]> git.ipfire.org Git - thirdparty/freeradius-server.git/log
thirdparty/freeradius-server.git
6 years agolink the "certs" directory
Alan T. DeKok [Wed, 10 Jun 2020 01:46:15 +0000 (21:46 -0400)] 
link the "certs" directory

so that modules which need certificates can use them

6 years agoMinor spelling fixes in man files (#3490)
Alexander Scheel [Tue, 9 Jun 2020 18:12:05 +0000 (14:12 -0400)] 
Minor spelling fixes in man files (#3490)

Signed-off-by: Alexander Scheel <ascheel@redhat.com>
6 years agoTypo
Arran Cudbard-Bell [Tue, 9 Jun 2020 14:24:26 +0000 (09:24 -0500)] 
Typo

6 years agotrunk: Fix assert
Arran Cudbard-Bell [Tue, 9 Jun 2020 13:37:39 +0000 (08:37 -0500)] 
trunk: Fix assert

6 years agofix typo
Alan T. DeKok [Mon, 8 Jun 2020 19:10:01 +0000 (15:10 -0400)] 
fix typo

6 years agodon't busy-loop if there are child requests to run
Alan T. DeKok [Mon, 8 Jun 2020 18:52:57 +0000 (14:52 -0400)] 
don't busy-loop if there are child requests to run

6 years agoremove unused assignment
Alan T. DeKok [Mon, 8 Jun 2020 18:52:21 +0000 (14:52 -0400)] 
remove unused assignment

6 years agoupdated from the old raddb/certs/README
Alan T. DeKok [Mon, 8 Jun 2020 18:31:57 +0000 (14:31 -0400)] 
updated from the old raddb/certs/README

6 years agofirst draft of new radclient
Alan T. DeKok [Thu, 4 Jun 2020 19:28:47 +0000 (15:28 -0400)] 
first draft of new radclient

* doesn't parse "server:port"
* doesn't pass packet type to unit_test_module
* the debug output is horrible (i.e. full server)
* many command line options don't work
-c, -p, -n, -i, ...

6 years agoallow casting for FR_TYPE_COMBO_IP
Alan T. DeKok [Thu, 4 Jun 2020 19:16:55 +0000 (15:16 -0400)] 
allow casting for FR_TYPE_COMBO_IP

so that we can do

ipaddr = <ipv4addr>radius.example.com

instead of

ipv4addr = radius.example.com

it's just a nice little touch

6 years agoef should be passed when deleting. man kevent states kevents are identified by ident...
Arran Cudbard-Bell [Sun, 7 Jun 2020 22:13:27 +0000 (17:13 -0500)] 
ef should be passed when deleting.  man kevent states kevents are identified by ident/filter/udata tuple

6 years agoApparently the travis user can do binds even without CAP_NET_BIND_SERVICE
Arran Cudbard-Bell [Fri, 5 Jun 2020 23:50:04 +0000 (18:50 -0500)] 
Apparently the travis user can do binds even without CAP_NET_BIND_SERVICE

6 years agoBe more verbose
Arran Cudbard-Bell [Fri, 5 Jun 2020 23:22:06 +0000 (18:22 -0500)] 
Be more verbose

6 years agotrunk: Zero out all the fields...
Arran Cudbard-Bell [Fri, 5 Jun 2020 23:02:17 +0000 (18:02 -0500)] 
trunk: Zero out all the fields...

6 years agocap: dedup cap functions
Arran Cudbard-Bell [Fri, 5 Jun 2020 22:26:54 +0000 (17:26 -0500)] 
cap: dedup cap functions

6 years agoMore efficient field initialisation
Arran Cudbard-Bell [Fri, 5 Jun 2020 16:35:20 +0000 (11:35 -0500)] 
More efficient field initialisation

6 years agoDon't include sys/un.h in individual compilation units
Arran Cudbard-Bell [Fri, 5 Jun 2020 01:03:49 +0000 (20:03 -0500)] 
Don't include sys/un.h in individual compilation units

6 years agoThis is why consistency is important...
Arran Cudbard-Bell [Thu, 4 Jun 2020 20:45:24 +0000 (15:45 -0500)] 
This is why consistency is important...

