]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Wed, 2 Aug 2017 12:46:26 +0000 (08:46 -0400)]
Merge pull request #2030 from pwdng/ldap_openssl
For CentOS, use LTB's openldap packages intead of the distribution ve…
Michael Ducharme [Wed, 2 Aug 2017 04:39:57 +0000 (23:39 -0500)]
Fix for mikrotik_snmp not working
If the interface table is above a certain size, MikroTik does not show the ppp interfaces in the interfae table, but they are always in the mtxrInterfaceStatsName in MIKROTIK-MIB. Simply switching to that fixes the problem and should not break compatibility for current users.
Philippe Wooding [Mon, 31 Jul 2017 12:35:17 +0000 (14:35 +0200)]
For CentOS, use LTB's openldap packages intead of the distribution version to avoid linking against NSS. Linking against NSS & OpenLDAP causes issues that manifest themselves through unexplained connection errors.
Alan T. DeKok [Sun, 30 Jul 2017 15:15:17 +0000 (11:15 -0400)]
more documentation
Alan T. DeKok [Sun, 30 Jul 2017 14:58:29 +0000 (10:58 -0400)]
add more notes on TODO's
Alan T. DeKok [Sun, 30 Jul 2017 14:51:23 +0000 (10:51 -0400)]
Add 'replicate' flag.
This allows the module to send packets and not expect responses
Alan T. DeKok [Sat, 29 Jul 2017 15:44:32 +0000 (11:44 -0400)]
implement packet retransmissions
Alan T. DeKok [Sat, 29 Jul 2017 14:07:35 +0000 (10:07 -0400)]
rename fields to be clearer
Alan T. DeKok [Fri, 28 Jul 2017 21:44:34 +0000 (17:44 -0400)]
rename configuration items
Alan T. DeKok [Fri, 28 Jul 2017 18:07:43 +0000 (14:07 -0400)]
sample config for RADIUS module
Alan T. DeKok [Fri, 28 Jul 2017 18:05:41 +0000 (14:05 -0400)]
add support for 'type = Access-Request'
Alan T. DeKok [Fri, 28 Jul 2017 17:58:31 +0000 (13:58 -0400)]
use correct offsets for non-void types
Alan T. DeKok [Fri, 28 Jul 2017 16:53:46 +0000 (12:53 -0400)]
notes on what to do for finishing the module
Alan T. DeKok [Fri, 28 Jul 2017 14:15:07 +0000 (10:15 -0400)]
insert localized message into the local queue
Alan T. DeKok [Fri, 28 Jul 2017 13:33:43 +0000 (09:33 -0400)]
increment 'i' when we use it, not at the end of the loop
which means if we 'continue' after running a callback, it will
still increment 'i'
Alan T. DeKok [Fri, 28 Jul 2017 13:23:28 +0000 (09:23 -0400)]
more assertions
Alan T. DeKok [Fri, 28 Jul 2017 13:16:48 +0000 (09:16 -0400)]
remove "namespace" parameter
Alan T. DeKok [Fri, 28 Jul 2017 13:15:46 +0000 (09:15 -0400)]
assert that listeners are loaded, and they're not old-style
Alan T. DeKok [Fri, 28 Jul 2017 02:28:03 +0000 (22:28 -0400)]
typo
Arran Cudbard-Bell [Fri, 28 Jul 2017 12:54:23 +0000 (08:54 -0400)]
Typo
Arran Cudbard-Bell [Fri, 28 Jul 2017 12:54:16 +0000 (08:54 -0400)]
Formatting
Alan T. DeKok [Fri, 28 Jul 2017 01:36:45 +0000 (21:36 -0400)]
remove unused macro
Alan T. DeKok [Thu, 27 Jul 2017 20:35:26 +0000 (16:35 -0400)]
return "no data"
Alan T. DeKok [Thu, 27 Jul 2017 20:33:41 +0000 (16:33 -0400)]
