]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
James Jones [Thu, 5 May 2022 21:56:46 +0000 (16:56 -0500)]
fr_strerror_clear() needn't bother with fr_strerror_init() (CID #
1503892 ) (#4486)
fr_strerror_clear() already refers to fr_strerror_buffer, and might as
well use it rather than the return value from fr_sterror_init().
James Jones [Thu, 5 May 2022 21:55:00 +0000 (16:55 -0500)]
Move use of out after check for NULL (CID #
1503900 ) (#4482)
James Jones [Thu, 5 May 2022 21:54:42 +0000 (16:54 -0500)]
Wrap malloc() in MEM() to reassure coverity (CID #
1469117 ) (#4485)
James Jones [Thu, 5 May 2022 21:54:22 +0000 (16:54 -0500)]
Add check for NULL from fr_value_box_alloc_null() (CID #
1469164 ) (#4487)
James Jones [Thu, 5 May 2022 21:53:54 +0000 (16:53 -0500)]
Add another check of fr_event_fd_insert() for trunk tests (CID #
1469165 ) (#4492)
James Jones [Thu, 5 May 2022 21:53:40 +0000 (16:53 -0500)]
Correct memory leak in add_dynlink_noinstall (CID #
1504301 , #
1504306 ) (#4493)
The single check for either install_path or current_path being NULL
will leak if it passes; load_noinstall_path() doesn't have an explicit
return NULL and lt_malloc() exits if malloc() returns NULL, so
current_path is guaranteed non-NULL if load_noinstall_path() returns.
To avoid the issue, check install_path immediately after it's set.
(If coverity doesn't recognize that current_path will never be NULL,
it may need MEM().)
At the end of add_dynlink_noinstall(), both are freed.
Alan T. DeKok [Thu, 5 May 2022 12:57:38 +0000 (14:57 +0200)]
rename variables to something which makes sense
James Jones [Thu, 5 May 2022 12:52:12 +0000 (07:52 -0500)]
Remove needless check for FR_TYPE_NULL (CID #
1503975 ) (#4488)
Alan T. DeKok [Thu, 5 May 2022 12:47:05 +0000 (14:47 +0200)]
set target-specific C compiler via SRC_CC
which should usually be $(COMPILE.c)
Arran Cudbard-Bell [Thu, 5 May 2022 10:21:20 +0000 (12:21 +0200)]
More minor jlibtool fixes
Arran Cudbard-Bell [Thu, 5 May 2022 09:50:46 +0000 (11:50 +0200)]
Pass through our compiler and ranlib defs when building jlibtool
Arran Cudbard-Bell [Thu, 5 May 2022 09:36:59 +0000 (11:36 +0200)]
prctrl definitel isn't posix
Arran Cudbard-Bell [Thu, 5 May 2022 09:36:13 +0000 (11:36 +0200)]
LIBRARY_EXT is really TARGET_LIBRARY_EXT which isn't what we want
Arran Cudbard-Bell [Thu, 5 May 2022 09:35:36 +0000 (11:35 +0200)]
Set HOST_CC TARGET_CC HOST_RANLIB TARGET_RANLIB from their respective environmental variables
Arran Cudbard-Bell [Thu, 5 May 2022 09:26:34 +0000 (11:26 +0200)]
If the target is empty or set to "default" just use the host target
This makes it a little easier in the build framework
Arran Cudbard-Bell [Thu, 5 May 2022 09:14:25 +0000 (11:14 +0200)]
Don't build semaphore code under emscripten, it's not available
Arran Cudbard-Bell [Wed, 4 May 2022 23:41:34 +0000 (01:41 +0200)]
Do partial target matching in jlibtool
Arran Cudbard-Bell [Wed, 4 May 2022 13:48:48 +0000 (15:48 +0200)]
Typo
Arran Cudbard-Bell [Wed, 4 May 2022 12:25:22 +0000 (14:25 +0200)]
Avoid spurious output
Arran Cudbard-Bell [Wed, 4 May 2022 12:00:04 +0000 (14:00 +0200)]
Fixup jlibtool to allow crosscompilation without needing multiple versions of the binary
Arran Cudbard-Bell [Wed, 4 May 2022 10:14:08 +0000 (12:14 +0200)]
Control verbose flag passed to jlibtool
Alan T. DeKok [Tue, 3 May 2022 13:27:57 +0000 (15:27 +0200)]
print out build for jlibtool, etc.
Alan T. DeKok [Tue, 3 May 2022 12:20:28 +0000 (14:20 +0200)]
add "counter" flag for unsigned integers.
to enable more SNMP, and internal stats, along with automatic
aggregation of stats.
