]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Mon, 11 Apr 2022 17:43:14 +0000 (12:43 -0500)]
Switch the compiler used for coverity builds
Arran Cudbard-Bell [Mon, 11 Apr 2022 16:41:28 +0000 (11:41 -0500)]
More ndebug fixes
Nick Porter [Mon, 11 Apr 2022 16:40:08 +0000 (17:40 +0100)]
Match function to prototype
Arran Cudbard-Bell [Sun, 10 Apr 2022 20:19:21 +0000 (15:19 -0500)]
Don't print file loacations if this isn't a debug build
Arran Cudbard-Bell [Sat, 9 Apr 2022 19:56:10 +0000 (14:56 -0500)]
Allow atexit handlers to return error codes so we can track when they fail
Arran Cudbard-Bell [Sat, 9 Apr 2022 17:44:46 +0000 (12:44 -0500)]
Avoid infinite loop in module_root(), oops...
Arran Cudbard-Bell [Sat, 9 Apr 2022 17:36:20 +0000 (12:36 -0500)]
Cleanup more atexit stuff, and explicitly trigger atexit handlers to avoid ordering issues with libraries which install their own atexit handlers
Arran Cudbard-Bell [Sat, 9 Apr 2022 05:19:53 +0000 (00:19 -0500)]
Add the prefix of the root module onto the name of thread instance data
Arran Cudbard-Bell [Sat, 9 Apr 2022 02:22:58 +0000 (21:22 -0500)]
Use separate module lists for proto, process and rlm modules
James Jones [Fri, 8 Apr 2022 19:52:08 +0000 (14:52 -0500)]
Enforce that [length], if present, must be at the end of the type field (#4457)
Alan T. DeKok [Fri, 8 Apr 2022 19:44:55 +0000 (15:44 -0400)]
remove cursor code
now that it's wholly unused
Alan T. DeKok [Fri, 8 Apr 2022 19:41:55 +0000 (15:41 -0400)]
remove use of cursors
Alan T. DeKok [Fri, 8 Apr 2022 19:36:47 +0000 (15:36 -0400)]
remove dependency on fr_cursor_t
we're not going to fix everything right now, but this is easy,
fast, and not wrong.
Alan T. DeKok [Fri, 8 Apr 2022 17:32:17 +0000 (13:32 -0400)]
dcursor.c is only for testing
Alan T. DeKok [Fri, 8 Apr 2022 17:31:59 +0000 (13:31 -0400)]
update comments to clarify
Alan T. DeKok [Fri, 8 Apr 2022 14:28:38 +0000 (10:28 -0400)]
correct list checks
we're checking if the input VP is already in ANY list, and not
if it's in the TARGET list.
For insert after/before, we're checking if the "pos" item is
in the TARGET list, and the to_insert item is not in ANY list
Alan T. DeKok [Fri, 8 Apr 2022 14:28:18 +0000 (10:28 -0400)]
expose "in A list" API, versus "in THIS list" API
and fix up tlist functions with better checks for "in a list",
versus "in this list", which we can't do for dlists.
Alan T. DeKok [Fri, 8 Apr 2022 14:27:17 +0000 (10:27 -0400)]
comment out broken code which re-wrote OIDs to be broken.
We now trust the author to write the OIDs correctly, and don't
try to "fix" them.
Alan T. DeKok [Fri, 8 Apr 2022 14:00:00 +0000 (10:00 -0400)]
add a few things, pretty-print the strings, sort alphabetically
James Jones [Fri, 8 Apr 2022 19:23:44 +0000 (14:23 -0500)]
change error message to match the types that can have explicit length (#4456)
Alan T. DeKok [Fri, 8 Apr 2022 13:17:00 +0000 (09:17 -0400)]
add APIs to get parent list / vp from an input vp
which means that the caller has to do less tracking, if it wants
Nick Porter [Fri, 8 Apr 2022 12:54:10 +0000 (13:54 +0100)]
Remove unused variable
Nick Porter [Fri, 8 Apr 2022 10:13:38 +0000 (11:13 +0100)]
Define PAIR_VERIFY_WITH_LIST for non WITH_VERIFY_PTR build
Alan T. DeKok [Fri, 8 Apr 2022 01:27:46 +0000 (21:27 -0400)]
don't "return" in function returning void
Alan T. DeKok [Fri, 8 Apr 2022 00:19:46 +0000 (20:19 -0400)]
don't even bother checking for operators
Alan T. DeKok [Fri, 8 Apr 2022 00:00:04 +0000 (20:00 -0400)]
argument 2 can be NULL
Alan T. DeKok [Thu, 7 Apr 2022 23:04:53 +0000 (19:04 -0400)]
add NDEBUG_UNUSED
Alan T. DeKok [Thu, 7 Apr 2022 22:58:00 +0000 (18:58 -0400)]
just hard-code '=' in output
Alan T. DeKok [Thu, 7 Apr 2022 22:09:54 +0000 (18:09 -0400)]
op is always T_OP_EQ
Alan T. DeKok [Thu, 7 Apr 2022 22:08:39 +0000 (18:08 -0400)]
no need to set vp->op in map callback functions
map_to_request() ignores it, and just uses map->op
Alan T. DeKok [Thu, 7 Apr 2022 21:59:51 +0000 (17:59 -0400)]
always print '=' for vp->op
nothing here should need it or use it.
