]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Fri, 9 May 2025 01:34:47 +0000 (19:34 -0600)]
Another minor typo
Arran Cudbard-Bell [Fri, 9 May 2025 01:32:57 +0000 (19:32 -0600)]
Minor cleanups to finally docs
Arran Cudbard-Bell [Thu, 8 May 2025 23:47:10 +0000 (17:47 -0600)]
Unify elements of child request handling for parallel, subrequests, and manual subrequests
Fix signalling in parallel. Both timeouts triggered on the parent side, and timeouts triggered on the child side now work.
Use relative timers for the retry keyword so synthetic time works correctly, same with the synchronous interpreter.
Arran Cudbard-Bell [Thu, 8 May 2025 21:38:32 +0000 (15:38 -0600)]
Reduce spurious debug messages
Alan T. DeKok [Thu, 8 May 2025 18:18:48 +0000 (14:18 -0400)]
rename functions for consistencyo
%debug_attr() %pairs.debug()
%pairs() %pairs.print()
perl -p -i -e 's/\%pairs/\%pairs.print/g' $(git grep -l '%pairs' src)
perl -p -i -e 's/\%debug_attr/\%pairs.debug/g' $(git grep -l '%debug_attr' src)
Alan T. DeKok [Thu, 8 May 2025 17:50:55 +0000 (13:50 -0400)]
rename %client() to %request.client()
the old name is still supported for compatibility
Nick Porter [Thu, 8 May 2025 13:49:10 +0000 (14:49 +0100)]
Free tracking structure for ignored packet
Nick Porter [Thu, 8 May 2025 13:40:21 +0000 (14:40 +0100)]
Remove config options which controlled use of packet-authenticator in dedup
Nick Porter [Thu, 8 May 2025 13:39:22 +0000 (14:39 +0100)]
The same for proto_radius_tcp
Nick Porter [Thu, 8 May 2025 13:32:17 +0000 (14:32 +0100)]
Remove packet authenticator from tracking table comparison
If a packet arrives with the same ID and type, from the same client
(which is already deduped on source IP and port), then there is no need
to keep the old entry - there is no way for us to re-reply.
nolade [Wed, 23 Apr 2025 15:35:46 +0000 (11:35 -0400)]
Update winbind troublshooting (PR 5553 comments)
nolade [Fri, 11 Apr 2025 19:20:13 +0000 (15:20 -0400)]
docs: Import customer doc Debugging winbind errors into Antora docs Introduction >> Troubleshooting >> Connectivity section
nolade [Tue, 6 May 2025 21:45:29 +0000 (17:45 -0400)]
docs: make docsite build fixes - updated xrefs and adjusted heading levels
Nick Porter [Thu, 8 May 2025 09:33:46 +0000 (10:33 +0100)]
Add listener names
Nick Porter [Thu, 8 May 2025 08:40:31 +0000 (09:40 +0100)]
Correct tracking structure talloc pool sizes
Nick Porter [Thu, 8 May 2025 07:20:25 +0000 (08:20 +0100)]
Remove pair that didn't get populated
Nick Porter [Wed, 7 May 2025 19:36:19 +0000 (20:36 +0100)]
Remove unused argument from format
Nick Porter [Wed, 7 May 2025 19:35:19 +0000 (20:35 +0100)]
Handle erros parsing call_env parser rules with name of CF_IDENT_ANY
Alan T. DeKok [Thu, 8 May 2025 01:53:04 +0000 (21:53 -0400)]
mash void to string
if the argument to 'switch' is a function returning void, then
update the tmpl to mash it to "string"
Alan T. DeKok [Thu, 8 May 2025 01:46:43 +0000 (21:46 -0400)]
parse %func() as argument to 'switch'
Alan T. DeKok [Thu, 8 May 2025 01:45:18 +0000 (21:45 -0400)]
xlats can't be constant, either
Alan T. DeKok [Thu, 8 May 2025 01:40:06 +0000 (21:40 -0400)]
quote strings are strings
Arran Cudbard-Bell [Thu, 8 May 2025 01:02:25 +0000 (19:02 -0600)]
...and this
Arran Cudbard-Bell [Thu, 8 May 2025 01:01:52 +0000 (19:01 -0600)]
Remove debugging crap
Arran Cudbard-Bell [Thu, 8 May 2025 00:59:34 +0000 (18:59 -0600)]
Can't check for presence of the event
Arran Cudbard-Bell [Wed, 7 May 2025 19:24:09 +0000 (13:24 -0600)]
Insert timer event for resumption before altering time
Arran Cudbard-Bell [Wed, 7 May 2025 19:23:26 +0000 (13:23 -0600)]
Tests for nested timeouts
Alan T. DeKok [Wed, 7 May 2025 15:03:23 +0000 (11:03 -0400)]
start splitting out %time(everything)
into use-specific cases
Alan T. DeKok [Wed, 7 May 2025 14:49:00 +0000 (10:49 -0400)]
move %nextime() to %time.next()
Alan T. DeKok [Wed, 7 May 2025 14:33:50 +0000 (10:33 -0400)]
clean up time_delta handling
check if the fractions overflow the scale.
