mech_eap.git
10 years agoGAS: Send error response if no room for pending dialog context
Jouni Malinen [Sun, 25 May 2014 15:34:10 +0000 (18:34 +0300)]
GAS: Send error response if no room for pending dialog context

Instead of silently dropping the response that requires fragmentation,
send an error response to the station to make it aware that no full
response will be available.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Verify SQLite DB as hostapd EAP user database
Jouni Malinen [Sun, 25 May 2014 14:55:37 +0000 (17:55 +0300)]
tests: Verify SQLite DB as hostapd EAP user database

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoHS 2.0: Include OSU client sample in wpa_supplicant release package
Jouni Malinen [Fri, 23 May 2014 20:35:42 +0000 (23:35 +0300)]
HS 2.0: Include OSU client sample in wpa_supplicant release package

This adds the new hs20 subdirectory into release tarballs.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoReserve QCA vendor specific nl80211 commands 14..19
Jouni Malinen [Thu, 22 May 2014 09:15:16 +0000 (12:15 +0300)]
Reserve QCA vendor specific nl80211 commands 14..19

These are reserved for QCA use.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Avoid resetting pending_listen_freq if p2p_listen is pending
Jithu Jance [Thu, 22 May 2014 06:55:53 +0000 (12:25 +0530)]
P2P: Avoid resetting pending_listen_freq if p2p_listen is pending

If p2p_listen is called while previous listen command's
remain_on_channel event is pending, the p2p_listen would fail
and it used to clear pending_listen_freq. Now when the remain-
on-channel event comes from the driver, the pending_listen_freq
doesn't match and gets ignored. This was leading to a case
where listen state was getting stuck (in case of WAIT_PEER_CONNECT
state).

Signed-off-by: Jithu Jance <jithu@broadcom.com>
10 years agonl80211: Indicate SHA256-based AKM suites in CONNECT/ASSOCIATE
Jithu Jance [Tue, 20 May 2014 05:35:15 +0000 (11:05 +0530)]
nl80211: Indicate SHA256-based AKM suites in CONNECT/ASSOCIATE

Previously, the NL80211_ATTR_AKM_SUITES was skipped if either of these
SHA256-based AKMs was negotiated.

Signed-off-by: Jithu Jance <jithu@broadcom.com>
10 years agotests: gitignore TNC library files
Jouni Malinen [Wed, 21 May 2014 15:47:57 +0000 (18:47 +0300)]
tests: gitignore TNC library files

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Add -rdynamic to fix TNC IMV/IMC loading on some platforms
Jouni Malinen [Wed, 21 May 2014 11:56:34 +0000 (14:56 +0300)]
tests: Add -rdynamic to fix TNC IMV/IMC loading on some platforms

The example IMV and IMC used for TNC testing has references to
wpa_printf and other functions from hostapd/wpa_supplicant. Link the
binaries in a way that allows these symbols to be resolved while loading
the libraries at run time.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Add CAVP test vectors for RSA/PKCS #1 v1.5 signature validation
Jouni Malinen [Mon, 19 May 2014 20:22:49 +0000 (23:22 +0300)]
tests: Add CAVP test vectors for RSA/PKCS #1 v1.5 signature validation

This allow the PKCS #1 and RSA implementation to be validated against
the test vectors from
http://csrc.nist.gov/groups/STM/cavp/documents/dss/186-2rsatestvectors.zip
and
http://csrc.nist.gov/groups/STM/cavp/documents/dss/SigVer15EMTest.txt.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoPKCS 1: Add function for checking v1.5 RSA signature
Jouni Malinen [Mon, 19 May 2014 20:21:55 +0000 (23:21 +0300)]
PKCS 1: Add function for checking v1.5 RSA signature

This could be used as a step towards replacing more specific functions
used in X.509 and TLS processing.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoRSA: Add OID definitions and helper function for hash algorithms
Jouni Malinen [Mon, 19 May 2014 20:21:07 +0000 (23:21 +0300)]
RSA: Add OID definitions and helper function for hash algorithms

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoAdd function for building RSA public key from n and e parameters
Jouni Malinen [Mon, 19 May 2014 20:19:08 +0000 (23:19 +0300)]
Add function for building RSA public key from n and e parameters

