]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Tue, 3 Oct 2017 15:53:48 +0000 (23:53 +0800)]
Restore previous state
Arran Cudbard-Bell [Tue, 3 Oct 2017 15:52:18 +0000 (23:52 +0800)]
Follow the normal state transition if we have a ->failed callback and it returns FR_CONNECTION_STATE_INIT
Alan T. DeKok [Tue, 3 Oct 2017 15:43:20 +0000 (11:43 -0400)]
remove it from the thread queue, too
Alan T. DeKok [Tue, 3 Oct 2017 15:37:11 +0000 (11:37 -0400)]
remove packet from c->queued on timeout or reply
Alan T. DeKok [Tue, 3 Oct 2017 15:29:08 +0000 (11:29 -0400)]
don't access u->rr if it doesn't exist
Alan T. DeKok [Tue, 3 Oct 2017 15:25:40 +0000 (11:25 -0400)]
print time after the timer has been initialized
Alan T. DeKok [Tue, 3 Oct 2017 15:24:13 +0000 (11:24 -0400)]
access u->timer directly
Alan T. DeKok [Tue, 3 Oct 2017 15:24:06 +0000 (11:24 -0400)]
notes for todo
Arran Cudbard-Bell [Tue, 3 Oct 2017 15:21:03 +0000 (23:21 +0800)]
Don't need this in _conn_failed as _conn_close is always called
Alan T. DeKok [Tue, 3 Oct 2017 15:12:10 +0000 (11:12 -0400)]
remove connection from list / heap
Alan T. DeKok [Tue, 3 Oct 2017 12:49:42 +0000 (08:49 -0400)]
don't use one-character variable names
or prefixes. or _p where it's not necessary
simplify some logic to make it clearer, and update comments
to describe what the code is doing
Alan T. DeKok [Tue, 3 Oct 2017 12:39:36 +0000 (08:39 -0400)]
formatting
so that it's readable in a terminal
Alan T. DeKok [Tue, 3 Oct 2017 12:38:53 +0000 (08:38 -0400)]
fix documentation
Alan T. DeKok [Mon, 2 Oct 2017 20:14:47 +0000 (16:14 -0400)]
more gluing worker into proto_detail
if "transport != work", then we load the worker, as the various
other modules are intended to use it, too
Alan T. DeKok [Mon, 2 Oct 2017 19:58:14 +0000 (15:58 -0400)]
more preparations for auto-create detail worker
Arran Cudbard-Bell [Tue, 3 Oct 2017 14:55:03 +0000 (22:55 +0800)]
Package proto_detail_work
Alan T. DeKok [Mon, 2 Oct 2017 19:45:46 +0000 (15:45 -0400)]
added proto_detail_file
and started on integrating it. But it looks like NOTE_EXTEND
for directories isn't supported on OSX. <sigh>
Alan T. DeKok [Mon, 2 Oct 2017 19:12:56 +0000 (15:12 -0400)]
NOTE_EXTEND can be applied to directories
Alan T. DeKok [Mon, 2 Oct 2017 19:10:30 +0000 (15:10 -0400)]
include type name in message
Alan T. DeKok [Mon, 2 Oct 2017 18:59:49 +0000 (14:59 -0400)]
instance isn't const
Alan T. DeKok [Mon, 2 Oct 2017 18:54:09 +0000 (14:54 -0400)]
add fr_schedule_directory_add()
Alan T. DeKok [Mon, 2 Oct 2017 18:44:36 +0000 (14:44 -0400)]
rename proto_detail_file to proto_detail_work
in preparation for new proto_detail_file
Alan T. DeKok [Mon, 2 Oct 2017 18:41:48 +0000 (14:41 -0400)]
move common "work" header to proto_detail header
Alan T. DeKok [Mon, 2 Oct 2017 18:12:50 +0000 (14:12 -0400)]
ifdef out unused function
Alan T. DeKok [Mon, 2 Oct 2017 17:50:41 +0000 (13:50 -0400)]
add support for unmounted file systems
Alan T. DeKok [Mon, 2 Oct 2017 17:38:21 +0000 (13:38 -0400)]
fix comments
Alan T. DeKok [Mon, 2 Oct 2017 17:33:52 +0000 (13:33 -0400)]
add "track_progress" configuration
Alan T. DeKok [Mon, 2 Oct 2017 17:31:06 +0000 (13:31 -0400)]
minor cleanups
Alan DeKok [Mon, 2 Oct 2017 17:56:04 +0000 (13:56 -0400)]
Merge pull request #2084 from herwinw/allow_expired_crl_docs
Added some docs for allow_expired_crl
Alan DeKok [Mon, 2 Oct 2017 17:55:48 +0000 (13:55 -0400)]
Merge pull request #2085 from herwinw/docs_2.1.10_and_later
Removed references to version 2.1.10 in the config docs
Herwin Weststrate [Mon, 2 Oct 2017 17:17:13 +0000 (19:17 +0200)]
Removed references to version 2.1.10 in the config docs
That version is years out of date, there's no added value in keeping
references to that version in the docs.
