]> git.ipfire.org Git - thirdparty/freeradius-server.git/log
thirdparty/freeradius-server.git
5 months agomash leading '.', too.
Alan T. DeKok [Wed, 4 Mar 2026 22:31:36 +0000 (17:31 -0500)] 
mash leading '.', too.

and set safe_for

5 months agofix: 1. initialise thread mutex before use
Catalina Pineros [Wed, 4 Mar 2026 19:46:35 +0000 (14:46 -0500)] 
fix: 1. initialise thread mutex before use
(rlm_stats: initialise thread mutex before use 415)

5 months agofix: 1. typo
Catalina Pineros [Wed, 4 Mar 2026 19:33:52 +0000 (14:33 -0500)] 
fix: 1. typo
(rlm_sqlcounter: fix copy-paste bug in find_prev_reset() for "never" reset 413)

5 months agofix: 1. function parameter typo fixed
Catalina Pineros [Wed, 4 Mar 2026 19:18:15 +0000 (14:18 -0500)] 
fix: 1. function parameter typo fixed
(Fix copy-paste error: response_window never validated in rlm_tacacs and rlm_radius 416)

5 months agofix: 1. exact duplicate code removed
Catalina Pineros [Wed, 4 Mar 2026 19:00:45 +0000 (14:00 -0500)] 
fix: 1. exact duplicate code removed
2. return 0  -->  return -1   (for invalid discriminator/state)
(protocols/bfd: fix duplicate detect_multi check and wrong return value in decode_proto 433)

5 months agofix: 1. missing 'false' argument when calling function
Catalina Pineros [Wed, 4 Mar 2026 18:39:25 +0000 (13:39 -0500)] 
fix: 1. missing 'false' argument when calling function
2. num_events == EINTR  -->  errno == EINTR     (typo)
3. if (*start > MY_ARRAY_SIZE)   -->   if (*start >= MY_ARRAY_SIZE)   (wraparound sometimes skipped)
(tests/util: fix missing 5th arg to fr_message_set_create, wrong EINTR check, and off-by-one 225)

5 months agofix: 1. calling wrong function, is now pop_tail (copy paste typo)
Catalina Pineros [Wed, 4 Mar 2026 16:48:15 +0000 (11:48 -0500)] 
fix: 1. calling wrong function, is now pop_tail (copy paste typo)
2. fix undefined behaviour
(lib/util/dlist: fix free_tail pop direction and sort-merge NULL deref 156)

5 months agofix: 1. pesudo --> pseudo (typo)
Catalina Pineros [Wed, 4 Mar 2026 15:40:06 +0000 (10:40 -0500)] 
fix: 1. pesudo --> pseudo (typo)
2. added outer parenthesis
(net.h/net.c: fix 'pesudo' typo in function name; add missing outer parens to IP_VHL macro 283)

5 months agojust use sprintf to match the rest of the code
Alan T. DeKok [Wed, 4 Mar 2026 22:05:46 +0000 (17:05 -0500)] 
just use sprintf to match the rest of the code

and other minor tweaks

5 months agogreatly simplify function
Alan T. DeKok [Wed, 4 Mar 2026 22:01:10 +0000 (17:01 -0500)] 
greatly simplify function

as it should only be taking unsafe inputs

5 months agothe function is only used by one caller
Alan T. DeKok [Wed, 4 Mar 2026 21:52:44 +0000 (16:52 -0500)] 
the function is only used by one caller

and doesn't have to be exposed

5 months agoadd callback for escaping value-box with sbuff erules
Alan T. DeKok [Wed, 4 Mar 2026 16:39:20 +0000 (11:39 -0500)] 
add callback for escaping value-box with sbuff erules

5 months agouse real type, and not uintptr_t
Alan T. DeKok [Wed, 4 Mar 2026 16:29:42 +0000 (11:29 -0500)] 
use real type, and not uintptr_t

5 months agoremove functionality which is useless and never worked
Alan T. DeKok [Wed, 4 Mar 2026 16:19:05 +0000 (11:19 -0500)] 
remove functionality which is useless and never worked

5 months agoadd simple tests for value-box functions
Alan T. DeKok [Wed, 4 Mar 2026 15:29:18 +0000 (10:29 -0500)] 
add simple tests for value-box functions

5 months agouse standard APIs for filename escapes, and check for errors
Alan T. DeKok [Wed, 4 Mar 2026 15:24:34 +0000 (10:24 -0500)] 
use standard APIs for filename escapes, and check for errors

5 months agoadd value-box escape with erules
Alan T. DeKok [Wed, 4 Mar 2026 15:03:12 +0000 (10:03 -0500)] 
add value-box escape with erules

there are a number of places which need it.