This is similar to the existing functionality that parsed ASN.1-encoded
RSA public key by generating a similar public key instance from already
parsed n and e parameters.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoPKCS #1: Enforce minimum padding for decryption in internal TLS
Jouni Malinen [Mon, 19 May 2014 20:26:43 +0000 (23:26 +0300)]
PKCS #1: Enforce minimum padding for decryption in internal TLS

Follow the PKCS #1 v1.5, 8.1 constraint of at least eight octets long PS
for the case where the internal TLS implementation decrypts PKCS #1
formatted data. Similar limit was already in place for signature
validation, but not for this decryption routine.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoPKCS #1: Allow only BT=01 for signature in internal TLS
Jouni Malinen [Mon, 19 May 2014 20:26:19 +0000 (23:26 +0300)]
PKCS #1: Allow only BT=01 for signature in internal TLS

Based on PKCS #1, v1.5, 10.1.3, the block type shall be 01 for a
signature. This avoids a potential attack vector for internal TLS/X.509
implementation.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoX.509: Fix internal TLS/X.509 validation of PKCS#1 signature
Jouni Malinen [Mon, 19 May 2014 20:25:38 +0000 (23:25 +0300)]
X.509: Fix internal TLS/X.509 validation of PKCS#1 signature

Verify that there is no extra data after the hash field. This is needed
to avoid potential attacks using additional data to construct a value
that passes the RSA operation and allows the hash value to be forged.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Add CAVP test vectors for byte-oriented SHA-1 and SHA-256
Jouni Malinen [Sun, 18 May 2014 16:47:59 +0000 (19:47 +0300)]
tests: Add CAVP test vectors for byte-oriented SHA-1 and SHA-256

This allows the SHA-1 implementation to be validated against the
SHA1ShortMsg.rsp and SHA1LongMsg.rsp test vectors from
http://csrc.nist.gov/groups/STM/cavp/documents/shs/shabytetestvectors.zip.
Similarly, the SHA-256 can be validated against the SHA256ShortMsg.rsp
and SHA256LongMsg.rsp.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-PEAP parameters
Jouni Malinen [Sat, 17 May 2014 21:44:39 +0000 (00:44 +0300)]
tests: EAP-PEAP parameters

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-PEAP crypto_binding=0/1
Jouni Malinen [Sat, 17 May 2014 21:33:59 +0000 (00:33 +0300)]
tests: EAP-PEAP crypto_binding=0/1

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-FAST using PAC file
Jouni Malinen [Sat, 17 May 2014 21:28:31 +0000 (00:28 +0300)]
tests: EAP-FAST using PAC file

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-AKA' and EAP-AKA both enabled (bidding mechanism)
Jouni Malinen [Sat, 17 May 2014 21:04:18 +0000 (00:04 +0300)]
tests: EAP-AKA' and EAP-AKA both enabled (bidding mechanism)

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-TTLS/EAP-AKA, EAP-PEAP/EAP-AKA, EAP-FAST/EAP-AKA
Jouni Malinen [Sat, 17 May 2014 20:39:16 +0000 (23:39 +0300)]
tests: EAP-TTLS/EAP-AKA, EAP-PEAP/EAP-AKA, EAP-FAST/EAP-AKA

These add some more EAP-TTLS/PEAP/FAST coverage to test pending Phase 2
response re-processing.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Incorrect/missing password with TTLS non-EAP methods
Jouni Malinen [Sat, 17 May 2014 19:52:40 +0000 (22:52 +0300)]
tests: Incorrect/missing password with TTLS non-EAP methods

Incorrect password was already tested with TTLS/MSCHAPv2, but the other
non-EAP inner methods in TTLS use their own implementation of password
validation, so check each and also verify the case of no matching EAP
user entry for the specific method.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: TNC testing
Jouni Malinen [Fri, 16 May 2014 21:47:37 +0000 (00:47 +0300)]
tests: TNC testing

This implements minimal IMC and IMV to allow TNC testing with PEAP (SoH)
and TTLS/FAST with EAP-TNC.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoTNC: Allow TNC to be enabled dynamically
Jouni Malinen [Fri, 16 May 2014 21:01:45 +0000 (00:01 +0300)]
TNC: Allow TNC to be enabled dynamically

