]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Matthew Newton [Fri, 20 Feb 2026 19:50:19 +0000 (19:50 +0000)]
ci: remove unmaintained action
addnab/docker-run-action is broken and hasn't been updated in five years
Ethan Thompson [Fri, 20 Feb 2026 22:35:46 +0000 (17:35 -0500)]
Feat: Allow octets boxes to be encoded as base16 (hex) in JSON blobs (#5750)
Arran Cudbard-Bell [Fri, 20 Feb 2026 05:31:56 +0000 (22:31 -0700)]
Allow octets boxes to be encoded as base64 in JSON blobs
Arran Cudbard-Bell [Sat, 14 Feb 2026 01:37:32 +0000 (17:37 -0800)]
Similar fixups in radclient-ng
Arran Cudbard-Bell [Sat, 14 Feb 2026 00:40:39 +0000 (16:40 -0800)]
Minor cleanups in radclient
Nick Porter [Fri, 20 Feb 2026 09:59:56 +0000 (09:59 +0000)]
Annotate return overflow (CID #
1680684 )
Coverity doesn't have the context of how negative slen is built to
understand that these will not overflow.
Nick Porter [Fri, 20 Feb 2026 09:27:54 +0000 (09:27 +0000)]
Pacify Coverity (CID #
1520867 , #
1520875 )
It doesn't identify that paths which set request->reply->code only set
to valid packet codes.
Nick Porter [Fri, 20 Feb 2026 08:54:58 +0000 (08:54 +0000)]
Re-sequence code to check before dereference (CID #
1681438 )
Nick Porter [Thu, 19 Feb 2026 19:29:26 +0000 (19:29 +0000)]
Current FreeBSD 14 is 14.3
Nick Porter [Fri, 20 Feb 2026 08:11:41 +0000 (08:11 +0000)]
Update cross platform action version
Nick Porter [Fri, 20 Feb 2026 08:32:50 +0000 (08:32 +0000)]
Pass pipe fd array to pipe() not a pointer to the array
Nick Porter [Thu, 19 Feb 2026 19:25:07 +0000 (19:25 +0000)]
Tidy control_test
github-actions[bot] [Fri, 20 Feb 2026 05:00:16 +0000 (05:00 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/der.tar
github-actions[bot] [Fri, 20 Feb 2026 04:59:40 +0000 (04:59 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/cbor.tar
Alan T. DeKok [Thu, 19 Feb 2026 20:04:31 +0000 (15:04 -0500)]
various bug fixes in master.c
Alan T. DeKok [Thu, 19 Feb 2026 19:51:10 +0000 (14:51 -0500)]
don't use smart quotes, etc.
98manu [Thu, 19 Feb 2026 14:29:13 +0000 (19:59 +0530)]
update EAP-MD5 tutorial
doc:add new files to the eap-md5 tutorial
doc:add overview section to the eap-md5 tutorial
doc:add diagram to the eap-md5 tutorial
doc:add configuration steps to the eap-md5 tutorial
doc:add expected output section to the eap-md5 tutorial
doc:add eapol-test to the eap-md5 tutorial
Alan T. DeKok [Thu, 19 Feb 2026 19:31:58 +0000 (14:31 -0500)]
use fr_cloexec()
Alan T. DeKok [Thu, 19 Feb 2026 16:52:04 +0000 (11:52 -0500)]
various bug fixes in network.c
Alan T. DeKok [Thu, 19 Feb 2026 19:31:10 +0000 (14:31 -0500)]
better handle pipe full
Alan T. DeKok [Thu, 19 Feb 2026 13:22:41 +0000 (08:22 -0500)]
only free message once we're done with it
we should really delay the decode until after the various other
checks have been done.
