]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Wed, 6 Sep 2017 12:38:33 +0000 (08:38 -0400)]
WARN on using %{Response-Packet-Type}
they should really use %{reply:Packet-Type}
Alan T. DeKok [Wed, 6 Sep 2017 12:38:11 +0000 (08:38 -0400)]
remove expansion for %{Response-Packet-Type}
it's already handled by the server core
Alan T. DeKok [Tue, 5 Sep 2017 15:50:46 +0000 (11:50 -0400)]
move run-time checks to asserts
Alan T. DeKok [Tue, 5 Sep 2017 15:47:21 +0000 (11:47 -0400)]
manually mark things as allow empty group or not
Alan T. DeKok [Tue, 5 Sep 2017 15:07:18 +0000 (11:07 -0400)]
remove no longer needed "todo"
Alan T. DeKok [Tue, 5 Sep 2017 15:06:40 +0000 (11:06 -0400)]
"create" takes a static string
Alan T. DeKok [Tue, 5 Sep 2017 15:05:56 +0000 (11:05 -0400)]
run the CHILDREN, not the parent
Alan T. DeKok [Tue, 5 Sep 2017 14:37:17 +0000 (10:37 -0400)]
remove more unused fields from REQUEST
Alan T. DeKok [Tue, 5 Sep 2017 14:30:23 +0000 (10:30 -0400)]
rename "fork" to "create"
It's clearer for normal humans
Alan T. DeKok [Tue, 5 Sep 2017 14:30:14 +0000 (10:30 -0400)]
one less todo
Alan T. DeKok [Tue, 5 Sep 2017 13:49:57 +0000 (09:49 -0400)]
set request->reply->code from the reply code
Alan T. DeKok [Mon, 4 Sep 2017 17:39:20 +0000 (13:39 -0400)]
Revert "add instantiate / detach for timers"
This reverts commit
1d2c0352ada27c281f83a279532bf3a4d6a37167.
need a better way to do this
Arran Cudbard-Bell [Wed, 6 Sep 2017 16:47:14 +0000 (23:47 +0700)]
More doxygen
Arran Cudbard-Bell [Wed, 6 Sep 2017 16:28:41 +0000 (23:28 +0700)]
Doxygen
Arran Cudbard-Bell [Wed, 6 Sep 2017 16:07:47 +0000 (23:07 +0700)]
Fix some connection.c function calls
Arran Cudbard-Bell [Wed, 6 Sep 2017 14:53:07 +0000 (21:53 +0700)]
Start of async LDAP connection state machine
Arran Cudbard-Bell [Tue, 29 Aug 2017 13:17:25 +0000 (21:17 +0800)]
Not necessarily a socket
klemens [Tue, 5 Sep 2017 17:25:44 +0000 (19:25 +0200)]
URL fixes
Arran Cudbard-Bell [Tue, 5 Sep 2017 09:22:44 +0000 (16:22 +0700)]
Same fix for debian unit files
Arran Cudbard-Bell [Tue, 5 Sep 2017 09:21:43 +0000 (16:21 +0700)]
Runtime dir should be relative
Alan T. DeKok [Sun, 3 Sep 2017 20:40:02 +0000 (16:40 -0400)]
add instantiate / detach for timers
when a socket is added to a network event loop, there may be
timers which it needs. So, have callbacks to allow them
Arran Cudbard-Bell [Sun, 3 Sep 2017 12:22:52 +0000 (19:22 +0700)]
There may be legitimate reasons for not providing information
Easier just to keep the "Refusal to follow instructions for solving a problem"
Alan T. DeKok [Fri, 1 Sep 2017 13:24:10 +0000 (09:24 -0400)]
remove unused macro
