]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Thu, 14 May 2020 22:17:14 +0000 (17:17 -0500)]
radsniff: Fix crash on exit.
Arran Cudbard-Bell [Thu, 14 May 2020 20:29:08 +0000 (15:29 -0500)]
Add HEX_MARKER macros
Arran Cudbard-Bell [Thu, 14 May 2020 20:23:13 +0000 (15:23 -0500)]
radsniff: Fix event expiry, and "Waking up in" debug
Arran Cudbard-Bell [Thu, 14 May 2020 19:53:20 +0000 (14:53 -0500)]
event: Only call talloc_get_type_abort when we know we have an event to check *sigh*
Arran Cudbard-Bell [Thu, 14 May 2020 18:26:08 +0000 (13:26 -0500)]
dbuff: Move the condition inside of the compound literal in FR_DBUFF_RESERVE
James Jones [Thu, 14 May 2020 18:25:40 +0000 (13:25 -0500)]
Correct FR_DBUF_MAX macro and add tests for it. (#3452)
Arran Cudbard-Bell [Thu, 14 May 2020 18:17:50 +0000 (13:17 -0500)]
event: Tidy up report output, and stop reports being quite so badly interleaved
Arran Cudbard-Bell [Thu, 14 May 2020 17:46:04 +0000 (12:46 -0500)]
event: Fix various issues with reporting, closes #3451
Arran Cudbard-Bell [Thu, 14 May 2020 16:49:00 +0000 (11:49 -0500)]
s/rbtree_create/rbtree_alloc/ s/heap_create/heap_alloc/
Arran Cudbard-Bell [Thu, 14 May 2020 16:44:28 +0000 (11:44 -0500)]
event: report function - remove magic numbers, fix build issue
Alan T. DeKok [Thu, 14 May 2020 16:17:47 +0000 (12:17 -0400)]
add information on timers in the heap
Jorge Pereira [Thu, 14 May 2020 13:59:47 +0000 (10:59 -0300)]
heap: Typo (#3450)
Arran Cudbard-Bell [Wed, 13 May 2020 20:49:53 +0000 (15:49 -0500)]
dbuff: Dedup dbuff code
James Jones [Wed, 13 May 2020 19:07:04 +0000 (14:07 -0500)]
Add the ability to move signed integers into dbuffs. (#3446)
Alan T. DeKok [Wed, 13 May 2020 14:26:52 +0000 (10:26 -0400)]
add missing FALL-THROUGH
Arran Cudbard-Bell [Tue, 12 May 2020 21:07:25 +0000 (16:07 -0500)]
trunk: Don't call trunk_manage directly
Arran Cudbard-Bell [Tue, 12 May 2020 18:52:01 +0000 (13:52 -0500)]
trunk: re-calculate the ratios when we're rebalancing
Arran Cudbard-Bell [Tue, 12 May 2020 17:36:51 +0000 (12:36 -0500)]
trunk: Change the way we assert in trunk_requests_per_conn
Arran Cudbard-Bell [Tue, 12 May 2020 17:12:19 +0000 (12:12 -0500)]
trunk: Add comments
Arran Cudbard-Bell [Tue, 12 May 2020 17:07:07 +0000 (12:07 -0500)]
Revert "Sync conditions with trunk_manage"
This reverts commit
21da2ae9a510b95d10f1300730188ffbbc255dd1 .
Jorge Pereira [Tue, 12 May 2020 01:35:34 +0000 (22:35 -0300)]
trunk.c: s/_per_connnection/_per_connection/g
Arran Cudbard-Bell [Tue, 12 May 2020 14:18:07 +0000 (09:18 -0500)]
Sync conditions with trunk_manage
Alan T. DeKok [Mon, 11 May 2020 21:30:31 +0000 (17:30 -0400)]
fix scan warnings
Alan T. DeKok [Mon, 11 May 2020 21:26:52 +0000 (17:26 -0400)]
shut up compiler about alignment
Arran Cudbard-Bell [Mon, 11 May 2020 18:35:44 +0000 (13:35 -0500)]
value: Don't zero out structures and then immdiately write data to them
...and as for uninitialised fields in compound literals containing garbage, c99 states that shouldn't be the case.
http://c0x.coding-guidelines.com/6.7.8.html 1677.
