]>
git.ipfire.org Git - thirdparty/freeradius-server.git/log
Arran Cudbard-Bell [Fri, 15 Dec 2017 12:55:23 +0000 (12:55 +0000)]
Bump clang other docker images to clang 5.0
Arran Cudbard-Bell [Fri, 15 Dec 2017 12:36:59 +0000 (12:36 +0000)]
Bump trusty to clang 5.0
Arran Cudbard-Bell [Fri, 15 Dec 2017 12:34:27 +0000 (12:34 +0000)]
Fix various compiler clang scan warnings
Arran Cudbard-Bell [Thu, 14 Dec 2017 22:21:42 +0000 (22:21 +0000)]
Silently discard encrypted attributes if we're sending them too early
Arran Cudbard-Bell [Thu, 14 Dec 2017 21:06:39 +0000 (21:06 +0000)]
Functions don't need to write anywhere
Arran Cudbard-Bell [Thu, 14 Dec 2017 21:00:54 +0000 (21:00 +0000)]
talloc_ctx can be NULL
Arran Cudbard-Bell [Thu, 14 Dec 2017 20:30:22 +0000 (20:30 +0000)]
Bump xenial's clang version to 5.0
Arran Cudbard-Bell [Thu, 14 Dec 2017 19:39:11 +0000 (19:39 +0000)]
Revert "patch libkqueue in dockerfile for centos6"
This reverts commit
047df3289e347ff2e21a3997247d7736e80819ca .
Matthew Newton [Thu, 14 Dec 2017 19:33:24 +0000 (19:33 +0000)]
patch libkqueue in dockerfile for centos6
Arran Cudbard-Bell [Thu, 14 Dec 2017 18:56:57 +0000 (18:56 +0000)]
We now use trusty (not precise), and may as well bump clang to v5.0
Arran Cudbard-Bell [Thu, 14 Dec 2017 18:40:42 +0000 (18:40 +0000)]
data can be null
Arran Cudbard-Bell [Thu, 14 Dec 2017 18:37:10 +0000 (18:37 +0000)]
Revert "'out' can't be NULL either..."
This reverts commit
dbcabeb1116cf15ec8418eae834837226e684ed8 .
Arran Cudbard-Bell [Thu, 14 Dec 2017 18:35:02 +0000 (18:35 +0000)]
Revert "require that CONF_PARSER entries put their data somewhere"
This reverts commit
d40748b6c7572e79b50f988f096dfaa117ab3bd5 .
Alan T. DeKok [Thu, 14 Dec 2017 18:14:44 +0000 (13:14 -0500)]
require that CONF_PARSER entries put their data somewhere
Alan T. DeKok [Thu, 14 Dec 2017 18:10:15 +0000 (13:10 -0500)]
check for NULL in tls_global_version_check()
Arran Cudbard-Bell [Thu, 14 Dec 2017 13:01:48 +0000 (13:01 +0000)]
Toggle whether we send protected success messages
Alan T. DeKok [Thu, 14 Dec 2017 14:00:42 +0000 (09:00 -0500)]
set nonnull for parameters here, too
Alan T. DeKok [Thu, 14 Dec 2017 14:00:32 +0000 (09:00 -0500)]
'out' can't be NULL either...
Arran Cudbard-Bell [Wed, 13 Dec 2017 22:40:36 +0000 (22:40 +0000)]
How about we free the reply *AFTER* copying its contents
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:39:15 +0000 (21:39 +0000)]
Use the correct width field accessors
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:19:12 +0000 (21:19 +0000)]
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:17:09 +0000 (21:17 +0000)]
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:16:21 +0000 (21:16 +0000)]
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:14:46 +0000 (21:14 +0000)]
Arran Cudbard-Bell [Wed, 13 Dec 2017 21:13:54 +0000 (21:13 +0000)]
Alan T. DeKok [Wed, 13 Dec 2017 21:16:52 +0000 (16:16 -0500)]
Alan T. DeKok [Wed, 13 Dec 2017 21:15:50 +0000 (16:15 -0500)]
close file on return. CID #
1185992
Arran Cudbard-Bell [Wed, 13 Dec 2017 19:51:41 +0000 (19:51 +0000)]
Fixup protected success/failure notifications
Arran Cudbard-Bell [Wed, 13 Dec 2017 17:55:39 +0000 (17:55 +0000)]
Minor fixes
Arran Cudbard-Bell [Tue, 12 Dec 2017 22:23:11 +0000 (22:23 +0000)]
Need lynx
Arran Cudbard-Bell [Tue, 12 Dec 2017 22:08:05 +0000 (22:08 +0000)]
Tab not space
Arran Cudbard-Bell [Tue, 12 Dec 2017 22:00:09 +0000 (22:00 +0000)]
...and now travis doesn't run clang scan at all
Arran Cudbard-Bell [Tue, 12 Dec 2017 21:50:29 +0000 (21:50 +0000)]
Emit more clang scan information
Arran Cudbard-Bell [Tue, 12 Dec 2017 21:15:30 +0000 (21:15 +0000)]
Only pair rules need to have a destination
Arran Cudbard-Bell [Tue, 12 Dec 2017 20:28:38 +0000 (20:28 +0000)]
More clang scan fixes
Arran Cudbard-Bell [Tue, 12 Dec 2017 19:56:05 +0000 (19:56 +0000)]
Various fixes for clan scan
Seems to have been caused by a Travis upgrade to llvm 5.0
Arran Cudbard-Bell [Tue, 12 Dec 2017 19:39:59 +0000 (19:39 +0000)]
