]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Wed, 20 Jan 2021 18:31:00 +0000 (18:31 +0000)]
Add DHCPv6 test server
Arran Cudbard-Bell [Wed, 20 Jan 2021 18:30:33 +0000 (18:30 +0000)]
Take Packet-Type from the input file
We can't just default to Access-Request anymore
Arran Cudbard-Bell [Wed, 20 Jan 2021 18:04:43 +0000 (18:04 +0000)]
Slightly better way of initialising
Alan T. DeKok [Wed, 20 Jan 2021 17:17:06 +0000 (12:17 -0500)]
move RADIUS virtual servers to radius.conf
Alan T. DeKok [Wed, 20 Jan 2021 17:16:30 +0000 (12:16 -0500)]
export PROTOCOL for use in tests, if needed
Alan T. DeKok [Wed, 20 Jan 2021 16:50:47 +0000 (11:50 -0500)]
no quotes needed
Alan T. DeKok [Wed, 20 Jan 2021 16:49:05 +0000 (11:49 -0500)]
fix typo
Alan T. DeKok [Wed, 20 Jan 2021 16:30:36 +0000 (11:30 -0500)]
add support for PROTOCOL in keyword tests
Alan T. DeKok [Wed, 20 Jan 2021 16:18:14 +0000 (11:18 -0500)]
use internal attribute for tests, not RADIUS Filter-Id
Alan T. DeKok [Wed, 20 Jan 2021 16:03:00 +0000 (11:03 -0500)]
remove unused attributes
Alan T. DeKok [Wed, 20 Jan 2021 14:30:08 +0000 (09:30 -0500)]
don't manually skip internal attributes.
rely on the iterator for that. And as a side effect, remove
support for Raw-Attribute
Alan T. DeKok [Wed, 20 Jan 2021 14:07:01 +0000 (09:07 -0500)]
aggregate tags across the entire packet
not just one attribute
the previous tests were decoding lists of pairs, which worked.
updating the tests to decoding packets should the issue.
note that packet encode with tags is still failing. That will be
fixed next.
Alan T. DeKok [Wed, 20 Jan 2021 13:48:39 +0000 (08:48 -0500)]
remove end[], and just use sizeof()
we should probably add CC_HINT(__packed__) to the structures,
but that creates additional compiler warnings. One thing at
a time.
Alan T. DeKok [Wed, 20 Jan 2021 13:39:26 +0000 (08:39 -0500)]
none of the keyword tests use rlm_cache
Arran Cudbard-Bell [Wed, 20 Jan 2021 16:23:54 +0000 (16:23 +0000)]
Fix bad field access in json_object_from_value_box, and encode size
Arran Cudbard-Bell [Wed, 20 Jan 2021 14:10:21 +0000 (14:10 +0000)]
Remember parent's log lvl
Arran Cudbard-Bell [Wed, 20 Jan 2021 14:10:12 +0000 (14:10 +0000)]
Remove magic number
Arran Cudbard-Bell [Wed, 20 Jan 2021 13:30:52 +0000 (13:30 +0000)]
More ctx fixes (dhcpv4)
Arran Cudbard-Bell [Wed, 20 Jan 2021 13:12:52 +0000 (13:12 +0000)]
Parent DHCPv6 pairs correctly
Arran Cudbard-Bell [Wed, 20 Jan 2021 13:02:57 +0000 (13:02 +0000)]
Revert "At least _try_ to resolve unresolved attributes in xlats"
This reverts commit
db9dd99009666a3df032964c3f2d00b0e44bc296 .
Arran Cudbard-Bell [Wed, 20 Jan 2021 13:02:48 +0000 (13:02 +0000)]
Unknown attributes get marked up as XLAT_VIRTUAL_UNRESOLVED
Arran Cudbard-Bell [Wed, 20 Jan 2021 12:08:56 +0000 (12:08 +0000)]
Don't print unresolved - unresolved
Arran Cudbard-Bell [Wed, 20 Jan 2021 12:05:59 +0000 (12:05 +0000)]
At least _try_ to resolve unresolved attributes in xlats
lyubomirtraykov [Wed, 20 Jan 2021 09:44:49 +0000 (11:44 +0200)]
Update dictionary.fortinet (#3862)
Add Host-Port-AVPair attribute.
Alan T. DeKok [Tue, 19 Jan 2021 21:27:31 +0000 (16:27 -0500)]
some fixes
allow output directory to be set with FUZZER_ARTIFACTS.
