Aug 2000 14:55:56 +0000 (14:55 +0000)]
Patch from "Alan Curry" <pacman-radius@cqc.com>
Allow ';' to denote empty conf-sections.
aland [Fri, 25 Aug 2000 14:48:14 +0000 (14:48 +0000)]
Patches from "Alan Curry" <pacman-radius@cqc.com>
Move the module auth/acct/autz decisions from raddb/modules
to raddb/radiusd.conf.
Add module instantiation, to have future allowances for
multiple versions of the same module, with different configurations.
Associated minor cleanups
aland [Fri, 25 Aug 2000 14:29:20 +0000 (14:29 +0000)]
cleaned up the handling of proxy replies, so that we don't have
un-initialized variables.
Bug found by "Alan Curry" <pacman-radius@cqc.com>
aland [Thu, 24 Aug 2000 21:52:29 +0000 (21:52 +0000)]
converted proxy_setuptimeout() to a *generic* setuptimeout()
function, which waits the maximum amount of time before it wakes
up.
This code is a lot better than the commit from an hour ago.
aland [Thu, 24 Aug 2000 20:00:00 +0000 (20:00 +0000)]
fix the server so that it sleeps for as long as possible
between requests.
i.e. If the proxy code needs servicing in N seconds, sleep for N
seconds. If the master request list needs servicing sooner, then
sleep for less time.
If nothing needs servicing, then sleep forever, until a new
packet comes in.
aland [Thu, 24 Aug 2000 19:38:59 +0000 (19:38 +0000)]
if the packet code is well-known (i.e. a standard), then
rad_recv() and rad_send() now print out the RFC names for the
packet code, instead of 'code=4', etc.
This makes more sense to me.
aland [Thu, 24 Aug 2000 19:23:59 +0000 (19:23 +0000)]
minor re-arrangements
aland [Tue, 22 Aug 2000 14:57:51 +0000 (14:57 +0000)]
added more value names for integer attributes, as taken from rfc2865
aland [Tue, 22 Aug 2000 14:35:27 +0000 (14:35 +0000)]
corrected typo in trusted stuff
aland [Mon, 21 Aug 2000 19:30:15 +0000 (19:30 +0000)]
corrected typos
aland [Mon, 21 Aug 2000 19:28:42 +0000 (19:28 +0000)]
don't smash the operators any more. pairmove() now does the
right thing.
aland [Mon, 21 Aug 2000 19:28:04 +0000 (19:28 +0000)]
cleaned up code a little.
Removed overflow/invalid memory access from pairmove2().
aland [Mon, 21 Aug 2000 18:58:48 +0000 (18:58 +0000)]
allow += in check item lists, too
aland [Mon, 21 Aug 2000 18:49:55 +0000 (18:49 +0000)]
when doing a pairmove(), the relational operators >=, <=, etc.
are all equivalent to +=. This isn't perfect, but it's better
than what we had before, I think.