Apply prefix check for non-debug builds to quiet clang scan
Arran Cudbard-Bell [Tue, 12 Dec 2017 18:46:22 +0000 (18:46 +0000)]
Plumb in OPc generation
Arran Cudbard-Bell [Tue, 12 Dec 2017 17:32:32 +0000 (17:32 +0000)]
Add (op, ki)->opc derivation
Matthew Newton [Tue, 12 Dec 2017 13:39:18 +0000 (13:39 +0000)]
bomb out immediately rather than silently fail
if we can't generate passwords.mk, we're stuffed anyway
Matthew Newton [Tue, 12 Dec 2017 13:37:56 +0000 (13:37 +0000)]
install inner-server.cnf config file
Alan T. DeKok [Mon, 11 Dec 2017 15:32:25 +0000 (10:32 -0500)]
add "inner-server" scripts and configuration
Arran Cudbard-Bell [Tue, 12 Dec 2017 14:08:53 +0000 (14:08 +0000)]
One more missed ki
Arran Cudbard-Bell [Tue, 12 Dec 2017 13:48:55 +0000 (13:48 +0000)]
Set SQN to 2, seems to be the minimum number eapol_test accepts
Arran Cudbard-Bell [Tue, 12 Dec 2017 13:48:29 +0000 (13:48 +0000)]
s/k/ki/
Arran Cudbard-Bell [Tue, 12 Dec 2017 13:48:17 +0000 (13:48 +0000)]
Document EAP-SIM
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:33:56 +0000 (12:33 +0000)]
Move OPC check
rename variables to be inline with other *_from_ki functions
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:26:20 +0000 (12:26 +0000)]
Add tests for sim methods
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:25:38 +0000 (12:25 +0000)]
Rewrite large chunks of EAP-SIM so it behaves like the RFC says it should
I don't know if this was written pre RFC 4186 but large parts of it don't follow the RFC.
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:17:22 +0000 (12:17 +0000)]
Shuffle identity logic around in preparation for doing identity selection properly
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:16:08 +0000 (12:16 +0000)]
Check arguments to fr_sim_crypto_sign_packet
...and add helful debugging output for developers.
eapol_test provides similar debugging output, so when something's not working it makes it very easy to see which cryptographic input/output was incorrect.
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:14:19 +0000 (12:14 +0000)]
Add SIM method hint table so we can warn when we think the wrong SIM/AKA flavour is being used
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:13:47 +0000 (12:13 +0000)]
Document kdf field
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:13:32 +0000 (12:13 +0000)]
Cleanup and document the milenage code some more
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:11:08 +0000 (12:11 +0000)]
No start
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:10:30 +0000 (12:10 +0000)]
Only include an AMF attribute if we're doing AKA'
Arran Cudbard-Bell [Tue, 12 Dec 2017 12:09:58 +0000 (12:09 +0000)]
There is no 'Start' state in EAP-AKA[']
Alan T. DeKok [Mon, 11 Dec 2017 21:52:51 +0000 (16:52 -0500)]
sample config for EAP-SIM
Alan T. DeKok [Mon, 11 Dec 2017 21:50:44 +0000 (16:50 -0500)]
don't enable AKA until such time as we have a sample config for it
Alan T. DeKok [Mon, 11 Dec 2017 21:49:14 +0000 (16:49 -0500)]
opc may be NULL
Alan T. DeKok [Mon, 11 Dec 2017 15:33:33 +0000 (10:33 -0500)]
added disconnect-reason
and reformatted
Arran Cudbard-Bell [Mon, 11 Dec 2017 17:12:20 +0000 (17:12 +0000)]
clang scan
Arran Cudbard-Bell [Mon, 11 Dec 2017 17:10:36 +0000 (17:10 +0000)]
Doxygen
Arran Cudbard-Bell [Mon, 11 Dec 2017 11:18:31 +0000 (11:18 +0000)]
Fix CK' IK' derivation
Switch to newer EVP interface so we get hardware acceleration.
Arran Cudbard-Bell [Sun, 10 Dec 2017 23:51:47 +0000 (23:51 +0000)]
Initial implementation of EAP-AKA-Prime
Doesn't work yet
Arran Cudbard-Bell [Sun, 10 Dec 2017 20:34:51 +0000 (20:34 +0000)]
Indent keys
Arran Cudbard-Bell [Sun, 10 Dec 2017 20:30:40 +0000 (20:30 +0000)]
Doxygen
Arran Cudbard-Bell [Sun, 10 Dec 2017 20:28:00 +0000 (20:28 +0000)]
Fixup EAP-AKA state machine to match RFC 4187
Which in most cases means sending an AKA-Notification instead of just an EAP-Failure, which is why the wpa_supplicant state machine seemed to hang.
