]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Fri, 27 May 2016 00:58:19 +0000 (20:58 -0400)]
notes on use_tunneled_reply and copy_request_to_tunnel
Tho we still have to write those policies.
Alan T. DeKok [Fri, 27 May 2016 00:54:44 +0000 (20:54 -0400)]
Allow manual setting of MS-CHAP-Peer-Challenge
Alan T. DeKok [Fri, 27 May 2016 00:46:59 +0000 (20:46 -0400)]
rename challenge to auth_challenge
Arran Cudbard-Bell [Thu, 26 May 2016 22:09:06 +0000 (18:09 -0400)]
Remove support for use_tunneled_reply and copy_request_to_tunnel
It's time for them to die as their functionality can be emulated with unlang.
Arran Cudbard-Bell [Thu, 26 May 2016 22:07:50 +0000 (18:07 -0400)]
Only print deprecated substitutes if the offset field is not 0
Arran Cudbard-Bell [Thu, 26 May 2016 17:43:45 +0000 (13:43 -0400)]
Arran Cudbard-Bell [Thu, 26 May 2016 15:54:37 +0000 (11:54 -0400)]
Fix magic numbers
Alan T. DeKok [Thu, 26 May 2016 15:25:14 +0000 (11:25 -0400)]
Use old PRF functions if SSL_export_keying_material() fails.
Because https://tools.ietf.org/html/rfc5705#section-6
says that certain labels are prohibited, and OpenSSL enforces
that prohibition.
We need the keys for EAP, so we have to *always* implement the
functions. Because SSL_export_keying_material() might fail.
We'll need a later commit to implement the TLS 1.2 PRF.
Alan T. DeKok [Thu, 26 May 2016 15:02:34 +0000 (11:02 -0400)]
Expose prf_label in eap_tls_gen_challenge()
and rename it from eap_ttls_gen_challenge()
Alan T. DeKok [Wed, 25 May 2016 21:33:35 +0000 (17:33 -0400)]
Remove modcall.h, it was only 4 lines long
Alan T. DeKok [Wed, 25 May 2016 21:26:23 +0000 (17:26 -0400)]
rename modcall_compile_section() to unlang_compile()
and fix the arguments to it
Alan T. DeKok [Wed, 25 May 2016 21:19:06 +0000 (17:19 -0400)]
more functions use group_allocate
Alan T. DeKok [Wed, 25 May 2016 20:48:44 +0000 (16:48 -0400)]
Remove all knowledge of components from the interpreter.
Alan T. DeKok [Wed, 25 May 2016 20:37:47 +0000 (16:37 -0400)]
Don't set request->server_cs. it's done in modcall
Alan T. DeKok [Wed, 25 May 2016 20:38:31 +0000 (16:38 -0400)]
always set request->server_cs
Alan T. DeKok [Wed, 25 May 2016 20:37:28 +0000 (16:37 -0400)]
set foo->proxy->server, too. With notes for the future
Alan T. DeKok [Wed, 25 May 2016 16:17:33 +0000 (12:17 -0400)]
make unlang_interpret() take a default action
which is used if there's no CONF_SECTION or compile unlang
Alan T. DeKok [Wed, 25 May 2016 16:03:37 +0000 (12:03 -0400)]
Allow empty "map" sections sometimes
Alan T. DeKok [Tue, 24 May 2016 21:57:26 +0000 (17:57 -0400)]
Make EAP modules require a virtual_server configuration
Alan T. DeKok [Tue, 24 May 2016 21:57:14 +0000 (17:57 -0400)]
Set request->server_cs if not already set
Alan T. DeKok [Tue, 24 May 2016 21:07:50 +0000 (17:07 -0400)]
remove indexed_modcall and virtual_server_t
Which makes for fewer data structures, and allows for everything
to be keyed off of CONF_SECTIONs
It *might* be a bit slower, because we're now doing string lookups
on section names, instead of integer lookups. But that should be
lost in the noise of EAP and SQL
Alan T. DeKok [Tue, 24 May 2016 21:00:29 +0000 (17:00 -0400)]