Previously, hostapd had to be started with at least one of the
configuration files enabling TNC for TNC to be usable. Change this to
allow TNC to be enabled when the first interface with TNC enabled gets
added during runtime.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoTNC: Move common definitions into a shared header file
Jouni Malinen [Fri, 16 May 2014 21:01:04 +0000 (00:01 +0300)]
TNC: Move common definitions into a shared header file

No need to duplicate these in multiple places.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoTNC: Allow tnc_config file path to be replaced
Jouni Malinen [Fri, 16 May 2014 19:37:54 +0000 (22:37 +0300)]
TNC: Allow tnc_config file path to be replaced

This is for enabling easier testing of TNCS/TNCC functionality as part
of the test scripts without having to use the fixed /etc/tnc_config
location that could be used by the main system and would require changes
within /etc.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: FT negative test with mismatching R0KH-ID
Jouni Malinen [Sat, 17 May 2014 07:44:41 +0000 (10:44 +0300)]
tests: FT negative test with mismatching R0KH-ID

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-AKA using external USIM processing for UMTS auth
Jouni Malinen [Fri, 16 May 2014 17:37:39 +0000 (20:37 +0300)]
tests: EAP-AKA using external USIM processing for UMTS auth

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Add more details to help with initial setup configuration
Jouni Malinen [Fri, 16 May 2014 16:24:47 +0000 (19:24 +0300)]
tests: Add more details to help with initial setup configuration

This documents some more steps needed during initial test setup
configuration to make it easier to get this running even without
thorough knowledge of the network setup used by the operating system.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoeloop: Add epoll option for better performance
Masashi Honma [Tue, 13 May 2014 00:35:48 +0000 (09:35 +0900)]
eloop: Add epoll option for better performance

This patch adds epoll option for the eloop implementation. This can be
selected with the CONFIG_ELOOP_EPOLL=y build option.

[merit]
See Table1.

Table1. comparison table
+--------+--------+-----------+------------+-------------+
|        | add fd | remove fd | prepare fd | dispatch fd |
+--------+--------+-----------+------------+-------------+
| select | O(1)   | O(1)      | O(N)       | O(N)        |
+--------+--------+-----------+------------+-------------+
| poll   | O(1)   | O(1)      | O(N)       | O(N)        |
+--------+--------+-----------+------------+-------------+
| epoll  | O(1)   | O(1)      | 0          | O(M)        |
+--------+--------+-----------+------------+-------------+
"add fd" is addition of fd by eloop_sock_table_add_sock().
"remove fd" is removal of fd by eloop_sock_table_remove_sock().
"prepare fd" is preparation of fds before wait in eloop_run().
"dispatch fd" is dispatchment of fds by eloop_sock_table_dispatch().
"N" is all watching fds.
"M" is fds which could be dispatched after waiting.

As shown in Table1, epoll option has better performance on "prepare fd" column.
Because select/poll option requires setting fds before every select()/poll().
But epoll_wait() doesn't need it.

And epoll option has also better performance on "dispatch fd" column.
Because select/poll option needs to check all registered fds to find out
dispatchable fds. But epoll option doesn't require checking all registered fds.
Because epoll_wait() returns dispatchable fd set.

So epoll option is effective for GO/AP functionality.

[demerit]
The epoll option requires additional heap memory. In case of P2P GO, it is
about 8K bytes.

Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
10 years agoeloop: Separate event loop select/poll implementation
Masashi Honma [Fri, 2 May 2014 13:33:44 +0000 (22:33 +0900)]
eloop: Separate event loop select/poll implementation

This allows yet another eloop.c option to be added.

Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
10 years agodbus: No need to recompute group object path on GroupStarted signal
Tomasz Bursztyka [Wed, 14 May 2014 13:10:41 +0000 (16:10 +0300)]
dbus: No need to recompute group object path on GroupStarted signal

The group object is already registered on DBus at that point, thus wpa_s
structure holds its path already.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
10 years agodbus: Provide the P2P Device Address from the relevant structure
Tomasz Bursztyka [Wed, 14 May 2014 13:10:40 +0000 (16:10 +0300)]
dbus: Provide the P2P Device Address from the relevant structure

Fixes a minor mistake: the p2p_info structure should be used here
instead of the peer_handler_args one.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
10 years agodbus: Fix interface DeviceFound signal specification
Tomasz Bursztyka [Wed, 14 May 2014 13:10:39 +0000 (16:10 +0300)]
dbus: Fix interface DeviceFound signal specification

DeviceFound does not provide any properties, just the peer object path.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
10 years agodbus: Declare GONegotiation signals properly
Tomasz Bursztyka [Wed, 14 May 2014 13:10:38 +0000 (16:10 +0300)]
dbus: Declare GONegotiation signals properly

These signals delivers an array of key/value pairs, thus declaring those
as it should.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
10 years agoSend authentication failure reason in wpas_auth_failed()
Dmitry Shmidt [Thu, 15 May 2014 22:55:21 +0000 (15:55 -0700)]
Send authentication failure reason in wpas_auth_failed()

"WRONG_KEY"   - possibly wrong psk
"AUTH_FAILED" - authentication failure
"CONN_FAILED" - continiuos connection failure

Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
10 years agoP2P: Iterate through full pref_chan list in search of a valid channel
Jouni Malinen [Fri, 16 May 2014 13:49:17 +0000 (16:49 +0300)]
P2P: Iterate through full pref_chan list in search of a valid channel

p2p_get_pref_freq() went through the full list only if the channels
arguments was provided. If no channel list contraint was in place, the
first pref_chan item was picked regardless of whether it is valid
channel and as such, a later valid entry could have been ignored. Allow
this to loop through all the entries until a valid channel is found or
the end of the pref_chan list is reached. As an extra bonus, this
simplifies the p2p_get_pref_freq() implementation quite a bit.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Verify HT40 coex scan with 20 MHz legacy BSS
Jouni Malinen [Thu, 15 May 2014 22:23:29 +0000 (01:23 +0300)]
tests: Verify HT40 coex scan with 20 MHz legacy BSS

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agohostapd: Check for overlapping 20 MHz BSS before starting 20/40 MHz BSS
Rajkumar Manoharan [Wed, 14 May 2014 07:41:44 +0000 (13:11 +0530)]
hostapd: Check for overlapping 20 MHz BSS before starting 20/40 MHz BSS

Before starting a 20/40 MHz BSS on the 2.4 GHz band, a 40-MHz-capable HT
AP is required by the rules defined in IEEE Std 802.11-2012 10.15.5 to
examine the channels of the current operational regulatory domain to
determine whether the operation of a 20/40 MHz BSS might unfairly
interfere with the operation of existing 20 MHz BSSs. The AP (or some of
its associated HT STAs) is required to scan all of the channels of the
current regulatory domain in order to ascertain the operating channels
of any existing 20 MHz BSSs and 20/40 MHz BSSs. (IEEE 802.11-2012 S.5.2
Establishing a 20/40 MHz BSS).

Add the check for an overlapping 20 MHz BSS to the initial AP scan for
the P == OT_i case in 10.15.3.2.

Signed-off-by: Rajkumar Manoharan <rmanohar@qti.qualcomm.com>
10 years agoWPS: Deinit before wpas_p2p_disconnect()
Eduardo Abinader [Wed, 14 May 2014 13:08:47 +0000 (09:08 -0400)]
WPS: Deinit before wpas_p2p_disconnect()

When there is a pending WPS negotiation for P2P and signal interrupt is
triggered, invalid read occurs in wpas_wps_terminate_pending() if
wpas_p2p_disconnect() removed the interface. Inverting deinit order
solves the issue.

Signed-off-by: Eduardo Abinader <eduardo.abinader@openbossa.org>
10 years agoSCARD: Fix GSM authentication on USIM
Simon Baatz [Thu, 15 May 2014 18:10:39 +0000 (20:10 +0200)]
SCARD: Fix GSM authentication on USIM

scard_gsm_auth() used SIM_CMD_GET_RESPONSE for both SIM and USIM. Convert
the command into USIM_CMD_GET_RESPONSE for USIM.

Since commit eb324600295 ("Fix switching from EAP-SIM to EAP-AKA/AKA'")
EAP-SIM is using the USIM if available. This triggers a probably ancient
bug in scard_gsm_auth(), which results in sending the wrong get response
command to the USIM. Thus, EAP-SIM stopped to work after this change on
USIMs that expect the proper command.

