]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Wed, 25 Aug 2021 16:43:19 +0000 (11:43 -0500)]
Add cf_log_pwarn
Arran Cudbard-Bell [Wed, 25 Aug 2021 02:50:03 +0000 (21:50 -0500)]
Our version of unix time can't be negaitve
Arran Cudbard-Bell [Wed, 25 Aug 2021 02:37:28 +0000 (21:37 -0500)]
Simpler way of getting the serial
Arran Cudbard-Bell [Wed, 25 Aug 2021 02:37:00 +0000 (21:37 -0500)]
Correct uses of fr_time_from_utc
Arran Cudbard-Bell [Wed, 25 Aug 2021 02:11:57 +0000 (21:11 -0500)]
Verify certificates are valid
Check certificate validity periods explicitly for lib/tls users and rlm_cipher
Add %(<inst>_certificate:notBefore) and %(<inst>_certificate:notAfter)
Alan T. DeKok [Wed, 25 Aug 2021 12:57:48 +0000 (08:57 -0400)]
move actions to struct
Alan T. DeKok [Wed, 25 Aug 2021 12:54:42 +0000 (08:54 -0400)]
better error messages. Fixes #4204
Alan T. DeKok [Wed, 25 Aug 2021 12:26:35 +0000 (08:26 -0400)]
move actions to struct
Alan T. DeKok [Wed, 25 Aug 2021 12:23:23 +0000 (08:23 -0400)]
use typedef to simplify things
Alan T. DeKok [Wed, 25 Aug 2021 12:16:05 +0000 (08:16 -0400)]
this is "abinary"
ported manually from v3
Jorge Pereira [Wed, 25 Aug 2021 12:06:59 +0000 (09:06 -0300)]
Fix typo (#4208)
James Jones [Tue, 24 Aug 2021 13:10:34 +0000 (08:10 -0500)]
Add assert to reassure (or shut up) "make scan" (#4206)
James Jones [Tue, 24 Aug 2021 12:50:44 +0000 (07:50 -0500)]
Use CMP() in _trunk_connection_order_by_shortest_queue() (#4205)
Alan T. DeKok [Mon, 23 Aug 2021 20:27:47 +0000 (16:27 -0400)]
enforce response_window only for synchronous requests
Alan T. DeKok [Mon, 23 Aug 2021 20:22:41 +0000 (16:22 -0400)]
remove unused variables
Alan T. DeKok [Mon, 23 Aug 2021 19:49:59 +0000 (15:49 -0400)]
rework to respect response_window, etc.
Alan T. DeKok [Mon, 23 Aug 2021 18:41:25 +0000 (14:41 -0400)]
add and document response_window
Alan T. DeKok [Mon, 23 Aug 2021 18:17:14 +0000 (14:17 -0400)]
fix typo
Alan T. DeKok [Mon, 23 Aug 2021 17:19:47 +0000 (13:19 -0400)]
cleanups and corner conditions for deferred transitions
Alan T. DeKok [Mon, 23 Aug 2021 15:27:28 +0000 (11:27 -0400)]
add support for multiple deferred state transitions
Alan T. DeKok [Thu, 19 Aug 2021 14:23:51 +0000 (10:23 -0400)]
add signal handler
Terry Burton [Fri, 20 Aug 2021 18:49:54 +0000 (19:49 +0100)]
Data usage for MySQL: Run SP as invoker not definer (#4201)
Terry Burton [Fri, 20 Aug 2021 18:49:37 +0000 (19:49 +0100)]
Usage schema extension for SQLite: Typo (#4199)
Typo in
4c67d412956d2efed8b6ab4f216eb49abf3119c5 component of PR #3635.
Arran Cudbard-Bell [Thu, 19 Aug 2021 13:14:56 +0000 (08:14 -0500)]
Fix fr_machine_s
Arran Cudbard-Bell [Thu, 19 Aug 2021 02:54:05 +0000 (21:54 -0500)]
Revert "LSTs likely slower when used here"
This reverts commit
6a88d68e3e5c4398e7feee501a855ea84a57633d .
Arran Cudbard-Bell [Thu, 19 Aug 2021 01:13:47 +0000 (20:13 -0500)]
Free the heap
Arran Cudbard-Bell [Thu, 19 Aug 2021 00:44:33 +0000 (19:44 -0500)]
Needs to be alloced in hp not ctx
Arran Cudbard-Bell [Thu, 19 Aug 2021 00:04:52 +0000 (19:04 -0500)]
Shouldn't need to specify ctx... but maybe this will fix the leak?