Arran Cudbard-Bell [Sun, 10 Dec 2017 18:35:38 +0000 (18:35 +0000)]
Quiet clang scan
Arran Cudbard-Bell [Sun, 10 Dec 2017 18:33:44 +0000 (18:33 +0000)]
Add support for checkcodes
Add support for result_ind
Make signing digests flexible in preparation for EAP-AKA'
Add success notification state for result_ind
Arran Cudbard-Bell [Sun, 10 Dec 2017 18:31:01 +0000 (18:31 +0000)]
Formatting
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:56:54 +0000 (12:56 +0000)]
Remove fr_sim_crypto_mac_verify and use fr_sim_crypto_sign_packet
for signing and validating
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:51:03 +0000 (12:51 +0000)]
Add macro for SIM hash size
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:50:28 +0000 (12:50 +0000)]
Add identity negotiation to EAP_AKA
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:49:07 +0000 (12:49 +0000)]
Build reply list just containing AKA attribute
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:45:25 +0000 (12:45 +0000)]
Emit error if the encoder fails
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:40:29 +0000 (12:40 +0000)]
Allow the decoder to decode empty packets
These are valid, as SIM/AKA Notifications are just ACKs of a previous notification sent be the server and contain no attributes.
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:39:42 +0000 (12:39 +0000)]
Enable EAP-SIM and SIM/USIM simulators in eapol_test builds
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:22:47 +0000 (12:22 +0000)]
Print milenage inputs at DEBUG3
These are as (if not more so) sensitive as passwords
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:22:07 +0000 (12:22 +0000)]
Better output when an IMSI starting with zero isn't in the format we expected
Arran Cudbard-Bell [Sun, 10 Dec 2017 12:21:47 +0000 (12:21 +0000)]
Move identity request types and put them in a table
Arran Cudbard-Bell [Sat, 9 Dec 2017 12:31:43 +0000 (12:31 +0000)]
The EAP-AKA actually contains examples for ID tags, so use those
Arran Cudbard-Bell [Sat, 9 Dec 2017 11:52:33 +0000 (11:52 +0000)]
Fix vector generation so it selects the algorithm by which attributes are present
Defaults to compl128v3/milenage for GSM depending on whether OPc is present, and milenage for UMTS.
Arran Cudbard-Bell [Sat, 9 Dec 2017 07:30:49 +0000 (07:30 +0000)]
Plumb in milenage
Arran Cudbard-Bell [Fri, 8 Dec 2017 22:17:32 +0000 (22:17 +0000)]
Cleanup comp128
Arran Cudbard-Bell [Fri, 8 Dec 2017 22:08:29 +0000 (22:08 +0000)]
Fixup mileage and run test cases 1 and 19 from 3GPP TS 35.208
Arran Cudbard-Bell [Fri, 8 Dec 2017 22:07:29 +0000 (22:07 +0000)]
Don't need to include radius headers in proto.h
Alan T. DeKok [Fri, 8 Dec 2017 20:49:12 +0000 (15:49 -0500)]
debian/ubuntu pkgs need dh-systemd. Fixes #2136
Alan T. DeKok [Fri, 8 Dec 2017 17:07:43 +0000 (12:07 -0500)]
and another copyright typo
Alan T. DeKok [Fri, 8 Dec 2017 16:35:20 +0000 (11:35 -0500)]
copyright is lowercase
Alan T. DeKok [Fri, 8 Dec 2017 15:47:39 +0000 (10:47 -0500)]
remove tree_num and client stats
Arran Cudbard-Bell [Fri, 8 Dec 2017 11:49:18 +0000 (11:49 +0000)]
Add the requisite dh stuff to install the systemd unit files correctly
Arran Cudbard-Bell [Fri, 8 Dec 2017 11:47:54 +0000 (11:47 +0000)]
Shamelessly copy the milenage code from wpa_supplicant
This lets us simulate an AS, and hopefully have eapol_test simulate a usim, meaning we can add sim,aka,aka' to CIT in future.
Alan T. DeKok [Thu, 7 Dec 2017 21:39:18 +0000 (16:39 -0500)]
move assignment for debug purposes
Alan T. DeKok [Thu, 7 Dec 2017 21:00:52 +0000 (16:00 -0500)]
NDEBUG
Alan T. DeKok [Thu, 7 Dec 2017 20:26:33 +0000 (15:26 -0500)]
ring buffer alloc may fail
Arran Cudbard-Bell [Thu, 7 Dec 2017 13:54:38 +0000 (13:54 +0000)]
Recommit license so GitHub recognises we have a license...?
Arran Cudbard-Bell [Thu, 7 Dec 2017 07:28:00 +0000 (07:28 +0000)]
Fix various SIM packet encoder issues
Arran Cudbard-Bell [Wed, 6 Dec 2017 22:39:57 +0000 (22:39 +0000)]
Change xlats to work on the username portion only
as required by the 3gpp standard
Alan T. DeKok [Wed, 6 Dec 2017 16:52:55 +0000 (11:52 -0500)]
ignore DELETE events for the directory
we didn't ask for them, but libkqueue delivers them to us