Signed-off-by: Simon Baatz <gmbnomis@gmail.com>
10 years agotests: Verify that WPS operation report failure if WPS is disabled
Jouni Malinen [Thu, 15 May 2014 21:57:40 +0000 (00:57 +0300)]
tests: Verify that WPS operation report failure if WPS is disabled

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoWPS: Fix return value when context is not valid
Petar Koretic [Wed, 14 May 2014 13:51:04 +0000 (13:51 +0000)]
WPS: Fix return value when context is not valid

If WPS isn't enabled, hostapd_cli returns 'OK' even though WPS doesn't
get activated because WPS context is not valid:

$ hostapd_cli wps_pbc
Selected interface 'wlan0'
OK

$ hostapd_cli wps_cancel
Selected interface 'wlan0'
OK

Fix this by returning appropriate error when WPS fails to activate:

$ hostapd_cli wps_pbc
Selected interface 'wlan0'
FAIL

$ hostapd_cli wps_cancel
Selected interface 'wlan0'
FAIL

Signed-off-by: Petar Koretic <petar.koretic@sartura.hr>
CC: Luka Perkov <luka.perkov@sartura.hr>
10 years agotests: Make GAS test cases more robust
Jouni Malinen [Thu, 15 May 2014 21:27:47 +0000 (00:27 +0300)]
tests: Make GAS test cases more robust

Scan explicitly for the specific AP to work around issues where under
heavy CPU load, the single active scan round may miss the delayed Probe
Response from the AP. In addition, verify that ANQP_GET commands succeed
to make error cases clearer in the log.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Verify P2P GO Negotiation wait-for-ready timeout
Jouni Malinen [Thu, 15 May 2014 20:58:10 +0000 (23:58 +0300)]
tests: Verify P2P GO Negotiation wait-for-ready timeout

This verifies that the wait for peer to be ready for GO Negotiation is
timed out properly at no less than 120 seconds. Since this is a long
test case, it is disabled by default without the --long option.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Modify the timeout for GO Negotiation on no concurrent session
Rashmi Ramanna [Wed, 14 May 2014 09:05:02 +0000 (14:35 +0530)]
P2P: Modify the timeout for GO Negotiation on no concurrent session

Peer should handle a GO Negotiation exchange correctly when the
responding device does not have WSC credentials available at the
time of receiving the GO Negotiation Request. WSC Credentials
(e.g., Pushbutton) can be entered within the 120 second timeout.

Presently, if concurrent session is not active, the peer would wait for
GO Negotiation Request frame from the other device for approximately one
minute due to the earlier optimization change in commit
a2d63657603b8f0714274f34bea45cb5d0c0a7b9. To meet the two minute
requirement, replace this design based on number of iterations with a
more appropriate wait for the required number of seconds.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Refrain from performing extended listen during PD
Jouni Malinen [Thu, 15 May 2014 19:16:19 +0000 (22:16 +0300)]
P2P: Refrain from performing extended listen during PD

Extend the previous commit 0f1034e3889e7b8f54ed59317f1234db8167d12e to
skip extended listen also based on ongoing provision discovery operation
(which does not show up as a separate P2P module state and as such, was
not coveraged by the previous commit).

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Verify P2P_FIND operation when interface gets disabled
Jouni Malinen [Thu, 15 May 2014 18:35:09 +0000 (21:35 +0300)]
tests: Verify P2P_FIND operation when interface gets disabled

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Reject P2P_FIND and P2P_LISTEN on disabled interface
Jouni Malinen [Thu, 15 May 2014 18:34:32 +0000 (21:34 +0300)]
P2P: Reject P2P_FIND and P2P_LISTEN on disabled interface

This makes the P2P operations behave more consistently with the SCAN
command.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Clear P2P state if active interface is disabled
Jouni Malinen [Thu, 15 May 2014 18:32:54 +0000 (21:32 +0300)]
P2P: Clear P2P state if active interface is disabled