Add / sub fractions as needed
Alan T. DeKok [Wed, 7 May 2025 13:29:53 +0000 (09:29 -0400)]
print out error if reference is missing
Nick Porter [Wed, 7 May 2025 14:04:27 +0000 (15:04 +0100)]
Correct TLS version numbers
Nick Porter [Wed, 7 May 2025 07:25:58 +0000 (08:25 +0100)]
We don't build against OpenSSL < 3
So alt-libs doesn't need to switch from OpenSSL 1 to 3
Nick Porter [Wed, 7 May 2025 07:22:21 +0000 (08:22 +0100)]
Ensure Homebrew installed OpenSSL is used
Alan T. DeKok [Wed, 7 May 2025 13:28:10 +0000 (09:28 -0400)]
remove debugging message
Alan T. DeKok [Wed, 7 May 2025 13:25:37 +0000 (09:25 -0400)]
add cbor crashing test
Alan T. DeKok [Wed, 7 May 2025 13:23:03 +0000 (09:23 -0400)]
check for PANIC_ACTION, and do fr_fault_setup() if so
so we can get gdb back traces from fuzzer crashes
Alan T. DeKok [Wed, 7 May 2025 13:12:10 +0000 (09:12 -0400)]
add -g3 when fuzzing
which should hopefully get us some more debug information on
fuzzing failures
Gerald Vogt [Wed, 7 May 2025 12:56:13 +0000 (14:56 +0200)]
Adding support for 389-ds/FreeIPA PBKDF2 passwords (#5564)
* Add support for PBKDF2-SHA{1,256,512} as implemented in 389ds
* implement support for {PBKDF2_SHA256} 389ds legacy password hash
---------
Co-authored-by: Gerald Vogt <12151414+gvde@users.noreply.github.com>
Alan T. DeKok [Wed, 7 May 2025 11:12:10 +0000 (07:12 -0400)]
rewrite cf_reference_item()
while it's now more code, the code is at least clearer, and also
returns descriptive errors sayng exactly what went wrong.
Update the callers to print out these errors
Alan T. DeKok [Wed, 7 May 2025 11:04:16 +0000 (07:04 -0400)]
tweaks and more examples
Alan T. DeKok [Tue, 6 May 2025 20:41:28 +0000 (16:41 -0400)]
allow references in EAP such as 'tls = server[tls].config'
Alan T. DeKok [Tue, 6 May 2025 20:13:06 +0000 (16:13 -0400)]
don't allow in-place `...`
we are gradually getting rid of all uses of back-ticks. And
there is really no reason to allow `...` by itself on a line.
We still allow %foo() on a line by itself.
Arran Cudbard-Bell [Wed, 7 May 2025 00:52:37 +0000 (18:52 -0600)]
Add a special subrequest_child operation type
We used to use functions for this, but we need the frame to be non-cancellable.