Alan T. DeKok [Tue, 3 May 2022 12:18:18 +0000 (14:18 +0200)]
show more structs
Arran Cudbard-Bell [Tue, 3 May 2022 07:52:24 +0000 (09:52 +0200)]
Revert "Fixup jlibtool to allow crosscompilation without needing multiple versions of the binary"
This reverts commit
83892e9c37da2de186afdc775e6438b02d5a35ad .
Alan T. DeKok [Tue, 3 May 2022 07:26:32 +0000 (09:26 +0200)]
export VERBOSE
Arran Cudbard-Bell [Mon, 2 May 2022 12:55:19 +0000 (14:55 +0200)]
FreeBSD uses the .so extension too
Arran Cudbard-Bell [Thu, 28 Apr 2022 18:20:12 +0000 (13:20 -0500)]
Fixup jlibtool to allow crosscompilation without needing multiple versions of the binary
github-actions[bot] [Mon, 2 May 2022 09:54:04 +0000 (09:54 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dhcpv4.tar
github-actions[bot] [Mon, 2 May 2022 09:54:01 +0000 (09:54 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/radius.tar
github-actions[bot] [Mon, 2 May 2022 09:53:57 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/util.tar
github-actions[bot] [Mon, 2 May 2022 09:53:55 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/tftp.tar
github-actions[bot] [Mon, 2 May 2022 09:53:52 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/vmps.tar
github-actions[bot] [Mon, 2 May 2022 09:53:48 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/tacacs.tar
github-actions[bot] [Mon, 2 May 2022 09:53:46 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dns.tar
github-actions[bot] [Mon, 2 May 2022 09:53:43 +0000 (09:53 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dhcpv6.tar
James Jones [Fri, 29 Apr 2022 19:16:56 +0000 (14:16 -0500)]
Make minmax_heap_burn_in() exercise all operations (CID #
1503996 ) (#4478)
random choice of operation wasn't hitting all cases
James Jones [Fri, 29 Apr 2022 19:16:37 +0000 (14:16 -0500)]
Move assignment out of fr_assert_msg() to pacify Coverity (CID #
1504180 ) (#4479)
James Jones [Fri, 29 Apr 2022 19:15:36 +0000 (14:15 -0500)]
remove needless check of parent in parse_input() (CID #
1469153 (#4480)
Earlier in the code, parent->allow_unlang is tested; if parent
is NULL, it will break thre, so the later check is unnecessary.
James Jones [Fri, 29 Apr 2022 19:14:50 +0000 (14:14 -0500)]
Move use of name after NULL check in dict_by_protocol_substr (CID #
1503920 ) (#4481)
James Jones [Fri, 29 Apr 2022 12:34:30 +0000 (07:34 -0500)]
Remove needless NULL checks in cf_pair_parse_internal() (CID #
1469127 , #
1469144 ) (#4476)
array is checked for NULL, and an error returned if it is, before
the loop that fills the array, so the checks of it in the loop (and
also of entry, which can't be NULL either).
James Jones [Fri, 29 Apr 2022 12:34:07 +0000 (07:34 -0500)]
Quiet coverity (CID #
1503969 , #
1504008 ) (#4475)
timestr_match() converts time from the beginning of the week in
a struct tm to minutes. Coverity doesn't consider the bounds
on those values (tm_wday and tm_hour), and thus warns of possible
overflow.
James Jones [Fri, 29 Apr 2022 12:33:43 +0000 (07:33 -0500)]
silence coverity complaint about possible overflow (CID #
1469150 ) (#4474)
NotAProfessionalDeveloper [Fri, 29 Apr 2022 12:32:24 +0000 (14:32 +0200)]
Update link to Cisco website (#4473)
James Jones [Fri, 29 Apr 2022 12:32:12 +0000 (07:32 -0500)]
Correction to description of function return values (#4472)
Nick Porter [Fri, 29 Apr 2022 09:39:48 +0000 (10:39 +0100)]
Add ciena dictionary
Alan T. DeKok [Thu, 28 Apr 2022 15:39:22 +0000 (11:39 -0400)]
set head->flags correctly, and other minor cleanups
don't set the flags repeatedly in xlat_tokenize_argv()
merge head->flags into input flags
Alan T. DeKok [Thu, 28 Apr 2022 14:41:31 +0000 (10:41 -0400)]
just call flatten_argv()
and update it's API to re-parent the nodes and free the input
list, so that there's no confusion about who owns what.