The default is $(BUILD_DIR)/fuzzer/$(PROTOCOL)
change -timeout=x to -max_total_time=$(FUZZER_TIMEOUT)
allow FUZZER_TIMEOUT to be set, too. Default is 10s
for "test.fuzzer". For "make fuzzer.radius", there's no
timeout. Perhaps we could unify that.
Add FUZZER_ARGUMENTS for extra arguments which may be needed
by some process.
Arran Cudbard-Bell [Tue, 19 Jan 2021 19:45:50 +0000 (19:45 +0000)]
Fix stupid argument name
Alan T. DeKok [Tue, 19 Jan 2021 18:55:29 +0000 (13:55 -0500)]
add assertion that nothing is left unresolved
Alan T. DeKok [Tue, 19 Jan 2021 18:40:19 +0000 (13:40 -0500)]
add test for Packet-Src-IP-Address
so it fails here instead of in the radclient tests
Arran Cudbard-Bell [Tue, 19 Jan 2021 18:19:47 +0000 (18:19 +0000)]
Don't build the trunk_test functions either
Arran Cudbard-Bell [Tue, 19 Jan 2021 18:16:01 +0000 (18:16 +0000)]
Split out virtual attribute resolution login into a separate function so we can enumerate the cases
Arran Cudbard-Bell [Tue, 19 Jan 2021 18:06:48 +0000 (18:06 +0000)]
Disable for now
Alan T. DeKok [Tue, 19 Jan 2021 17:38:54 +0000 (12:38 -0500)]
free detached child on error
Alan T. DeKok [Tue, 19 Jan 2021 17:38:18 +0000 (12:38 -0500)]
cleanups and fixing logic
Alan T. DeKok [Tue, 19 Jan 2021 16:10:17 +0000 (11:10 -0500)]
clean up logic
Alan T. DeKok [Tue, 19 Jan 2021 15:01:31 +0000 (10:01 -0500)]
clean up logic
add brackets so the && applies to the virtual attributes.
check for allow unresolved before checking for attr / virtual
Alan T. DeKok [Tue, 19 Jan 2021 14:49:06 +0000 (09:49 -0500)]
do compile-time checks for unresolved attributes
Arran Cudbard-Bell [Tue, 19 Jan 2021 15:38:18 +0000 (15:38 +0000)]
Set the unresolved flag in xlat_tokenize_attribute for unresolved attribute references
Arran Cudbard-Bell [Tue, 19 Jan 2021 14:30:20 +0000 (14:30 +0000)]
We produce asserts on bad arguments
Alan T. DeKok [Tue, 19 Jan 2021 14:14:41 +0000 (09:14 -0500)]
convert assertions to run-time checks.
with descriptive error messages, and tests
Alan T. DeKok [Tue, 19 Jan 2021 14:08:15 +0000 (09:08 -0500)]
add "test.unit.PROTOCOL" target
for utter laziness
Arran Cudbard-Bell [Tue, 19 Jan 2021 13:38:20 +0000 (13:38 +0000)]
Update dictionary.juniper (#3857)
Add Juniper-CWA-Redirect to dictionary.juniper
lyubomirtraykov [Tue, 19 Jan 2021 13:29:37 +0000 (15:29 +0200)]
remove the "Juniper-" prefix
Jorge Pereira [Tue, 19 Jan 2021 11:12:16 +0000 (08:12 -0300)]
wimax: Fix sql queries.conf (#3854)
Jorge Pereira [Tue, 19 Jan 2021 11:11:56 +0000 (08:11 -0300)]
eapol_test: Fix OSX build using clang >= 11 (#3855)
Fixes the eapol_test build using clang >= 11
../src/radius/radius_client.c:817:24: error: cast to smaller integer type 'RadiusType' from 'void *' [-Werror,-Wvoid-pointer-to-enum-cast]
RadiusType msg_type = (RadiusType) sock_ctx;
lyubomirtraykov [Tue, 19 Jan 2021 08:02:05 +0000 (10:02 +0200)]
Update dictionary.juniper
Add Juniper-CWA-Redirect to dictionary.juniper
Arran Cudbard-Bell [Mon, 18 Jan 2021 22:12:51 +0000 (22:12 +0000)]
Not used
Arran Cudbard-Bell [Mon, 18 Jan 2021 22:11:58 +0000 (22:11 +0000)]
Fix various spec file issues
Arran Cudbard-Bell [Mon, 18 Jan 2021 21:36:48 +0000 (21:36 +0000)]
Add clang format markers
Alan T. DeKok [Mon, 18 Jan 2021 17:48:02 +0000 (12:48 -0500)]
minor cleanups
Arran Cudbard-Bell [Mon, 18 Jan 2021 18:53:01 +0000 (18:53 +0000)]
Missing parens
Arran Cudbard-Bell [Mon, 18 Jan 2021 14:00:47 +0000 (14:00 +0000)]
CI is slow
Arran Cudbard-Bell [Mon, 18 Jan 2021 13:42:14 +0000 (13:42 +0000)]
Fix issue in JSON bool evaluation
Arran Cudbard-Bell [Mon, 18 Jan 2021 13:33:06 +0000 (13:33 +0000)]
We'll fix this another way...