Alan T. DeKok [Mon, 11 May 2020 15:42:07 +0000 (11:42 -0400)]
if the dictionary has octets[n], then it's invalid for length != n
Matthew Newton [Mon, 11 May 2020 10:29:36 +0000 (11:29 +0100)]
Add ubuntu20 Dockerfiles
Alan T. DeKok [Sun, 10 May 2020 20:01:04 +0000 (16:01 -0400)]
cleanup
Alan T. DeKok [Sun, 10 May 2020 19:10:34 +0000 (15:10 -0400)]
one more UNUSED
Alan T. DeKok [Sun, 10 May 2020 16:27:53 +0000 (12:27 -0400)]
fix UNUSED
Alan T. DeKok [Sun, 10 May 2020 16:17:56 +0000 (12:17 -0400)]
updated comments
Alan T. DeKok [Sun, 10 May 2020 16:06:10 +0000 (12:06 -0400)]
for ICMPV6, also calculate checksum over the IPv6 pseudo-header
Alan T. DeKok [Sun, 10 May 2020 15:57:47 +0000 (11:57 -0400)]
add fr_ip6_pesudo_header_checksum()
Alan T. DeKok [Sun, 10 May 2020 15:08:02 +0000 (11:08 -0400)]
make rlm_icmp support ICMPv6
not (quite) tested yet. But ICMPv4 still works
Alan T. DeKok [Sun, 10 May 2020 14:04:11 +0000 (10:04 -0400)]
add fr_cap_net_raw() and use it
This function sets CAP_NET_RAW if possible.
Alan T. DeKok [Sun, 10 May 2020 12:57:12 +0000 (08:57 -0400)]
use fr_socket_bind() which does all necessary magic
Alan T. DeKok [Sun, 10 May 2020 12:56:36 +0000 (08:56 -0400)]
emulate SO_BINDTODEVICE on systems which do not support it
Alan T. DeKok [Sun, 10 May 2020 12:18:14 +0000 (08:18 -0400)]
allow fr_socket_bind() to take NULL for src_ipaddr.
In which case it just tries to bind to the interface via
system-agnostic methods
Alan T. DeKok [Sat, 9 May 2020 19:29:36 +0000 (15:29 -0400)]
added warning for unsupported configurations
Alan T. DeKok [Sat, 9 May 2020 19:26:48 +0000 (15:26 -0400)]
call destroy functions before talloc_free
Alan T. DeKok [Sat, 9 May 2020 19:04:53 +0000 (15:04 -0400)]
allow icmp to bind to a source IP, too
Alan T. DeKok [Sat, 9 May 2020 15:58:02 +0000 (11:58 -0400)]
add a unique counter
so that multiple requests can ping the same IP address without
interfering with each other.
Alan T. DeKok [Sat, 9 May 2020 15:36:56 +0000 (11:36 -0400)]
and HTML-ize this, too
Alan T. DeKok [Sat, 9 May 2020 15:36:31 +0000 (11:36 -0400)]
fix asciidoc typo
Alan T. DeKok [Sat, 9 May 2020 15:35:31 +0000 (11:35 -0400)]
typo
Alan T. DeKok [Sat, 9 May 2020 15:32:38 +0000 (11:32 -0400)]
Add ICMP "ping" support
Alan T. DeKok [Sat, 9 May 2020 15:25:11 +0000 (11:25 -0400)]
ensure that _all_ fields of vb_ip are initialized
otherwise the struct has random garbage in it, and then
fr_value_box_cmp() doesn't work. Because it just uses memcmp()
on the struct, and doesn't look at the fields.
Arran Cudbard-Bell [Sat, 9 May 2020 15:16:23 +0000 (10:16 -0500)]
No reason to be lazy...
