]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Thu, 26 Sep 2019 18:03:38 +0000 (14:03 -0400)]
split "map" into init / apply
Alan T. DeKok [Thu, 26 Sep 2019 17:30:24 +0000 (13:30 -0400)]
add frame_inst_size to unlang_op_t and start using it
also add frame_setup() function which sets up a new frame.
Alan T. DeKok [Thu, 26 Sep 2019 16:27:44 +0000 (12:27 -0400)]
put state into frame->state, and let it manage the lifetime
Arran Cudbard-Bell [Wed, 25 Sep 2019 20:02:21 +0000 (15:02 -0500)]
s/USERLOCK/DISALLOW/g
Arran Cudbard-Bell [Tue, 3 Sep 2019 22:38:01 +0000 (18:38 -0400)]
Hacks to allow more EAP methods to function
Arran Cudbard-Bell [Tue, 3 Sep 2019 17:42:09 +0000 (13:42 -0400)]
Create unified library for EAP-SIM/AKA/AKA' and make those modules fully async
Arran Cudbard-Bell [Tue, 3 Sep 2019 21:07:12 +0000 (17:07 -0400)]
Call some EAP methods with a subrequest
Arran Cudbard-Bell [Thu, 26 Sep 2019 18:25:48 +0000 (13:25 -0500)]
Add fallback for non-zeroed talloc pools
Alan T. DeKok [Thu, 26 Sep 2019 16:16:06 +0000 (12:16 -0400)]
try to quiet analyzer
Alan T. DeKok [Thu, 26 Sep 2019 16:12:35 +0000 (12:12 -0400)]
use g->num_children instead of manually counting things
Alan T. DeKok [Thu, 26 Sep 2019 15:45:53 +0000 (11:45 -0400)]
always set presult
Alan T. DeKok [Thu, 26 Sep 2019 14:46:11 +0000 (10:46 -0400)]
set presult for non-debug builds, too
Alan T. DeKok [Thu, 26 Sep 2019 14:45:21 +0000 (10:45 -0400)]
note that load-balance etc. returns the return code of the child
Alan T. DeKok [Thu, 26 Sep 2019 14:39:16 +0000 (10:39 -0400)]
don't set frame->result, it's set by the interpreter
Alan T. DeKok [Thu, 26 Sep 2019 13:43:00 +0000 (09:43 -0400)]
quiet static analyzer
Alan T. DeKok [Thu, 26 Sep 2019 02:35:05 +0000 (22:35 -0400)]
try to quiet static analyzer
Arran Cudbard-Bell [Thu, 26 Sep 2019 15:37:54 +0000 (10:37 -0500)]
Rename and document some functions in map.c
Arran Cudbard-Bell [Thu, 26 Sep 2019 02:46:12 +0000 (21:46 -0500)]
Revert correct part
Arran Cudbard-Bell [Thu, 26 Sep 2019 02:45:04 +0000 (21:45 -0500)]
Partially Revert "Minor tweaks in the interpreter"
This reverts commit
60a53e88a54e31e98db6e8bbe40e10bd0608b21c .
Alan T. DeKok [Thu, 26 Sep 2019 01:49:16 +0000 (21:49 -0400)]
remove priority from unlang_op_call_t
it is now handled by the main interpret loop
Alan T. DeKok [Thu, 26 Sep 2019 01:16:22 +0000 (21:16 -0400)]
hoist priority setting to the main interpret loop
which not only minimizes code, but makes it so that the priority
isn't really needed.
Alan T. DeKok [Thu, 26 Sep 2019 01:05:25 +0000 (21:05 -0400)]
set an rcode if we return "calculate_result"
Otherwise how do we calculate results that don't have results?
unlang keywords shouldn't be returning RLM_MODULE_UNKNOWN.