Nicolas C [Mon, 18 Jan 2021 13:14:07 +0000 (14:14 +0100)]
Better error messages in fr_pair_list_afrom_substr (#3852)
* Better error messages in fr_pair_list_afrom_substr
Helps seeing what we did wrong when using dhcpclient or radclient
* Not const
Terry Burton [Mon, 18 Jan 2021 01:10:11 +0000 (01:10 +0000)]
CI: Build with PCRE 2; Add "LIBS_ALT:yes" target; Tidier naming (#3851)
Now build using PCRE 2 by default.
LIBS_ALT: yes will replace default libraries as follows:
PCRE 2 -> PCRE 1
MIT Kerberos -> Heimdal Kerberos
Add custom names for the matrix entries so that the workflow output is easier
to interpret. Reorganise the entries so that the variable components are first.
Terry Burton [Mon, 18 Jan 2021 01:09:58 +0000 (01:09 +0000)]
Debian packaging: Use PCRE 2 in preference to PCRE 1, if known (#3850)
James Jones [Mon, 18 Jan 2021 01:09:33 +0000 (19:09 -0600)]
Turn floating point comparison into memcmp() to silence compiler (#3849)
In numerical methods, it is indeed a bad idea to compare
floating point numbers for equality, but this just tests
that fr_dbuff_in() and fr_dbuff_out(), barring lack of space
breaking fr_dbuff_in(), are inverses for float and double.
No floating point arithmetic involved, so it's reasonable to
compare the raw bits. We'll use memcmp() to (1) emphasize that
and (2) keep the compiler from warning about comparing floating
point values for equality.
Arran Cudbard-Bell [Sun, 17 Jan 2021 23:52:34 +0000 (23:52 +0000)]
No longer need to do anything special with state when detaching
Arran Cudbard-Bell [Sun, 17 Jan 2021 23:39:29 +0000 (23:39 +0000)]
Remove (now) invalid asserts
Arran Cudbard-Bell [Sun, 17 Jan 2021 22:33:45 +0000 (22:33 +0000)]
Don't leak session state
Arran Cudbard-Bell [Sun, 17 Jan 2021 22:05:10 +0000 (22:05 +0000)]
Bad pointer to terminal elements
Arran Cudbard-Bell [Sun, 17 Jan 2021 21:41:21 +0000 (21:41 +0000)]
re-enable this one again for now... they can often be fixed and it's causing issues with GCC
Alan T. DeKok [Sun, 17 Jan 2021 20:03:41 +0000 (15:03 -0500)]
use clang diagnostics only on clang
Alan T. DeKok [Sun, 17 Jan 2021 18:37:13 +0000 (13:37 -0500)]
clear out array when it's freed
Alan T. DeKok [Sun, 17 Jan 2021 18:30:45 +0000 (13:30 -0500)]
use TEST_BIN_DIR instead of hard-coded paths
Alan T. DeKok [Sun, 17 Jan 2021 18:24:06 +0000 (13:24 -0500)]
raw attributes have no tag
Alan T. DeKok [Sun, 17 Jan 2021 15:44:00 +0000 (10:44 -0500)]
clear out data structures that we're using
Alan T. DeKok [Sun, 17 Jan 2021 15:26:29 +0000 (10:26 -0500)]
dhcpv4 encode / decode does not support array of dns_label
once the code is fixed, we will update the dictionaries
Alan T. DeKok [Sun, 17 Jan 2021 15:20:46 +0000 (10:20 -0500)]
don't hit assert if the dictionary parser is broken
we should also fix the dictionary parser...