Arran Cudbard-Bell [Sat, 9 May 2020 15:13:03 +0000 (10:13 -0500)]
tmpl: More macro renaming
Alan T. DeKok [Sat, 9 May 2020 12:47:20 +0000 (08:47 -0400)]
just use talloc. Fixes #3444
And use inst->filter for 'and %s', instead of the "filter"
Alan T. DeKok [Fri, 8 May 2020 19:26:10 +0000 (15:26 -0400)]
IP header length is in 32-bit words
Arran Cudbard-Bell [Fri, 8 May 2020 23:54:54 +0000 (18:54 -0500)]
tmpl: Fix other weirdly named accessors
Arran Cudbard-Bell [Fri, 8 May 2020 23:52:52 +0000 (18:52 -0500)]
tmpl: Fix type check macro name
Arran Cudbard-Bell [Fri, 8 May 2020 23:52:00 +0000 (18:52 -0500)]
tmpl: quiet clang
Arran Cudbard-Bell [Fri, 8 May 2020 21:35:35 +0000 (16:35 -0500)]
tmpl: Account for TMPL_TYPE_MAX in switch statements
Arran Cudbard-Bell [Fri, 8 May 2020 21:31:52 +0000 (16:31 -0500)]
tmpl: Fix consistency check
Arran Cudbard-Bell [Fri, 8 May 2020 21:24:05 +0000 (16:24 -0500)]
tmpl: Remove bad asserts
Arran Cudbard-Bell [Fri, 8 May 2020 21:20:45 +0000 (16:20 -0500)]
s/FR_TOKEN/fr_token_t/
Arran Cudbard-Bell [Fri, 8 May 2020 21:10:22 +0000 (16:10 -0500)]
one more...
Arran Cudbard-Bell [Fri, 8 May 2020 21:09:03 +0000 (16:09 -0500)]
rename more tmpl types to be consistent
Arran Cudbard-Bell [Fri, 8 May 2020 20:49:55 +0000 (15:49 -0500)]
TMPL_TYPE_XLAT is the normal state, TMPL_TYPE_XLAT_UNPARSED is the exceptional one
Arran Cudbard-Bell [Fri, 8 May 2020 19:43:46 +0000 (14:43 -0500)]
remove more proxy stuff
Alan T. DeKok [Fri, 8 May 2020 18:54:26 +0000 (14:54 -0400)]
include header file
Terry Burton [Fri, 8 May 2020 15:40:14 +0000 (16:40 +0100)]
cURL: Fix type conflation with curl_multi_add_handle()
Alan T. DeKok [Fri, 8 May 2020 18:12:23 +0000 (14:12 -0400)]
move to unlang_interpret_frame_talloc_ctx()
instead of unlang_module_frame_talloc_ctx()
the new function is more generic, and can be used in more places.
Alan T. DeKok [Fri, 8 May 2020 18:05:20 +0000 (14:05 -0400)]
add timeouts to child ctx
so that we can delete it when the frame is popped, or when the
xlat is resumed or cancelled.
There isn't really any good reason to keep the timeouts for
longer than that.
Terry Burton [Fri, 8 May 2020 12:30:38 +0000 (13:30 +0100)]
MySQL: Use correct syntax for creating the radius user
Arran Cudbard-Bell [Fri, 8 May 2020 18:07:27 +0000 (13:07 -0500)]
Start stripping proxy lists
Alan T. DeKok [Fri, 8 May 2020 17:27:54 +0000 (13:27 -0400)]
attribute 999 is invalid for RADIUS
there isn't much point in creating unknown attributes which
can't be sent on the wire.
Alan T. DeKok [Fri, 8 May 2020 14:16:46 +0000 (10:16 -0400)]
missed this
Alan T. DeKok [Fri, 8 May 2020 14:09:33 +0000 (10:09 -0400)]
decode attributes for unknown vendors
the code was there, but no test case
Alan T. DeKok [Fri, 8 May 2020 12:47:23 +0000 (08:47 -0400)]
rerun the tests if `radiusd.conf` changes, too
Alan T. DeKok [Fri, 8 May 2020 12:39:20 +0000 (08:39 -0400)]
add tests for actually sending invalid attributes
If an attribute is unknown, it's type MUST be octets. AND we can
encode it. In fact, we generally MUST encode it, as the admin
is trying to work around stupid vendor issues.
<cough> SUPERMICRO </cough>
Alan T. DeKok [Thu, 7 May 2020 22:20:13 +0000 (18:20 -0400)]
use standard names
Alan T. DeKok [Thu, 7 May 2020 19:43:59 +0000 (15:43 -0400)]
remove the "touch" file on error
Alan T. DeKok [Thu, 7 May 2020 19:35:28 +0000 (15:35 -0400)]
fix typo
Alan T. DeKok [Thu, 7 May 2020 19:23:28 +0000 (15:23 -0400)]
add output "touch" files for each test
which should permit inter-test dependencies
Arran Cudbard-Bell [Thu, 7 May 2020 19:24:06 +0000 (14:24 -0500)]
Rename fr_dict_parent_common
Arran Cudbard-Bell [Thu, 7 May 2020 19:04:51 +0000 (14:04 -0500)]
dlist: Add fr_dlist_talloc_reverse_free
Useful for freeing items which were parented by the previous item in the list.
