]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Sun, 20 Jan 2019 09:59:03 +0000 (16:59 +0700)]
Use SSL_is_init_finished instead of checking the last handshake_type
This fixes some compatibility issues with TLS 1.3. We still disagree on MPPE keys with eapol_test though...
Arran Cudbard-Bell [Sun, 20 Jan 2019 09:58:10 +0000 (16:58 +0700)]
Use other, less weird macros for testing content_type
Arran Cudbard-Bell [Sun, 20 Jan 2019 09:56:56 +0000 (16:56 +0700)]
Make the TLS alert record generate code less crappy...
Alert fixup
Arran Cudbard-Bell [Sun, 20 Jan 2019 08:06:44 +0000 (15:06 +0700)]
Minor debugging and formatting
Arran Cudbard-Bell [Sun, 20 Jan 2019 07:25:33 +0000 (14:25 +0700)]
Allow eapol_test to build where OpenSSL functions are deprecated
Arran Cudbard-Bell [Sun, 20 Jan 2019 07:21:41 +0000 (14:21 +0700)]
Allow eapol_test build dir to be overriden in case we want to use the same copy of the src multiple times
Arran Cudbard-Bell [Wed, 16 Jan 2019 08:23:34 +0000 (15:23 +0700)]
Use the British spelling
Arran Cudbard-Bell [Wed, 16 Jan 2019 08:19:23 +0000 (15:19 +0700)]
Typo
Arran Cudbard-Bell [Tue, 15 Jan 2019 09:42:47 +0000 (16:42 +0700)]
Unclear if/how the TLS1_RT_CRYPTO* macros are actually used
Arran Cudbard-Bell [Tue, 15 Jan 2019 09:40:05 +0000 (16:40 +0700)]
Check for more handhsake macros
Arran Cudbard-Bell [Tue, 15 Jan 2019 09:08:06 +0000 (16:08 +0700)]
Too soon!
Arran Cudbard-Bell [Tue, 15 Jan 2019 09:02:07 +0000 (16:02 +0700)]
Move OpenSSL macro -> string conversions to arrays
Much faster, and fewer LoC.
Arran Cudbard-Bell [Tue, 15 Jan 2019 09:01:29 +0000 (16:01 +0700)]
Same thing for unit_test_module
Arran Cudbard-Bell [Tue, 15 Jan 2019 02:59:05 +0000 (09:59 +0700)]
Only load TLS attributes if we have OpenSSL
Arran Cudbard-Bell [Tue, 15 Jan 2019 02:58:26 +0000 (09:58 +0700)]
Print detailed identity at debug level 3
Arran Cudbard-Bell [Tue, 15 Jan 2019 02:30:36 +0000 (09:30 +0700)]
Fix relatively severe memory leak in EAP-PWD
Arran Cudbard-Bell [Tue, 15 Jan 2019 02:30:15 +0000 (09:30 +0700)]
Swap out memory allocation failure checks for MEM() macro in EAP-PWD
Arran Cudbard-Bell [Tue, 15 Jan 2019 02:29:47 +0000 (09:29 +0700)]
Whitespace
Arran Cudbard-Bell [Tue, 15 Jan 2019 01:57:11 +0000 (08:57 +0700)]
Missed declaration
Arran Cudbard-Bell [Mon, 14 Jan 2019 18:18:10 +0000 (01:18 +0700)]
Call tls_init() earlier
...and error out if we can't set our own memory allocation functions. It's now critical.
Arran Cudbard-Bell [Mon, 14 Jan 2019 16:31:12 +0000 (23:31 +0700)]
Fix minor warning with OpenSSL 1.1.1
Arran Cudbard-Bell [Sun, 13 Jan 2019 06:59:34 +0000 (14:59 +0800)]
Packet-Authentication-Vector, Packet-Type and Response-Packet-Type are RADIUS specific
They should probably be added by the radius proto modules instead of being virtual attributes
Arran Cudbard-Bell [Sat, 12 Jan 2019 11:02:42 +0000 (19:02 +0800)]
Bump copright - Woo 20th Aniversary
Alan T. DeKok [Fri, 11 Jan 2019 18:59:02 +0000 (13:59 -0500)]
remove (mostly) duplicate configs.