Arran Cudbard-Bell [Wed, 7 May 2025 00:51:47 +0000 (18:51 -0600)]
Sanity checks
Arran Cudbard-Bell [Wed, 7 May 2025 00:43:36 +0000 (18:43 -0600)]
Say what we're marking as runnable
Arran Cudbard-Bell [Wed, 7 May 2025 00:43:18 +0000 (18:43 -0600)]
Set the initial frame action correctly for resumed requests
Arran Cudbard-Bell [Wed, 7 May 2025 00:43:00 +0000 (18:43 -0600)]
Fall-Through to deal with pop
Arran Cudbard-Bell [Tue, 6 May 2025 23:04:49 +0000 (17:04 -0600)]
Cancelling and unwinding are now really separate things
Arran Cudbard-Bell [Tue, 6 May 2025 22:31:27 +0000 (16:31 -0600)]
Don't allow signals to be delivered after a frame is cancelled
Arran Cudbard-Bell [Tue, 6 May 2025 22:31:08 +0000 (16:31 -0600)]
Immediately cancel timeout section timers on timeout
Arran Cudbard-Bell [Tue, 6 May 2025 21:57:24 +0000 (15:57 -0600)]
Apparently this gets auto freed
Arran Cudbard-Bell [Tue, 6 May 2025 21:13:02 +0000 (15:13 -0600)]
Set prereqs for new tests
Arran Cudbard-Bell [Tue, 6 May 2025 21:11:22 +0000 (15:11 -0600)]
Give cancelled requests extremely high priority
This ensures they're cleaned ASAP and release any resources.
Arran Cudbard-Bell [Tue, 6 May 2025 20:04:01 +0000 (14:04 -0600)]
Ensure retry times out
Arran Cudbard-Bell [Tue, 6 May 2025 19:49:00 +0000 (13:49 -0600)]
Make fr_timer_armed a static inline to avoid compiler errors
Arran Cudbard-Bell [Tue, 6 May 2025 19:43:00 +0000 (13:43 -0600)]
Hack in request->timeout to unit_test_module
Arran Cudbard-Bell [Tue, 6 May 2025 19:07:16 +0000 (13:07 -0600)]
Add docs for finally
Arran Cudbard-Bell [Tue, 6 May 2025 17:58:13 +0000 (11:58 -0600)]
proto_control_unix needs to link to the util submodule
Arran Cudbard-Bell [Tue, 6 May 2025 17:55:07 +0000 (11:55 -0600)]
Don't trash the rcodes in process state machine sections
Arran Cudbard-Bell [Tue, 6 May 2025 17:17:43 +0000 (11:17 -0600)]
Failing to create attributes in an error
Arran Cudbard-Bell [Tue, 6 May 2025 17:17:25 +0000 (11:17 -0600)]
Default priority for timeout in module calls should be return
Arran Cudbard-Bell [Tue, 6 May 2025 17:16:55 +0000 (11:16 -0600)]
Functions shouldn't have braces
Arran Cudbard-Bell [Tue, 6 May 2025 04:14:56 +0000 (22:14 -0600)]
Add talloc_realloc_zero to do zeroed reallocs
Arran Cudbard-Bell [Mon, 5 May 2025 23:18:34 +0000 (17:18 -0600)]
Add finally virtual server sections
Arran Cudbard-Bell [Mon, 5 May 2025 23:17:59 +0000 (17:17 -0600)]
Top level timeout sections are going away rework retry-timeout
Arran Cudbard-Bell [Mon, 5 May 2025 23:13:24 +0000 (17:13 -0600)]
Test parent cancellation in subrequest
Arran Cudbard-Bell [Mon, 5 May 2025 23:12:57 +0000 (17:12 -0600)]
WS
Arran Cudbard-Bell [Mon, 5 May 2025 22:56:52 +0000 (16:56 -0600)]
Function to return how long until an event fires
Arran Cudbard-Bell [Mon, 5 May 2025 22:51:38 +0000 (16:51 -0600)]
Have process modules export their packet types
Arran Cudbard-Bell [Mon, 5 May 2025 22:29:14 +0000 (16:29 -0600)]
Bail early if there's no stack
Arran Cudbard-Bell [Mon, 5 May 2025 22:22:16 +0000 (16:22 -0600)]
Explicitly cancel retry timers when the frame is cancelled
It's not enough to wait for the frame to be popped, because the cancellation and popping can occur at different times, which is enough for the retry event to fire, and things to go boom.
Arran Cudbard-Bell [Mon, 5 May 2025 22:21:19 +0000 (16:21 -0600)]
Fix typos in timer.c that caused it not to build WITH_EVENT_DEBUG
Arran Cudbard-Bell [Mon, 5 May 2025 20:42:01 +0000 (14:42 -0600)]
Add time.advance() for unit_test_module
This lets us artificially advance time for testing timeouts
Arran Cudbard-Bell [Mon, 5 May 2025 19:50:12 +0000 (13:50 -0600)]
Forward cancellations to children, and don't cancel the parent until the child is done
Arran Cudbard-Bell [Sun, 4 May 2025 03:41:27 +0000 (23:41 -0400)]
Asynchronously deliver signals, synchronously cancel frames
This lets us stop subrequests immediately when their parents are cancelled.