Alan T. DeKok [Fri, 1 Sep 2017 13:17:00 +0000 (09:17 -0400)]
clarify code. Helps with #2051
Alan T. DeKok [Fri, 1 Sep 2017 12:34:24 +0000 (08:34 -0400)]
fix typo, and use memcpy for buffers of known size. Should fix #2051
Alan T. DeKok [Fri, 1 Sep 2017 12:25:37 +0000 (08:25 -0400)]
Include trailing \0 in "used" count. Part of #2051
Alan T. DeKok [Thu, 31 Aug 2017 14:52:09 +0000 (10:52 -0400)]
make it clearer we're updating the actual CPU time
Alan T. DeKok [Thu, 31 Aug 2017 14:50:46 +0000 (10:50 -0400)]
move list of workers to an array
So we can use "power of two" to get decent load-balancing,
without O(nlgn) issues
Alan T. DeKok [Thu, 31 Aug 2017 14:06:31 +0000 (10:06 -0400)]
make number of network / worker threads configurable
Alan T. DeKok [Thu, 31 Aug 2017 13:58:00 +0000 (09:58 -0400)]
remove "listen" and "checkrad" from mainconfig
Alan T. DeKok [Thu, 31 Aug 2017 13:53:05 +0000 (09:53 -0400)]
more notes
Alan T. DeKok [Thu, 31 Aug 2017 13:40:13 +0000 (09:40 -0400)]
no longer needed
Alan T. DeKok [Wed, 30 Aug 2017 17:09:14 +0000 (13:09 -0400)]
notes on max_connections
Alan T. DeKok [Wed, 30 Aug 2017 17:07:53 +0000 (13:07 -0400)]
re-arrange
Alan T. DeKok [Wed, 30 Aug 2017 17:07:01 +0000 (13:07 -0400)]
limit the number of outgoing connections
Alan T. DeKok [Wed, 30 Aug 2017 01:49:33 +0000 (21:49 -0400)]
updated docs on status checks
Alan T. DeKok [Wed, 30 Aug 2017 01:38:22 +0000 (21:38 -0400)]
create VPs for status-checks as normal packets
Alan T. DeKok [Wed, 30 Aug 2017 01:37:12 +0000 (21:37 -0400)]
add provisions for status-checks as normal packets
Alan T. DeKok [Wed, 30 Aug 2017 01:36:52 +0000 (21:36 -0400)]
subsections can be required
Alan T. DeKok [Tue, 29 Aug 2017 23:41:42 +0000 (19:41 -0400)]
update notes
Alan T. DeKok [Tue, 29 Aug 2017 20:45:01 +0000 (16:45 -0400)]
this is no longer needed
Alan T. DeKok [Tue, 29 Aug 2017 20:42:28 +0000 (16:42 -0400)]
remove in_request_hash and in_proxy_hash
Alan T. DeKok [Tue, 29 Aug 2017 16:12:20 +0000 (12:12 -0400)]
add no_connection_fail
which immediately fails the request if there are no active
connections
Alan T. DeKok [Tue, 29 Aug 2017 14:49:01 +0000 (10:49 -0400)]
allow MOD_ACTION_RETURN for parallel
so that we can return as soon as one child says so
Alan T. DeKok [Tue, 29 Aug 2017 14:48:51 +0000 (10:48 -0400)]
add destructor for parallel state
Alan T. DeKok [Tue, 29 Aug 2017 14:40:21 +0000 (10:40 -0400)]
notes on parallel
Alan T. DeKok [Tue, 29 Aug 2017 14:35:55 +0000 (10:35 -0400)]
more alphabetization
Alan T. DeKok [Tue, 29 Aug 2017 14:32:44 +0000 (10:32 -0400)]
note name change of sections
Alan T. DeKok [Tue, 29 Aug 2017 13:08:49 +0000 (09:08 -0400)]
nope, don't need this...