Use CONF_SECTIONs directly in unlang_interpret
Alan T. DeKok [Tue, 24 May 2016 20:57:36 +0000 (16:57 -0400)]
set request->server_cs, too
Arran Cudbard-Bell [Tue, 24 May 2016 19:23:20 +0000 (15:23 -0400)]
Remember the original pointer
Arran Cudbard-Bell [Tue, 24 May 2016 18:50:10 +0000 (14:50 -0400)]
Fix issues in EAP sim
Alan T. DeKok [Fri, 20 May 2016 13:59:03 +0000 (09:59 -0400)]
move modcall() to unlang_interpret()
and various other minor cleanups
Arran Cudbard-Bell [Tue, 24 May 2016 12:47:59 +0000 (08:47 -0400)]
Use value_data_cast to translate between native JSON types and attribute types Closes #1525
Arran Cudbard-Bell [Tue, 24 May 2016 12:47:26 +0000 (08:47 -0400)]
copy if types are the same in value_data_cast
Arran Cudbard-Bell [Sun, 22 May 2016 14:28:38 +0000 (10:28 -0400)]
Intermediary pointers must be const qualified
Arran Cudbard-Bell [Sun, 22 May 2016 14:24:18 +0000 (10:24 -0400)]
Always use OpenSSL's free functions Closes #1610
Arran Cudbard-Bell [Sun, 22 May 2016 14:13:16 +0000 (10:13 -0400)]
Free basic response fire, maybe Fixes #1610
Arran Cudbard-Bell [Sun, 22 May 2016 14:11:06 +0000 (10:11 -0400)]
Formatting
Arran Cudbard-Bell [Sun, 22 May 2016 04:16:31 +0000 (00:16 -0400)]
Don't use fields
Arran Cudbard-Bell [Sun, 22 May 2016 03:23:21 +0000 (23:23 -0400)]
Print out autoconf.h for debugging
Arran Cudbard-Bell [Sun, 22 May 2016 02:25:25 +0000 (22:25 -0400)]
Document arguments
Arran Cudbard-Bell [Sun, 22 May 2016 02:16:24 +0000 (22:16 -0400)]
GCC on Travis doesn't support _Generic
Arran Cudbard-Bell [Sat, 21 May 2016 15:19:15 +0000 (11:19 -0400)]
Fix issue with the control list being (possibly) empty in EAP-SIM
Arran Cudbard-Bell [Sat, 21 May 2016 14:50:55 +0000 (10:50 -0400)]
Fix warning with rad_cond_assert
Arran Cudbard-Bell [Sat, 21 May 2016 13:27:36 +0000 (09:27 -0400)]
Fix line break
Arran Cudbard-Bell [Sat, 21 May 2016 13:01:33 +0000 (09:01 -0400)]
Fix build issue
Matthew Newton [Sat, 21 May 2016 04:41:57 +0000 (05:41 +0100)]
raddb: move driver-specific config out from mods-available
Symlinking them into mods-enabled, the location INCLUDEd in mods-available/sql,
breaks the server as it tries to load rlm_<driver>.so, which doesn't exist.
They're only part of a module config, so don't really fit that well in
mods-available.
Additionally, turn the $INCLUDE into an $-INCLUDE to pull the config in if it
exists, rather than having to be locally uncommented. This makes it easier to
configure sql for the first time (especially to get up and running with sqlite).
Arran Cudbard-Bell [Sat, 21 May 2016 04:28:42 +0000 (00:28 -0400)]
Use C11 _Generic for type checks
Mantas Mikulėnas [Fri, 20 May 2016 16:01:50 +0000 (19:01 +0300)]
Fix build error in checking whether staple is enabled
Make #1613 actually compile.
Philippe Wooding [Fri, 20 May 2016 09:42:32 +0000 (11:42 +0200)]
Only set stable cb is staple is enabled
Matthew Newton [Thu, 19 May 2016 10:27:30 +0000 (11:27 +0100)]
rlm_sql: don't add SQL-User-Name in xlat expansion
All attributes are now expanded before passing to the sql xlat function, so
it's too late to add the attribute to be of any use. Except for things after
this xlat, which is just confusing.