Herwin Weststrate [Mon, 2 Oct 2017 17:11:16 +0000 (19:11 +0200)]
Added some docs for allow_expired_crl
As mentioned on the mailing list today: omission might confuse people
Alan T. DeKok [Mon, 2 Oct 2017 16:31:38 +0000 (12:31 -0400)]
fix assertion for uninitialized variable
Alan T. DeKok [Mon, 2 Oct 2017 16:31:02 +0000 (12:31 -0400)]
GCC definitions for sanitize-address
Alan T. DeKok [Mon, 2 Oct 2017 15:27:09 +0000 (11:27 -0400)]
keep timers around even when removing packets from a connection
so that the timers keep running, and the packets are retransmitted
when the timers hit.
This shouldn't be an issue normally. If a new connection opens,
any pending t->queued are drained, and sent immediately. BUT
there may be some pending packets in the queue which didn't make
it to the connection. So we have to handle that case, too.
If there is no connection, any timed-out packets are sent only when
a connection is available. This means that the timers aren't
perfect in that case, but there's no perfect situation here.
Alan T. DeKok [Mon, 2 Oct 2017 15:14:00 +0000 (11:14 -0400)]
add thread to udp_request_t
Alan T. DeKok [Mon, 2 Oct 2017 15:08:16 +0000 (11:08 -0400)]
move retry into udp_request_t
so we can access it when there's no connection
Alan T. DeKok [Mon, 2 Oct 2017 14:57:41 +0000 (10:57 -0400)]
use talloc_ctx, as it doesn't need the actual ID struct
Alan T. DeKok [Mon, 2 Oct 2017 14:30:39 +0000 (10:30 -0400)]
clear timers and remove (now wrong) assert
Alan T. DeKok [Mon, 2 Oct 2017 13:13:25 +0000 (09:13 -0400)]
remove todo as it's done
Alan T. DeKok [Mon, 2 Oct 2017 13:11:11 +0000 (09:11 -0400)]
re-initialize the status check packet on connection failed
and remove any pending writes, events, etc.
Arran Cudbard-Bell [Mon, 2 Oct 2017 09:29:18 +0000 (17:29 +0800)]
Add address-sanitizer 'feature' to version.c
...so we can check if the code was really built with support
Alan T. DeKok [Sun, 1 Oct 2017 22:48:28 +0000 (18:48 -0400)]
doxygen
Alan T. DeKok [Sun, 1 Oct 2017 21:58:32 +0000 (17:58 -0400)]
move retransmission timers to rlm_radius_udp_t
so that they operate across connections
Alan T. DeKok [Sun, 1 Oct 2017 21:24:53 +0000 (17:24 -0400)]
doxygen
Alan T. DeKok [Sun, 1 Oct 2017 15:00:56 +0000 (11:00 -0400)]
typo
Alan T. DeKok [Sun, 1 Oct 2017 14:58:37 +0000 (10:58 -0400)]
set pending flag
Alan T. DeKok [Sun, 1 Oct 2017 14:37:13 +0000 (10:37 -0400)]
reset connection state to INIT on failure
and allow active -> init states for reconnect
Alan T. DeKok [Sun, 1 Oct 2017 14:16:45 +0000 (10:16 -0400)]
move packets back to thread queue, not connection queue
on connection failure. the thread queue is for packets
without IDs allocated. The connection queue is for packets
with IDs allocated
Alan T. DeKok [Sun, 1 Oct 2017 14:16:32 +0000 (10:16 -0400)]
stop if we're at end of buffer
Alan T. DeKok [Sun, 1 Oct 2017 14:16:02 +0000 (10:16 -0400)]
transition to FAILED on conneciton failure before reconnect
Alan T. DeKok [Sun, 1 Oct 2017 13:47:45 +0000 (09:47 -0400)]
set last_search = 0
Alan T. DeKok [Sat, 30 Sep 2017 12:44:06 +0000 (08:44 -0400)]
notes on dynamic clients
Alan T. DeKok [Fri, 29 Sep 2017 15:07:46 +0000 (11:07 -0400)]
detail file packets are lower priority than packets from the network
Arran Cudbard-Bell [Sun, 1 Oct 2017 13:27:33 +0000 (21:27 +0800)]
Move most of conn_error to a conn_failed function to reconnections are handled properly
The connection state machine calls conn_failed on signal_reconnect() and when there's issues opening the connection.