Arran Cudbard-Bell [Thu, 26 Sep 2019 01:11:51 +0000 (20:11 -0500)]
Should use authparams as the ctx
Arran Cudbard-Bell [Wed, 25 Sep 2019 22:25:52 +0000 (17:25 -0500)]
Fix various binary safety issues in rlm_mschap
Arran Cudbard-Bell [Wed, 25 Sep 2019 21:50:45 +0000 (16:50 -0500)]
Fix talloc_pooled_object check and add talloc_zero_pooled_object
Arran Cudbard-Bell [Wed, 25 Sep 2019 20:13:14 +0000 (15:13 -0500)]
Minor tweaks in the interpreter
Alan T. DeKok [Wed, 25 Sep 2019 20:11:02 +0000 (16:11 -0400)]
make "update" use frame->process
Alan T. DeKok [Wed, 25 Sep 2019 19:16:29 +0000 (15:16 -0400)]
make "load-balance" and friends use frame->process
Alan T. DeKok [Wed, 25 Sep 2019 18:57:11 +0000 (14:57 -0400)]
make "foreach" use frame->process
Alan T. DeKok [Wed, 25 Sep 2019 18:55:34 +0000 (14:55 -0400)]
run the main "foreach" test before this one
Alan T. DeKok [Wed, 25 Sep 2019 18:36:58 +0000 (14:36 -0400)]
add process / signal to the unlang stack frame
so that the process functions can set / reset them as necessary
which then means that we can get rid of the repeat / resume
handling. And just use the "process" functions
Arran Cudbard-Bell [Wed, 25 Sep 2019 18:42:40 +0000 (13:42 -0500)]
Minor typo
Arran Cudbard-Bell [Wed, 25 Sep 2019 18:42:32 +0000 (13:42 -0500)]
Minor typos
Alan T. DeKok [Wed, 25 Sep 2019 17:16:18 +0000 (13:16 -0400)]
turn assert into run-time check. Fixes #3008
Arran Cudbard-Bell [Wed, 25 Sep 2019 14:40:37 +0000 (09:40 -0500)]
Fix the rcode lookup table to be in the right order
Alan T. DeKok [Wed, 25 Sep 2019 14:12:11 +0000 (10:12 -0400)]
cp->value can be NULL. Fixes #3008
Arran Cudbard-Bell [Wed, 25 Sep 2019 13:56:12 +0000 (08:56 -0500)]
Flesh out the return code tutorial
Arran Cudbard-Bell [Wed, 25 Sep 2019 13:56:00 +0000 (08:56 -0500)]
Fix some issues with the update_blocks_and_conditions tutorial
Alan T. DeKok [Wed, 25 Sep 2019 13:48:52 +0000 (09:48 -0400)]
allow encoding of ipv4prefix, and skip scope for ipv6
This makes the encoder match the decoder.
Arran Cudbard-Bell [Wed, 25 Sep 2019 10:57:28 +0000 (05:57 -0500)]
More documentation for return codes
Arran Cudbard-Bell [Wed, 25 Sep 2019 10:51:54 +0000 (05:51 -0500)]
s/userlock/disallow/g
A more generic "authorization failed" rcode.
Arran Cudbard-Bell [Wed, 25 Sep 2019 00:26:44 +0000 (19:26 -0500)]
The prefix types are internal, definitely not specific to RADIUS
Alan T. DeKok [Wed, 25 Sep 2019 00:57:07 +0000 (20:57 -0400)]
the cast will automatically truncate the variable
no need for a mask
Alan T. DeKok [Wed, 25 Sep 2019 00:45:04 +0000 (20:45 -0400)]
add a cast
Arran Cudbard-Bell [Tue, 24 Sep 2019 21:54:20 +0000 (16:54 -0500)]
Specify minimum buffer sizes
Alan T. DeKok [Tue, 24 Sep 2019 18:29:41 +0000 (14:29 -0400)]
add fr_get/put_16/32/64 and use then in from/to network code
TBH, we should probably get rid of htonl()+memcpy in the rest of
the code, as these macros seem better.
Alan T. DeKok [Tue, 24 Sep 2019 13:26:33 +0000 (09:26 -0400)]
add fr_unix_time to distinguish from fr_time_t
Alan T. DeKok [Tue, 24 Sep 2019 12:33:39 +0000 (08:33 -0400)]
remove use of time_t and time()
Arran Cudbard-Bell [Tue, 24 Sep 2019 11:00:01 +0000 (06:00 -0500)]
Various fixups
Alan T. DeKok [Tue, 24 Sep 2019 01:27:58 +0000 (21:27 -0400)]
fix date / time_delta handling
check for time_delta < 0 when converting to uint32/uint64
call time_to_Xsec() when converting dates to network, instead of
just dividing by 10^N
but time deltas can just be divided by 10^N
fix cast date to uint64 so that it casts based on precision,
and not always casts to seconds
Alan T. DeKok [Tue, 24 Sep 2019 00:33:56 +0000 (20:33 -0400)]
typo. I guess no one uses this...