Alan T. DeKok [Tue, 29 Aug 2017 12:48:00 +0000 (08:48 -0400)]
added more notes
Alan T. DeKok [Tue, 29 Aug 2017 12:46:21 +0000 (08:46 -0400)]
set state_ctx so we don't leak
Alan T. DeKok [Tue, 29 Aug 2017 12:14:09 +0000 (08:14 -0400)]
final travis issue
Alan T. DeKok [Tue, 29 Aug 2017 11:54:07 +0000 (07:54 -0400)]
UNUSED issues
Alan T. DeKok [Tue, 29 Aug 2017 11:38:23 +0000 (07:38 -0400)]
better debug output, and merge priorities in
Alan T. DeKok [Tue, 29 Aug 2017 11:23:33 +0000 (07:23 -0400)]
parallel{} now works.
Next step is to merge all of the results in...
Alan T. DeKok [Tue, 29 Aug 2017 11:22:25 +0000 (07:22 -0400)]
more notes
Alan T. DeKok [Tue, 29 Aug 2017 11:15:28 +0000 (07:15 -0400)]
remove 'const' from resume_ctx
The whole point of resuming is to modify the ctx...
Alan T. DeKok [Tue, 29 Aug 2017 11:15:01 +0000 (07:15 -0400)]
use *presult, and set priority, too
Alan T. DeKok [Tue, 29 Aug 2017 11:13:44 +0000 (07:13 -0400)]
move unlang_child_alloc() to it's own function
Alan T. DeKok [Mon, 28 Aug 2017 13:44:36 +0000 (09:44 -0400)]
clean up unlang_resume() a bit
Alan T. DeKok [Mon, 28 Aug 2017 13:40:52 +0000 (09:40 -0400)]
add ops callback for unlang_resume()
so that it's data driven, instead of if / then / else
which lets us have resumable operations for any kind of unlang_t
Alan T. DeKok [Mon, 28 Aug 2017 13:14:55 +0000 (09:14 -0400)]
don't check for STOP in modules, and check top-most parent
Alan T. DeKok [Sun, 27 Aug 2017 20:58:21 +0000 (16:58 -0400)]
it's max_packet_size, not default_message_size
and use num_messages...
Alan T. DeKok [Sun, 27 Aug 2017 20:57:48 +0000 (16:57 -0400)]
make write() take non-const 'inst'
Arran Cudbard-Bell [Tue, 29 Aug 2017 08:16:41 +0000 (16:16 +0800)]
Not sure we actually *need* to source rvm scripts
Arran Cudbard-Bell [Tue, 29 Aug 2017 08:16:28 +0000 (16:16 +0800)]
Remove ruby version pring
Arran Cudbard-Bell [Tue, 29 Aug 2017 08:04:57 +0000 (16:04 +0800)]
Using sudo might be messing up rvm
Arran Cudbard-Bell [Tue, 29 Aug 2017 08:00:21 +0000 (16:00 +0800)]
gem install via rvm
Arran Cudbard-Bell [Mon, 28 Aug 2017 15:59:49 +0000 (23:59 +0800)]
Print version
Arran Cudbard-Bell [Mon, 28 Aug 2017 15:36:53 +0000 (23:36 +0800)]
Needs to be bash
Arran Cudbard-Bell [Mon, 28 Aug 2017 15:29:08 +0000 (23:29 +0800)]
No double brackets
Arran Cudbard-Bell [Mon, 28 Aug 2017 15:16:36 +0000 (23:16 +0800)]
Source RVM
Arran Cudbard-Bell [Mon, 28 Aug 2017 14:24:41 +0000 (22:24 +0800)]
Without the sudo
Arran Cudbard-Bell [Mon, 28 Aug 2017 12:23:14 +0000 (20:23 +0800)]
Stupid ruby
Arran Cudbard-Bell [Mon, 28 Aug 2017 10:48:18 +0000 (18:48 +0800)]
Set default ruby
Arran Cudbard-Bell [Mon, 28 Aug 2017 10:05:09 +0000 (18:05 +0800)]
Switch out ruby version
Alan T. DeKok [Sat, 26 Aug 2017 15:15:17 +0000 (11:15 -0400)]
Revert "let's see if rwquiring ruby 2.3 fixes the build"
This reverts commit
aca61c20ecb95614fe3eb6f6959a3c4925f1af5e.