Alan T. DeKok [Thu, 19 Feb 2026 13:19:49 +0000 (08:19 -0500)]
remove unused variables
Alan T. DeKok [Thu, 19 Feb 2026 13:18:28 +0000 (08:18 -0500)]
check return code of set timeout
Alan T. DeKok [Thu, 19 Feb 2026 04:21:52 +0000 (23:21 -0500)]
fix various bugs in worker.c
use MEM() rather than assert
release request on failure
unlink request from listen entry when channel is closed
only call time once.
stop / clean up all requests.
don't divide by zero in corner cases
use correct channel message
Alan T. DeKok [Thu, 19 Feb 2026 04:05:01 +0000 (23:05 -0500)]
use size_t rather than uint32_t to over overflow
Alan T. DeKok [Thu, 19 Feb 2026 04:02:30 +0000 (23:02 -0500)]
tweak control message send
arguably we should use a method which is better at high performance
Alan T. DeKok [Thu, 19 Feb 2026 03:47:11 +0000 (22:47 -0500)]
use util functions for nonblock
Alan T. DeKok [Thu, 19 Feb 2026 03:44:09 +0000 (22:44 -0500)]
set FD_CLOEXEC properly
for F_GETFD / F_SETFD, and not F_GETFL / F_SETFL
Alan T. DeKok [Thu, 19 Feb 2026 03:40:33 +0000 (22:40 -0500)]
typo
Alan T. DeKok [Thu, 19 Feb 2026 03:39:59 +0000 (22:39 -0500)]
we can use talloc_set_name_const
Alan T. DeKok [Thu, 19 Feb 2026 03:39:17 +0000 (22:39 -0500)]
destroy semaphore on failure
Alan T. DeKok [Thu, 19 Feb 2026 03:34:40 +0000 (22:34 -0500)]
export thread pool configuration to scheduler
so that it can find worker-specific configuration
Alan T. DeKok [Thu, 19 Feb 2026 03:27:36 +0000 (22:27 -0500)]
various bug fixes in network.c
fix assertion
fix arguments to memmove(), and NULL out now-unused entry.
return negative for error
Alan T. DeKok [Wed, 18 Feb 2026 20:35:22 +0000 (15:35 -0500)]
various bug fixes in dedup.c
Alan T. DeKok [Wed, 18 Feb 2026 20:23:59 +0000 (15:23 -0500)]
various bug fixes in packet.c
Alan T. DeKok [Wed, 18 Feb 2026 20:20:55 +0000 (15:20 -0500)]
various bug fixes in retry.c
Alan T. DeKok [Wed, 18 Feb 2026 20:13:31 +0000 (15:13 -0500)]
various bug fixes in haproxy.c
Alan T. DeKok [Wed, 18 Feb 2026 19:44:42 +0000 (14:44 -0500)]
various bug fixes in queue.c
Alan T. DeKok [Wed, 18 Feb 2026 19:41:50 +0000 (14:41 -0500)]
various bug fixes in fd.c
Alan T. DeKok [Wed, 18 Feb 2026 18:56:44 +0000 (13:56 -0500)]
keep blocked flag if we can't flush when already blocked
Alan T. DeKok [Wed, 18 Feb 2026 18:55:08 +0000 (13:55 -0500)]
stop on error, too
Alan T. DeKok [Wed, 18 Feb 2026 18:53:22 +0000 (13:53 -0500)]
various buf fixes in mem.c
Alan T. DeKok [Wed, 18 Feb 2026 18:33:49 +0000 (13:33 -0500)]
clean up on failure, and ensure that "next" is always an FD bio
Alan T. DeKok [Wed, 18 Feb 2026 18:28:54 +0000 (13:28 -0500)]
reset buffer just in case
Alan T. DeKok [Wed, 18 Feb 2026 18:27:19 +0000 (13:27 -0500)]
varius bug fixes in fd_open.c
Alan T. DeKok [Wed, 18 Feb 2026 17:56:16 +0000 (12:56 -0500)]
return dirfd and filename if we have to create it
Alan T. DeKok [Wed, 18 Feb 2026 15:58:40 +0000 (10:58 -0500)]
check for errors when reading the password from a file
Alan T. DeKok [Wed, 18 Feb 2026 15:47:16 +0000 (10:47 -0500)]
various bug fixes in cache.c
free resources on failure. Don't set "request=child", as that
just confuses things.
Alan T. DeKok [Wed, 18 Feb 2026 15:22:20 +0000 (10:22 -0500)]
fix typo to use correct function
Alan T. DeKok [Wed, 18 Feb 2026 15:16:32 +0000 (10:16 -0500)]
free dh on error
Alan T. DeKok [Wed, 18 Feb 2026 14:14:26 +0000 (09:14 -0500)]
rewrite to cache stateless isn't an error for TLS 1.3
Alan T. DeKok [Wed, 18 Feb 2026 14:13:17 +0000 (09:13 -0500)]
fix not_before -> not_after typo
Alan T. DeKok [Wed, 18 Feb 2026 12:48:32 +0000 (07:48 -0500)]
OpenSSL requires calls to engine finish/free
for all engines, as the ENGINE* pointer is reference counted.
