]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Mon, 2 Jun 2025 03:25:45 +0000 (21:25 -0600)]
Fix minor issue in rlm_lua
Arran Cudbard-Bell [Mon, 2 Jun 2025 03:25:25 +0000 (21:25 -0600)]
p_result may be NULL
Arran Cudbard-Bell [Mon, 2 Jun 2025 02:30:30 +0000 (20:30 -0600)]
Add variant of functions which do not produce a result
Arran Cudbard-Bell [Sun, 1 Jun 2025 21:22:13 +0000 (15:22 -0600)]
Can't use stack variables for p_result
Arran Cudbard-Bell [Sun, 1 Jun 2025 21:21:53 +0000 (15:21 -0600)]
Only resolve priorities if the rcode is actually set
Arran Cudbard-Bell [Sun, 1 Jun 2025 20:25:34 +0000 (14:25 -0600)]
Frames always have an instruction
Arran Cudbard-Bell [Sun, 1 Jun 2025 19:52:07 +0000 (13:52 -0600)]
More p_result fixes
Arran Cudbard-Bell [Sun, 1 Jun 2025 18:33:58 +0000 (12:33 -0600)]
Allow p_result to be set for rlm_ldap_find_user_async
Arran Cudbard-Bell [Sun, 1 Jun 2025 18:20:12 +0000 (12:20 -0600)]
Switch ldap modifications to using process functions
Arran Cudbard-Bell [Sun, 1 Jun 2025 18:12:24 +0000 (12:12 -0600)]
Add sanity check in unlang_module_yield
Why? Because if the yield is done _AFTER_ a function has been pushed onto the stack, we get undefined reuslts
Arran Cudbard-Bell [Sun, 1 Jun 2025 17:52:41 +0000 (11:52 -0600)]
Run the ldap.authorize method as module methods, not functions
Arran Cudbard-Bell [Sun, 1 Jun 2025 17:36:02 +0000 (11:36 -0600)]
Audit more uses of unlang_function_push
Arran Cudbard-Bell [Sun, 1 Jun 2025 17:08:05 +0000 (11:08 -0600)]
Results for call_env expansion are written to env_result
Arran Cudbard-Bell [Sun, 1 Jun 2025 17:05:02 +0000 (11:05 -0600)]
Notes on unlang_function_push calls in virtual servers
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:49:23 +0000 (10:49 -0600)]
Nothing seems to use the rcode result from ldap_trunk_query_results, so remove the return codes as they're confusing
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:44:39 +0000 (10:44 -0600)]
Return UNLANG_ACTION_FAIL from RETURN_UNLANG_FAIL
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:28:06 +0000 (10:28 -0600)]
Explicitly store the result of edit get password operations
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:24:53 +0000 (10:24 -0600)]
Explicitly store the result of bind operations in the calling frame in LDAP
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:08:48 +0000 (10:08 -0600)]
Add note on using unlang_interpret_result to make unlang functions transparent
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:04:41 +0000 (10:04 -0600)]
Make the eap_virtual_server_resume function transparent in terms of rcodes
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:04:00 +0000 (10:04 -0600)]
Handshake resume function writes out results to eap_tls_session->state so we don't need to pass through the rcode
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:03:23 +0000 (10:03 -0600)]
Make the logging destination resumption function transparent in terms of rcodes
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:02:43 +0000 (10:02 -0600)]
Make the result pointer for the current frame available
Arran Cudbard-Bell [Sun, 1 Jun 2025 16:01:58 +0000 (10:01 -0600)]
s/unlang_interpret_result/unlang_interpret_rcode/
Arran Cudbard-Bell [Sun, 1 Jun 2025 15:45:44 +0000 (09:45 -0600)]
We discard the return codes from triggers because they're fully async
