]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Wed, 8 Apr 2020 13:26:00 +0000 (09:26 -0400)]
cleanups and add fr_pair_ctx_reset
in preparation for adding similar functions for tmpls and maps
Arran Cudbard-Bell [Wed, 8 Apr 2020 00:11:33 +0000 (19:11 -0500)]
track: Initialise file/line
Arran Cudbard-Bell [Wed, 8 Apr 2020 00:09:39 +0000 (19:09 -0500)]
track: Add debugging code
Arran Cudbard-Bell [Tue, 7 Apr 2020 22:51:10 +0000 (17:51 -0500)]
trunk: Initialise tconn
Arran Cudbard-Bell [Tue, 7 Apr 2020 22:20:52 +0000 (17:20 -0500)]
trunk: Always do checks for inst->recv_buff range if we're no on Linux
Arran Cudbard-Bell [Tue, 7 Apr 2020 22:18:10 +0000 (17:18 -0500)]
track: We don't need to track num_free now the dlists have counters
Arran Cudbard-Bell [Tue, 7 Apr 2020 21:51:03 +0000 (16:51 -0500)]
trunk: Rate limit enqueue failure log messages
Arran Cudbard-Bell [Tue, 7 Apr 2020 21:43:11 +0000 (16:43 -0500)]
log: Improve rate limiting macros
Arran Cudbard-Bell [Tue, 7 Apr 2020 20:34:03 +0000 (15:34 -0500)]
rlm_radius: Assert if we can't allocate a tracking entry
We should only allow as many requests to be assigned to a connection as there are tracking entries available
Alan T. DeKok [Tue, 7 Apr 2020 19:59:35 +0000 (15:59 -0400)]
put queued treqs into the array
if we signal fail on one packet in the middle of looping over
'i', then 'queued != i'. So track them separately.
Arran Cudbard-Bell [Tue, 7 Apr 2020 19:54:30 +0000 (14:54 -0500)]
trunk: Assign different debug levels to different connection state transitions
Arran Cudbard-Bell [Tue, 7 Apr 2020 17:49:37 +0000 (12:49 -0500)]
shouldn't set inst->recv_buff_is_set either
Arran Cudbard-Bell [Tue, 7 Apr 2020 17:47:57 +0000 (12:47 -0500)]
rlm_radius: Only set recv buff to zero on Linux Closes #3355
Arran Cudbard-Bell [Tue, 7 Apr 2020 17:22:09 +0000 (12:22 -0500)]
trunk: Ensure requests are in the right states fo the lists they're in
Arran Cudbard-Bell [Tue, 7 Apr 2020 17:15:53 +0000 (12:15 -0500)]
rlm_radius: More asserts
Arran Cudbard-Bell [Tue, 7 Apr 2020 14:20:30 +0000 (09:20 -0500)]
cursor: store/restore fr_cursor_intersect_next's iterator function and uctx
Arran Cudbard-Bell [Tue, 7 Apr 2020 14:19:36 +0000 (09:19 -0500)]
cursor: s/ctx/uctx/
Jorge Pereira [Tue, 7 Apr 2020 14:17:10 +0000 (11:17 -0300)]
trunk: Fix missing 'return' (#3357)
Arran Cudbard-Bell [Tue, 7 Apr 2020 04:55:50 +0000 (23:55 -0500)]
trunk: Inactive-Draining->Draining is fine
Arran Cudbard-Bell [Tue, 7 Apr 2020 01:16:41 +0000 (20:16 -0500)]
whitespace
Arran Cudbard-Bell [Tue, 7 Apr 2020 01:01:32 +0000 (20:01 -0500)]
scheduler: Better identifiers when we're running in single threaded mode
Arran Cudbard-Bell [Tue, 7 Apr 2020 00:54:30 +0000 (19:54 -0500)]
network: More orderly exit
Arran Cudbard-Bell [Tue, 7 Apr 2020 00:52:52 +0000 (19:52 -0500)]
trunk: Better assert for unexpected post cancellation state
Arran Cudbard-Bell [Mon, 6 Apr 2020 23:28:03 +0000 (18:28 -0500)]
rlm_radius: Fix debug message
Arran Cudbard-Bell [Mon, 6 Apr 2020 23:27:53 +0000 (18:27 -0500)]
event: Add function to move events between event loops
Arran Cudbard-Bell [Mon, 6 Apr 2020 23:27:35 +0000 (18:27 -0500)]
no perror here...