descriptive messages on timeouts
Alan T. DeKok [Thu, 27 Jul 2017 20:33:29 +0000 (16:33 -0400)]
APP may not have set NAK
Alan T. DeKok [Thu, 27 Jul 2017 20:01:24 +0000 (16:01 -0400)]
comments go at the END of lines
Alan T. DeKok [Thu, 27 Jul 2017 19:35:26 +0000 (15:35 -0400)]
set connection in udp link
Alan T. DeKok [Thu, 27 Jul 2017 19:14:44 +0000 (15:14 -0400)]
reset was_sleeping when we're not sleeping
Alan T. DeKok [Thu, 27 Jul 2017 19:05:22 +0000 (15:05 -0400)]
fix debug code
Alan T. DeKok [Thu, 27 Jul 2017 19:03:01 +0000 (15:03 -0400)]
enforce max_request_time once per second
Alan T. DeKok [Thu, 27 Jul 2017 18:09:04 +0000 (14:09 -0400)]
more cleanups on exit case, and track num_active
Alan T. DeKok [Thu, 27 Jul 2017 16:07:13 +0000 (12:07 -0400)]
simplify loop over dlist
Alan T. DeKok [Thu, 27 Jul 2017 15:46:19 +0000 (11:46 -0400)]
resume the time tracking so that it's removed from the right lists
Alan T. DeKok [Thu, 27 Jul 2017 15:36:19 +0000 (11:36 -0400)]
suppress multiple sleeping signals.
If we're running single-threaded. then the network && worker
are in the same event list. So the worker sending a signal
saying "I'm sleeping" will cause the event loop to get run
again, including the worker pre-handler, which signals the
event loop that it's sleeping.
It's best to remove infinite loops.
Alan T. DeKok [Thu, 27 Jul 2017 15:09:29 +0000 (11:09 -0400)]
use the allocated ID, not a hard-coded 0
Arran Cudbard-Bell [Thu, 27 Jul 2017 19:32:08 +0000 (15:32 -0400)]
Support passlib's alt-64
We already did the work to make padding bytes optional in a previous commit
Alan T. DeKok [Thu, 27 Jul 2017 13:56:58 +0000 (09:56 -0400)]
notes on 'todo's
Alan T. DeKok [Thu, 27 Jul 2017 13:55:36 +0000 (09:55 -0400)]
just make authenticate actions the same as authorize
Alan T. DeKok [Thu, 27 Jul 2017 13:53:37 +0000 (09:53 -0400)]
fix authenticate hacks
the thing in "authenticate" are no longer bare modules.
They are now always subsections
Alan T. DeKok [Thu, 27 Jul 2017 13:44:09 +0000 (09:44 -0400)]
OK is higher priority than NOOP
Alan T. DeKok [Thu, 27 Jul 2017 13:29:43 +0000 (09:29 -0400)]
clear pending flags as necessary
Alan T. DeKok [Thu, 27 Jul 2017 13:17:33 +0000 (09:17 -0400)]
fix assertion
Alan T. DeKok [Thu, 27 Jul 2017 13:08:23 +0000 (09:08 -0400)]
move assert to run-time check
Alan T. DeKok [Thu, 27 Jul 2017 13:08:07 +0000 (09:08 -0400)]
check for allowed packets
Alan T. DeKok [Thu, 27 Jul 2017 13:07:54 +0000 (09:07 -0400)]
doxygen
Alan T. DeKok [Thu, 27 Jul 2017 02:07:50 +0000 (22:07 -0400)]
signing packets would be a good idea...
Alan T. DeKok [Thu, 27 Jul 2017 01:59:43 +0000 (21:59 -0400)]
initialize queues
Alan T. DeKok [Thu, 27 Jul 2017 01:59:10 +0000 (21:59 -0400)]
increment num_free when adding to the free list
Alan T. DeKok [Thu, 27 Jul 2017 01:58:51 +0000 (21:58 -0400)]
-Wunused
Alan T. DeKok [Thu, 27 Jul 2017 00:08:46 +0000 (20:08 -0400)]
make ldap optional...