Arran Cudbard-Bell [Wed, 18 Aug 2021 22:17:48 +0000 (17:17 -0500)]
Move array pop code to function and prevent it being inlined
Arran Cudbard-Bell [Wed, 18 Aug 2021 22:14:15 +0000 (17:14 -0500)]
Not sure how that got there...
Arran Cudbard-Bell [Wed, 18 Aug 2021 20:34:46 +0000 (15:34 -0500)]
Record first pop
James Jones [Wed, 18 Aug 2021 22:12:30 +0000 (17:12 -0500)]
"init" parameter isn't in LST alloc functions yet (#4195)
Arran Cudbard-Bell [Wed, 18 Aug 2021 18:33:26 +0000 (13:33 -0500)]
More fixes
Arran Cudbard-Bell [Wed, 18 Aug 2021 18:25:07 +0000 (13:25 -0500)]
Less confusing way of representing struct fr_heap_s **
Arran Cudbard-Bell [Wed, 18 Aug 2021 18:10:09 +0000 (13:10 -0500)]
LSTs likely faster here
Arran Cudbard-Bell [Wed, 18 Aug 2021 18:06:18 +0000 (13:06 -0500)]
Fix heap tests
Arran Cudbard-Bell [Wed, 18 Aug 2021 17:27:43 +0000 (12:27 -0500)]
LSTs likely slower when used here
This reverts commit
c59a612b93a55b130ea5fea768c76f6c661b17e0 .
Arran Cudbard-Bell [Wed, 18 Aug 2021 17:26:55 +0000 (12:26 -0500)]
Fix up queue cmp tests
Arran Cudbard-Bell [Wed, 18 Aug 2021 16:57:29 +0000 (11:57 -0500)]
Higher memory locality gives significant performance gains for small numbers of heap elements
Alan T. DeKok [Wed, 18 Aug 2021 15:51:35 +0000 (11:51 -0400)]
force link errors on OSX for unknown symbols
When linking binaries. So that we get errors at compile time,
and not at run-time, or even later in CI tests.
Arran Cudbard-Bell [Wed, 18 Aug 2021 15:36:50 +0000 (10:36 -0500)]
Link to the crypt library
Arran Cudbard-Bell [Wed, 18 Aug 2021 15:11:49 +0000 (10:11 -0500)]
Minor tweaks
Alan T. DeKok [Wed, 18 Aug 2021 14:42:45 +0000 (10:42 -0400)]
cast function pointers to void*
because we're using them as placeholders. We don't actually
care what they are
Alan T. DeKok [Wed, 18 Aug 2021 14:42:16 +0000 (10:42 -0400)]
use [] instead of [0]
Arran Cudbard-Bell [Wed, 18 Aug 2021 14:39:37 +0000 (09:39 -0500)]
Add lst/heap comprison test
Arran Cudbard-Bell [Wed, 18 Aug 2021 14:39:02 +0000 (09:39 -0500)]
Typo
Alan T. DeKok [Wed, 18 Aug 2021 14:06:55 +0000 (10:06 -0400)]
initial stab at generic state machine handler.
Alan T. DeKok [Mon, 16 Aug 2021 20:24:36 +0000 (16:24 -0400)]
should not have been committed
Arran Cudbard-Bell [Wed, 18 Aug 2021 03:05:57 +0000 (22:05 -0500)]
Move crypt code into rlm_pap
Arran Cudbard-Bell [Wed, 18 Aug 2021 02:59:30 +0000 (21:59 -0500)]
Unused
Arran Cudbard-Bell [Wed, 18 Aug 2021 02:57:42 +0000 (21:57 -0500)]
Fix indent
James Jones [Wed, 18 Aug 2021 02:27:00 +0000 (21:27 -0500)]
Switch to unsigned type for fr_lst_index_t (#4190)
* Switch to unsigned type for fr_lst_index_t
This involves adding one to the true index when storing it in an
item, and subtracting one when reading it for use.
* Switch to unsigned type for fr_lst_index_t
This involves adding one to the true index when storing it in an
item, and subtracting one when reading it for use.