Alan T. DeKok [Sun, 17 Jan 2021 15:19:15 +0000 (10:19 -0500)]
zero-length attributes are raw, not asserts
Alan T. DeKok [Sun, 17 Jan 2021 15:19:00 +0000 (10:19 -0500)]
this is a DNS label, not an array
Alan T. DeKok [Sun, 17 Jan 2021 15:12:05 +0000 (10:12 -0500)]
check for empty packets
Alan T. DeKok [Sun, 17 Jan 2021 15:01:44 +0000 (10:01 -0500)]
add "test.fuzzer" target in order to do fuzzer tests
Alan T. DeKok [Sun, 17 Jan 2021 14:50:18 +0000 (09:50 -0500)]
be more forgiving about fuzzer flags
Alan T. DeKok [Sun, 17 Jan 2021 14:39:37 +0000 (09:39 -0500)]
fix typo
Alan T. DeKok [Sun, 17 Jan 2021 14:38:29 +0000 (09:38 -0500)]
remove extra semicolon
Alan T. DeKok [Sun, 17 Jan 2021 14:22:46 +0000 (09:22 -0500)]
use pre-existing DIAG_OFF macros
Alan T. DeKok [Sun, 17 Jan 2021 14:07:34 +0000 (09:07 -0500)]
turn off clang-specific complaints. The code is fine
Alan T. DeKok [Sun, 17 Jan 2021 14:07:14 +0000 (09:07 -0500)]
wrap macro in `do { ... } while (0)` to shut up compiler
Alan T. DeKok [Sun, 17 Jan 2021 13:38:29 +0000 (08:38 -0500)]
better error messages
Alan T. DeKok [Sun, 17 Jan 2021 13:38:09 +0000 (08:38 -0500)]
add missing "break".
so we don't hit unknown packet type
Terry Burton [Fri, 15 Jan 2021 00:02:41 +0000 (00:02 +0000)]
CI: Add MacOS build to the matrix
No fixtures, just ordinary "make test" for now.
We disable clang static analysis scan for now to keep the runtime within that
of the other jobs.
Terry Burton [Sun, 17 Jan 2021 01:15:30 +0000 (01:15 +0000)]
CI: More complete FreeBSD build and tests
Terry Burton [Sat, 16 Jan 2021 21:52:28 +0000 (21:52 +0000)]
Tests: There's no bash at /bin/bash on FreeBSD
Alan T. DeKok [Sun, 17 Jan 2021 12:51:16 +0000 (07:51 -0500)]
shut up compiler.
tho arguably we should be using memcmp() i.e. *bit* comparison
instead of "==", which is more *semantic* comparison
Alan T. DeKok [Sun, 17 Jan 2021 12:40:30 +0000 (07:40 -0500)]
shut up clang about system libraries
Alan T. DeKok [Sun, 17 Jan 2021 12:35:04 +0000 (07:35 -0500)]
variable is used
Arran Cudbard-Bell [Sun, 17 Jan 2021 01:17:16 +0000 (01:17 +0000)]
Missing break
Arran Cudbard-Bell [Sun, 17 Jan 2021 01:12:29 +0000 (01:12 +0000)]
Revert back to previous way of unsetting UNENCRYPTED flag even though it shouldn't be necessary?
Arran Cudbard-Bell [Sun, 17 Jan 2021 00:29:20 +0000 (00:29 +0000)]
Should be static
Arran Cudbard-Bell [Sun, 17 Jan 2021 00:19:13 +0000 (00:19 +0000)]
More float comparisons
Arran Cudbard-Bell [Sun, 17 Jan 2021 00:06:00 +0000 (00:06 +0000)]
More stupid flag set/unset warnings
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:28:30 +0000 (23:28 +0000)]
Extra comma
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:25:32 +0000 (23:25 +0000)]
Should be static
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:12:48 +0000 (23:12 +0000)]
Shouldn't be marked as inline
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:03:52 +0000 (23:03 +0000)]
Unused
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:03:47 +0000 (23:03 +0000)]
Fix more optional args
Arran Cudbard-Bell [Sat, 16 Jan 2021 23:03:36 +0000 (23:03 +0000)]
Should probably be exported
Arran Cudbard-Bell [Sat, 16 Jan 2021 22:53:01 +0000 (22:53 +0000)]
Make more things static
Arran Cudbard-Bell [Sat, 16 Jan 2021 22:52:54 +0000 (22:52 +0000)]
Bad comma