6 years agoUse our own SUN_LEN macro
Arran Cudbard-Bell [Thu, 4 Jun 2020 20:37:57 +0000 (15:37 -0500)] 
Use our own SUN_LEN macro

6 years agoFix ubsan issues in talloc_bstrdup and talloc_strndup
Arran Cudbard-Bell [Thu, 4 Jun 2020 20:26:41 +0000 (15:26 -0500)] 
Fix ubsan issues in talloc_bstrdup and talloc_strndup

6 years agotrunk: Ensure all trunk request fields are properly initialised, and unsed bound_to_c...
Arran Cudbard-Bell [Thu, 4 Jun 2020 20:10:30 +0000 (15:10 -0500)] 
trunk: Ensure all trunk request fields are properly initialised, and unsed bound_to_conn when the request is freed

6 years agoUpdate comitted adoc files
Arran Cudbard-Bell [Thu, 4 Jun 2020 19:33:52 +0000 (14:33 -0500)] 
Update comitted adoc files

6 years agoFix ubsan issues in hash update functions
Arran Cudbard-Bell [Thu, 4 Jun 2020 19:01:06 +0000 (14:01 -0500)] 
Fix ubsan issues in hash update functions

6 years agoInitialise res
Arran Cudbard-Bell [Thu, 4 Jun 2020 18:38:56 +0000 (13:38 -0500)] 
Initialise res

6 years agoFix weird whitespace in misc.c
Arran Cudbard-Bell [Thu, 4 Jun 2020 18:35:37 +0000 (13:35 -0500)] 
Fix weird whitespace in misc.c

6 years agoradsniff: The test need files from LFS (#3481)
Jorge Pereira [Thu, 4 Jun 2020 18:20:42 +0000 (15:20 -0300)] 
radsniff: The test need files from LFS (#3481)

The radsniff test needs some files on LFS. therefore, if the system
doesn't have the 'git lfs' installed properly. just skip it and keep
going.

Scald :battery:

6 years agoRunning with ubsan works locally... Lets see how much it breaks CIT
Arran Cudbard-Bell [Thu, 4 Jun 2020 18:15:30 +0000 (13:15 -0500)] 
Running with ubsan works locally... Lets see how much it breaks CIT

6 years agoAdd mul + mod without overflow and fixup rlm_passwd hash function
Arran Cudbard-Bell [Thu, 4 Jun 2020 18:12:25 +0000 (13:12 -0500)] 
Add mul + mod without overflow and fixup rlm_passwd hash function

6 years agofiles: Only copy pairs if there are pairs to copy (ubsan)
Arran Cudbard-Bell [Thu, 4 Jun 2020 17:42:07 +0000 (12:42 -0500)] 
files: Only copy pairs if there are pairs to copy (ubsan)

6 years agoFix ubsan complaint with fr_sha1_update
Arran Cudbard-Bell [Thu, 4 Jun 2020 17:36:31 +0000 (12:36 -0500)] 
Fix ubsan complaint with fr_sha1_update

6 years agoBetter warning message
Arran Cudbard-Bell [Thu, 4 Jun 2020 17:09:10 +0000 (12:09 -0500)] 
Better warning message

6 years agoDisable the radsniff test because it still doesn't pass
Arran Cudbard-Bell [Thu, 4 Jun 2020 17:07:58 +0000 (12:07 -0500)] 
Disable the radsniff test because it still doesn't pass

6 years agoFix uses of BIN_PATH and DONTRUNVARIABLENAMESTOGETHERBECAUSEITMAKESTHEMREALLYHARDTOREAD
Arran Cudbard-Bell [Thu, 4 Jun 2020 17:05:18 +0000 (12:05 -0500)] 
Fix uses of BIN_PATH and DONTRUNVARIABLENAMESTOGETHERBECAUSEITMAKESTHEMREALLYHARDTOREAD

6 years agoatomic queue: Fix queue talloc chunk verification
Arran Cudbard-Bell [Thu, 4 Jun 2020 16:47:55 +0000 (11:47 -0500)] 
atomic queue: Fix queue talloc chunk verification

6 years agotests: Actually emit the server output
Arran Cudbard-Bell [Thu, 4 Jun 2020 16:27:33 +0000 (11:27 -0500)] 
tests: Actually emit the server output

6 years agoalways clean up FDs, timers, etc. on Linux
Alan T. DeKok [Thu, 4 Jun 2020 15:39:47 +0000 (11:39 -0400)] 
always clean up FDs, timers, etc. on Linux

6 years agofree children before freeing the parent
Alan T. DeKok [Thu, 4 Jun 2020 15:20:05 +0000 (11:20 -0400)] 
free children before freeing the parent

so that `*(ev->parent) = NULL` will work properly

6 years agoneed sys/wait.h here, too
Alan T. DeKok [Thu, 4 Jun 2020 14:39:01 +0000 (10:39 -0400)] 
need sys/wait.h here, too

6 years agoinclude necessary header file
Alan T. DeKok [Thu, 4 Jun 2020 14:13:26 +0000 (10:13 -0400)] 
include necessary header file

6 years agoremove rad_waitpid()
Alan T. DeKok [Thu, 4 Jun 2020 13:57:17 +0000 (09:57 -0400)] 
remove rad_waitpid()

which is now just an empty wrapper around waitpid()

6 years agoworkarounds for Linux, where libkqueue can't wait for a PID
Alan T. DeKok [Thu, 4 Jun 2020 13:54:44 +0000 (09:54 -0400)] 
workarounds for Linux, where libkqueue can't wait for a PID

in the optimistic case, we can beat the race condition, and just
call waitpid(pid, &status, WNOHANG).  This works often enough
that it's worth doing for now.