Arran Cudbard-Bell [Sun, 1 Oct 2017 13:26:13 +0000 (21:26 +0800)]
Rename CONN_UNUSED to CONN_INIT
Arran Cudbard-Bell [Sun, 1 Oct 2017 12:09:04 +0000 (20:09 +0800)]
Add missing get_type_abort in conn_free
Arran Cudbard-Bell [Sun, 1 Oct 2017 12:06:48 +0000 (20:06 +0800)]
Don't double free status_u
Already freed by talloc_free_children
Arran Cudbard-Bell [Sun, 1 Oct 2017 11:56:31 +0000 (19:56 +0800)]
Use correct argument order for fr_pair_afrom_num
The VPs should always be added (not just at L_DBG_LVL_2)
Arran Cudbard-Bell [Sun, 1 Oct 2017 11:53:36 +0000 (19:53 +0800)]
Name arguments in rdebug* declarations
Arran Cudbard-Bell [Sun, 1 Oct 2017 11:43:15 +0000 (19:43 +0800)]
Callback functions have _<func>
Arran Cudbard-Bell [Sun, 1 Oct 2017 11:42:43 +0000 (19:42 +0800)]
Break long lines
Arran Cudbard-Bell [Sun, 1 Oct 2017 11:38:38 +0000 (19:38 +0800)]
Use correct doxygen
Arran Cudbard-Bell [Sun, 1 Oct 2017 10:21:43 +0000 (18:21 +0800)]
Make it clear the connection is now open
Arran Cudbard-Bell [Sun, 1 Oct 2017 10:21:12 +0000 (18:21 +0800)]
Don't leak memory on reconnection
Arran Cudbard-Bell [Sun, 1 Oct 2017 10:13:02 +0000 (18:13 +0800)]
Obligatory log message pass
Arran Cudbard-Bell [Sun, 1 Oct 2017 09:45:29 +0000 (17:45 +0800)]
Obligatory formatting pass
Arran Cudbard-Bell [Sun, 1 Oct 2017 09:23:34 +0000 (17:23 +0800)]
Add fr_asprintf
Arran Cudbard-Bell [Sun, 1 Oct 2017 09:08:08 +0000 (17:08 +0800)]
Allow RADIUS module calls in accounting/authorize
Arran Cudbard-Bell [Sun, 1 Oct 2017 08:09:26 +0000 (16:09 +0800)]
Specify path to config item
Arran Cudbard-Bell [Sun, 1 Oct 2017 07:40:44 +0000 (15:40 +0800)]
Dup the pool CONF_SECTION when connecting new nodes
CONF_SECTIONs can now only be parsed once so we get pools with no settings otherwise
Arran Cudbard-Bell [Sun, 1 Oct 2017 05:51:55 +0000 (13:51 +0800)]
Typo
Arran Cudbard-Bell [Sun, 1 Oct 2017 05:31:07 +0000 (13:31 +0800)]
Allow for node ID and other elements in cluster slots output
Arran Cudbard-Bell [Sat, 30 Sep 2017 14:06:30 +0000 (22:06 +0800)]
Switch to using WITH_SASL to indicate SASL support
Arran Cudbard-Bell [Sat, 30 Sep 2017 10:17:27 +0000 (18:17 +0800)]
Package proto_detail_file and proto_detail_process
Arran Cudbard-Bell [Sat, 30 Sep 2017 10:08:39 +0000 (18:08 +0800)]
Fix doxygen
Arran Cudbard-Bell [Sat, 30 Sep 2017 10:00:54 +0000 (18:00 +0800)]
Add support for toggling filters on/off without removing the callbacks or ef
Arran Cudbard-Bell [Sat, 30 Sep 2017 10:00:26 +0000 (18:00 +0800)]
Store the map used for function/filter mapping against the ef (less branchy)
Fix formatting in some places
Expose the union containing all the functions
Have fr_event_build_evset write out the active functions
Arran Cudbard-Bell [Fri, 29 Sep 2017 14:01:53 +0000 (22:01 +0800)]
Formatting
Alan T. DeKok [Fri, 29 Sep 2017 14:46:33 +0000 (10:46 -0400)]
doxygen
Alan T. DeKok [Fri, 29 Sep 2017 13:59:44 +0000 (09:59 -0400)]
Revert "add read_pause / read_continue"
This reverts commit
702f9a8b7c7546bdd9d3f49ce700050ff860fade .