Alan T. DeKok [Thu, 7 Apr 2022 21:56:40 +0000 (17:56 -0400)]
don't smash vp->op to T_OP_EQ
the rest of the code in the server should take care of either
not setting vp->op, or setting it to T_OP_EQ after using the
operator.
Alan T. DeKok [Thu, 7 Apr 2022 21:56:16 +0000 (17:56 -0400)]
we don't need to set vp->op
it's set by fr_pair_afrom_da()
Alan T. DeKok [Thu, 7 Apr 2022 21:52:17 +0000 (17:52 -0400)]
this is no longer needed
Alan T. DeKok [Thu, 7 Apr 2022 21:06:59 +0000 (17:06 -0400)]
imove fr_pair_list_t to use tlists
and do various other cleanups and fixes as necessitated by that.
* doing actual work inside of the dcursor insert/remove callbacks
* adding fr_pair_list_t* to fr_pair_verify()
* adding PAIR_VERIFY_WITH_LIST() to verify that a VP is in a list
* ensure that pair_init_from_da() is always used, instead of
manually doing things in multiple places
Alan T. DeKok [Thu, 7 Apr 2022 20:55:58 +0000 (16:55 -0400)]
remove unnecessary struct/union
this doesn't change anything other than formatting
Alan T. DeKok [Thu, 7 Apr 2022 20:54:24 +0000 (16:54 -0400)]
convert error-prone code to inline typed functions
James Jones [Thu, 7 Apr 2022 20:59:42 +0000 (15:59 -0500)]
Add VALUEs for Datum in Geoconf-Option and change test to match (#4454)
James Jones [Thu, 7 Apr 2022 20:35:51 +0000 (15:35 -0500)]
Use SET_FLAG(subtype) rather than explicitly writing the equivalent code (#4453)
It *may* be that the explicit version is there to avoid the compiler
kvetching about a needless bit++ at the end. If that's the case, one
can switch to initializing bit to -1 and preincrementing rather than
postincrementing.
James Jones [Thu, 7 Apr 2022 20:33:28 +0000 (15:33 -0500)]
Convert GeoConf-Option (code 123) to struct, with test (#4450)
James Jones [Thu, 7 Apr 2022 20:33:18 +0000 (15:33 -0500)]
Convert GeoConf-Civic to use struct and add test (#4449)
James Jones [Thu, 7 Apr 2022 20:33:11 +0000 (15:33 -0500)]
Add test for RFC 3495; this exercises nested TLVs (#4448)
Arran Cudbard-Bell [Thu, 7 Apr 2022 20:28:07 +0000 (15:28 -0500)]
Typo
Arran Cudbard-Bell [Thu, 7 Apr 2022 19:34:27 +0000 (14:34 -0500)]
Use the standard instantiate bootstrap callbacks for proto and proc modules
Alan T. DeKok [Thu, 7 Apr 2022 18:03:58 +0000 (14:03 -0400)]
use offset from value box, not from datum
Alan T. DeKok [Thu, 7 Apr 2022 17:42:58 +0000 (13:42 -0400)]
copy at the correct offset
Alan T. DeKok [Thu, 7 Apr 2022 14:02:49 +0000 (10:02 -0400)]
might as well make this "const"
Alan T. DeKok [Wed, 6 Apr 2022 02:43:31 +0000 (22:43 -0400)]
use API instead of poking fields directly
Alan T. DeKok [Tue, 5 Apr 2022 22:35:29 +0000 (18:35 -0400)]
make the dlist offset point to the parent structure
instead of pointing to the fr_tlist_t. That way we can just
use the dlist functions, and get the enclosing structure.