5 months agocontinue after \r, \n, \t
Alan T. DeKok [Wed, 4 Mar 2026 12:46:50 +0000 (07:46 -0500)] 
continue after \r, \n, \t

5 months agothe sbuff escape / unescape rules should be const
Alan T. DeKok [Wed, 4 Mar 2026 12:38:33 +0000 (07:38 -0500)] 
the sbuff escape / unescape rules should be const

5 months agomove escape sbuff sequences to utility library
Alan T. DeKok [Wed, 4 Mar 2026 12:33:55 +0000 (07:33 -0500)] 
move escape sbuff sequences to utility library

so that others can use them

5 months agocheck for initialize failure
Alan T. DeKok [Wed, 4 Mar 2026 12:06:19 +0000 (07:06 -0500)] 
check for initialize failure

5 months agotypo
Alan T. DeKok [Wed, 4 Mar 2026 12:02:48 +0000 (07:02 -0500)] 
typo

5 months agoquiet static analyzer
Alan T. DeKok [Wed, 4 Mar 2026 12:02:12 +0000 (07:02 -0500)] 
quiet static analyzer

5 months agofix: 1. return 01 --> return -1 (it's meant to be an error return val)
Catalina Pineros [Tue, 3 Mar 2026 21:22:18 +0000 (16:22 -0500)] 
fix:    1. return 01 --> return -1 (it's meant to be an error return val)
2. uid_str --> gid_str  (orelse it's checking the same condition twice)
3. uid_str --> gid_str typo (orelse it's running a gid function on a uid_str)
(fix(sem): octal return value, wrong NULL check variable, and wrong gid_t type in header 168)

5 months agoformatting bug: changed '%zu' to '%zuk' to stay consisten in error message
Catalina Pineros [Tue, 3 Mar 2026 20:59:31 +0000 (15:59 -0500)] 
formatting bug: changed '%zu' to '%zuk' to stay consisten in error message
(rlm_sql_sqlite: add missing 'k' unit to bootstrap file size error message 411)

5 months agoprotocols/der: fix fr_der_flag_has_default_value macro — trailing semicolon and spuri...
Catalina Pineros [Tue, 3 Mar 2026 19:38:07 +0000 (14:38 -0500)] 
protocols/der: fix fr_der_flag_has_default_value macro — trailing semicolon and spurious != NULL comparison

5 months agorlm_wimax: fix trailing dash overwriting null terminator in Calling-Station-Id reform...
Catalina Pineros [Tue, 3 Mar 2026 19:37:21 +0000 (14:37 -0500)] 
rlm_wimax: fix trailing dash overwriting null terminator in Calling-Station-Id reformatter

5 months agocheck proper return code
Alan T. DeKok [Tue, 3 Mar 2026 22:04:44 +0000 (17:04 -0500)] 
check proper return code

the function returns 0 for "more data", and 1 for "no data",
which is different from the other APIs

5 months agocheck for invalid hex digits
Alan T. DeKok [Tue, 3 Mar 2026 22:01:13 +0000 (17:01 -0500)] 
check for invalid hex digits

5 months agouse correct IP when prefix=0
Alan T. DeKok [Tue, 3 Mar 2026 21:55:30 +0000 (16:55 -0500)] 
use correct IP when prefix=0

5 months agovarious bug fixes in command.c
Alan T. DeKok [Tue, 3 Mar 2026 21:55:20 +0000 (16:55 -0500)] 
various bug fixes in command.c

5 months agoremove unused macro
Alan T. DeKok [Tue, 3 Mar 2026 19:51:16 +0000 (14:51 -0500)] 
remove unused macro

5 months agoturn off more warning messages
Alan T. DeKok [Tue, 3 Mar 2026 19:51:26 +0000 (14:51 -0500)] 
turn off more warning messages

5 months agoignore AI test files
Alan T. DeKok [Tue, 3 Mar 2026 19:41:20 +0000 (14:41 -0500)] 
ignore AI test files

5 months agoadd basic tests for calc.c
Alan T. DeKok [Tue, 3 Mar 2026 15:39:21 +0000 (10:39 -0500)] 
add basic tests for calc.c