The radio works for the interface get removed if interface is disabled.
This could have left P2P module in invalid state if the interface got
disabled during a p2p_find or p2p_listen operation. Clear the state in
such a case to avoid blocking following operations due to P2P module
assuming it is still in progress of doing something.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Verify scan behavior when interface gets disabled
Jouni Malinen [Thu, 15 May 2014 18:10:52 +0000 (21:10 +0300)]
tests: Verify scan behavior when interface gets disabled

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoAdd DRIVER_EVENT ctrl_iface command for testing purposes
Jouni Malinen [Thu, 15 May 2014 18:09:48 +0000 (21:09 +0300)]
Add DRIVER_EVENT ctrl_iface command for testing purposes

This new command can be used to simulate driver events without having to
go through the driver wrapper or kernel code for this. This enables more
testing coverage with hwsim.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Show number of remaining parallel VMs
Jouni Malinen [Thu, 15 May 2014 14:50:51 +0000 (17:50 +0300)]
tests: Show number of remaining parallel VMs

Make the parallel-vm.sh output a bit more helpful by showing a count of
running VMs.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make ap_acl_accept and ap_acl_deny more robust
Jouni Malinen [Thu, 15 May 2014 15:04:59 +0000 (18:04 +0300)]
tests: Make ap_acl_accept and ap_acl_deny more robust

Use scan_for_bss() to make sure the AP is seen even under heavy CPU
load.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make wpas_ctrl_bssid_filter more robust
Jouni Malinen [Thu, 15 May 2014 13:57:10 +0000 (16:57 +0300)]
tests: Make wpas_ctrl_bssid_filter more robust

Use scan_for_bss() to make sure the AP is seen even under heavy CPU
load.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make discovery_group_client more robust
Jouni Malinen [Thu, 15 May 2014 11:13:39 +0000 (14:13 +0300)]
tests: Make discovery_group_client more robust

Allow three P2P_FIND attempts for discovering the GO on a non-social
channels since the single Probe Response frame can be missed easily
under heavy CPU load.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agowlantest: Add support for OSEN
Jouni Malinen [Thu, 15 May 2014 09:40:02 +0000 (12:40 +0300)]
wlantest: Add support for OSEN

This allows Hotspot 2.0 OSEN connection to be analyzed more
conveniently. The frames from an OSEN association can now be decrypted
using an MSK file.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agotests: Make WPS test cases more robust
Jouni Malinen [Wed, 14 May 2014 22:02:03 +0000 (01:02 +0300)]
tests: Make WPS test cases more robust

Scan explicitly for the AP that may be started during the test case
execution. This is needed to work around issues where under heavy CPU
load, the single active scan round may miss the delayed Probe Response
from the second AP.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make ap_wpa2_eap_ttls_server_cert_hash_invalid more robust
Jouni Malinen [Wed, 14 May 2014 21:31:30 +0000 (00:31 +0300)]
tests: Make ap_wpa2_eap_ttls_server_cert_hash_invalid more robust

Instead of checking for multiple EAP starts (which can occur if
EAPOL-Start from supplicant goes out quickly enough, e.g., due to CPU
load), look for the explicit message indicating that TTLS method
initialization failed.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make FT test cases more robust
Jouni Malinen [Wed, 14 May 2014 14:02:32 +0000 (17:02 +0300)]
tests: Make FT test cases more robust

Scan explicitly for the AP that may be started during the test case
execution. This is needed to work around issues where under heavy CPU
load, the single active scan round may miss the delayed Probe Response
from the second AP. In addition, check for ROAM/FT_DS failures to be
able to report errors more clearly.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make HS 2.0 test cases more robust
Jouni Malinen [Wed, 14 May 2014 10:35:32 +0000 (13:35 +0300)]
tests: Make HS 2.0 test cases more robust

Scan explicitly for the AP that may be started during the test case
execution. This is needed to work around issues where under heavy CPU
load, the single active scan round may miss the delayed Probe Response
from the second AP.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make scan and scan_only more robust
Jouni Malinen [Wed, 14 May 2014 09:54:13 +0000 (12:54 +0300)]
tests: Make scan and scan_only more robust

These can fail during heavy CPU load due to active scan dwell time not
being long enough to catch the delayed Probe Response frame from the AP.
Work around this by allowing multiple scan attempts to see the response.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Verify P2P GO start when scan_req = MANUAL_SCAN_REQ
Jouni Malinen [Tue, 13 May 2014 23:31:55 +0000 (02:31 +0300)]
tests: Verify P2P GO start when scan_req = MANUAL_SCAN_REQ

