]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Thu, 9 Sep 2021 20:54:42 +0000 (16:54 -0400)]
print an error, too
Alan T. DeKok [Thu, 9 Sep 2021 20:33:51 +0000 (16:33 -0400)]
respect SO_SNDBUF
Alan T. DeKok [Thu, 9 Sep 2021 15:34:56 +0000 (11:34 -0400)]
cleanups and fixes
bind to interface
set REUSEPORT
read and discard any replies
Alan T. DeKok [Thu, 9 Sep 2021 14:27:24 +0000 (10:27 -0400)]
let's not leak FDs
Alan T. DeKok [Thu, 9 Sep 2021 13:44:03 +0000 (09:44 -0400)]
API to write (or enqueue) UDP packets
for use with "fire and forget" UDP messaging, such as DHCP relay,
RADIUS replication, etc.
Arran Cudbard-Bell [Thu, 9 Sep 2021 00:03:38 +0000 (19:03 -0500)]
Less fragile way of defferring event FD freeing
Arran Cudbard-Bell [Wed, 8 Sep 2021 23:42:57 +0000 (18:42 -0500)]
Initialise temporary event list if we're not passed one in unlang_interpret_synchronous
Arran Cudbard-Bell [Wed, 8 Sep 2021 23:30:10 +0000 (18:30 -0500)]
Add basic trigger tests
Arran Cudbard-Bell [Wed, 8 Sep 2021 23:29:47 +0000 (18:29 -0500)]
Fix logic bug in synchronous interpreter where it'd free yielded detached subrequests
Arran Cudbard-Bell [Wed, 8 Sep 2021 23:29:25 +0000 (18:29 -0500)]
Include triggers file in modules test
Arran Cudbard-Bell [Wed, 8 Sep 2021 23:29:07 +0000 (18:29 -0500)]
Remove request->el
We now use the one associated with the interpreter. Requests don't have event lists... interpreters do. There were some instances where request->el wasn't being inherited correctly, this fixes that.
Arran Cudbard-Bell [Wed, 8 Sep 2021 22:54:54 +0000 (17:54 -0500)]
More event loop debugging
Arran Cudbard-Bell [Wed, 8 Sep 2021 20:51:10 +0000 (15:51 -0500)]
Fix wrong signed/unsigned value_box set (#4223)
Co-authored-by: Jorge Pereira <jpereiran@gmail.com>
Alan T. DeKok [Wed, 8 Sep 2021 19:00:38 +0000 (15:00 -0400)]
set sequence number for Access-Request packets which contain State
Alan T. DeKok [Wed, 8 Sep 2021 18:52:31 +0000 (14:52 -0400)]
allow clients to set dedup_authenticator, too.
We should really either (a) move clients to RADIUS only, or
(b) allow for protocol-specific extensions to clients
Alan T. DeKok [Wed, 8 Sep 2021 17:18:28 +0000 (13:18 -0400)]
bail from functions early
so that the code has fewer levels of indentation, and is somewhat
easier to read
Arran Cudbard-Bell [Wed, 8 Sep 2021 18:48:37 +0000 (13:48 -0500)]
More trigger fixes
Arran Cudbard-Bell [Wed, 8 Sep 2021 17:41:23 +0000 (12:41 -0500)]
Fix test module test
Arran Cudbard-Bell [Wed, 8 Sep 2021 17:41:07 +0000 (12:41 -0500)]
Strip out code from triggers... making them better
Triggers now use the same codepath for sync/async
Arran Cudbard-Bell [Wed, 8 Sep 2021 17:37:15 +0000 (12:37 -0500)]
If we can't setup a reaper, kill the process
zombie children are a PITA
Arran Cudbard-Bell [Wed, 8 Sep 2021 17:36:52 +0000 (12:36 -0500)]
Formatting
Arran Cudbard-Bell [Wed, 8 Sep 2021 17:34:40 +0000 (12:34 -0500)]
Allow for reaper callbacks
Alan T. DeKok [Wed, 8 Sep 2021 17:14:58 +0000 (13:14 -0400)]
Set sequence so that we can prioritize ongoing multi-packet sessions.
