]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Alan T. DeKok [Mon, 4 Oct 2021 16:35:01 +0000 (12:35 -0400)]
update expected output
Alan T. DeKok [Mon, 4 Oct 2021 15:54:50 +0000 (11:54 -0400)]
reset to correct parent
Alan T. DeKok [Mon, 4 Oct 2021 15:54:31 +0000 (11:54 -0400)]
remove extra whitespace
Arran Cudbard-Bell [Sat, 2 Oct 2021 23:21:44 +0000 (18:21 -0500)]
Bump macOS and hostap versions
Arran Cudbard-Bell [Sat, 2 Oct 2021 22:04:37 +0000 (17:04 -0500)]
Fix eapol_test picking up the wrong OpenSSL version?
Arran Cudbard-Bell [Sat, 2 Oct 2021 18:53:19 +0000 (13:53 -0500)]
Build eapol_test against OpenSSL 1.1.0
Building against 3.0 produces many errors and the resulting eapol_test utility seems to fail in mschapv2. This is likely because it's not explicitly loading the "legacy" provider, and so md4 is not available.
Arran Cudbard-Bell [Sat, 2 Oct 2021 18:18:15 +0000 (13:18 -0500)]
Use OpenSSL 3.0 for macOS tests
Arran Cudbard-Bell [Sat, 2 Oct 2021 16:26:44 +0000 (11:26 -0500)]
Work around bug in OpenSSL 3.0 that causes handshakes to fail
Arran Cudbard-Bell [Sat, 2 Oct 2021 16:26:15 +0000 (11:26 -0500)]
Improve fr_tls_log_io_error
It now prints out the string identifier for the error, and produces debugging output for none fatal errors.
Arran Cudbard-Bell [Sat, 2 Oct 2021 00:37:51 +0000 (19:37 -0500)]
OpenSSL 3.0 PBKDF2 functions don't like zero iterations
Arran Cudbard-Bell [Fri, 1 Oct 2021 18:51:44 +0000 (13:51 -0500)]
Use newer EC_POINT_*affine_coordinates functions and EVP API in rlm_eap_pwd
Arran Cudbard-Bell [Fri, 1 Oct 2021 18:00:12 +0000 (13:00 -0500)]
Use EVP API for eap_fast_crypto.c
Arran Cudbard-Bell [Fri, 1 Oct 2021 17:35:03 +0000 (12:35 -0500)]
Use newer certificate functions to work with OpenSSL 3.0
Arran Cudbard-Bell [Fri, 1 Oct 2021 22:58:15 +0000 (17:58 -0500)]
Don't build engine.c if we're using OpenSSL 3.0
Arran Cudbard-Bell [Fri, 1 Oct 2021 17:34:13 +0000 (12:34 -0500)]
Update hmac_sha1 and hmac_md5 to use the EVP API
Calling the HMAC functions directly is deprecated in OpenSSL 3.0
Arran Cudbard-Bell [Fri, 1 Oct 2021 18:52:29 +0000 (13:52 -0500)]
Use EVP api in rlm_wimax
Arran Cudbard-Bell [Fri, 1 Oct 2021 17:18:30 +0000 (12:18 -0500)]
FIPS_mode() has been removed in OpenSSL 3.0
Load "legacy" provider for MD4
Arran Cudbard-Bell [Fri, 1 Oct 2021 17:18:02 +0000 (12:18 -0500)]
RSA_SSLV23_PADDING is not available in OpenSSL 3.0
We didn't do anything with it anyway...