Alan T. DeKok [Thu, 28 Apr 2022 14:41:04 +0000 (10:41 -0400)]
parent things properly
Nick Porter [Thu, 28 Apr 2022 08:04:37 +0000 (09:04 +0100)]
Add covaro dictionary
Jorge Pereira [Thu, 28 Apr 2022 00:37:23 +0000 (21:37 -0300)]
Fix potential leak (#4471)
Alan T. DeKok [Wed, 27 Apr 2022 22:37:03 +0000 (18:37 -0400)]
filter out for CFLAGS, too
Alan T. DeKok [Wed, 27 Apr 2022 22:07:37 +0000 (18:07 -0400)]
filter out warnings for local targets, we don't care
Alan T. DeKok [Wed, 27 Apr 2022 21:32:17 +0000 (17:32 -0400)]
use HOST_CC
Alan T. DeKok [Wed, 27 Apr 2022 21:31:08 +0000 (17:31 -0400)]
Revert "Don't try and build the make support libraries if this is a WASM build... we'll figure this out later"
This reverts commit
932bf41d78ab0c463446cdf46a8ac82812f0e2fc .
Alan T. DeKok [Wed, 27 Apr 2022 21:28:39 +0000 (17:28 -0400)]
use HOST_CC for jlibtool
as it's a local binary
Alan T. DeKok [Wed, 27 Apr 2022 17:18:22 +0000 (13:18 -0400)]
move xlat_exp_head_t to it's own data structure
fmt and flags are not yet filled out
Alan T. DeKok [Wed, 27 Apr 2022 15:52:45 +0000 (11:52 -0400)]
node->fmt is already set above. Re-setting it mangles things
Arran Cudbard-Bell [Wed, 27 Apr 2022 20:31:54 +0000 (15:31 -0500)]
Alternative source of sockaddr_ll for emscripten/musl
Arran Cudbard-Bell [Wed, 27 Apr 2022 20:30:43 +0000 (15:30 -0500)]
If we use the cache, then we don't populate CPPFLAGS on the cache run...
Arran Cudbard-Bell [Wed, 27 Apr 2022 19:48:15 +0000 (14:48 -0500)]
Ignore wasm build products
Arran Cudbard-Bell [Wed, 27 Apr 2022 19:48:05 +0000 (14:48 -0500)]
Add support to jlibtool for emscripten, it's basically the same as Linux, right?
Arran Cudbard-Bell [Wed, 27 Apr 2022 19:47:27 +0000 (14:47 -0500)]
Emscripten provides mallopt, but not the same knobs
Arran Cudbard-Bell [Wed, 27 Apr 2022 19:31:01 +0000 (14:31 -0500)]
Don't try and build the make support libraries if this is a WASM build... we'll figure this out later
It involves using two compilers in the build process, and as EMSCRIPTEN mostly uses static linking, using dlopen to figure out symbol presence probably isn't going to work. We'll need some higher level function that execs the wasm linker like configure does.
Arran Cudbard-Bell [Wed, 27 Apr 2022 19:29:20 +0000 (14:29 -0500)]
Minor fixes for emscripten
James Jones [Wed, 27 Apr 2022 19:17:12 +0000 (14:17 -0500)]
Remove needless check for NULL fp (CID #
1503900 ) (#4468)
Nick Porter [Wed, 27 Apr 2022 14:34:07 +0000 (15:34 +0100)]
Remove spurious debugging
Nick Porter [Wed, 27 Apr 2022 11:06:37 +0000 (12:06 +0100)]
Add simultaneous use queries for MS SQL
Nick Porter [Wed, 27 Apr 2022 06:47:41 +0000 (07:47 +0100)]
Fix decoding of single TLV
- use fr_dcursor_next_peek to check for multiple children so that the
cursor doesn't move
- extract the pair from the source list to append to the destination
Arran Cudbard-Bell [Tue, 26 Apr 2022 13:58:31 +0000 (08:58 -0500)]
New config.sub and config.guess
Alan T. DeKok [Tue, 26 Apr 2022 13:00:34 +0000 (09:00 -0400)]
fix locking. Helps with #4465
Alan T. DeKok [Tue, 26 Apr 2022 13:00:07 +0000 (09:00 -0400)]
release mutex before destroying it. Helps with #4465
Arran Cudbard-Bell [Mon, 25 Apr 2022 17:06:04 +0000 (12:06 -0500)]
Fix potential leak on error
James Jones [Mon, 25 Apr 2022 15:23:12 +0000 (10:23 -0500)]
Correct UB_EV_PERSIST check in _unbound_io_service_writable() (CID #
1469157 ) (#4466)
Matthew Newton [Sat, 23 Apr 2022 10:53:27 +0000 (11:53 +0100)]
crossbuild: update image prefix to remove clashes with v3
Matthew Newton [Sat, 23 Apr 2022 10:26:11 +0000 (11:26 +0100)]
docker: tidy unused jenkins dockerfiles
Arran Cudbard-Bell [Sat, 23 Apr 2022 02:25:30 +0000 (21:25 -0500)]
Add args back
Arran Cudbard-Bell [Sat, 23 Apr 2022 01:18:41 +0000 (20:18 -0500)]
Just use fopen...