Alan T. DeKok [Wed, 18 May 2016 21:30:08 +0000 (17:30 -0400)]
parent compiled groups from CS
Alan T. DeKok [Wed, 18 May 2016 21:29:01 +0000 (17:29 -0400)]
remove modcall_compile() and modcall_append()
They are no longer needed
Alan T. DeKok [Wed, 18 May 2016 21:24:04 +0000 (17:24 -0400)]
Simplify load component section.
So it no longer needs to loop over the children
Alan T. DeKok [Wed, 18 May 2016 21:17:04 +0000 (17:17 -0400)]
move complaints about 'ignoring' to modcall.c
Alan T. DeKok [Wed, 18 May 2016 21:11:15 +0000 (17:11 -0400)]
skip precompile blocks
Alan T. DeKok [Wed, 18 May 2016 21:04:57 +0000 (17:04 -0400)]
move unlang cache to modcall_compile_section()
Alan T. DeKok [Wed, 18 May 2016 20:35:39 +0000 (16:35 -0400)]
set component in group_allocate
Alan T. DeKok [Wed, 18 May 2016 20:32:19 +0000 (16:32 -0400)]
call group_allocate() in more places
Alan T. DeKok [Wed, 18 May 2016 20:02:11 +0000 (16:02 -0400)]
be more liberal with [...] in attr definitions
and add tests for struct
Arran Cudbard-Bell [Wed, 18 May 2016 21:08:29 +0000 (17:08 -0400)]
Fix is_set for subsections
Arran Cudbard-Bell [Wed, 18 May 2016 21:08:10 +0000 (17:08 -0400)]
Don't need to recurse in cf_section_parse_init
Arran Cudbard-Bell [Tue, 17 May 2016 21:02:25 +0000 (17:02 -0400)]
Reset back to default Fall-Through if the group has no reply items
Arran Cudbard-Bell [Tue, 17 May 2016 20:41:22 +0000 (16:41 -0400)]
Show the keys as they're added
Alan T. DeKok [Tue, 17 May 2016 15:13:07 +0000 (11:13 -0400)]
rows can be zero. Fixes #1611
Alan T. DeKok [Tue, 17 May 2016 15:07:47 +0000 (11:07 -0400)]
always initialize filename on alloc
Arran Cudbard-Bell [Mon, 16 May 2016 20:01:13 +0000 (16:01 -0400)]
Initialise pointers
Alan T. DeKok [Mon, 16 May 2016 20:01:03 +0000 (16:01 -0400)]
Catch duplicates and error out
Alan T. DeKok [Mon, 16 May 2016 19:52:54 +0000 (15:52 -0400)]
Convert authenticate modules to sections.
So that we can cache compiled "unlang" blocks in the sections
Alan T. DeKok [Mon, 16 May 2016 16:33:09 +0000 (12:33 -0400)]
CID #
1361571
Check value pointed to, not pointer
Arran Cudbard-Bell [Sun, 15 May 2016 16:03:02 +0000 (12:03 -0400)]
Add base type macro to ensure we don't get tripped up by flags
Arran Cudbard-Bell [Sun, 15 May 2016 12:59:03 +0000 (08:59 -0400)]
Mask before checking whether it's a subsection
Arran Cudbard-Bell [Sun, 15 May 2016 12:51:28 +0000 (08:51 -0400)]
Type can't be signed
Arran Cudbard-Bell [Sun, 15 May 2016 12:43:44 +0000 (08:43 -0400)]
Evaluate _t first
Arran Cudbard-Bell [Sun, 15 May 2016 12:36:12 +0000 (08:36 -0400)]
Print the number of the unknown type
Arran Cudbard-Bell [Sun, 15 May 2016 06:01:01 +0000 (02:01 -0400)]
Must be NULL
Arran Cudbard-Bell [Sun, 15 May 2016 05:35:57 +0000 (01:35 -0400)]
Add is_set macros
Arran Cudbard-Bell [Sun, 15 May 2016 01:02:22 +0000 (21:02 -0400)]
Add support for writing out whether CONF_PAIRs were provided
Arran Cudbard-Bell [Fri, 13 May 2016 23:12:25 +0000 (19:12 -0400)]
Add support for deriving triplets from quintuplets
Arran Cudbard-Bell [Fri, 13 May 2016 23:11:09 +0000 (19:11 -0400)]
Fix stupid function names in EAP-SIM
Arran Cudbard-Bell [Fri, 13 May 2016 13:40:50 +0000 (09:40 -0400)]
Add EAP-AKA attributes
Don't get excited... No EAP-AKA yet.