Alan T. DeKok [Wed, 18 Feb 2026 12:43:04 +0000 (07:43 -0500)]
call pthread_attr_destroy() for completeness.
Some platforms may dynamically allocate memory for attr
Alan T. DeKok [Wed, 18 Feb 2026 02:21:17 +0000 (21:21 -0500)]
hoist extension decode into common routine
Alan T. DeKok [Wed, 18 Feb 2026 01:39:19 +0000 (20:39 -0500)]
zero is a valid number of bytes to copy
Alan T. DeKok [Wed, 18 Feb 2026 01:38:19 +0000 (20:38 -0500)]
make function static, and use thread-local variable
to match the rest of the code
Alan T. DeKok [Wed, 18 Feb 2026 01:35:20 +0000 (20:35 -0500)]
various bug fixes in ctx.c
Alan T. DeKok [Wed, 18 Feb 2026 01:31:26 +0000 (20:31 -0500)]
various bug fixes in session.c
Alan T. DeKok [Wed, 18 Feb 2026 01:20:56 +0000 (20:20 -0500)]
request can be NULL
if fr_tls_session_msg_cb() is called via SSL_shutdown() in the
session destructor.
Alan T. DeKok [Wed, 18 Feb 2026 01:18:00 +0000 (20:18 -0500)]
fix typo to use correct variable
Alan T. DeKok [Wed, 18 Feb 2026 01:10:06 +0000 (20:10 -0500)]
fix typo, to use the correct log bio
Alan T. DeKok [Wed, 18 Feb 2026 01:06:42 +0000 (20:06 -0500)]
insert / remove engine into tree, and fix list pointer
the code finds the engine in the tree, but never inserts it.
Alan T. DeKok [Wed, 18 Feb 2026 00:59:23 +0000 (19:59 -0500)]
distinguish "can't encode data" from failure.
Alan T. DeKok [Wed, 18 Feb 2026 00:59:03 +0000 (19:59 -0500)]
hoist clear cache, so we do it even in non-debug mode
Alan T. DeKok [Wed, 18 Feb 2026 00:58:45 +0000 (19:58 -0500)]
use equality check
Alan T. DeKok [Wed, 18 Feb 2026 00:38:47 +0000 (19:38 -0500)]
check return code of function, not TLS connection status
Alan T. DeKok [Tue, 17 Feb 2026 19:50:14 +0000 (14:50 -0500)]
check return code and fail on error
Alan T. DeKok [Tue, 17 Feb 2026 19:49:30 +0000 (14:49 -0500)]
use enum names instead of hard-coded constants
nolade [Wed, 11 Feb 2026 19:35:54 +0000 (14:35 -0500)]
docs-v4: update & reorg developers/bugs section HIVE 3722 / 3727
Alan T. DeKok [Tue, 17 Feb 2026 16:32:13 +0000 (11:32 -0500)]
set EAP fail if eap_tls_request() fails
Alan T. DeKok [Tue, 17 Feb 2026 16:15:35 +0000 (11:15 -0500)]
minor tweaks
Alan T. DeKok [Tue, 17 Feb 2026 16:14:52 +0000 (11:14 -0500)]
there's no need for this to be a uint16_t
Alan T. DeKok [Tue, 17 Feb 2026 16:12:19 +0000 (11:12 -0500)]
zero-length fragments with 'M'ore bit set are invalid
Alan T. DeKok [Tue, 17 Feb 2026 16:09:16 +0000 (11:09 -0500)]
free buffer on error
this buffer should really be on the stack?
Alan T. DeKok [Tue, 17 Feb 2026 16:07:46 +0000 (11:07 -0500)]
various bug fixes for chbind.c
Alan T. DeKok [Tue, 17 Feb 2026 16:03:47 +0000 (11:03 -0500)]
we don't accept TLS records of larger than 64K.
These are very unlikely to actually work.