Alan T. DeKok [Mon, 6 Apr 2020 21:34:02 +0000 (17:34 -0400)]
add fr_pair_ctx_afrom_str() API
for parsing nested pairs with parent / child references
Alan T. DeKok [Mon, 6 Apr 2020 21:22:31 +0000 (17:22 -0400)]
zero length names are not allowed, too
Alan T. DeKok [Mon, 6 Apr 2020 16:44:52 +0000 (12:44 -0400)]
publish skip_string API for quoted strings
Arran Cudbard-Bell [Mon, 6 Apr 2020 21:25:20 +0000 (16:25 -0500)]
unused args
Arran Cudbard-Bell [Mon, 6 Apr 2020 19:38:04 +0000 (14:38 -0500)]
strerror: Use PERROR() where available
Jorge Pereira [Mon, 6 Apr 2020 21:22:46 +0000 (18:22 -0300)]
Suppress leak warnings in talloc_set_memlimit() (#3354)
Arran Cudbard-Bell [Mon, 6 Apr 2020 16:55:20 +0000 (11:55 -0500)]
network: Re-arrange functions to place "create" functions nearer the bottom
Prefix worker messages with the worker name
Arran Cudbard-Bell [Mon, 6 Apr 2020 16:54:45 +0000 (11:54 -0500)]
scheduler: Debugging improvements
Arran Cudbard-Bell [Mon, 6 Apr 2020 16:54:15 +0000 (11:54 -0500)]
Interpreter: Debugging improvements
Arran Cudbard-Bell [Mon, 6 Apr 2020 16:53:55 +0000 (11:53 -0500)]
worker: Debugging improvements
Arran Cudbard-Bell [Mon, 6 Apr 2020 02:55:49 +0000 (21:55 -0500)]
channel: Expose signal names table
Arran Cudbard-Bell [Mon, 6 Apr 2020 02:55:34 +0000 (21:55 -0500)]
typo in comments
Arran Cudbard-Bell [Mon, 6 Apr 2020 02:26:25 +0000 (21:26 -0500)]
event: Defer event loop exit
Only process the exit on the next call to fr_event_corral to make exits more orderly and predictable.
Arran Cudbard-Bell [Mon, 6 Apr 2020 02:25:44 +0000 (21:25 -0500)]
rlm_radius: Remove invalid assert
Arran Cudbard-Bell [Mon, 6 Apr 2020 02:25:24 +0000 (21:25 -0500)]
rlm_radius: There's a macro for that...
Alan T. DeKok [Mon, 6 Apr 2020 16:36:46 +0000 (12:36 -0400)]
dots and dashes are allowed
Alan T. DeKok [Mon, 6 Apr 2020 15:30:08 +0000 (11:30 -0400)]
disallow '.' in attribute names
except for unknown attrs with OIDs: Attr-1.2.2.3
Alan T. DeKok [Mon, 6 Apr 2020 13:14:06 +0000 (09:14 -0400)]
we no longer need RADIUS attributes
Alan T. DeKok [Sat, 4 Apr 2020 19:46:53 +0000 (15:46 -0400)]
skip sending reply if we're exiting
Alan T. DeKok [Fri, 3 Apr 2020 22:20:01 +0000 (18:20 -0400)]
fix typo and simplify
Arran Cudbard-Bell [Sat, 4 Apr 2020 03:03:52 +0000 (21:03 -0600)]
trunk: Allow management whilst in a handler, but don't allow connections to be freed
Arran Cudbard-Bell [Sat, 4 Apr 2020 02:07:58 +0000 (20:07 -0600)]
trunk: Implement deferred connection freeing, and report if a connection has been moved to the halted state from within the pop functions.
Arran Cudbard-Bell [Sat, 4 Apr 2020 02:06:19 +0000 (20:06 -0600)]
trunk: Move request state log printing into a callable function
Arran Cudbard-Bell [Sat, 4 Apr 2020 02:05:31 +0000 (20:05 -0600)]
connection: Allow modification of watchers without re-insertion
Prevent multiple nested calls to the deferred signal processing function.
Fix issue with memory usage after stack frame returned
Arran Cudbard-Bell [Fri, 3 Apr 2020 23:23:58 +0000 (17:23 -0600)]
global: Explicitly initialise instance counts
It really doesn't matter, but it means all the code looks the same...