Arran Cudbard-Bell [Fri, 13 May 2016 03:55:13 +0000 (23:55 -0400)]
... and again
Arran Cudbard-Bell [Fri, 13 May 2016 03:53:53 +0000 (23:53 -0400)]
Fix arguments to rlm_ldap_result in sasl.c
Arran Cudbard-Bell [Fri, 13 May 2016 03:41:04 +0000 (23:41 -0400)]
Fix radlog_hex
Arran Cudbard-Bell [Thu, 12 May 2016 19:37:22 +0000 (15:37 -0400)]
Add radlog_hex function
Alan T. DeKok [Wed, 11 May 2016 20:07:34 +0000 (16:07 -0400)]
make rlm_ldap_sasl_interactive() take / use timeout
Alan T. DeKok [Wed, 11 May 2016 20:06:00 +0000 (16:06 -0400)]
add timeout here, too
Alan T. DeKok [Wed, 11 May 2016 19:23:21 +0000 (15:23 -0400)]
move "goto error" into error block.
Alan T. DeKok [Wed, 11 May 2016 18:57:02 +0000 (14:57 -0400)]
Pass timeout to rlm_ldap_bind when creating a connection
Alan T. DeKok [Wed, 11 May 2016 18:54:56 +0000 (14:54 -0400)]
const'ify timeouts
and fix uninitialized memory issue with memset
Alan T. DeKok [Wed, 11 May 2016 18:49:20 +0000 (14:49 -0400)]
add timeout to rlm_ldap_bind
Arran Cudbard-Bell [Wed, 11 May 2016 18:01:21 +0000 (14:01 -0400)]
Add result timeout argument
Arran Cudbard-Bell [Wed, 11 May 2016 18:01:08 +0000 (14:01 -0400)]
Formatting
Arran Cudbard-Bell [Mon, 9 May 2016 19:48:36 +0000 (15:48 -0400)]
Merge pull request #1608 from alanbuxey/patch-5
fix week adjustment
Alan Buxey [Mon, 9 May 2016 19:44:10 +0000 (20:44 +0100)]
fix week adjustment
as proposed by dipixx
Arran Cudbard-Bell [Fri, 6 May 2016 13:24:18 +0000 (09:24 -0400)]
Merge pull request #1605 from gyx1000/fix_issue_1384
Remove deprecated configuration from default configuration #1384
Guillaume Pannatier [Fri, 6 May 2016 13:21:25 +0000 (15:21 +0200)]
Remove deprecated log configuration from default configuration file #1384
Alan T. DeKok [Fri, 6 May 2016 13:11:10 +0000 (09:11 -0400)]
Ignore "client", too. Fixes #1603
Matthew Newton [Fri, 6 May 2016 09:49:58 +0000 (10:49 +0100)]
Bounds check typo for max_spare_threads
Fixes #1602
Arran Cudbard-Bell [Fri, 6 May 2016 03:31:56 +0000 (04:31 +0100)]
Helps to actually include the header...
Arran Cudbard-Bell [Thu, 5 May 2016 17:22:33 +0000 (13:22 -0400)]
Fixup example certs
Alan T. DeKok [Thu, 5 May 2016 16:08:44 +0000 (12:08 -0400)]
Arran Cudbard-Bell [Thu, 5 May 2016 13:40:54 +0000 (09:40 -0400)]
Merge pull request #1600 from gyx1000/python_tests_more_clear
Use radiusd constants and comment rlm_python tests
Guillaume Pannatier [Thu, 5 May 2016 08:51:56 +0000 (10:51 +0200)]
Test python module instance configuration
Guillaume Pannatier [Thu, 5 May 2016 08:37:12 +0000 (10:37 +0200)]
Test cext_compat vs sub interpreters
Guillaume Pannatier [Thu, 5 May 2016 07:34:53 +0000 (09:34 +0200)]
Use radiusd constants for the returns and comment rlm_python tests