Arran Cudbard-Bell [Fri, 1 Oct 2021 15:29:01 +0000 (10:29 -0500)]
*OpenSSL 1.1.*
Arran Cudbard-Bell [Fri, 1 Oct 2021 15:25:28 +0000 (10:25 -0500)]
Explicitly install openssl for CI
Arran Cudbard-Bell [Fri, 1 Oct 2021 14:52:37 +0000 (09:52 -0500)]
Various log fixes
Arran Cudbard-Bell [Fri, 1 Oct 2021 13:27:41 +0000 (08:27 -0500)]
Set macosx-version-min to be the current version
Arran Cudbard-Bell [Thu, 30 Sep 2021 23:49:16 +0000 (18:49 -0500)]
Better names for log destination table
Arran Cudbard-Bell [Thu, 30 Sep 2021 23:37:03 +0000 (18:37 -0500)]
Standardise on _not_ returning an errcode from logging functions
Arran Cudbard-Bell [Thu, 30 Sep 2021 19:51:11 +0000 (14:51 -0500)]
rename L_DST_EXTRA to L_DST_FUNC
Arran Cudbard-Bell [Thu, 30 Sep 2021 19:44:15 +0000 (14:44 -0500)]
New logging init functions
Arran Cudbard-Bell [Thu, 30 Sep 2021 19:36:44 +0000 (14:36 -0500)]
Simpler user/group parsing
Alan T. DeKok [Tue, 28 Sep 2021 14:49:06 +0000 (10:49 -0400)]
advance the correct buffer
Alan T. DeKok [Tue, 28 Sep 2021 14:22:52 +0000 (10:22 -0400)]
this should be const
Alan T. DeKok [Fri, 24 Sep 2021 18:37:26 +0000 (14:37 -0400)]
allow decoding part-way through a buffer
Alan T. DeKok [Fri, 24 Sep 2021 18:33:39 +0000 (14:33 -0400)]
tests are in GMT, so print that
Alan T. DeKok [Fri, 24 Sep 2021 15:52:38 +0000 (11:52 -0400)]
allow 'length=uint16' as a flag for STRUCT
because struct.c supports it, so why not...
Arran Cudbard-Bell [Fri, 24 Sep 2021 23:22:21 +0000 (18:22 -0500)]
typo
Nick Porter [Fri, 24 Sep 2021 20:43:43 +0000 (21:43 +0100)]
v4: Async LDAP connection fixes (#4240)
* Improve log message for successful LDAP async bind
* typo
* net_timeout is still used
* ldap_get_option can return LDAP_SUCCESS before the fd is known
* If _ldap_bind_io_write was called without an fd look it up now
* Store the connection fd for setting trunk events
* Signal the connection once the bind has succeeded
* Setting LDAP_OPT_NETWORK_TIMEOUT actually stops async calls
* log_prefix should be const
* Associate connection state handle with ldap connection
* Add error message if LDAP connection does not allocate
* Explain use of net_timeout
* Explain behavior of ldap_get_option
* Explain why we may call _ldap_bind_io_write() without a writable fd
Co-authored-by: Arran Cudbard-Bell <a.cudbardb@freeradius.org>
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:39:55 +0000 (15:39 -0500)]
Fix request timeout in rlm_sql_cassandra
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:27:29 +0000 (15:27 -0500)]
Init lhs/rhs in fr_value_box_print
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:27:15 +0000 (15:27 -0500)]
Fix overflow
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:21:59 +0000 (15:21 -0500)]
Fix another timeout arg
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:18:26 +0000 (15:18 -0500)]
Fix passing keepalive config items
Arran Cudbard-Bell [Fri, 24 Sep 2021 20:08:10 +0000 (15:08 -0500)]
Fix time delta issues in ldap related code
Arran Cudbard-Bell [Fri, 24 Sep 2021 19:34:48 +0000 (14:34 -0500)]
Fix watchdog time check
Arran Cudbard-Bell [Fri, 24 Sep 2021 19:31:23 +0000 (14:31 -0500)]
Set sd_watchdog_interval
Arran Cudbard-Bell [Fri, 24 Sep 2021 19:11:00 +0000 (14:11 -0500)]
Wrap fr_time_delta_t and fr_unix_time_t in structs
Fix invalid timeouts passed to redis and memcached
Alan T. DeKok [Fri, 24 Sep 2021 14:17:56 +0000 (10:17 -0400)]
update interpreter docs
Alan T. DeKok [Thu, 23 Sep 2021 18:15:48 +0000 (14:15 -0400)]
add EAP-FAST dictionary and make it run
it starts up, but it doesn't work, because the underlying
EAP handlers only do EAP-TLS for now.