Arran Cudbard-Bell [Fri, 3 Apr 2020 23:22:41 +0000 (17:22 -0600)]
global: Have instance_count underflow if the free function is called too many times...
Arran Cudbard-Bell [Fri, 3 Apr 2020 23:10:41 +0000 (17:10 -0600)]
tls: Don't need more indirection...
Arran Cudbard-Bell [Fri, 3 Apr 2020 22:40:57 +0000 (16:40 -0600)]
tls: Rename OpenSSL specific functions to fr_openssl_ and generic tls functions to fr_tls_
OpenSSL isn't great about prefixing its internal functions, and this helps when we're looking at stack traces.
James Jones [Fri, 3 Apr 2020 16:01:26 +0000 (11:01 -0500)]
add tests for fr_cursor_intersect* functions
Alan T. DeKok [Fri, 3 Apr 2020 14:07:52 +0000 (10:07 -0400)]
run miscellaneous tests early in the game.
right now it's just cursors, which are needed for the
unit tests.
Alan T. DeKok [Fri, 3 Apr 2020 13:54:31 +0000 (09:54 -0400)]
added wrapper for miscellaneous tests
Arran Cudbard-Bell [Thu, 2 Apr 2020 23:45:39 +0000 (17:45 -0600)]
rlm_radius: Disassociate the REQUEST * with its treq on completion
Issues with treqs being signalled after they're no longer being bound to a REQUEST *.
Arran Cudbard-Bell [Thu, 2 Apr 2020 23:22:47 +0000 (17:22 -0600)]
trunk: Write out a message if the request state log was empty
Arran Cudbard-Bell [Thu, 2 Apr 2020 23:15:41 +0000 (17:15 -0600)]
connection: make the deferred signal processor immune to the connection being freed out from under it...
Arran Cudbard-Bell [Thu, 2 Apr 2020 22:47:48 +0000 (16:47 -0600)]
trunk_tests: Fix clang scan issue
Arran Cudbard-Bell [Thu, 2 Apr 2020 22:15:38 +0000 (16:15 -0600)]
cursor: Fix merge tests
Arran Cudbard-Bell [Thu, 2 Apr 2020 21:52:25 +0000 (15:52 -0600)]
trunk: Typo
James Jones [Thu, 2 Apr 2020 14:48:54 +0000 (09:48 -0500)]
make dhcpv4 use an iterator. This includes making dhcpv4_encode_xlat() only
pass dhcpv4-encodable items to fr_dhcpv4_encode_option().
Arran Cudbard-Bell [Thu, 2 Apr 2020 21:17:58 +0000 (15:17 -0600)]
connection: Fix missed state transition
Arran Cudbard-Bell [Thu, 2 Apr 2020 21:14:53 +0000 (15:14 -0600)]
connection: Pause all signals, not just the deferred ones
Arran Cudbard-Bell [Thu, 2 Apr 2020 20:38:45 +0000 (14:38 -0600)]
trunk: Add a log of state transitions to debug builds
Arran Cudbard-Bell [Thu, 2 Apr 2020 20:09:25 +0000 (14:09 -0600)]
cursor: Deal with the case where a and b are out of sync
This allows something else to advance a, and then fr_cursor_intersect_next, to select the next pair in both sets.
Arran Cudbard-Bell [Thu, 2 Apr 2020 20:07:06 +0000 (14:07 -0600)]
trunk: Fix unassigned/init assert
Arran Cudbard-Bell [Thu, 2 Apr 2020 16:34:31 +0000 (10:34 -0600)]
cursor: Add note and guard for fr_cursor_intersect_* functions
Arran Cudbard-Bell [Thu, 2 Apr 2020 16:12:03 +0000 (10:12 -0600)]
test_point: Add missing field
Arran Cudbard-Bell [Thu, 2 Apr 2020 15:06:00 +0000 (09:06 -0600)]
cursor: Add intersection functions
Herwin Weststrate [Thu, 2 Apr 2020 08:37:01 +0000 (10:37 +0200)]
Fix uint128_to_64 on systems without uint128_t
Alan T. DeKok [Wed, 1 Apr 2020 21:33:05 +0000 (17:33 -0400)]
remove extraneous dependency
the PRE notation is for tests in the same directory, and we
don't need that here
Alan T. DeKok [Wed, 1 Apr 2020 21:26:19 +0000 (17:26 -0400)]
fix encode / decode of boolean
Rapid-Commit should be boolean, not empty octets.
Encoding was correct, in that it skipped encoding the value.