6 years agokeep analyzer happy
Alan T. DeKok [Thu, 4 Jun 2020 13:38:42 +0000 (09:38 -0400)] 
keep analyzer happy

6 years agofixes for libkqueue
Alan T. DeKok [Thu, 4 Jun 2020 12:55:01 +0000 (08:55 -0400)] 
fixes for libkqueue

6 years agoactually kill the child process on cancel signal
Alan T. DeKok [Thu, 4 Jun 2020 12:43:08 +0000 (08:43 -0400)] 
actually kill the child process on cancel signal

6 years agoclean ups and fixes.
Alan T. DeKok [Thu, 4 Jun 2020 12:28:27 +0000 (08:28 -0400)] 
clean ups and fixes.

hoist the "wait for PID" call to immediately after the fork.
The previous code waited until the process was done writing,
which meant that the waitpid was sometimes missed.

Add "state->failed" so that we can distinguish administrative
failure (timeout or too much output) from program failure.

allow the kill timer to run until waitpid() returns. This change
lets the program continue to run after closing it's STDOUT.
But also enforces that the program does not run forever.

distinguish "no more data" from "too much data".  When there's
no more data, we close the pipe and remove the IO handler.
But we let the program keep running.

When there's too much data, we remove the reader callback, close
the pipe FD, kill the program, and set the "failed" status.

6 years agoapparently we need NOTE_EXITSTATUS, too
Alan T. DeKok [Thu, 4 Jun 2020 12:26:49 +0000 (08:26 -0400)] 
apparently we need NOTE_EXITSTATUS, too

otherwise we get told that the process exited, but not what
the status is.

6 years agodhcpv6: Fix alignment issue writing out the length
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:39:53 +0000 (22:39 -0500)] 
dhcpv6: Fix alignment issue writing out the length

6 years agoinet: Fix unaligned access in fr_in6addr_mask
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:32:26 +0000 (22:32 -0500)] 
inet: Fix unaligned access in fr_in6addr_mask

6 years agohash: Fix invalid shift in hash table code
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:18:25 +0000 (22:18 -0500)] 
hash: Fix invalid shift in hash table code

6 years agoaq: Fix atomic queue to use correct cache line alignment
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:02:58 +0000 (22:02 -0500)] 
aq: Fix atomic queue to use correct cache line alignment

6 years agos/fr_atomic_queue_create/fr_atomic_queue_alloc/
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:02:32 +0000 (22:02 -0500)] 
s/fr_atomic_queue_create/fr_atomic_queue_alloc/

6 years agohw: Add function to get cache line sizes
Arran Cudbard-Bell [Thu, 4 Jun 2020 03:00:58 +0000 (22:00 -0500)] 
hw: Add function to get cache line sizes

6 years agotalloc: Add talloc aligned array
Arran Cudbard-Bell [Thu, 4 Jun 2020 02:59:14 +0000 (21:59 -0500)] 
talloc: Add talloc aligned array

6 years agobuild: attribute can take multiple arguments
Arran Cudbard-Bell [Thu, 4 Jun 2020 02:58:25 +0000 (21:58 -0500)] 
build: attribute can take multiple arguments

6 years agoadd "discard" as a synonym for "do not respond"
Alan T. DeKok [Wed, 3 Jun 2020 23:16:11 +0000 (19:16 -0400)] 
add "discard" as a synonym for "do not respond"

6 years agomove trigger_exec_free() to server_free()
Alan T. DeKok [Wed, 3 Jun 2020 23:14:06 +0000 (19:14 -0400)] 
move trigger_exec_free() to server_free()

because it's set up in server_init()

6 years agominor cleanups
Alan T. DeKok [Wed, 3 Jun 2020 22:48:22 +0000 (18:48 -0400)] 
minor cleanups

6 years agosetup debugging functions for libcurl
Arran Cudbard-Bell [Wed, 3 Jun 2020 21:21:54 +0000 (16:21 -0500)] 
setup debugging functions for libcurl