Alan T. DeKok [Wed, 8 Sep 2021 15:46:51 +0000 (11:46 -0400)]
add test for "foo.bar.baz = stuff"
Alan T. DeKok [Wed, 8 Sep 2021 15:45:54 +0000 (11:45 -0400)]
el might be NULL if we failed to parse the config
Alan T. DeKok [Wed, 8 Sep 2021 15:42:23 +0000 (11:42 -0400)]
better name
Alan T. DeKok [Wed, 8 Sep 2021 15:39:23 +0000 (11:39 -0400)]
tests for local templates
Alan T. DeKok [Wed, 8 Sep 2021 15:32:40 +0000 (11:32 -0400)]
allow in-line templates
and skip them when defining modules.
And when referencing an item, look in the template if it can't
be found in the section.
Alan T. DeKok [Wed, 8 Sep 2021 14:18:11 +0000 (10:18 -0400)]
clarify error messages
Alan T. DeKok [Wed, 8 Sep 2021 14:16:41 +0000 (10:16 -0400)]
we need offer_duration, too
Alan T. DeKok [Wed, 8 Sep 2021 14:08:29 +0000 (10:08 -0400)]
pool_name is needed by the queries.
and use ${pool_name} consistently everywhere.
Apparently we don't even use this module during the tests :(
Alan T. DeKok [Wed, 8 Sep 2021 14:07:35 +0000 (10:07 -0400)]
if we reference ${skip_locked}, we must use it
Alan T. DeKok [Wed, 8 Sep 2021 13:48:10 +0000 (09:48 -0400)]
functions here take fr_time_t, not time_t
Alan T. DeKok [Wed, 8 Sep 2021 12:56:50 +0000 (08:56 -0400)]
whoops, convert to time_delta
Alan T. DeKok [Wed, 8 Sep 2021 12:53:55 +0000 (08:53 -0400)]
use fr_time()
Alan T. DeKok [Wed, 8 Sep 2021 12:52:49 +0000 (08:52 -0400)]
notes on time
Alan T. DeKok [Wed, 8 Sep 2021 12:50:55 +0000 (08:50 -0400)]
Uh... connection_spawn() takes fr_time_t, not time(NULL)
Alan T. DeKok [Wed, 8 Sep 2021 12:50:10 +0000 (08:50 -0400)]
use fr_time()
Alan T. DeKok [Wed, 8 Sep 2021 12:49:25 +0000 (08:49 -0400)]
use fr_time()
Alan T. DeKok [Wed, 8 Sep 2021 12:47:06 +0000 (08:47 -0400)]
use request->packet->timestamp instead of time(NULL)
Alan T. DeKok [Wed, 8 Sep 2021 12:43:39 +0000 (08:43 -0400)]
use fr_time()
Alan T. DeKok [Wed, 8 Sep 2021 12:39:30 +0000 (08:39 -0400)]
use fr_time_t
Alan T. DeKok [Wed, 8 Sep 2021 12:36:34 +0000 (08:36 -0400)]
use packet timestamp as "now"
instead of whenever we got around to processing the request.