Alan T. DeKok [Wed, 26 Jul 2017 20:48:51 +0000 (16:48 -0400)]
remove unnecessary scheduler messages
Alan T. DeKok [Wed, 26 Jul 2017 20:47:43 +0000 (16:47 -0400)]
add more notes
Alan T. DeKok [Wed, 26 Jul 2017 20:42:09 +0000 (16:42 -0400)]
fill in most of conn_read()
Alan T. DeKok [Wed, 26 Jul 2017 20:20:11 +0000 (16:20 -0400)]
more cleanups and finalization
Alan T. DeKok [Wed, 26 Jul 2017 20:17:44 +0000 (16:17 -0400)]
remove io_inst_size
Alan T. DeKok [Wed, 26 Jul 2017 19:49:28 +0000 (15:49 -0400)]
more detach / destructor calls
Alan T. DeKok [Wed, 26 Jul 2017 19:25:13 +0000 (15:25 -0400)]
notes on why we manually do talloc_free
Alan T. DeKok [Wed, 26 Jul 2017 18:53:52 +0000 (14:53 -0400)]
more of the new / better method of managing connections
Alan T. DeKok [Wed, 26 Jul 2017 15:47:48 +0000 (11:47 -0400)]
re-do rlm_radius.
The old method was way too complicated. So the new method
is better.
The connection code will be re-added in the next few commits.
Alan T. DeKok [Wed, 26 Jul 2017 14:05:19 +0000 (10:05 -0400)]
rename
Alan T. DeKok [Wed, 26 Jul 2017 13:56:07 +0000 (09:56 -0400)]
simplify the thread config
Alan T. DeKok [Wed, 26 Jul 2017 13:05:57 +0000 (09:05 -0400)]
fix checks for ipaddr is set
Arran Cudbard-Bell [Wed, 26 Jul 2017 23:09:03 +0000 (19:09 -0400)]
Add support for PBKDF2
Arran Cudbard-Bell [Wed, 26 Jul 2017 23:06:30 +0000 (19:06 -0400)]
Fix base64 decode
Code is now:
- Less branchy
- Easier to understand
- Calculates the correct number of bytes needed for the output buffer instead of estimating.
- Returns negative offsets for errors.
- Writes errors using fr_strerror_printf
Matthew Newton [Wed, 26 Jul 2017 22:03:02 +0000 (23:03 +0100)]
typo
Matthew Newton [Wed, 26 Jul 2017 21:57:59 +0000 (22:57 +0100)]
Correct category names
Matthew Newton [Wed, 26 Jul 2017 21:49:00 +0000 (22:49 +0100)]
README files for modules that were missing or incorrect
Alan T. DeKok [Tue, 25 Jul 2017 19:56:15 +0000 (15:56 -0400)]
typos
Alan T. DeKok [Tue, 25 Jul 2017 19:52:56 +0000 (15:52 -0400)]
notes and typos
Alan T. DeKok [Tue, 25 Jul 2017 19:48:49 +0000 (15:48 -0400)]
add unlang_signal callbacks
Alan T. DeKok [Tue, 25 Jul 2017 19:47:23 +0000 (15:47 -0400)]
be gracious in signaling
Alan T. DeKok [Tue, 25 Jul 2017 19:44:12 +0000 (15:44 -0400)]
remove the "waiting_to_die" list.
If the module can't deal with this, well, too bad
Alan T. DeKok [Tue, 25 Jul 2017 19:42:33 +0000 (15:42 -0400)]
call FR_IO_DONE when exiting
Alan T. DeKok [Tue, 25 Jul 2017 19:25:12 +0000 (15:25 -0400)]
re-work rlm_radius to be much simpler
Alan T. DeKok [Tue, 25 Jul 2017 19:01:43 +0000 (15:01 -0400)]
clean up the requests on fr_worker_destroy()
We still need to check ordering of network / worker deletion,
as they call each other via RPC messages.
And we need a way to destroy all requests which came from a
particular network thread.
We probably also need to get rid of FR_IO_ACTION_DONE, as it
isn't used anywhere. Instead, just rely on the talloc destructors
to clean up timers, FD events, etc.