Alan T. DeKok [Tue, 17 Feb 2026 14:49:17 +0000 (09:49 -0500)]
tweak nonnull settings
if dict is NULL, it just defaults to internal
Alan T. DeKok [Tue, 17 Feb 2026 14:53:25 +0000 (09:53 -0500)]
use Asciidoc headings in configuration files
not Markdown ones. the output is slightly different.
perl -p -i -e 's/^# # /# = /;s/^# ## /# == /;s/^# ### /# === /;s/^# #### /# ==== /;' $(git grep -l '^# #' raddb/)
Alan T. DeKok [Tue, 17 Feb 2026 14:30:07 +0000 (09:30 -0500)]
delay increment until after everything succeeds
which doesn't make a lot of difference practically, as if the
adding fails, we just stop processing all of the dictionaries.
but it keeps the static analyser happy
Alan T. DeKok [Tue, 17 Feb 2026 12:58:16 +0000 (07:58 -0500)]
bare END is argc=0not argc=1
Alan T. DeKok [Tue, 17 Feb 2026 12:58:04 +0000 (07:58 -0500)]
add more NONNULL
Alan T. DeKok [Tue, 17 Feb 2026 12:51:16 +0000 (07:51 -0500)]
check return of dict_fixup_init, and remove unused ctx
Alan T. DeKok [Tue, 17 Feb 2026 00:16:42 +0000 (19:16 -0500)]
clean up on error
Alan T. DeKok [Tue, 17 Feb 2026 00:14:15 +0000 (19:14 -0500)]
don't free da after it was added to the dictionaries
Alan T. DeKok [Tue, 17 Feb 2026 00:06:18 +0000 (19:06 -0500)]
correct return codes in dict_attr_allow_dup
Alan T. DeKok [Mon, 16 Feb 2026 23:51:02 +0000 (18:51 -0500)]
various bug fixes in dhcpv6 processor.
* don't free rctx, the framework will clean it up when the stack
frame pops. And the rctx ptr is cached in mctx->rctx, so freeing
it will cause errors
* hoist "allocate sbuff" outside of loop
Alan T. DeKok [Mon, 16 Feb 2026 23:47:24 +0000 (18:47 -0500)]
typo - use decline config, not request
Alan T. DeKok [Mon, 16 Feb 2026 23:45:19 +0000 (18:45 -0500)]
fr_pair_list_copy() returns 0 if there are no elements in the list
Alan T. DeKok [Mon, 16 Feb 2026 23:41:54 +0000 (18:41 -0500)]
set datalen from str length. CID #
1681389
Alan T. DeKok [Mon, 16 Feb 2026 17:20:55 +0000 (12:20 -0500)]
fail on invalid packet codes, instead of dereferencing NULL ptr
this won't happen as the rest of the code is sane, but it's good
to double-check the values to be sure.
Alan T. DeKok [Mon, 16 Feb 2026 17:13:18 +0000 (12:13 -0500)]
add missing Cookie-Load-Fail
Alan T. DeKok [Mon, 16 Feb 2026 17:08:05 +0000 (12:08 -0500)]
MAX is last valid+1
Alan T. DeKok [Mon, 16 Feb 2026 17:03:16 +0000 (12:03 -0500)]
check for DO_NOT_RESPOND, and for rcode in authentication continue
Alan T. DeKok [Mon, 16 Feb 2026 16:48:21 +0000 (11:48 -0500)]
ensure Disconnect-Request and Protocol-Error save/restore Proxy-State
Alan T. DeKok [Mon, 16 Feb 2026 16:36:00 +0000 (11:36 -0500)]
add notes for what to do when modifying rlm_rcode_t
because TIMEOUT was added, and some things were missed
Alan T. DeKok [Mon, 16 Feb 2026 16:05:18 +0000 (11:05 -0500)]
correct logic and add new functionality.
* RLM_RCODE_TIMEOUT was missing from authenticate handler
* add pointer to documentation if the shared secret is wrong
* allow 'authenticate' section to set reply.Packet-Type
* various other minor tweaks and logic corrections
Alan T. DeKok [Mon, 16 Feb 2026 16:03:30 +0000 (11:03 -0500)]
don't add Event-Timestamp in code
it should be a policy. And ignore Acct-Delay-Time if its value
is unreasonable
Alan T. DeKok [Mon, 16 Feb 2026 15:31:47 +0000 (10:31 -0500)]
check if state_tree_init fails
Alan T. DeKok [Mon, 16 Feb 2026 15:31:20 +0000 (10:31 -0500)]
typo