Arran Cudbard-Bell [Fri, 22 Apr 2022 22:08:21 +0000 (17:08 -0500)]
radmin isn't thread safe
Arran Cudbard-Bell [Fri, 22 Apr 2022 21:38:42 +0000 (16:38 -0500)]
Passing back port via variables is not thread safe, so set a unique variable containing the port for a given test suite
Arran Cudbard-Bell [Fri, 22 Apr 2022 21:06:53 +0000 (16:06 -0500)]
MODULE_TEST_UNLANG is a path
Alan T. DeKok [Fri, 22 Apr 2022 20:53:46 +0000 (16:53 -0400)]
not yet for this
Arran Cudbard-Bell [Fri, 22 Apr 2022 20:49:05 +0000 (15:49 -0500)]
Horrible hacks to get EAPOL tests to run in parallel
Arran Cudbard-Bell [Fri, 22 Apr 2022 17:52:29 +0000 (12:52 -0500)]
Fix radclient tests for parallel execution
Arran Cudbard-Bell [Fri, 22 Apr 2022 17:26:56 +0000 (12:26 -0500)]
Allow the secret and base port to be overridden from the shell
Arran Cudbard-Bell [Fri, 22 Apr 2022 17:21:24 +0000 (12:21 -0500)]
Use the PID file to identify unique instances of FreeRADIUS if it's set
James Jones [Fri, 22 Apr 2022 20:23:51 +0000 (15:23 -0500)]
Get rid of superfluous test in loop in command_encode_dns_label (#4463)
By virtue of CC_HINT(nonnull), in, and hence p, will be non-NULL
the first time through the loop. To take another iteration, next
must be non-NULL, so that p, set to next + 1, will also be non-NULL;
hence the "while (p)" is effectively "while (true)".
Alan T. DeKok [Fri, 22 Apr 2022 20:21:41 +0000 (16:21 -0400)]
typo
Alan T. DeKok [Fri, 22 Apr 2022 19:37:17 +0000 (15:37 -0400)]
use correct data type
Alan T. DeKok [Fri, 22 Apr 2022 19:14:02 +0000 (15:14 -0400)]
move encode_array() to common function
which further shrinks and simplifies the protocol-specific
encoders.
Alan T. DeKok [Fri, 22 Apr 2022 18:58:58 +0000 (14:58 -0400)]
none of this is used, so change it to void*
Alan T. DeKok [Fri, 22 Apr 2022 14:49:21 +0000 (10:49 -0400)]
catch the case of zero returns
it's either allowed, and we do nothing, or it means that we didn't
decode any data, and the input is malformed
Alan T. DeKok [Fri, 22 Apr 2022 14:48:42 +0000 (10:48 -0400)]
options can have zero data
Alan T. DeKok [Thu, 21 Apr 2022 17:43:29 +0000 (13:43 -0400)]
add fake xlat_exp_head_t, which is just xlat_exp_t for now
and update function signatures, etc. to use it. In preparation
for moving to tlists
Alan T. DeKok [Thu, 21 Apr 2022 17:33:01 +0000 (13:33 -0400)]
don't do N^2 checks
we just need to check if it's a value box,
xlat_is_literal() will check all subsequent nodes, and we don't
want that.
Arran Cudbard-Bell [Fri, 22 Apr 2022 14:42:10 +0000 (09:42 -0500)]
Fix parallel tests for sqlite, and enable parallel testing
James Jones [Thu, 21 Apr 2022 19:18:38 +0000 (14:18 -0500)]
Check fr_event_fd_insert() return value. CID #
1469154 (#4461)
Arran Cudbard-Bell [Thu, 21 Apr 2022 16:48:10 +0000 (11:48 -0500)]
Add temporary thread local dbuff buffers too
Alan T. DeKok [Thu, 21 Apr 2022 13:56:28 +0000 (09:56 -0400)]
typo
Alan T. DeKok [Wed, 20 Apr 2022 21:16:13 +0000 (17:16 -0400)]
for length fields, just limit work_dbuff
instead of encoding "too much" and checking for it later