6 years agoFix skip on eapol_test build failure
Arran Cudbard-Bell [Wed, 3 Jun 2020 19:00:59 +0000 (14:00 -0500)] 
Fix skip on eapol_test build failure

6 years agoFix ODR issue
Arran Cudbard-Bell [Wed, 3 Jun 2020 18:52:18 +0000 (13:52 -0500)] 
Fix ODR issue

6 years agoFix copy paste issue
Arran Cudbard-Bell [Wed, 3 Jun 2020 15:23:39 +0000 (10:23 -0500)] 
Fix copy paste issue

6 years agoippool - mysql: simplified allocate_find query.
Thomas Faller [Tue, 12 May 2020 18:56:45 +0000 (20:56 +0200)] 
ippool - mysql: simplified allocate_find query.

6 years agoradsniff: Add test suite
Jorge Pereira [Wed, 20 May 2020 16:46:52 +0000 (13:46 -0300)] 
radsniff: Add test suite

6 years agotravis: Add steps to install git-lfs
Jorge Pereira [Tue, 19 May 2020 20:42:03 +0000 (17:42 -0300)] 
travis: Add steps to install git-lfs

6 years agoconfigure: support enabling multiple sanitizers
Arran Cudbard-Bell [Tue, 2 Jun 2020 21:06:36 +0000 (16:06 -0500)] 
configure: support enabling multiple sanitizers

6 years agoEliminate extra increment of _dbuff->p in the fr_dbuff_<type>_in() functions
James Jones [Tue, 2 Jun 2020 18:40:30 +0000 (13:40 -0500)] 
Eliminate extra increment of _dbuff->p in the fr_dbuff_<type>_in() functions

6 years agoparse fields into value boxes at startup, instead of at run-time
Alan T. DeKok [Tue, 2 Jun 2020 16:06:13 +0000 (12:06 -0400)] 
parse fields into value boxes at startup, instead of at run-time

We can't do this always, but we can do it much of the time.

6 years agostart of skeleton for parsing value boxes for fields
Alan T. DeKok [Tue, 2 Jun 2020 15:26:27 +0000 (11:26 -0400)] 
start of skeleton for parsing value boxes for fields

6 years agoif we can't parse the key field, error out
Alan T. DeKok [Tue, 2 Jun 2020 15:25:02 +0000 (11:25 -0400)] 
if we can't parse the key field, error out

instead of silently ignoring the row

6 years agoupdate to correct documentation, and make it clearer
Alan T. DeKok [Tue, 2 Jun 2020 15:00:54 +0000 (11:00 -0400)] 
update to correct documentation, and make it clearer

6 years agofree detached children, even if request did not yield
Alan T. DeKok [Tue, 2 Jun 2020 13:41:05 +0000 (09:41 -0400)] 
free detached children, even if request did not yield

6 years agoclean up detached children
Alan T. DeKok [Tue, 2 Jun 2020 12:47:09 +0000 (08:47 -0400)] 
clean up detached children

we don't need to run them, so we just delete them.

6 years agoremove unused assignment
Alan T. DeKok [Tue, 2 Jun 2020 00:28:55 +0000 (20:28 -0400)] 
remove unused assignment

6 years agoremove functions no longer needed by unit_test_module
Alan T. DeKok [Mon, 1 Jun 2020 23:26:01 +0000 (19:26 -0400)] 
remove functions no longer needed by unit_test_module

6 years agoswitch to using virtual_server_get_process_by_name()
Alan T. DeKok [Mon, 1 Jun 2020 23:22:13 +0000 (19:22 -0400)] 
switch to using virtual_server_get_process_by_name()

instead of unlang_interpret_synchronous()

This change means that we have to run our own event loop.  But,
it also means that we just call the proto modules to do all of
the processing.  Which means that unit_test_module no longer
needs to know about any magic packet codes.  The proto_radius_auth
etc. functions take care of all of that.

6 years agoremove old cruft
Alan T. DeKok [Mon, 1 Jun 2020 23:21:08 +0000 (19:21 -0400)] 
remove old cruft

most of these had manual "update control cleartext-password"
which isn't necessary, as that's done in unit_test_module.conf

Also added "accept" in a few places, as it will be needed soon.