We will need to fix this for v4, in order to:
* define and use src/lib/process/fast
* define a real dictionary (not an internal one)
* use the struct encoder where possible
* fix the encoder / decoder to use the newer API
Alan T. DeKok [Thu, 23 Sep 2021 17:45:13 +0000 (13:45 -0400)]
remove EAP-FAST aliases. they're really not needed
Alan T. DeKok [Thu, 23 Sep 2021 17:26:10 +0000 (13:26 -0400)]
renew at 60% of PAC lifetime
Arran Cudbard-Bell [Thu, 23 Sep 2021 18:13:59 +0000 (13:13 -0500)]
Fix overflow in rlm_cache
Arran Cudbard-Bell [Thu, 23 Sep 2021 17:32:30 +0000 (12:32 -0500)]
Quiet warning about pragma
Arran Cudbard-Bell [Thu, 23 Sep 2021 17:24:41 +0000 (12:24 -0500)]
Add `fr_time_max()`
Arran Cudbard-Bell [Thu, 23 Sep 2021 16:59:48 +0000 (11:59 -0500)]
Fix time addition in proto_ldap_sync
Arran Cudbard-Bell [Thu, 23 Sep 2021 16:54:40 +0000 (11:54 -0500)]
Wrap time when calling sd_watchdog_event manually
Arran Cudbard-Bell [Thu, 23 Sep 2021 16:35:06 +0000 (11:35 -0500)]
Switch fr_time_t to use a struct to detect misuses
Alan T. DeKok [Wed, 22 Sep 2021 19:04:28 +0000 (15:04 -0400)]
pac_type is uint16, not uint32
Alan T. DeKok [Wed, 22 Sep 2021 17:19:48 +0000 (13:19 -0400)]
fix data type here, too
Alan T. DeKok [Wed, 22 Sep 2021 17:15:55 +0000 (13:15 -0400)]
convert query timeout to time_delta
Alan T. DeKok [Wed, 22 Sep 2021 16:59:51 +0000 (12:59 -0400)]
use time_delta
Alan T. DeKok [Wed, 22 Sep 2021 16:54:06 +0000 (12:54 -0400)]
let's finish the conversion
Alan T. DeKok [Wed, 22 Sep 2021 16:46:57 +0000 (12:46 -0400)]
convert to using fr_time_t
Alan T. DeKok [Wed, 22 Sep 2021 16:16:40 +0000 (12:16 -0400)]
use time_delta instead of uint32
Alan T. DeKok [Wed, 22 Sep 2021 16:13:56 +0000 (12:13 -0400)]
it helps to honor these, too
Alan T. DeKok [Wed, 22 Sep 2021 16:09:59 +0000 (12:09 -0400)]
use fr_time_delta_t for lifetimes and timeouts
Alan T. DeKok [Wed, 22 Sep 2021 00:46:49 +0000 (20:46 -0400)]
normalize error messages. fixes #4239
and return on failure.
Alan T. DeKok [Tue, 21 Sep 2021 20:44:31 +0000 (16:44 -0400)]
updates and notes from Nokia
https://infocenter.nokia.com/public/7750SR150R5A/index.jsp?topic=%2Fcom.sr.radius%2Fhtml%2Fsros_radius_attrib.html
Alan T. DeKok [Tue, 21 Sep 2021 20:03:19 +0000 (16:03 -0400)]
let's do unlang thread initialization *after* loading virtual servers.
Alan T. DeKok [Tue, 21 Sep 2021 19:13:58 +0000 (15:13 -0400)]
comment out perf stuff until we find out why it crashes
Alan T. DeKok [Tue, 21 Sep 2021 17:24:40 +0000 (13:24 -0400)]
fix format options
Alan T. DeKok [Tue, 21 Sep 2021 16:53:56 +0000 (12:53 -0400)]
print out stats in a hokey format
we can always change it later
Alan T. DeKok [Tue, 21 Sep 2021 14:53:57 +0000 (10:53 -0400)]
start numbering at 1
some "internal" instructions are build dynamically, as with
unlang_module_push(). The instruction number there will be
zero, as there's currently no way to track these dynamic instructions.
As a result, we want to reserve array entry 0 as the canonical
"dynamic" instruction number, which won't get their stats tracked.