5 months agoremove unused variables
Alan T. DeKok [Tue, 3 Mar 2026 18:13:04 +0000 (13:13 -0500)] 
remove unused variables

5 months agonot all packet codes are valid
Alan T. DeKok [Tue, 3 Mar 2026 18:10:52 +0000 (13:10 -0500)] 
not all packet codes are valid

5 months agocheck for errors and set flags
Alan T. DeKok [Tue, 3 Mar 2026 18:08:40 +0000 (13:08 -0500)] 
check for errors and set flags

5 months agovarious bug fixes in eap-md5
Alan T. DeKok [Tue, 3 Mar 2026 16:27:38 +0000 (11:27 -0500)] 
various bug fixes in eap-md5

we don't use "name".

check password length

add more sanity checks

5 months agoctime() can return NULL
Alan T. DeKok [Tue, 3 Mar 2026 15:20:01 +0000 (10:20 -0500)] 
ctime() can return NULL

5 months agotweak bounds
Alan T. DeKok [Tue, 3 Mar 2026 15:06:57 +0000 (10:06 -0500)] 
tweak bounds

5 months agomore fix typos
Alan T. DeKok [Tue, 3 Mar 2026 14:43:40 +0000 (09:43 -0500)] 
more fix typos

to finish the changes from bcf136dd0

5 months agotypo in macro
Alan T. DeKok [Tue, 3 Mar 2026 14:29:04 +0000 (09:29 -0500)] 
typo in macro

5 months agouse "goto fail" in more situations
Alan T. DeKok [Tue, 3 Mar 2026 14:27:42 +0000 (09:27 -0500)] 
use "goto fail" in more situations

5 months agominor tweaks and updates to the use of AI
Alan T. DeKok [Tue, 3 Mar 2026 13:39:16 +0000 (08:39 -0500)] 
minor tweaks and updates to the use of AI

5 months agoAdd client name to error message
Nick Porter [Tue, 3 Mar 2026 17:30:07 +0000 (17:30 +0000)] 
Add client name to error message

5 months agoFallback to attribute parsing if cast doesn't work
Nick Porter [Mon, 2 Mar 2026 17:58:31 +0000 (17:58 +0000)] 
Fallback to attribute parsing if cast doesn't work

This allows call_env parsing for types such as uint32 to support

foo = 10

and

foo = attr

with the value being a bare word in both cases.

5 months agomanually do dir + filename
Alan T. DeKok [Mon, 2 Mar 2026 21:51:03 +0000 (16:51 -0500)] 
manually do dir + filename

to get rid of all double '/', and to avoid walking over the
input strings repeatedly.

5 months agorelax checks in corner conditions
Alan T. DeKok [Mon, 2 Mar 2026 21:49:58 +0000 (16:49 -0500)] 
relax checks in corner conditions

5 months agoassert that the entry is in the heap
Alan T. DeKok [Mon, 2 Mar 2026 19:50:26 +0000 (14:50 -0500)] 
assert that the entry is in the heap

5 months agofree group on error
Alan T. DeKok [Mon, 2 Mar 2026 19:48:20 +0000 (14:48 -0500)] 
free group on error

5 months agoshut up analyzer
Alan T. DeKok [Mon, 2 Mar 2026 16:24:00 +0000 (11:24 -0500)] 
shut up analyzer

5 months agowhen shutdown fails, enter failed state, not closed state
Alan T. DeKok [Mon, 2 Mar 2026 16:15:11 +0000 (11:15 -0500)] 
when shutdown fails, enter failed state, not closed state

5 months agovarious bug fixes in limit.c
Alan T. DeKok [Mon, 2 Mar 2026 16:14:57 +0000 (11:14 -0500)] 
various bug fixes in limit.c

5 months agovarious bug fixes in hash.c
Alan T. DeKok [Mon, 2 Mar 2026 16:04:32 +0000 (11:04 -0500)] 
various bug fixes in hash.c

5 months agoadd locks when accessing entry
Alan T. DeKok [Mon, 2 Mar 2026 12:32:44 +0000 (07:32 -0500)] 
add locks when accessing entry

5 months agofix typo
Alan T. DeKok [Mon, 2 Mar 2026 12:29:27 +0000 (07:29 -0500)] 
fix typo

5 months agofix name comparison callbacks to use correct values
Alan T. DeKok [Sun, 1 Mar 2026 21:17:55 +0000 (16:17 -0500)] 
fix name comparison callbacks to use correct values