Arran Cudbard-Bell [Sun, 1 Jun 2025 15:45:03 +0000 (09:45 -0600)]
Add better notes on sigmasks
Arran Cudbard-Bell [Sun, 1 Jun 2025 15:41:16 +0000 (09:41 -0600)]
Check unlang_module_yield return code in ttls.c
Arran Cudbard-Bell [Sun, 1 Jun 2025 15:39:32 +0000 (09:39 -0600)]
Notes on how subrequests are handled in peap.c
Arran Cudbard-Bell [Sun, 1 Jun 2025 15:39:08 +0000 (09:39 -0600)]
peap: Cancel child request's frames
Nick Porter [Tue, 17 Jun 2025 19:42:06 +0000 (20:42 +0100)]
Pass unlang_result_t to modules instead of rlm_rcode_t
Arran Cudbard-Bell [Sun, 1 Jun 2025 02:34:41 +0000 (20:34 -0600)]
Populate this frame's result with the results of the resolutions
Arran Cudbard-Bell [Sun, 1 Jun 2025 02:10:12 +0000 (20:10 -0600)]
Fix scan issues
Arran Cudbard-Bell [Sun, 1 Jun 2025 02:10:00 +0000 (20:10 -0600)]
Use a dummy instruction in frame[0]
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:42:18 +0000 (19:42 -0600)]
Allow the rcode from virtual servers to be redirected
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:42:01 +0000 (19:42 -0600)]
Make discarding function rcodes official
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:40:48 +0000 (19:40 -0600)]
Check we can find slapd
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:40:29 +0000 (19:40 -0600)]
Fix arg order to unlang_subrequest_child_push
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:39:41 +0000 (19:39 -0600)]
Only assert if the child request is runnable
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:38:01 +0000 (19:38 -0600)]
Set priority of the process_reply function high enough to bubble up
Arran Cudbard-Bell [Sun, 1 Jun 2025 01:37:28 +0000 (19:37 -0600)]
Help tracing the crazyness
Arran Cudbard-Bell [Sat, 31 May 2025 23:52:56 +0000 (17:52 -0600)]
Yielded makes more sense than resumable
Arran Cudbard-Bell [Sat, 31 May 2025 23:52:42 +0000 (17:52 -0600)]
Ensure that if the child hasn't completed, then its parent must be yielded if it's signalled
Arran Cudbard-Bell [Sat, 31 May 2025 22:14:48 +0000 (16:14 -0600)]
Don't crash on request timeout
Arran Cudbard-Bell [Sat, 31 May 2025 22:05:01 +0000 (16:05 -0600)]
Use section result from rctx in eap_aka_sim state machine
Arran Cudbard-Bell [Sat, 31 May 2025 21:08:59 +0000 (15:08 -0600)]
Use unlang_map_yield for LDAP maps
Arran Cudbard-Bell [Sat, 31 May 2025 19:26:42 +0000 (13:26 -0600)]
Fix header include in dict.h
Arran Cudbard-Bell [Sat, 31 May 2025 19:26:30 +0000 (13:26 -0600)]
Remove useless calls to unlang_function_push and just use unlang_module_yield
Adding new stack frames is expensive, and there's no need to do this as we can just mutate the module's frame
Arran Cudbard-Bell [Sat, 31 May 2025 19:25:12 +0000 (13:25 -0600)]
Add better support for async maps
We really don't want to push additional frames if we can help it
Arran Cudbard-Bell [Mon, 19 May 2025 04:19:11 +0000 (22:19 -0600)]
No reason to use a function to represent module state when processing accounting queries
Arran Cudbard-Bell [Mon, 19 May 2025 03:28:21 +0000 (21:28 -0600)]
Reduce stack dumpage
Arran Cudbard-Bell [Mon, 19 May 2025 03:28:09 +0000 (21:28 -0600)]
No need to use functions to resume in the rlm_files module
Arran Cudbard-Bell [Sun, 18 May 2025 21:16:52 +0000 (15:16 -0600)]
try-no-sibling test needs to set the rcode back to ok
Arran Cudbard-Bell [Sun, 18 May 2025 21:16:35 +0000 (15:16 -0600)]