Arran Cudbard-Bell [Mon, 23 Sep 2019 19:14:50 +0000 (14:14 -0500)]
Move exercises to Turtorials and add copyright notices
Arran Cudbard-Bell [Mon, 23 Sep 2019 18:51:11 +0000 (13:51 -0500)]
Add protocols doc dir
Arran Cudbard-Bell [Mon, 23 Sep 2019 18:48:52 +0000 (13:48 -0500)]
Remove the rlm dir
Alan T. DeKok [Mon, 23 Sep 2019 18:41:04 +0000 (14:41 -0400)]
tests for encode / decode of time_delta
Alan T. DeKok [Mon, 23 Sep 2019 18:39:18 +0000 (14:39 -0400)]
allow encoding time_delta in RADIUS
Alan T. DeKok [Mon, 23 Sep 2019 18:38:55 +0000 (14:38 -0400)]
time_delta is a numerical type, too
Arran Cudbard-Bell [Mon, 23 Sep 2019 18:14:35 +0000 (13:14 -0500)]
Quiet "implicit use of sequentially-consistent atomic may incur stronger memory barriers"
Alan T. DeKok [Mon, 23 Sep 2019 15:54:39 +0000 (11:54 -0400)]
allow encoding / decoding and casting of type 'time_delta'
Alan T. DeKok [Mon, 23 Sep 2019 15:44:24 +0000 (11:44 -0400)]
allow type 'time_delta' in dictionaries
with exactly the same syntax and variants as type 'date'
To be done: encode / decode time delta
Alan T. DeKok [Mon, 23 Sep 2019 12:48:15 +0000 (08:48 -0400)]
use rlm_rcode_t instead of fr_io_final_t
in preparation for moving app_process_t function signatures
to match module_method_t
perl -p -i -e 's/fr_io_final_t/rlm_rcode_t/g;s/FR_IO_YIELD/RLM_MODULE_YIELD/g;s/FR_IO_REPLY/RLM_MODULE_OK/g;s/FR_IO_FAIL/RLM_MODULE_FAIL/g;s/FR_IO_DONE/RLM_MODULE_OK/g' $(find src/ -name "*.[ch]" -print)
along with some minor / manual editing.
Arran Cudbard-Bell [Mon, 23 Sep 2019 03:26:00 +0000 (22:26 -0500)]
Ignore build products
Arran Cudbard-Bell [Mon, 23 Sep 2019 03:25:52 +0000 (22:25 -0500)]
Merge training exercises with documentation
Arran Cudbard-Bell [Sun, 22 Sep 2019 23:56:09 +0000 (18:56 -0500)]
More reorg
Arran Cudbard-Bell [Sun, 22 Sep 2019 23:38:23 +0000 (18:38 -0500)]
Reorganise howto files
Arran Cudbard-Bell [Sun, 22 Sep 2019 21:47:43 +0000 (16:47 -0500)]
Tweaks
Arran Cudbard-Bell [Sun, 22 Sep 2019 21:16:04 +0000 (16:16 -0500)]
Finish edits to the ldapsearch doc
Arran Cudbard-Bell [Sat, 21 Sep 2019 23:19:46 +0000 (18:19 -0500)]
Revert "Put cert override in session-state to make it easier to configure"
This reverts commit
9c50531336a8caa157a73da7c225743e3b436949 .
Arran Cudbard-Bell [Sat, 21 Sep 2019 23:17:59 +0000 (18:17 -0500)]
Put cert override in session-state to make it easier to configure
Arran Cudbard-Bell [Sat, 21 Sep 2019 16:23:15 +0000 (11:23 -0500)]
Typos
Arran Cudbard-Bell [Sat, 21 Sep 2019 16:20:06 +0000 (11:20 -0500)]
Almost done...