this is more likely to be what the user intended
Alan T. DeKok [Wed, 8 Sep 2021 12:35:58 +0000 (08:35 -0400)]
use fr_time(), which doesn't jump around when NTP changes
Alan T. DeKok [Wed, 8 Sep 2021 12:30:58 +0000 (08:30 -0400)]
don't use time(0) for consistency, use time(NULL)
Alan T. DeKok [Wed, 8 Sep 2021 12:19:11 +0000 (08:19 -0400)]
notes and word smithing
Alan T. DeKok [Wed, 8 Sep 2021 11:59:03 +0000 (07:59 -0400)]
replace rlm_expiration with a ~10 line unlang policy
Alan T. DeKok [Wed, 8 Sep 2021 10:41:56 +0000 (06:41 -0400)]
Session-Timeout is uint32, not uint64
Alan T. DeKok [Wed, 8 Sep 2021 10:38:17 +0000 (06:38 -0400)]
limit is uint64, not uint32
Arran Cudbard-Bell [Tue, 7 Sep 2021 22:18:31 +0000 (17:18 -0500)]
Print out the current ime correctly
Arran Cudbard-Bell [Tue, 7 Sep 2021 21:59:15 +0000 (16:59 -0500)]
Add debian SNMP package
Arran Cudbard-Bell [Tue, 7 Sep 2021 21:50:47 +0000 (16:50 -0500)]
Add RHEL SNMP package
Arran Cudbard-Bell [Tue, 7 Sep 2021 21:26:24 +0000 (16:26 -0500)]
Install the mibs in <prefix>/share/snmp/mibs
Arran Cudbard-Bell [Tue, 7 Sep 2021 20:21:44 +0000 (15:21 -0500)]
Various eap-aka/sim doc cleanups
Alan T. DeKok [Tue, 7 Sep 2021 18:47:19 +0000 (14:47 -0400)]
cleanups and fixes
Alan T. DeKok [Tue, 7 Sep 2021 18:39:31 +0000 (14:39 -0400)]
allow 32-bit attributes for TACACS
because we don't care about the numbers
Alan T. DeKok [Tue, 7 Sep 2021 17:22:53 +0000 (13:22 -0400)]
encode named attributes, too
Alan T. DeKok [Tue, 7 Sep 2021 15:23:42 +0000 (11:23 -0400)]
check at compile time for invalid assignments
Alan T. DeKok [Tue, 7 Sep 2021 15:09:47 +0000 (11:09 -0400)]
sign promote the input only if the underlying data type is signed
Alan T. DeKok [Tue, 7 Sep 2021 12:51:10 +0000 (08:51 -0400)]
rearrange code in preparation for encoding other attributes
and also fix some bugs with encoding, where the argument count
was tracked incorrectly
Alan T. DeKok [Mon, 6 Sep 2021 20:03:55 +0000 (16:03 -0400)]
start of trying to decode as named attributes
and fix the test which somehow had an extraneous empty string
for Argument-List
Alan T. DeKok [Mon, 6 Sep 2021 19:59:33 +0000 (15:59 -0400)]
print out decoded packet, too
Alan T. DeKok [Mon, 6 Sep 2021 19:41:13 +0000 (15:41 -0400)]
normalize name
Alan T. DeKok [Mon, 6 Sep 2021 19:14:13 +0000 (15:14 -0400)]
fix typo
Alan T. DeKok [Mon, 6 Sep 2021 15:26:08 +0000 (11:26 -0400)]
add support for foo.bar[baz] = x
tho it doesn't _quite_ work yet, because the tokenizer will
stop at the '[', and will not pass the full name to the add_pair()
function.
Alan T. DeKok [Fri, 3 Sep 2021 18:40:56 +0000 (14:40 -0400)]
regenerate
Alan T. DeKok [Fri, 3 Sep 2021 18:38:40 +0000 (14:38 -0400)]
fix links
Alan T. DeKok [Fri, 3 Sep 2021 18:14:53 +0000 (14:14 -0400)]
use dots now, not :
Arran Cudbard-Bell [Tue, 7 Sep 2021 14:59:41 +0000 (09:59 -0500)]
Add toggle for requiring client certificates
Arran Cudbard-Bell [Tue, 7 Sep 2021 14:07:53 +0000 (09:07 -0500)]
Add TEST_CHECK_SLEN
James Jones [Tue, 7 Sep 2021 14:05:16 +0000 (09:05 -0500)]
Correct casts in TEST_CHECK_LEN() (#4221)
Arran Cudbard-Bell [Mon, 6 Sep 2021 22:32:10 +0000 (17:32 -0500)]
Display build result to try and figure out why coverity isn't working
Arran Cudbard-Bell [Mon, 6 Sep 2021 22:24:36 +0000 (17:24 -0500)]
LGTM is stupid and says we don't return a value in all code paths.
Arran Cudbard-Bell [Mon, 6 Sep 2021 21:54:58 +0000 (16:54 -0500)]
Fix a couple more issues...
Arran Cudbard-Bell [Mon, 6 Sep 2021 21:41:53 +0000 (16:41 -0500)]
Install the lsb-release package
Arran Cudbard-Bell [Mon, 6 Sep 2021 21:39:27 +0000 (16:39 -0500)]
Should be testing packaging not whether the upstream OS provides broke things...