It's better to define `skip_locked = "SKIP LOCKED"` once,
and then use it where necessary. Otherwise, we get near-infinite
variations of the same queries
Alejandro Perez [Thu, 10 Jan 2019 14:48:02 +0000 (15:48 +0100)]
Add libssl1.1 as patched for Heartbleed
Alan T. DeKok [Wed, 9 Jan 2019 21:47:49 +0000 (16:47 -0500)]
don't retransmit packet if we're not synchronous
Alan T. DeKok [Wed, 9 Jan 2019 21:40:32 +0000 (16:40 -0500)]
refactor to use common code
Arran Cudbard-Bell [Fri, 11 Jan 2019 11:31:46 +0000 (19:31 +0800)]
Make it easier to see what's going on
Arran Cudbard-Bell [Fri, 11 Jan 2019 11:01:49 +0000 (19:01 +0800)]
Document the negotiation parameters better
Document lists
Nathan Ward [Fri, 11 Jan 2019 03:23:44 +0000 (16:23 +1300)]
Add a note that ORDER BY RANDOM() is very slow.
This first gets all the free IPs, then randomises them, then selects the first one.
Nathan Ward [Fri, 11 Jan 2019 03:14:13 +0000 (16:14 +1300)]
Comment out sqlippool PostgreSQL allocate_find query, so only one allocate_find query is uncommented, mirroring MySQL and Oracle queries configs.
# Conflicts:
# raddb/mods-config/sql/ippool/postgresql/queries.conf
Nathan Ward [Fri, 11 Jan 2019 03:13:35 +0000 (16:13 +1300)]
Add SKIP LOCKED variations of sqlippool allocate_find queries as options.
# Conflicts:
# raddb/mods-config/sql/ippool/mysql/queries.conf
Arran Cudbard-Bell [Fri, 11 Jan 2019 00:23:31 +0000 (08:23 +0800)]
Didn't spot the loop condition...
Arran Cudbard-Bell [Fri, 11 Jan 2019 00:18:17 +0000 (08:18 +0800)]
Loop condition ensures REDIS_RCODE_TRY_AGAIN is true, but acked must be initialised
Arran Cudbard-Bell [Thu, 10 Jan 2019 09:47:19 +0000 (17:47 +0800)]
Missed extern
Arran Cudbard-Bell [Thu, 10 Jan 2019 01:05:04 +0000 (09:05 +0800)]
Revert "remove redundant check"
Not sure why CCP check flagged this, it's definitely not a redundant check.
Arran Cudbard-Bell [Thu, 10 Jan 2019 00:59:53 +0000 (08:59 +0800)]
Partially Revert "remove redundancies found by cppcheck"
We do need to check if out is NULL in the autoload struct.
Alan T. DeKok [Wed, 9 Jan 2019 17:31:04 +0000 (12:31 -0500)]
fix return code
Alan T. DeKok [Wed, 9 Jan 2019 17:29:46 +0000 (12:29 -0500)]
remove redundant check
Alan T. DeKok [Wed, 9 Jan 2019 17:28:24 +0000 (12:28 -0500)]
remove unused struct member
Alan T. DeKok [Wed, 9 Jan 2019 17:23:54 +0000 (12:23 -0500)]
remove redundant check
Alan T. DeKok [Wed, 9 Jan 2019 17:21:41 +0000 (12:21 -0500)]
format sanity checks
Matthew Newton [Wed, 9 Jan 2019 19:57:59 +0000 (19:57 +0000)]
Clean up OpenSSL manually
Not recommended by OpenSSL, but this is to stop segfaults in the
atexit handler because the memory has already been freed with the
main config.