* Switch to unsigned type for fr_lst_index_t
This involves adding one to the true index when storing it in an
item, and subtracting one when reading it for use.
Arran Cudbard-Bell [Wed, 18 Aug 2021 02:26:00 +0000 (21:26 -0500)]
Re-enable support for TLS 1.0 fixes #4191
Arran Cudbard-Bell [Wed, 18 Aug 2021 00:34:02 +0000 (19:34 -0500)]
Allocate the stack as part of the lst structure
Arran Cudbard-Bell [Tue, 17 Aug 2021 21:01:18 +0000 (16:01 -0500)]
Control the initial number of elements in a heap
Arran Cudbard-Bell [Tue, 17 Aug 2021 19:10:46 +0000 (14:10 -0500)]
Sprinkle more unlikelys
Arran Cudbard-Bell [Tue, 17 Aug 2021 19:06:23 +0000 (14:06 -0500)]
Sprinkle unlikely
Arran Cudbard-Bell [Tue, 17 Aug 2021 17:50:17 +0000 (12:50 -0500)]
Pre-allocate LST stack memory
This is mostly to ensure good locality
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:57:36 +0000 (11:57 -0500)]
Do talloc type checking if we have one set
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:52:47 +0000 (11:52 -0500)]
Use a consistent type for representing stack indexes
No reason for stack indexes to be signed. There are no negative indexes.
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:27:16 +0000 (11:27 -0500)]
Add explicit bracketing
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:22:05 +0000 (11:22 -0500)]
Minor style fixes
Doxygen function headers must start with /**
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:21:43 +0000 (11:21 -0500)]
Mark up iterators as nonnull and move doxygen comments to source file
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:20:49 +0000 (11:20 -0500)]
Number of elements and size tracking should be unsigned integers
Arran Cudbard-Bell [Tue, 17 Aug 2021 16:10:03 +0000 (11:10 -0500)]
Convert preprocess macros to static inlines
The emitted code should be identical, but we get added type checking, and it's easier to see what's being passed into the macros/functions and what they return.
We also remove macros used for the LHS of assignments, as that's just weird.
Arran Cudbard-Bell [Tue, 17 Aug 2021 15:33:53 +0000 (10:33 -0500)]
Cast UINTPTR_MAX
James Jones [Tue, 17 Aug 2021 15:14:02 +0000 (10:14 -0500)]
Correct success returns for fr_lst_{insert, extract}() (#4187)
This change makes them consistent with the heap functions,
simplifying migration from heap to LST.
Arran Cudbard-Bell [Tue, 17 Aug 2021 14:50:51 +0000 (09:50 -0500)]
Fix some style issues in the lst tests
Arran Cudbard-Bell [Tue, 17 Aug 2021 14:50:24 +0000 (09:50 -0500)]
Index heaps at 1 instead of 0
Allows us to have more elements in the heap, and fixes an issue where the element inserted at index 0 may have been counted as not inserted.
Arran Cudbard-Bell [Tue, 17 Aug 2021 13:09:36 +0000 (08:09 -0500)]
Free all the things
Arran Cudbard-Bell [Tue, 17 Aug 2021 02:28:59 +0000 (21:28 -0500)]
Item to insert shouldn't be NULL
Arran Cudbard-Bell [Tue, 17 Aug 2021 01:41:48 +0000 (20:41 -0500)]
Verify pop returns heap elements in order
Alan T. DeKok [Mon, 16 Aug 2021 20:23:01 +0000 (16:23 -0400)]
some do pass NULL as the first parameter
Alan T. DeKok [Mon, 16 Aug 2021 20:21:30 +0000 (16:21 -0400)]
Revert "don't pass NULL to insert after/before"
This reverts commit
320e4d4fd63497a0524ecaa3584eca89d06b6e5f .