Arran Cudbard-Bell [Mon, 6 Sep 2021 21:04:28 +0000 (16:04 -0500)]
typo
Arran Cudbard-Bell [Mon, 6 Sep 2021 21:00:57 +0000 (16:00 -0500)]
UNUSED if buidling with EVENT_DEBUG
Arran Cudbard-Bell [Mon, 6 Sep 2021 20:55:58 +0000 (15:55 -0500)]
Fix unlikely
Arran Cudbard-Bell [Mon, 6 Sep 2021 20:51:05 +0000 (15:51 -0500)]
Wait for five seconds after the process exits for any children to finish
Fix other various issues with cleaning up processes.
Arran Cudbard-Bell [Mon, 6 Sep 2021 20:31:59 +0000 (15:31 -0500)]
Fix fr_dlist_foreach_safe
Arran Cudbard-Bell [Mon, 6 Sep 2021 16:35:01 +0000 (11:35 -0500)]
Fix more issues
Arran Cudbard-Bell [Mon, 6 Sep 2021 16:10:08 +0000 (11:10 -0500)]
...and here
Arran Cudbard-Bell [Mon, 6 Sep 2021 16:06:20 +0000 (11:06 -0500)]
Typo
Arran Cudbard-Bell [Mon, 6 Sep 2021 16:02:20 +0000 (11:02 -0500)]
Maintain a list of "reap" requests, and make sure they're all cleaned up when the event loop exits
Arran Cudbard-Bell [Mon, 6 Sep 2021 14:00:02 +0000 (09:00 -0500)]
Make the repo commands prettier
Arran Cudbard-Bell [Mon, 6 Sep 2021 13:39:10 +0000 (08:39 -0500)]
Get CODENAME and DIST dynamically from lsb_release
Arran Cudbard-Bell [Mon, 6 Sep 2021 13:03:38 +0000 (08:03 -0500)]
...same but with an actions script
Arran Cudbard-Bell [Mon, 6 Sep 2021 12:52:59 +0000 (07:52 -0500)]
Add NR repo to main CI workflow
Arran Cudbard-Bell [Mon, 30 Aug 2021 17:59:30 +0000 (12:59 -0500)]
Note on what vps is.
Alan T. DeKok [Fri, 3 Sep 2021 18:01:52 +0000 (14:01 -0400)]
allow foo.bar.baz = 1 for laziness
with some caveats and qualifications
Alan T. DeKok [Thu, 2 Sep 2021 15:33:47 +0000 (11:33 -0400)]
time_delta is signed by default
add tests.
Catch the case (parsing and printing) of negative zero
Alan T. DeKok [Thu, 2 Sep 2021 15:33:26 +0000 (11:33 -0400)]
print out the strings we have, too
Alan T. DeKok [Thu, 2 Sep 2021 15:01:24 +0000 (11:01 -0400)]
parse 1h, 1m, etc. correctly
Alan T. DeKok [Thu, 2 Sep 2021 15:00:28 +0000 (11:00 -0400)]
use correct MOD macro.
Alan T. DeKok [Thu, 2 Sep 2021 14:48:24 +0000 (10:48 -0400)]
use Acct-Delay-Time when calculating Event-Timestamp
Alan T. DeKok [Wed, 1 Sep 2021 21:10:02 +0000 (17:10 -0400)]
clean up and check for overflows
Alan T. DeKok [Wed, 1 Sep 2021 20:37:06 +0000 (16:37 -0400)]
parse *signed* time deltas
and don't allow fractional minutes, hours, or days. That doesn't
make any sense.
There are still many possibilities for overflow, so fixing that
is next.
Alan T. DeKok [Wed, 1 Sep 2021 20:30:34 +0000 (16:30 -0400)]
print out signed time_delta, too
Alan T. DeKok [Wed, 1 Sep 2021 13:53:08 +0000 (09:53 -0400)]
allow time_delta to be _unsigned_ when we encode it in the network
e.g. for RADIUS, with Session-Timeout