Now that rcode works correctly, having OR or AND fail when their children fail causes tests to fail
Arran Cudbard-Bell [Sun, 18 May 2025 21:16:05 +0000 (15:16 -0600)]
Have xlat push functions take a result_p
Arran Cudbard-Bell [Sun, 18 May 2025 19:30:49 +0000 (13:30 -0600)]
Print what's being expanded in stack_dump
Arran Cudbard-Bell [Sun, 18 May 2025 02:10:08 +0000 (20:10 -0600)]
Transactions need to operate on the result of the child section, not the current frame result
Arran Cudbard-Bell [Sun, 18 May 2025 00:51:48 +0000 (18:51 -0600)]
Rework interpreter to make handling of rcodes sane, and to discard rcodes entirely when a frame is popped
Arran Cudbard-Bell [Sun, 18 May 2025 00:51:25 +0000 (18:51 -0600)]
Use common function to print closing section debug
Arran Cudbard-Bell [Sun, 18 May 2025 00:49:54 +0000 (18:49 -0600)]
Correct calculation of overall result for parallel keyword
Arran Cudbard-Bell [Sun, 18 May 2025 00:49:28 +0000 (18:49 -0600)]
Look in the correct place for transaction results
Arran Cudbard-Bell [Sun, 18 May 2025 00:48:37 +0000 (18:48 -0600)]
Fix auto-added headers
Arran Cudbard-Bell [Sun, 18 May 2025 00:06:35 +0000 (18:06 -0600)]
Use new rcode location in catch
Arran Cudbard-Bell [Sun, 18 May 2025 00:06:17 +0000 (18:06 -0600)]
Not finding a capture group should not be fail, it's just a NULL result
Arran Cudbard-Bell [Sun, 18 May 2025 00:05:54 +0000 (18:05 -0600)]
Global timeout is the only thing that produces a timeout rcode outside of a module now
Arran Cudbard-Bell [Sun, 18 May 2025 00:05:10 +0000 (18:05 -0600)]
Remove RETURN_MODULE_TRANSPARENT
Arran Cudbard-Bell [Sun, 18 May 2025 00:04:49 +0000 (18:04 -0600)]
Don't overwrite frame state in dhcpv6 state machine
Arran Cudbard-Bell [Sun, 18 May 2025 00:04:22 +0000 (18:04 -0600)]
Discard rcodes from finally
Arran Cudbard-Bell [Wed, 14 May 2025 23:55:05 +0000 (17:55 -0600)]
Don't include ourselves...
Arran Cudbard-Bell [Wed, 14 May 2025 22:32:35 +0000 (16:32 -0600)]
Switch to using unlang_result_t for process functions, and unlang functions
Arran Cudbard-Bell [Wed, 14 May 2025 19:43:31 +0000 (13:43 -0600)]
Add macros to populate unlang_result_t
Arran Cudbard-Bell [Wed, 14 May 2025 22:00:24 +0000 (16:00 -0600)]
Rework state machines so that they operate consistently with new result structure
Fix pre-existing issue in DHCPv4 where the default section rcode for send sections would cause the packet type to switch to Do-Not-Response
Add unlang_result_t to hold both a priority and a rcode
Arran Cudbard-Bell [Mon, 12 May 2025 20:10:55 +0000 (14:10 -0600)]
Allow modules to specify the size and type of the rctx to allocate
This mostly saves on boilerplate at the top of every module call.
Arran Cudbard-Bell [Sun, 11 May 2025 21:34:28 +0000 (15:34 -0600)]
Pass arguments to push functions in a structure
Makes it easier to add or remove args
Arran Cudbard-Bell [Sun, 11 May 2025 21:31:37 +0000 (15:31 -0600)]
We don't return yield anymore
Arran Cudbard-Bell [Sat, 10 May 2025 19:57:34 +0000 (13:57 -0600)]
Start simplifying rcode handling
Arran Cudbard-Bell [Sat, 10 May 2025 01:01:42 +0000 (19:01 -0600)]
Use proper type for the rcode priority
Arran Cudbard-Bell [Sat, 10 May 2025 00:25:04 +0000 (18:25 -0600)]
Minor interpreter cleanups
Arran Cudbard-Bell [Fri, 9 May 2025 21:34:03 +0000 (15:34 -0600)]
No customers here...