Arran Cudbard-Bell [Sat, 21 Sep 2019 16:19:51 +0000 (11:19 -0500)]
Add chain certificates to a trusted store. Apparently get issuer doesn't work with the untrusted args
Arran Cudbard-Bell [Sat, 21 Sep 2019 15:52:49 +0000 (10:52 -0500)]
Closing parens
Arran Cudbard-Bell [Sat, 21 Sep 2019 15:51:58 +0000 (10:51 -0500)]
Fix links. Github doesn't respect symlinks unfortunately
Arran Cudbard-Bell [Fri, 20 Sep 2019 20:06:04 +0000 (15:06 -0500)]
Goddamn OpenSSL...
Arran Cudbard-Bell [Fri, 20 Sep 2019 17:08:10 +0000 (12:08 -0500)]
unused things..
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:51:47 +0000 (11:51 -0500)]
More stupidity *sigh*
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:37:53 +0000 (11:37 -0500)]
Should return int
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:36:03 +0000 (11:36 -0500)]
Shut up stupid static analysis
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:30:23 +0000 (11:30 -0500)]
Doesn't do anything...
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:29:07 +0000 (11:29 -0500)]
Of course radwho is not threaded, so this was actually fine *sigh*
Arran Cudbard-Bell [Fri, 20 Sep 2019 16:23:18 +0000 (11:23 -0500)]
Redundant and apparently not used anywhere...
Arran Cudbard-Bell [Fri, 20 Sep 2019 14:55:53 +0000 (09:55 -0500)]
Add LGTM badge
Arran Cudbard-Bell [Fri, 20 Sep 2019 14:42:32 +0000 (09:42 -0500)]
Add header guards, fix typo in license boilerplate
Marcus Sundberg [Fri, 20 Sep 2019 10:14:55 +0000 (12:14 +0200)]
Sync dictionary.aptilo with latest official dictionary
Arran Cudbard-Bell [Thu, 19 Sep 2019 20:52:22 +0000 (15:52 -0500)]
More LDAP howto, WIP
Arran Cudbard-Bell [Thu, 19 Sep 2019 20:14:08 +0000 (15:14 -0500)]
Whitespace
Arran Cudbard-Bell [Thu, 19 Sep 2019 20:09:25 +0000 (15:09 -0500)]
Cleanup more ascend stuff
Arran Cudbard-Bell [Thu, 19 Sep 2019 19:56:35 +0000 (14:56 -0500)]
Doc update for LDAP howto
Arran Cudbard-Bell [Thu, 19 Sep 2019 17:38:54 +0000 (12:38 -0500)]
Always return instead of calling fr_cond_assert so stupid static analysers don't get confused
Arran Cudbard-Bell [Thu, 19 Sep 2019 17:26:44 +0000 (12:26 -0500)]
Fix some unsafe calls to snprintf, and change the way we handle truncation
Arran Cudbard-Bell [Thu, 19 Sep 2019 16:57:07 +0000 (11:57 -0500)]
Use localtime_r
Arran Cudbard-Bell [Thu, 19 Sep 2019 13:33:09 +0000 (08:33 -0500)]
Actually free the virtual server trees
Arran Cudbard-Bell [Thu, 19 Sep 2019 13:30:38 +0000 (08:30 -0500)]
Yes, technically if it's a normal build execution will continue *sigh*
Arran Cudbard-Bell [Thu, 19 Sep 2019 13:28:02 +0000 (08:28 -0500)]
Fix lgtm warnings in prepaid.py
Alan T. DeKok [Thu, 19 Sep 2019 01:08:00 +0000 (21:08 -0400)]
fsck it, we'll do it live
doing $(MAKE) ${BUILD_DIR}/coverage/index.html
hangs make. Which seems to be a bug in make. Or something stupid.
So just copy the danged build rules.
Alan T. DeKok [Wed, 18 Sep 2019 21:44:06 +0000 (17:44 -0400)]
add "clean.coverage" and try fixing dependencies
Matthew Newton [Wed, 18 Sep 2019 20:55:37 +0000 (20:55 +0000)]
crossbuild: debian9->10 in common image list
Matthew Newton [Wed, 18 Sep 2019 20:55:04 +0000 (20:55 +0000)]
crossbuild: update help
Arran Cudbard-Bell [Wed, 18 Sep 2019 19:45:42 +0000 (14:45 -0500)]
Refresh patch