nope
Alan T. DeKok [Sat, 26 Aug 2017 14:54:18 +0000 (10:54 -0400)]
let's see if rwquiring ruby 2.3 fixes the build
Alan T. DeKok [Sat, 26 Aug 2017 13:43:43 +0000 (09:43 -0400)]
notes on proxying and the RADIUS module
Alan T. DeKok [Fri, 25 Aug 2017 19:16:44 +0000 (15:16 -0400)]
remove one more todo
Alan T. DeKok [Fri, 25 Aug 2017 19:14:07 +0000 (15:14 -0400)]
check buffer length, not maximum packet size
Alan T. DeKok [Fri, 25 Aug 2017 18:51:22 +0000 (14:51 -0400)]
request is no longer NULL for Status-Server
Alan T. DeKok [Fri, 25 Aug 2017 18:31:45 +0000 (14:31 -0400)]
clean up and update todo's
Alan T. DeKok [Fri, 25 Aug 2017 18:31:35 +0000 (14:31 -0400)]
Add Error-Cause 601
Alan T. DeKok [Fri, 25 Aug 2017 18:20:44 +0000 (14:20 -0400)]
consistency for messages
Alan T. DeKok [Fri, 25 Aug 2017 18:20:30 +0000 (14:20 -0400)]
minor update
Alan T. DeKok [Fri, 25 Aug 2017 18:15:54 +0000 (14:15 -0400)]
do RFC 7930 negotiation on Protocol-Error and Status-Server
Alan T. DeKok [Fri, 25 Aug 2017 18:07:01 +0000 (14:07 -0400)]
fix incorrect assertion
Alan T. DeKok [Fri, 25 Aug 2017 17:53:58 +0000 (13:53 -0400)]
added support for Protocol-Error encoding
Alan T. DeKok [Fri, 25 Aug 2017 17:53:20 +0000 (13:53 -0400)]
add missing \t
Alan T. DeKok [Fri, 25 Aug 2017 17:40:04 +0000 (13:40 -0400)]
add Protocol-Error
Alan T. DeKok [Fri, 25 Aug 2017 15:15:33 +0000 (11:15 -0400)]
fix RDEBUG macro
which means we don't get the full behavior of radlog_request(),
but at least the test utilities will link correctly
Alan T. DeKok [Fri, 25 Aug 2017 14:33:42 +0000 (10:33 -0400)]
documentation for new features
Alan T. DeKok [Fri, 25 Aug 2017 14:23:45 +0000 (10:23 -0400)]
insert request into time_order heap AFTER doing de-dup detection
Alan T. DeKok [Fri, 25 Aug 2017 14:19:58 +0000 (10:19 -0400)]
set last_reply when we create a connection
Alan T. DeKok [Fri, 25 Aug 2017 13:59:43 +0000 (09:59 -0400)]
add mod_signal, which does retransmission on DUP action
Alan T. DeKok [Fri, 25 Aug 2017 13:59:30 +0000 (09:59 -0400)]
move retransmission to it's own function
Alan T. DeKok [Fri, 25 Aug 2017 13:50:55 +0000 (09:50 -0400)]
enable "track_duplicates"
Alan T. DeKok [Fri, 25 Aug 2017 13:50:24 +0000 (09:50 -0400)]
finish dup detection
with more debug output so we know what's going on
Alan T. DeKok [Fri, 25 Aug 2017 13:48:31 +0000 (09:48 -0400)]
don't check for messages which age out of the queue
we're SUPPOSED to service the queue quickly.
we probably need to differentiate between the recv_time and
original_recv_time. We do that in request->async, but not
in the channel_data_t