Alan T. DeKok [Wed, 9 Jan 2019 15:44:29 +0000 (10:44 -0500)]
use "end" to check end conditions
Alan T. DeKok [Wed, 9 Jan 2019 15:44:03 +0000 (10:44 -0500)]
remove redundant check
the "for" condition ensures that the test is always true
Alan T. DeKok [Wed, 9 Jan 2019 15:37:55 +0000 (10:37 -0500)]
parameter is NULL, not 0
Alan T. DeKok [Wed, 9 Jan 2019 15:37:33 +0000 (10:37 -0500)]
fix check in assertion, as 'c' may be NULL
Alan T. DeKok [Wed, 9 Jan 2019 15:33:43 +0000 (10:33 -0500)]
remove redundant checks
Alan T. DeKok [Wed, 9 Jan 2019 15:32:56 +0000 (10:32 -0500)]
format type checks
Alan T. DeKok [Wed, 9 Jan 2019 15:29:39 +0000 (10:29 -0500)]
64-bit signed numbers can only be shifted left by 62 bits
Alan T. DeKok [Wed, 9 Jan 2019 15:27:39 +0000 (10:27 -0500)]
data is always unsigned
Alan T. DeKok [Wed, 9 Jan 2019 15:27:02 +0000 (10:27 -0500)]
argument should be a pointer, not a number
Alan T. DeKok [Wed, 9 Jan 2019 15:24:26 +0000 (10:24 -0500)]
remove unnecessary function parameter
Alan T. DeKok [Wed, 9 Jan 2019 15:23:03 +0000 (10:23 -0500)]
remove unused code
Alan T. DeKok [Wed, 9 Jan 2019 15:21:41 +0000 (10:21 -0500)]
use the correct attribute for checks
Alan T. DeKok [Wed, 9 Jan 2019 15:11:59 +0000 (10:11 -0500)]
format type checks from cppcheck
Alan T. DeKok [Wed, 9 Jan 2019 15:11:22 +0000 (10:11 -0500)]
clarify precedence from cppcheck
Alan T. DeKok [Wed, 9 Jan 2019 15:08:46 +0000 (10:08 -0500)]
format type fixes from cppcheck
Alan T. DeKok [Wed, 9 Jan 2019 15:02:19 +0000 (10:02 -0500)]
data is unsigned, so it can't be less than zero
Alan T. DeKok [Wed, 9 Jan 2019 15:01:47 +0000 (10:01 -0500)]
remove redundant check
Alan T. DeKok [Tue, 8 Jan 2019 22:56:03 +0000 (17:56 -0500)]
fix comparison checks found by cppcheck
Alan T. DeKok [Tue, 8 Jan 2019 22:47:02 +0000 (17:47 -0500)]
remove redundancies found by cppcheck
Arran Cudbard-Bell [Wed, 9 Jan 2019 02:42:11 +0000 (10:42 +0800)]
Fix PID removal in tests that start RADIUSD
Arran Cudbard-Bell [Wed, 9 Jan 2019 02:41:36 +0000 (10:41 +0800)]
Don't make assumptions about the presence of the RADIUS state attribute
This doesn't work for nested EAP methods. As request data is handled outside of the EAP module, we just need to check if an eap_session_t is present in the request, and use that as an indication whether we're starting a new session or continuing a previoue one.
Matthew Newton [Tue, 8 Jan 2019 19:41:46 +0000 (19:41 +0000)]
Debian packaging statoverrides new cert location
Arran Cudbard-Bell [Tue, 8 Jan 2019 10:23:02 +0000 (18:23 +0800)]
If the EAP-Identity is > 253 bytes, then disable the User-Name EAP-Identity comparison
Arran Cudbard-Bell [Tue, 8 Jan 2019 10:22:14 +0000 (18:22 +0800)]
Prevent crash in EAP-Fast if CHAP-Challenge is not the length we expect
Arran Cudbard-Bell [Tue, 8 Jan 2019 10:18:51 +0000 (18:18 +0800)]
Qualify RADIUS macro names
fff
Arran Cudbard-Bell [Tue, 8 Jan 2019 08:19:52 +0000 (16:19 +0800)]
Fix EAP-Session initialisation
The presence of the state attribute in the request is what dictates whether we're starting a new session or not. *NOT* the fact that the initial packet is an EAP-Identity response.
Arran Cudbard-Bell [Tue, 8 Jan 2019 07:56:31 +0000 (15:56 +0800)]
Remove constraints on EAP-Identity
Identities containing \0 bytes are allowed. Zero length identities are explicitly supported by RFC 3748.