Decoding was wrong, in that it called the standard "from network"
code, which looked for one byte of a boolean value.
Alan T. DeKok [Wed, 1 Apr 2020 21:19:07 +0000 (17:19 -0400)]
export next_encodable, and use in tests
James Jones [Wed, 1 Apr 2020 18:10:33 +0000 (13:10 -0500)]
convert dhcpv6 to use iterator
Arran Cudbard-Bell [Wed, 1 Apr 2020 19:59:53 +0000 (13:59 -0600)]
internal: Decode multiple nested grouped VPs correctly
Arran Cudbard-Bell [Wed, 1 Apr 2020 19:59:37 +0000 (13:59 -0600)]
rest: Don't pass void *s around for the request handle
We now have fr_curl_io_request_t so it makes much more sense to use that. I'm not sure why we ever used void *.
Alan T. DeKok [Wed, 1 Apr 2020 17:42:59 +0000 (13:42 -0400)]
update comments
Alan T. DeKok [Wed, 1 Apr 2020 14:49:22 +0000 (10:49 -0400)]
add PUSH_CHILD_GROUP for XLAT_CHILD
Alexander Clouter [Wed, 1 Apr 2020 10:49:39 +0000 (11:49 +0100)]
support using a smaller DH_KEY_SIZE when building docker containers ('dev' mode)
Alexander Clouter [Sun, 29 Mar 2020 18:18:13 +0000 (19:18 +0100)]
make docker build faster and download less
Alexander Clouter [Mon, 30 Mar 2020 14:48:29 +0000 (15:48 +0100)]
dockerbuild support to pass in build args for source (and branch)
Alexander Clouter [Sun, 29 Mar 2020 14:15:28 +0000 (15:15 +0100)]
debian/sid uses libjson-c4
Alexander Clouter [Mon, 30 Mar 2020 14:47:04 +0000 (15:47 +0100)]
docker RIP Debian 'jessie' 8 (requires 1.0.2+ SSL for v4 build)
Alexander Clouter [Sun, 29 Mar 2020 14:01:53 +0000 (15:01 +0100)]
dockerbuild script should die on error and not just blindly continue
Arran Cudbard-Bell [Wed, 1 Apr 2020 03:07:59 +0000 (21:07 -0600)]
internal: Last of the decoder fixes
VP printing functions need fixing next....
Arran Cudbard-Bell [Wed, 1 Apr 2020 03:00:15 +0000 (21:00 -0600)]
internal: Missed break
Arran Cudbard-Bell [Wed, 1 Apr 2020 03:00:02 +0000 (21:00 -0600)]
internal: Encode correct length
Arran Cudbard-Bell [Wed, 1 Apr 2020 01:48:36 +0000 (19:48 -0600)]
internal: Various decode fixes
Arran Cudbard-Bell [Wed, 1 Apr 2020 01:34:29 +0000 (19:34 -0600)]
proto: Add hex marker macro
Arran Cudbard-Bell [Wed, 1 Apr 2020 01:34:14 +0000 (19:34 -0600)]
log: Fix marker indentation in fr_log_hex_marker
Arran Cudbard-Bell [Wed, 1 Apr 2020 00:58:05 +0000 (18:58 -0600)]
internal: Modify decode to produce TLV VPs sparingly...
Arran Cudbard-Bell [Tue, 31 Mar 2020 22:30:14 +0000 (16:30 -0600)]
Used
Arran Cudbard-Bell [Tue, 31 Mar 2020 22:24:58 +0000 (16:24 -0600)]
dhcpv6: Fix issue encoding groups (and reporting errors)
Arran Cudbard-Bell [Tue, 31 Mar 2020 20:59:51 +0000 (14:59 -0600)]
uint128: Add explicit cast function for 128bit -> 64bit conversion
Alan T. DeKok [Tue, 31 Mar 2020 20:43:35 +0000 (16:43 -0400)]
move xlat code to unlang
because it's only ever used dynamically at run time.
And its needed for handling XLAT_CHILD as FR_TYPE_GROUP
Arran Cudbard-Bell [Tue, 31 Mar 2020 18:41:23 +0000 (12:41 -0600)]
internal: More travis fixes....
Arran Cudbard-Bell [Tue, 31 Mar 2020 18:41:01 +0000 (12:41 -0600)]
net: Although, yes, technically we don't need any bytes to represent zero, it's probably still a good idea we write out one.