As a bonus, it's also a lot less code.
Alan T. DeKok [Tue, 5 Apr 2022 22:26:46 +0000 (18:26 -0400)]
update TLIST macro functions
Alan T. DeKok [Tue, 5 Apr 2022 22:25:32 +0000 (18:25 -0400)]
add function to get tlist from dlist head
because a lot of dlist callbacks give the dlist head, and we
sometimes wish to convert that back to a tlist
Alan T. DeKok [Tue, 5 Apr 2022 22:24:52 +0000 (18:24 -0400)]
remove unused / unneeded functions
Alan T. DeKok [Tue, 5 Apr 2022 22:22:27 +0000 (18:22 -0400)]
manually free the pair list, as done in the other test functions.
otherwise with the new tlist code, the pairs are freed via
the talloc_free_children() callback, and the PAIR_VERIFY() check
fails, because the pair is still in a list.
Alan T. DeKok [Tue, 5 Apr 2022 15:56:24 +0000 (11:56 -0400)]
allow "attribute" command to parse/print multiple pairs
Alan T. DeKok [Tue, 5 Apr 2022 14:46:15 +0000 (10:46 -0400)]
check for error before allocating the VP
which also makes the code a little cleaner.
Alan T. DeKok [Tue, 5 Apr 2022 14:44:26 +0000 (10:44 -0400)]
can't re-init structural types
where would the children go? Be freed?
Nick Porter [Wed, 6 Apr 2022 11:35:52 +0000 (12:35 +0100)]
Remove unused variables
Alan T. DeKok [Tue, 5 Apr 2022 14:27:21 +0000 (10:27 -0400)]
fix relative OIDs here, too
Alan T. DeKok [Tue, 5 Apr 2022 14:26:45 +0000 (10:26 -0400)]
fully qualify the dictionary OIDs
Alan T. DeKok [Tue, 5 Apr 2022 14:13:42 +0000 (10:13 -0400)]
add fr_tlist_talloc_free_to_tail()
similar to fr_dlist_talloc_free_to_tail()
Alan T. DeKok [Tue, 5 Apr 2022 14:04:09 +0000 (10:04 -0400)]
add fr_pair_list_dlist_head()
and use it so that no one looks at internal fields
Alan T. DeKok [Tue, 5 Apr 2022 13:59:36 +0000 (09:59 -0400)]
remove unused lavel
Alan T. DeKok [Tue, 5 Apr 2022 13:54:13 +0000 (09:54 -0400)]
update fr_pair_replace() and callers
it was previously never called. And documented as freeing the
input VP, even though it didn't. And documented as "finding"
a VP, even though it didn't.
Use fr_pair_replace() in pairmove() functions, instead of having
them poke the dlists themselves.
Alan T. DeKok [Tue, 5 Apr 2022 13:14:20 +0000 (09:14 -0400)]
use fewer memset
new VPs are always initialized to zero
Alan T. DeKok [Tue, 5 Apr 2022 12:54:05 +0000 (08:54 -0400)]
less code == better code
Alan T. DeKok [Tue, 5 Apr 2022 12:45:44 +0000 (08:45 -0400)]
fix fr_pair_reinit_from_da()
it may be called with a VP which has been created via
fr_pair_alloc_null(), in which case vp->da is NULL.