5 months agoreturn on failure
Alan T. DeKok [Sun, 1 Mar 2026 21:16:11 +0000 (16:16 -0500)] 
return on failure

5 months agocheck before memcpy
Alan T. DeKok [Sun, 1 Mar 2026 21:15:46 +0000 (16:15 -0500)] 
check before memcpy

5 months agouse digest_find_vp() for qop, too
Alan T. DeKok [Sun, 1 Mar 2026 21:13:51 +0000 (16:13 -0500)] 
use digest_find_vp() for qop, too

even though it's hex

5 months agopoint to correct attribute
Alan T. DeKok [Sun, 1 Mar 2026 21:11:18 +0000 (16:11 -0500)] 
point to correct attribute

5 months agotypos in lengths
Alan T. DeKok [Sun, 1 Mar 2026 21:11:08 +0000 (16:11 -0500)] 
typos in lengths

5 months agoclean up cs on error
Alan T. DeKok [Sun, 1 Mar 2026 21:09:25 +0000 (16:09 -0500)] 
clean up cs on error

5 months agoAdd test of redis_ippool with requested address
Nick Porter [Mon, 2 Mar 2026 14:51:34 +0000 (14:51 +0000)] 
Add test of redis_ippool with requested address

5 months agoRenumber lua script lines
Nick Porter [Mon, 2 Mar 2026 14:49:36 +0000 (14:49 +0000)] 
Renumber lua script lines

5 months agoHonour requested address in rlm_redis_ippool
Nick Porter [Mon, 2 Mar 2026 11:42:23 +0000 (11:42 +0000)] 
Honour requested address in rlm_redis_ippool

5 months agoCompare numbers rather than strings
Nick Porter [Mon, 2 Mar 2026 14:26:59 +0000 (14:26 +0000)] 
Compare numbers rather than strings

Since 10 > 2 but "10" < "2"

5 months agoMutex is protecting dlist maniplation, not entry->expires (CID #1682066)
Nick Porter [Mon, 2 Mar 2026 08:21:38 +0000 (08:21 +0000)] 
Mutex is protecting dlist maniplation, not entry->expires (CID #1682066)

5 months agovarious bug fixes in pool.c
Alan T. DeKok [Sun, 1 Mar 2026 20:04:52 +0000 (15:04 -0500)] 
various bug fixes in pool.c

5 months agofix pipe bio to use normal read/write blocked/resume callbacks
Alan T. DeKok [Sun, 1 Mar 2026 19:57:27 +0000 (14:57 -0500)] 
fix pipe bio to use normal read/write blocked/resume callbacks

5 months agojust use our own buffer for the pipes
Alan T. DeKok [Sun, 1 Mar 2026 19:31:52 +0000 (14:31 -0500)] 
just use our own buffer for the pipes

and tweak read/write logic to be better.

5 months agoremove duplicate print
Alan T. DeKok [Sun, 1 Mar 2026 18:55:00 +0000 (13:55 -0500)] 
remove duplicate print

5 months agomore tests for the edit code
Alan T. DeKok [Sun, 1 Mar 2026 16:45:03 +0000 (11:45 -0500)] 
more tests for the edit code

5 months agoFR_EDIT_CHILD doesn't have a vp
Alan T. DeKok [Sun, 1 Mar 2026 16:40:22 +0000 (11:40 -0500)] 
FR_EDIT_CHILD doesn't have a vp

move the check for it to before the assertions that VP is sane

5 months agocomment out talloc_get_type_abort_const() until we fix all callers
Alan T. DeKok [Sun, 1 Mar 2026 16:10:05 +0000 (11:10 -0500)] 
comment out talloc_get_type_abort_const() until we fix all callers

5 months agotypos in calc.c
Alan T. DeKok [Sun, 1 Mar 2026 16:00:20 +0000 (11:00 -0500)] 
typos in calc.c

5 months agodup the correct FD
Alan T. DeKok [Sun, 1 Mar 2026 15:58:47 +0000 (10:58 -0500)] 
dup the correct FD

5 months agotypo in cbor
Alan T. DeKok [Sun, 1 Mar 2026 15:57:23 +0000 (10:57 -0500)] 
typo in cbor

5 months agofix typos in time code
Alan T. DeKok [Sun, 1 Mar 2026 15:56:23 +0000 (10:56 -0500)] 
fix typos in time code

5 months agobetter checks for corner cases when parsing fractional seconds
Alan T. DeKok [Sun, 1 Mar 2026 15:55:02 +0000 (10:55 -0500)] 
better checks for corner cases when parsing fractional seconds