Alan T. DeKok [Mon, 16 Aug 2021 19:37:48 +0000 (15:37 -0400)]
don't pass NULL to insert after/before
the caller should instead use insert head/tail
Arran Cudbard-Bell [Mon, 16 Aug 2021 19:02:12 +0000 (14:02 -0500)]
store dlist element count as an unsigned int
Arran Cudbard-Bell [Mon, 16 Aug 2021 19:01:16 +0000 (14:01 -0500)]
Fix max outstanding type
James Jones [Tue, 10 Aug 2021 22:52:13 +0000 (17:52 -0500)]
Change type of lst_id to use the fr_lst_index_t typedef
James Jones [Tue, 10 Aug 2021 19:13:13 +0000 (14:13 -0500)]
Convert src/lib/io/master.c from heap to LST
Arran Cudbard-Bell [Mon, 16 Aug 2021 17:46:22 +0000 (12:46 -0500)]
Better timer structure docs
Arran Cudbard-Bell [Tue, 3 Aug 2021 21:50:45 +0000 (15:50 -0600)]
typo
Arran Cudbard-Bell [Tue, 3 Aug 2021 21:49:45 +0000 (15:49 -0600)]
s/maximum/max/
Alan T. DeKok [Mon, 16 Aug 2021 01:19:58 +0000 (21:19 -0400)]
use dict_freeradius for Packet-* attributes
Alan T. DeKok [Sun, 15 Aug 2021 20:40:39 +0000 (16:40 -0400)]
add tests for arrays of DNS labels in a struct
Alan T. DeKok [Sun, 15 Aug 2021 20:37:30 +0000 (16:37 -0400)]
dns labels have their own array decoder
Alan T. DeKok [Sun, 15 Aug 2021 20:31:10 +0000 (16:31 -0400)]
correct structure as per RFC 6731 Section 4.2
James Jones [Sun, 15 Aug 2021 13:20:00 +0000 (08:20 -0500)]
Remove reference to queue from comments on {runnable, time_order}_id (#4182)
Examining the code appears to indicate that the only places these
members are used is in a heap.
Matthew Newton [Sat, 14 Aug 2021 22:41:12 +0000 (23:41 +0100)]
Add Debian 11 Bullseye to CI
Matthew Newton [Sat, 14 Aug 2021 22:28:04 +0000 (23:28 +0100)]
Dockerfile and package updates for Debian 11 Bullseye
Alan T. DeKok [Fri, 13 Aug 2021 19:24:39 +0000 (15:24 -0400)]
point to correct byte for error
Alan T. DeKok [Fri, 13 Aug 2021 17:49:25 +0000 (13:49 -0400)]
correct encoding of vendor-class "data" field
Which should be repeated 16-bit length + data
The decoder already worked correctly, once the dictionary
was updated.
Alan T. DeKok [Fri, 13 Aug 2021 17:42:18 +0000 (13:42 -0400)]
arrays of string/octets are encoded as 16-bit length + data
which are fixed size.
And, clamp the maximum encoded value to 16-bits, too
Alan T. DeKok [Fri, 13 Aug 2021 17:32:47 +0000 (13:32 -0400)]
rearrange the order of checking for flag validity
so that the protocol-specific attr_valid() callback can muck with
the flags before the main validation function gets them.
Alan T. DeKok [Fri, 13 Aug 2021 15:49:14 +0000 (11:49 -0400)]
allow dns label decoding to stop at the 0x00 byte
so that when we're decoding DNS labels in a struct, or in an
array, we can stop when the labels end. The next field is then
whatever happens to be after the DNS labels.
However, when they're encoded in an option, they must exactly
fill the option, otherwise it's an error
Alan T. DeKok [Fri, 13 Aug 2021 15:15:16 +0000 (11:15 -0400)]
print offsets for large numbers
because fonts can only go so small
Alan T. DeKok [Fri, 13 Aug 2021 15:08:19 +0000 (11:08 -0400)]
set "is_known_width" flag for DNS labels
which will let us encode DNS labels inside of structs, as that
is used in DHCPv6
Alan T. DeKok [Fri, 13 Aug 2021 14:51:32 +0000 (10:51 -0400)]
ensure that we don't ask for length of DNS labels
the length is not fixed, but is instead dependent on the length
of the label. However, unlike normal "string" attributes, the
field is delimited, so we can figure out the length just by
looking at the data.
Alan T. DeKok [Fri, 13 Aug 2021 13:53:30 +0000 (09:53 -0400)]
allow string[n] to be used inside of structs, too
Jorge Pereira [Fri, 13 Aug 2021 13:01:37 +0000 (10:01 -0300)]
Fix typo (#4179)
Alan T. DeKok [Fri, 13 Aug 2021 00:41:14 +0000 (20:41 -0400)]
turn dlopen into a soft fail