Alan T. DeKok [Fri, 29 Sep 2017 13:58:17 +0000 (09:58 -0400)]
Revert "use new read_pause / read_continue routines"
This reverts commit
de5dd884c41db425896b83bb04c8bf9af48d812b .
Alan T. DeKok [Fri, 29 Sep 2017 13:57:50 +0000 (09:57 -0400)]
Revert "continue reading only if we've processed all outstanding packets"
This reverts commit
8856d7a291cd5e83711fd1d30cca07225627fd2c .
Arran Cudbard-Bell [Fri, 29 Sep 2017 13:52:10 +0000 (21:52 +0800)]
Pass filter to fr_event_fd_delete and other places
Alan T. DeKok [Fri, 29 Sep 2017 12:08:28 +0000 (08:08 -0400)]
set filter to find
Arran Cudbard-Bell [Fri, 29 Sep 2017 11:13:56 +0000 (19:13 +0800)]
Provide/embed RADIUSD_VERSION_RELEASE
Alan T. DeKok [Thu, 28 Sep 2017 22:23:06 +0000 (18:23 -0400)]
set src/dst ip/port if available
Alan T. DeKok [Wed, 27 Sep 2017 21:09:50 +0000 (17:09 -0400)]
notes on future work
Alan T. DeKok [Wed, 27 Sep 2017 21:09:35 +0000 (17:09 -0400)]
const
Alan T. DeKok [Wed, 27 Sep 2017 20:39:05 +0000 (16:39 -0400)]
debug messages aren't errors
Alan T. DeKok [Wed, 27 Sep 2017 20:14:06 +0000 (16:14 -0400)]
remove unused assignments
Alan T. DeKok [Wed, 27 Sep 2017 20:00:06 +0000 (16:00 -0400)]
fix code setting Done
by looking for \0 instead of \n, and allocating the tracking
entry only if we're processing this entry
Alan T. DeKok [Wed, 27 Sep 2017 19:39:12 +0000 (15:39 -0400)]
size_t casts for off_t
Alan T. DeKok [Wed, 27 Sep 2017 19:35:45 +0000 (15:35 -0400)]
continue reading only if we've processed all outstanding packets
to prevent overloading the system
Alan T. DeKok [Wed, 27 Sep 2017 19:13:53 +0000 (15:13 -0400)]
use new read_pause / read_continue routines
Alan T. DeKok [Wed, 27 Sep 2017 19:13:10 +0000 (15:13 -0400)]
add read_pause / read_continue
which allows the caller to pause reading from a socket,
and then continue reading from it.
With some magic that if there's no read AND write, the socket
is ignored
Alan T. DeKok [Wed, 27 Sep 2017 18:37:28 +0000 (14:37 -0400)]
allow write() to return 0, which means "close the socket"
There may be a better way, but this works for now
Alan T. DeKok [Wed, 27 Sep 2017 18:30:46 +0000 (14:30 -0400)]
add "priority" to detail listener config
and move it out of detail_file config
Alan T. DeKok [Wed, 27 Sep 2017 18:26:14 +0000 (14:26 -0400)]
round up ring buffer size to the nearest power of 2
Alan T. DeKok [Wed, 27 Sep 2017 16:25:07 +0000 (12:25 -0400)]
parse header && decode VPs