5 months agovarious bug fixes in timer.c
Alan T. DeKok [Sun, 1 Mar 2026 15:52:25 +0000 (10:52 -0500)] 
various bug fixes in timer.c

5 months agoremove duplicate code
Alan T. DeKok [Sun, 1 Mar 2026 15:50:06 +0000 (10:50 -0500)] 
remove duplicate code

5 months agoadd MEM()
Alan T. DeKok [Sun, 1 Mar 2026 15:47:45 +0000 (10:47 -0500)] 
add MEM()

5 months agouse vp->da instead of da
Alan T. DeKok [Sun, 1 Mar 2026 15:47:33 +0000 (10:47 -0500)] 
use vp->da instead of da

5 months agoclean up for alignment, overflow checks, etc.
Alan T. DeKok [Sun, 1 Mar 2026 15:09:37 +0000 (10:09 -0500)] 
clean up for alignment, overflow checks, etc.

arguably this code should just be removed, as pretty much every
modern OS doesn't need these wrappers

5 months agouse value-box API rather than hand-rolled code
Alan T. DeKok [Sun, 1 Mar 2026 14:42:26 +0000 (09:42 -0500)] 
use value-box API rather than hand-rolled code

5 months agouse talloc_strlen() instead of talloc_array_length() - 1
Alan T. DeKok [Sat, 28 Feb 2026 14:43:43 +0000 (09:43 -0500)] 
use talloc_strlen() instead of talloc_array_length() - 1

5 months agouse talloc_strdup() now that it's a define for talloc_typed_strdup()
Alan T. DeKok [Sat, 28 Feb 2026 14:44:05 +0000 (09:44 -0500)] 
use talloc_strdup() now that it's a define for talloc_typed_strdup()

5 months agooutputs depend on the libraries that we're using
Alan T. DeKok [Sat, 28 Feb 2026 16:30:51 +0000 (11:30 -0500)] 
outputs depend on the libraries that we're using

5 months agouse correct callback for copy, instead of recursing
Alan T. DeKok [Sat, 28 Feb 2026 15:37:57 +0000 (10:37 -0500)] 
use correct callback for copy, instead of recursing

5 months agotypos in pair_tokenize.c
Alan T. DeKok [Sat, 28 Feb 2026 15:36:21 +0000 (10:36 -0500)] 
typos in pair_tokenize.c

5 months agoset type for talloc_strdup() and friends
Alan T. DeKok [Sat, 28 Feb 2026 14:31:23 +0000 (09:31 -0500)] 
set type for talloc_strdup() and friends

The talloc string functions set the type of the string to the
string contents.  Which makes it very difficult to do type checking
on them.  As such, we wrap the various functions in ones which set
the name to something reasonable.

5 months agodon't pass "const" to talloc functions
Alan T. DeKok [Sat, 28 Feb 2026 14:25:40 +0000 (09:25 -0500)] 
don't pass "const" to talloc functions

5 months agosecret may be NULL, and make code more consistent
Alan T. DeKok [Sat, 28 Feb 2026 14:24:49 +0000 (09:24 -0500)] 
secret may be NULL, and make code more consistent

5 months agoparent filename off of thread, not inst
Alan T. DeKok [Sat, 28 Feb 2026 14:18:02 +0000 (09:18 -0500)] 
parent filename off of thread, not inst

5 months agouse correct attribute for duplicate checks
Alan T. DeKok [Sat, 28 Feb 2026 13:48:21 +0000 (08:48 -0500)] 
use correct attribute for duplicate checks

5 months agowarn if they're using a secret other than 'radsec' for TLS.
Alan T. DeKok [Sat, 28 Feb 2026 13:45:59 +0000 (08:45 -0500)] 
warn if they're using a secret other than 'radsec' for TLS.

@todo - also when we do DTLS, the secret should be different.
And when ALPN is used, we don't have a secret

5 months agoclean up on failure.
Alan T. DeKok [Sat, 28 Feb 2026 13:42:13 +0000 (08:42 -0500)] 
clean up on failure.

which doesn't matter in practice, as we don't dynamically read
users files.  But it's good practice

5 months agoclean up on corner case of failure
Alan T. DeKok [Sat, 28 Feb 2026 13:41:46 +0000 (08:41 -0500)] 
clean up on corner case of failure

and use "goto fail" in more places, to ensure a common cleanup
path