Arran Cudbard-Bell [Tue, 8 Jan 2019 07:49:47 +0000 (15:49 +0800)]
Note on zero length memcpy
Arran Cudbard-Bell [Tue, 8 Jan 2019 02:12:25 +0000 (10:12 +0800)]
Use md4/md5 calc functions where possibe
Arran Cudbard-Bell [Tue, 8 Jan 2019 02:53:15 +0000 (10:53 +0800)]
Allow local md5 functions to be selected at runtime depending on OpenSSL's FIPS setting
Arran Cudbard-Bell [Mon, 7 Jan 2019 11:59:47 +0000 (19:59 +0800)]
Allow local md4 functions to be selected at runtime depending on OpenSSL's FIPS setting
Arran Cudbard-Bell [Mon, 7 Jan 2019 11:46:27 +0000 (19:46 +0800)]
Include evp header in EAP base
Arran Cudbard-Bell [Mon, 7 Jan 2019 11:35:51 +0000 (19:35 +0800)]
libldap is deprecated in recent versions of macOS
Arran Cudbard-Bell [Mon, 7 Jan 2019 11:23:32 +0000 (19:23 +0800)]
Formatting in HMAC functions
Arran Cudbard-Bell [Mon, 7 Jan 2019 14:21:09 +0000 (22:21 +0800)]
Add toggle for OpenSSL FIPS mode
allow vulnerable
Arran Cudbard-Bell [Mon, 7 Jan 2019 01:34:42 +0000 (09:34 +0800)]
More static array bounds
Arran Cudbard-Bell [Sun, 6 Jan 2019 08:31:27 +0000 (16:31 +0800)]
More bounds checks
Arran Cudbard-Bell [Sun, 6 Jan 2019 04:29:27 +0000 (12:29 +0800)]
Missed semicolon
Arran Cudbard-Bell [Sun, 6 Jan 2019 02:48:20 +0000 (10:48 +0800)]
Remove CC_BOUNDED and use static array subscript for bounds enforcement
Should work with any compiler supporting C99.
Arran Cudbard-Bell [Sat, 5 Jan 2019 08:21:43 +0000 (16:21 +0800)]
Limit scope of buffers
Arran Cudbard-Bell [Sat, 5 Jan 2019 08:21:31 +0000 (16:21 +0800)]
Stupid compiler
Matthew Newton [Fri, 4 Jan 2019 22:06:13 +0000 (22:06 +0000)]
Fix up the debian "freeradius" name patch... again.
Arran Cudbard-Bell [Fri, 4 Jan 2019 14:46:52 +0000 (22:46 +0800)]
Correct format specifier
Arran Cudbard-Bell [Fri, 4 Jan 2019 14:01:31 +0000 (22:01 +0800)]
Catch bad secret values
fff
Arran Cudbard-Bell [Fri, 4 Jan 2019 09:00:56 +0000 (17:00 +0800)]
Disable foreign attribute check
Arran Cudbard-Bell [Fri, 4 Jan 2019 07:30:09 +0000 (15:30 +0800)]
Check rad_debug_lvl correctly, so the messages are enabled/disabled based on request specific debug levels
Reduce manual calls to hex printing functions
Arran Cudbard-Bell [Fri, 4 Jan 2019 07:07:19 +0000 (15:07 +0800)]
Code should always execute. WARNings are always emitted, and code applies fixup to VPs that shouldn't just be applied if debugging is enabled.
Alan T. DeKok [Mon, 31 Dec 2018 15:57:06 +0000 (10:57 -0500)]
clarify message
Alan T. DeKok [Mon, 31 Dec 2018 16:14:21 +0000 (11:14 -0500)]
remove unused assignment
Alan T. DeKok [Thu, 20 Dec 2018 16:30:30 +0000 (11:30 -0500)]
word smithing
Arran Cudbard-Bell [Wed, 26 Dec 2018 02:07:19 +0000 (10:07 +0800)]
Use newer ephemeral struct init syntax
Arran Cudbard-Bell [Wed, 26 Dec 2018 01:53:15 +0000 (09:53 +0800)]
Typo
Arran Cudbard-Bell [Wed, 26 Dec 2018 01:39:25 +0000 (09:39 +0800)]
Cleanup minor issues with attribute search functions
Arran Cudbard-Bell [Thu, 20 Dec 2018 00:53:32 +0000 (08:53 +0800)]
Probably not a useful test anymore
Arran Cudbard-Bell [Thu, 20 Dec 2018 00:46:06 +0000 (08:46 +0800)]
Make it easier to see the ref count in a debugger
Alan T. DeKok [Thu, 20 Dec 2018 12:44:41 +0000 (07:44 -0500)]
doxygen
Alan T. DeKok [Thu, 20 Dec 2018 12:04:43 +0000 (07:04 -0500)]
don't increase refcount on first autoload
because the manual load of the dictionaries doesn't result in
a manual unload. so the autoload has to auto-unload it for us