Alan T. DeKok [Tue, 25 Jul 2017 18:52:20 +0000 (14:52 -0400)]
destroy the scheduler before exiting
Alan T. DeKok [Tue, 25 Jul 2017 18:52:09 +0000 (14:52 -0400)]
clean up network before worker
Alan T. DeKok [Tue, 25 Jul 2017 18:51:46 +0000 (14:51 -0400)]
don't verify on exit
and don't clean up messages... just die
Arran Cudbard-Bell [Tue, 25 Jul 2017 18:53:33 +0000 (14:53 -0400)]
Use a free list to free file descriptor events
This fixes several potential corner cases, with file descriptor events being deleted in callbacks and causing SEGVs
Arran Cudbard-Bell [Tue, 25 Jul 2017 18:53:12 +0000 (14:53 -0400)]
Rename some structures and variables in event.c
Arran Cudbard-Bell [Tue, 25 Jul 2017 18:28:00 +0000 (14:28 -0400)]
Formatting
Arran Cudbard-Bell [Tue, 25 Jul 2017 17:57:31 +0000 (13:57 -0400)]
Add tests for group comparisons
Arran Cudbard-Bell [Tue, 25 Jul 2017 17:56:17 +0000 (13:56 -0400)]
Allow paircmp with expanded/reference RHS
Arran Cudbard-Bell [Tue, 25 Jul 2017 17:55:53 +0000 (13:55 -0400)]
Fix group checks to return updated when we add attributes
Alan T. DeKok [Tue, 25 Jul 2017 15:55:48 +0000 (11:55 -0400)]
better fix for parsing log{} sections
Alan T. DeKok [Tue, 25 Jul 2017 15:49:25 +0000 (11:49 -0400)]
remove calls to request_enqueue()
Alan T. DeKok [Tue, 25 Jul 2017 15:47:45 +0000 (11:47 -0400)]
Revert "move all logging configuration to initial parse pass"
This reverts commit
d03b2953b36311239f6059a66ec542a0998cf7a3 .
We still need to parse the OTHER log file entries... ugh
Alan T. DeKok [Tue, 25 Jul 2017 15:45:05 +0000 (11:45 -0400)]
free children before freeing the event list
Alan T. DeKok [Tue, 25 Jul 2017 15:40:31 +0000 (11:40 -0400)]
we no longer use / need the thread pool
Alan T. DeKok [Tue, 25 Jul 2017 15:40:31 +0000 (11:40 -0400)]
we no longer use / need the thread pool
Alan T. DeKok [Tue, 25 Jul 2017 15:39:17 +0000 (11:39 -0400)]
move all logging configuration to initial parse pass
Alan T. DeKok [Tue, 25 Jul 2017 15:20:30 +0000 (11:20 -0400)]
remove old log_file and log_destination
They've been deprecated for years
Alan T. DeKok [Tue, 25 Jul 2017 14:49:35 +0000 (10:49 -0400)]
clean up exfile code.
If it's not locking, skip all of the management logic
Alan T. DeKok [Tue, 25 Jul 2017 14:31:21 +0000 (10:31 -0400)]
entry->fd may be <0
Alan T. DeKok [Tue, 25 Jul 2017 14:30:45 +0000 (10:30 -0400)]
re-open and re-mkdir if the file was unlinked
Alan T. DeKok [Tue, 25 Jul 2017 14:28:29 +0000 (10:28 -0400)]
move open + mkdir into it's own function
Alan T. DeKok [Tue, 25 Jul 2017 14:26:51 +0000 (10:26 -0400)]
remove dup(). It's not necessary
Alan T. DeKok [Tue, 25 Jul 2017 14:25:42 +0000 (10:25 -0400)]
remove unused exfile_unlock()
Alan T. DeKok [Tue, 25 Jul 2017 14:24:52 +0000 (10:24 -0400)]
dup the FD before calling fdopen()
Alan T. DeKok [Tue, 25 Jul 2017 14:22:35 +0000 (10:22 -0400)]
remove append argument. It's always append