Arran Cudbard-Bell [Thu, 7 May 2020 19:04:19 +0000 (14:04 -0500)]
Add talloc bstrdup
Arran Cudbard-Bell [Thu, 7 May 2020 19:04:02 +0000 (14:04 -0500)]
Unused signals
Alan T. DeKok [Thu, 7 May 2020 13:39:30 +0000 (09:39 -0400)]
tests which use radiusd / radmin / radclient depend on the binaries, too
James Jones [Wed, 6 May 2020 22:28:09 +0000 (17:28 -0500)]
Make the actual fr_value_box_to_network_dbuff() visible (#3435)
We also get rid of the internal protocol's dbuff wrapper around
fr_value_box_to_network(); we leave fr_value_box_to_network()
(itself now a wrapper around the dbuff version) in place so we can
move other protocols to dbuff one at a time.
Alan T. DeKok [Wed, 6 May 2020 22:18:43 +0000 (18:18 -0400)]
allow the module to process ARP replies, too
and it then doesn't reply, and doesn't run "send do-not-respond"
James Jones [Wed, 6 May 2020 14:22:45 +0000 (09:22 -0500)]
Make ethernet.h the sole definer of ethernet_header_t
Member name references are now consistent with the now-sole definition.
Jorge Pereira [Wed, 6 May 2020 19:22:08 +0000 (16:22 -0300)]
test.conf: Typo
Jorge Pereira [Wed, 6 May 2020 18:29:17 +0000 (15:29 -0300)]
:gear: Fix 'make test.radiusd-c'
It fixes the file stamp related to that test.
James Jones [Wed, 6 May 2020 19:09:58 +0000 (14:09 -0500)]
Turn fr_value_box_to_network() into wrapper around dbuff version (#3425)
Jorge Pereira [Wed, 6 May 2020 18:40:21 +0000 (15:40 -0300)]
:gear: Fix 'make test.modules' verbosity
Lets keep the same behavior for all tests/modules/$name targets
whitout that 'OK' message.
Alan T. DeKok [Wed, 6 May 2020 12:13:09 +0000 (08:13 -0400)]
fix typo.
We want to re-run the tests if the radmin binary changes
Alan T. DeKok [Tue, 5 May 2020 18:38:36 +0000 (14:38 -0400)]
move "copy data from request" to arp encoder
which is really where it belongs
Jorge Pereira [Wed, 6 May 2020 00:00:30 +0000 (21:00 -0300)]
Fix 'make test.digest' build
Jorge Pereira [Tue, 5 May 2020 23:57:47 +0000 (20:57 -0300)]
Fix 'make test.radmin' build
Jorge Pereira [Tue, 5 May 2020 23:34:35 +0000 (20:34 -0300)]
Fix 'make test.radclient' build
Jorge Pereira [Tue, 5 May 2020 23:34:04 +0000 (20:34 -0300)]
radiusd.mk: Fix missing .PHONY for 'make test.${name}.radiusd_start'
Arran Cudbard-Bell [Tue, 5 May 2020 22:38:35 +0000 (17:38 -0500)]
tmpl: Have tmpl_value "return" a pointer to the box
Makes it behave more like a function
Arran Cudbard-Bell [Tue, 5 May 2020 22:27:12 +0000 (17:27 -0500)]
tmpl: Switch to "getter" style tmpl macros in preparation for nested attribute references
Terry Burton [Tue, 5 May 2020 19:10:57 +0000 (20:10 +0100)]
SQL IP pools howto: Reference the no-skip-locked SP for MariaDB and old MySQL
Jorge Pereira [Tue, 5 May 2020 19:11:18 +0000 (16:11 -0300)]
Fix digest dictionary attributes
Arran Cudbard-Bell [Tue, 5 May 2020 20:50:46 +0000 (15:50 -0500)]
Typo