Arran Cudbard-Bell [Tue, 21 Sep 2021 14:26:32 +0000 (09:26 -0500)]
Don't double advance
Alan T. DeKok [Tue, 21 Sep 2021 13:06:43 +0000 (09:06 -0400)]
add per-thread instantiation for unlang
and glue it into the various binaries
Arran Cudbard-Bell [Tue, 21 Sep 2021 13:02:32 +0000 (08:02 -0500)]
Unwrap fr_value_box_from_network
Arran Cudbard-Bell [Mon, 20 Sep 2021 23:09:41 +0000 (18:09 -0500)]
Fix redis ippool tests
Arran Cudbard-Bell [Mon, 20 Sep 2021 23:02:26 +0000 (18:02 -0500)]
Allowing copying NULL boxes
Arran Cudbard-Bell [Mon, 20 Sep 2021 23:02:13 +0000 (18:02 -0500)]
Remove debugging
Arran Cudbard-Bell [Mon, 20 Sep 2021 23:02:02 +0000 (18:02 -0500)]
redis: If we don't have an explicit dst type, then write directly to out
Arran Cudbard-Bell [Mon, 20 Sep 2021 22:27:19 +0000 (17:27 -0500)]
Need to clean up the tmp pool on the normal path too
Arran Cudbard-Bell [Mon, 20 Sep 2021 22:10:48 +0000 (17:10 -0500)]
...and remove the assert
Arran Cudbard-Bell [Mon, 20 Sep 2021 22:04:24 +0000 (17:04 -0500)]
Allow printing/parsing NULL
Arran Cudbard-Bell [Mon, 20 Sep 2021 21:42:58 +0000 (16:42 -0500)]
Enable redis tests again
Arran Cudbard-Bell [Mon, 20 Sep 2021 21:39:22 +0000 (16:39 -0500)]
Slave is already dead, kill the master
Arran Cudbard-Bell [Mon, 20 Sep 2021 21:27:23 +0000 (16:27 -0500)]
Use common code for casting groups
Arran Cudbard-Bell [Mon, 20 Sep 2021 21:27:07 +0000 (16:27 -0500)]
Use common code for printing groups
Arran Cudbard-Bell [Mon, 20 Sep 2021 21:24:59 +0000 (16:24 -0500)]
Rework list concat functions
Alan T. DeKok [Mon, 20 Sep 2021 15:49:02 +0000 (11:49 -0400)]
copy VALUEs too
Alan T. DeKok [Mon, 20 Sep 2021 15:43:21 +0000 (11:43 -0400)]
add enums, too
Alan T. DeKok [Mon, 20 Sep 2021 12:59:58 +0000 (08:59 -0400)]
add code and test cases for copying enums
even between attributes of different data types
Alan T. DeKok [Sun, 19 Sep 2021 13:40:56 +0000 (09:40 -0400)]
give better errors early, if we can
Alan T. DeKok [Sun, 19 Sep 2021 13:40:04 +0000 (09:40 -0400)]
ENUMs have values, but no children
and add tests
Nick Porter [Mon, 20 Sep 2021 12:58:44 +0000 (13:58 +0100)]
v4: Add a status to trunks and watchers on state changes (#4224)
* Define states of a trunk
* Add lookup table for names of trunk states
* Update trunk state on each call of the management timer
* Define structure and function type for trunk watchers
* Add lists of watchers to trunks
* Add functions to add / remove trunk state change watchers
* Call trunk state change watchers
Arran Cudbard-Bell [Sun, 19 Sep 2021 15:27:28 +0000 (10:27 -0500)]
FR_TYPE_VOID is better than FR_TYPE_NULL for indicating we want to use the type from Redis
Arran Cudbard-Bell [Sun, 19 Sep 2021 15:26:54 +0000 (10:26 -0500)]
environ isn't defined on FreeBSD either
Matthew Newton [Sat, 18 Sep 2021 17:00:43 +0000 (18:00 +0100)]
json: get full attribute name
Alan T. DeKok [Sat, 18 Sep 2021 19:43:14 +0000 (15:43 -0400)]
no need to check for 'ref==NULL'
Alan T. DeKok [Sat, 18 Sep 2021 18:22:59 +0000 (14:22 -0400)]
shut up static analyzer
Alan T. DeKok [Sat, 18 Sep 2021 13:56:44 +0000 (09:56 -0400)]
add and document ENUM
we'll add more complex tests later
Alan T. DeKok [Sat, 18 Sep 2021 13:50:09 +0000 (09:50 -0400)]
move self-allocated numbers to be dict-specific
which has fewer opportunities for collisions
Alan T. DeKok [Sat, 18 Sep 2021 13:46:11 +0000 (09:46 -0400)]
0 and 1 are valid boolean values.
Alan T. DeKok [Sat, 18 Sep 2021 13:19:06 +0000 (09:19 -0400)]
use local boolean.
by the time we get to checking if attr<0, it's already been updated
and allocated.
Alan T. DeKok [Fri, 17 Sep 2021 18:44:10 +0000 (14:44 -0400)]
s/fr_dict_enum_t/fr_dict_enum_value_t/