6 years agoremove assertion
Alan T. DeKok [Mon, 1 Jun 2020 23:05:09 +0000 (19:05 -0400)] 
remove assertion

it's no longer necessary, and unit_test_module doesn't always
add FDs to the event loop

6 years agomove error message to later in the code
Alan T. DeKok [Mon, 1 Jun 2020 23:04:21 +0000 (19:04 -0400)] 
move error message to later in the code

6 years agouse wait_for_event, to be the same as fr_worker()
Alan T. DeKok [Mon, 1 Jun 2020 22:48:03 +0000 (18:48 -0400)] 
use wait_for_event, to be the same as fr_worker()

6 years agoadd test case for "call"
Alan T. DeKok [Mon, 1 Jun 2020 22:33:22 +0000 (18:33 -0400)] 
add test case for "call"

6 years agomake "call" work correctly
Alan T. DeKok [Mon, 1 Jun 2020 22:29:31 +0000 (18:29 -0400)] 
make "call" work correctly

and fix virtual_server_get_process_by_name()

6 years agoset name / debug_name for "call"
Alan T. DeKok [Mon, 1 Jun 2020 20:44:41 +0000 (16:44 -0400)] 
set name / debug_name for "call"

6 years agoinitialize "dl" to shut up scanner
Alan T. DeKok [Mon, 1 Jun 2020 20:36:17 +0000 (16:36 -0400)] 
initialize "dl" to shut up scanner

6 years agoshould should be module_method_t*, not **
Alan T. DeKok [Mon, 1 Jun 2020 20:26:27 +0000 (16:26 -0400)] 
should should be module_method_t*, not **

6 years agofree this, too
Alan T. DeKok [Mon, 1 Jun 2020 20:17:38 +0000 (16:17 -0400)] 
free this, too

6 years agomove "get process function / ctx" to common API
Alan T. DeKok [Mon, 1 Jun 2020 17:07:46 +0000 (13:07 -0400)] 
move "get process function / ctx" to common API

where we may be able to use it with unit_test_module

6 years agoexplictly free it
Alan T. DeKok [Mon, 1 Jun 2020 16:41:33 +0000 (12:41 -0400)] 
explictly free it

6 years agoswtich to using our own dl_loader ctx
Alan T. DeKok [Mon, 1 Jun 2020 15:44:19 +0000 (11:44 -0400)] 
swtich to using our own dl_loader ctx

6 years agowe no longer need to link against libfreeradius-radius
Alan T. DeKok [Mon, 1 Jun 2020 15:13:28 +0000 (11:13 -0400)] 
we no longer need to link against libfreeradius-radius

6 years agofr_cap_net_raw: Don't return an error if the capability is already in the effective set
Arran Cudbard-Bell [Mon, 1 Jun 2020 15:35:57 +0000 (10:35 -0500)] 
fr_cap_net_raw: Don't return an error if the capability is already in the effective set

6 years agoOutput what script we're calling
Arran Cudbard-Bell [Sun, 31 May 2020 17:04:02 +0000 (12:04 -0500)] 
Output what script we're calling

6 years agouse standard names
Alan T. DeKok [Sun, 31 May 2020 13:40:46 +0000 (09:40 -0400)] 
use standard names

6 years agomore pedantic checks for re-registration
Alan T. DeKok [Sun, 31 May 2020 13:24:23 +0000 (09:24 -0400)] 
more pedantic checks for re-registration

if it's the same func / type / whatever, just return the same thing.
If it's different, error out.

6 years agothese need to link to the io library, too
Alan T. DeKok [Sat, 30 May 2020 23:02:38 +0000 (19:02 -0400)] 
these need to link to the io library, too

6 years agoverify that unit_test_module can send and receive dhcpv4 packets
Alan T. DeKok [Sat, 30 May 2020 22:46:32 +0000 (18:46 -0400)] 
verify that unit_test_module can send and receive dhcpv4 packets

6 years agoadd client function for unit_test_module
Alan T. DeKok [Sat, 30 May 2020 22:46:13 +0000 (18:46 -0400)] 
add client function for unit_test_module

6 years agomore sanity checks
Alan T. DeKok [Sat, 30 May 2020 22:45:29 +0000 (18:45 -0400)] 
more sanity checks

that the "default" virtual server exists, and has the same
namespace as we were told to use

6 years agocheck that unit_test_module can send CoA-Request packets
Alan T. DeKok [Sat, 30 May 2020 15:15:30 +0000 (11:15 -0400)] 
check that unit_test_module can send CoA-Request packets

6 years agoallow for "server dhcpv4"
Alan T. DeKok [Sat, 30 May 2020 15:01:16 +0000 (11:01 -0400)] 
allow for "server dhcpv4"

and fall back to "server default" if the protocol-specific
server does not exist.

6 years agolet's parse "-p"
Alan T. DeKok [Sat, 30 May 2020 14:15:12 +0000 (10:15 -0400)] 
let's parse "-p"