Arran Cudbard-Bell [Fri, 9 May 2025 21:33:51 +0000 (15:33 -0600)]
Make the delay module transparent for rcodes
Alan T. DeKok [Tue, 17 Jun 2025 17:58:05 +0000 (13:58 -0400)]
there is no 6th expansion, so we don't expand it
Nick Porter [Tue, 17 Jun 2025 20:07:41 +0000 (21:07 +0100)]
Move all rlm_crl tests to one file
To avoid parallel runs conflicting - OpenSSL appears to sometimes get in
a mess if there are parallel attempts to sign / revoke / create crl
Alan T. DeKok [Tue, 17 Jun 2025 17:11:56 +0000 (13:11 -0400)]
return fail on LHS expansion failed, too
Alan T. DeKok [Tue, 17 Jun 2025 15:44:23 +0000 (11:44 -0400)]
failed expanding RHS xlat is an rcode FAIL
Alan T. DeKok [Tue, 17 Jun 2025 11:20:19 +0000 (07:20 -0400)]
add test to match padding commit
b3f996f
Nick Porter [Tue, 17 Jun 2025 09:22:02 +0000 (10:22 +0100)]
Walk over DHCPv4 padding option rather than returning NULL
Some clients put padding in the middle of packets - so the option we're
looking for may be after the padding.
Nick Porter [Tue, 17 Jun 2025 08:55:46 +0000 (09:55 +0100)]
Look up interface packet was received on if inst->interface not set
Allows a DHCP server to be set to listen on 0.0.0.0 and not bound to a
specific interface to handle broadcast DHCP requests on multiple
interfaces.
Nick Porter [Mon, 16 Jun 2025 18:33:03 +0000 (19:33 +0100)]
Set IP_RECVIF on BSD sockets to get interface details
github-actions[bot] [Tue, 17 Jun 2025 04:34:25 +0000 (04:34 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/cbor.tar
github-actions[bot] [Tue, 17 Jun 2025 04:31:10 +0000 (04:31 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/der.tar
Alan T. DeKok [Mon, 16 Jun 2025 18:01:01 +0000 (14:01 -0400)]
set error only if it failed
Nick Porter [Mon, 16 Jun 2025 11:15:44 +0000 (12:15 +0100)]
Add CHAP to radclient tests
Nick Porter [Mon, 16 Jun 2025 10:35:56 +0000 (11:35 +0100)]
Create CHAP-Challenge attribute if not set
fr_packet_sign() re-calculates the request authenticator, so it can't be
used as CHAP-Challenge
Nick Porter [Mon, 16 Jun 2025 09:53:04 +0000 (10:53 +0100)]
CHAP-Challenge is not fixed to 16 octets
RFC2865 says min length is 7
Nick Porter [Mon, 16 Jun 2025 10:40:26 +0000 (11:40 +0100)]
Remove Password.Cleartext from expected radclient test output
Nick Porter [Mon, 16 Jun 2025 09:47:15 +0000 (10:47 +0100)]
No need to copy User-Password to Password.Cleartext
Nick Porter [Mon, 16 Jun 2025 08:02:05 +0000 (09:02 +0100)]
Set prefix length so address comparisons work
fr_ipaddr_cmp() uses prefix len to determine how many bits to compare
Nick Porter [Fri, 13 Jun 2025 16:08:59 +0000 (17:08 +0100)]
Associate a dictionary with %<proto>.decode xlats
And check that protocol decoders are only called in the correct
namespace.
Nick Porter [Fri, 13 Jun 2025 16:00:41 +0000 (17:00 +0100)]
Move fr_der_decode_ctx_t definition to header
So other modules can create a decode ctx when using der decoder
Nick Porter [Fri, 13 Jun 2025 15:43:29 +0000 (16:43 +0100)]
Move dict_der to attrs.h
So der.h can be included in modules with their own `dict_der` without
conflicts.
Nick Porter [Fri, 13 Jun 2025 15:35:40 +0000 (16:35 +0100)]
Remove unused variable