Add op flag to mask certain frames to _prevent_ their cancellation.
Arran Cudbard-Bell [Sat, 3 May 2025 00:28:47 +0000 (20:28 -0400)]
Constify the output of enumv functions
Add enumv iterators
Arran Cudbard-Bell [Thu, 1 May 2025 22:22:41 +0000 (18:22 -0400)]
Rename virtual server listener structure
Nick Porter [Tue, 6 May 2025 17:27:56 +0000 (18:27 +0100)]
Bump CI MacOS version
Alan T. DeKok [Tue, 6 May 2025 15:29:16 +0000 (11:29 -0400)]
better handle connected UDP sockets
and forbid reuse_port when we have connected UDP sockets, AND
the local IP is a wildcard. The kernel will send replies to a
random socket, and not to the correct one!
In order to do "connected" UDP sockets with a local wildcard IP,
we will need to use recvfromto(), and then manually direct the
packets to the correct destination socket.
This also means that an "accept" socket will be needed, even for
UDP. And, the "accept" socket will need to have a list of child
sockets.
Alan T. DeKok [Tue, 6 May 2025 14:23:56 +0000 (10:23 -0400)]
add fr_sockaddr_cmp()
Alan T. DeKok [Tue, 6 May 2025 13:20:53 +0000 (09:20 -0400)]
use consistent variable types and names for CMP() checks
Alan T. DeKok [Tue, 6 May 2025 13:03:21 +0000 (09:03 -0400)]
try a little more to make 'listen' names unique
and give better error messages where we can
nolade [Tue, 6 May 2025 16:03:00 +0000 (12:03 -0400)]
docs: add guidelines for documentation for contributers, coders, and me.
Nick Porter [Tue, 6 May 2025 12:43:54 +0000 (13:43 +0100)]
Remove errant policy
Alan T. DeKok [Tue, 6 May 2025 10:22:25 +0000 (06:22 -0400)]
add "finally"
Nick Porter [Tue, 6 May 2025 09:34:52 +0000 (10:34 +0100)]
Cast constants to float for correct comparison
Nick Porter [Mon, 5 May 2025 16:39:36 +0000 (17:39 +0100)]
Make the TLS session ID available in `establish session`
Nick Porter [Mon, 5 May 2025 11:06:34 +0000 (12:06 +0100)]
Call `new session` if configured when starting TLS for TTLS
Nick Porter [Mon, 5 May 2025 11:05:48 +0000 (12:05 +0100)]
Add attribute to record certificate chain depth
Nick Porter [Mon, 5 May 2025 08:54:36 +0000 (09:54 +0100)]
Call `new session` if configured when starting TLS session for PEAP
Nick Porter [Mon, 5 May 2025 07:55:26 +0000 (08:55 +0100)]
Include nested password parent name in warning
To avoid confusing users
Alan T. DeKok [Sun, 4 May 2025 13:19:47 +0000 (09:19 -0400)]
let's just commit the generated files, too
so that the rpm build phase doesn't need asciidoctor
Alan T. DeKok [Sun, 4 May 2025 13:12:10 +0000 (09:12 -0400)]
generate dictionary "man" page from Antora
Alan T. DeKok [Sun, 4 May 2025 12:50:25 +0000 (08:50 -0400)]
new "man unlang" page and cleanups
which is for now just a paragraph header, and then includes
the "keywords.adoc" file. The goal here is to provide a simple
offline reference, and not a complete documentation of the server
Alan T. DeKok [Sun, 4 May 2025 12:25:02 +0000 (08:25 -0400)]
remove "man unlang" and references to it.
The file is out of date, and has little more than automatic
changes made for a while. We will replace it with a file that
is generated from asciidoc, and which lives inside of antora
Alan T. DeKok [Sun, 4 May 2025 00:23:20 +0000 (20:23 -0400)]
notes on "don't use exec, it's slow"
Alan T. DeKok [Sat, 3 May 2025 13:54:26 +0000 (09:54 -0400)]
notes on v4
Alan T. DeKok [Sat, 3 May 2025 12:15:37 +0000 (08:15 -0400)]
add %regex.search()
Nick Porter [Fri, 2 May 2025 19:28:48 +0000 (20:28 +0100)]
Set start = 0 in SQLite test
To verify correct behaviour when initial query launches connection