Alan T. DeKok [Tue, 5 Apr 2022 12:46:40 +0000 (08:46 -0400)]
use fr_pair_init_null() instead of memset()
Alan T. DeKok [Tue, 5 Apr 2022 12:31:14 +0000 (08:31 -0400)]
add fr_pair_init_null()
because memset(vp, 0) is no longer appropriate
Alan T. DeKok [Tue, 5 Apr 2022 12:24:08 +0000 (08:24 -0400)]
use "free to tail", and don't keep tmp_list
Alan T. DeKok [Mon, 4 Apr 2022 19:29:23 +0000 (15:29 -0400)]
add "free to tail" function
Nick Porter [Tue, 5 Apr 2022 12:53:11 +0000 (13:53 +0100)]
Add apt retry configuration to second half of debian CI tests
Nick Porter [Tue, 5 Apr 2022 11:06:08 +0000 (12:06 +0100)]
Use correct apt option for retries
Alan T. DeKok [Mon, 4 Apr 2022 19:07:28 +0000 (15:07 -0400)]
minor typos and fixes
Alan T. DeKok [Mon, 4 Apr 2022 19:00:33 +0000 (15:00 -0400)]
add dlist_head function
Alan T. DeKok [Mon, 4 Apr 2022 18:55:14 +0000 (14:55 -0400)]
typo
Alan T. DeKok [Mon, 4 Apr 2022 13:21:44 +0000 (09:21 -0400)]
typo
Alan T. DeKok [Mon, 4 Apr 2022 12:43:11 +0000 (08:43 -0400)]
remove unused macros
Alan T. DeKok [Mon, 4 Apr 2022 12:42:30 +0000 (08:42 -0400)]
pair dlist functions are internal, not public
in preparation for changing the internal structure of pair lists
Alan T. DeKok [Fri, 1 Apr 2022 14:52:57 +0000 (10:52 -0400)]
add note about variable length time specs
Nick Porter [Fri, 1 Apr 2022 14:31:36 +0000 (15:31 +0100)]
Fix handling of affected_rows for FreeTDS SQL driver
After all the required calls to ct_results(), the internal counter of
affected rows is reset. So, retrieve it whilst still available and
store for later use.
Nick Porter [Fri, 1 Apr 2022 14:30:22 +0000 (15:30 +0100)]
Add AcctUpdateTime and AcctInterval to MS SQL accounting tables
To bring them in line with other databases
Arran Cudbard-Bell [Fri, 1 Apr 2022 21:13:09 +0000 (15:13 -0600)]
Try *really* hard to pull down packages
github-actions[bot] [Fri, 1 Apr 2022 09:37:16 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dhcpv4.tar
github-actions[bot] [Fri, 1 Apr 2022 09:37:13 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/radius.tar
github-actions[bot] [Fri, 1 Apr 2022 09:37:10 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/util.tar
github-actions[bot] [Fri, 1 Apr 2022 09:37:07 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/tftp.tar
github-actions[bot] [Fri, 1 Apr 2022 09:37:04 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/vmps.tar
github-actions[bot] [Fri, 1 Apr 2022 09:37:01 +0000 (09:37 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/tacacs.tar
github-actions[bot] [Fri, 1 Apr 2022 09:36:57 +0000 (09:36 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dns.tar
github-actions[bot] [Fri, 1 Apr 2022 09:36:55 +0000 (09:36 +0000)]
Scheduled fuzzing: Update src/tests/fuzzer-corpus/dhcpv6.tar
Alan T. DeKok [Thu, 31 Mar 2022 18:03:33 +0000 (14:03 -0400)]
correct references and numbers
apparently the formatter isn't quite perfect
Arran Cudbard-Bell [Thu, 31 Mar 2022 17:10:25 +0000 (11:10 -0600)]
Bump the timeout
This'll temporarily fix CI, but we should try and figure out exactly what's happening.
Arran Cudbard-Bell [Thu, 31 Mar 2022 17:09:09 +0000 (11:09 -0600)]
Return NULL on error
Alan T. DeKok [Thu, 31 Mar 2022 16:45:43 +0000 (12:45 -0400)]
run through the formatting script
which updates the copyright date, and changes whitespace
Alan T. DeKok [Thu, 31 Mar 2022 16:44:41 +0000 (12:44 -0400)]
this encoded as DNS
Arran Cudbard-Bell [Thu, 31 Mar 2022 16:35:02 +0000 (10:35 -0600)]
Check for NULL heap pointers
Arran Cudbard-Bell [Thu, 31 Mar 2022 16:12:05 +0000 (10:12 -0600)]
Remove one layer of heap derefs and shrink heaps as elements are removed
Alan T. DeKok [Thu, 31 Mar 2022 16:04:29 +0000 (12:04 -0400)]
update and rearrange
to reset TLV numbers
Alan T. DeKok [Thu, 31 Mar 2022 14:29:35 +0000 (10:29 -0400)]
the first 2 fields are mandatory, so we might as well define them :(
Alan T. DeKok [Thu, 31 Mar 2022 14:21:52 +0000 (10:21 -0400)]
these are "bool encode=exists", like Rapid-Commit