There was a bug in this code path that resulted in the
skip-scan-to-start-GO case to not actually skip the scan. It looks like
this could be hit at least when autoscan was enabled, but it is possible
that some other sequences could hit this as well.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoP2P: Make sure GO start does not miss connect_without_scan
Jouni Malinen [Tue, 13 May 2014 23:13:18 +0000 (02:13 +0300)]
P2P: Make sure GO start does not miss connect_without_scan

It looks like there was a possible sequence for wpa_s->scan_req to be
MANUAL_SCAN_REQ at the moment a GO is to be started. This could result
in the "Request scan (that will be skipped) to start GO" to actually not
skip the scan and end up stuck waiting for something external to trigger
a scan before the GO could be started. Fix this by clearing
wpa_s->scan_req when deciding to start the GO.

This issue could be hit at least by first enabling autoscan and then
issuing P2P_GROUP_ADD. Other sequences that set wpa_s->scan_req to
MANUAL_SCAN_REQ without going through wpa_supplicant_scan() to clear it
immediately could also have similar effect (and there is even a small
window for the wpa_supplicant_scan() call to happen only after the
P2P_GROUP_ADD command is processed, so this could potentially have
happened even with SCAN + P2P_GROUP_ADD).

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Work around grpform_pbc_overlap robustness issues
Jouni Malinen [Tue, 13 May 2014 22:29:55 +0000 (01:29 +0300)]
tests: Work around grpform_pbc_overlap robustness issues

Since P2P Client scan case is now optimzied to use a specific SSID, the
WPS AP will not reply to that and the scan after GO Negotiation can
quite likely miss the AP due to dwell time being short enoguh to miss
the Beaco frame. This has made the test case somewhat pointless, but
keep it here for now with an additional scan to confirm that PBC
detection works if there is a BSS entry for a overlapping AP.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agonl80211: Fix send_frame freq for IBSS
Jouni Malinen [Tue, 13 May 2014 21:47:42 +0000 (00:47 +0300)]
nl80211: Fix send_frame freq for IBSS

bss->freq was not updated for IBSS, so whatever old value was stored
from a previous AP mode operation could end up having been used as the
channel when trying to send Authentication frames in an RSN IBSS. This
resulted in the frame not sent (cfg80211 rejects it) and potentially not
being able to re-establish connection due to 4-way handshake failing
with replay counter mismatches. Fix this by learning the operating
channel of the IBSS both when join event is received and when a
management frame is being transmitted since the IBSS may have changed
channels due to merges.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Make pmksa_cache_on_roam_back more robust
Jouni Malinen [Tue, 13 May 2014 21:12:40 +0000 (00:12 +0300)]
tests: Make pmksa_cache_on_roam_back more robust

The single channel scan while associated to another AP and immediately
after starting the second AP can miss the Probe Response frame
especially under heavy CPU load. Avoid false error reports by allowing
multiple scan rounds to be performed. wpas_ctrl_bssid_filter is also
modified to take into account different get_bss() behavior.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoP2P: Fix scan optimization for GO during persistent group invocation
Rashmi Ramanna [Tue, 13 May 2014 10:44:18 +0000 (16:14 +0530)]
P2P: Fix scan optimization for GO during persistent group invocation

Commit 41d5ce9e0b7b37dd84fbf3c1aa5ed571c32321d4 was intended to scan for
GO on the negotiated channel for few iterations, but it did not work
correctly due to incorrect operator being used. Fix this by requiring
both conditions to be met for the single channel scan.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoP2P: Validate GO operating channel on channel list changes
Edhar, Mahesh Kumar [Mon, 5 May 2014 06:33:21 +0000 (12:03 +0530)]
P2P: Validate GO operating channel on channel list changes

On receiving CHANNEL_LIST_CHANGED event from driver, verify that local
GO (if any) is operating in valid frequency. If not, we should remove
the group and reform on valid frequency. Indicate this similarly to the
avoid-frequency notification (i.e., a control interface message for
upper layers to react to this for now; potentially CSA later).

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoRemove leftover timeouts on cleanup
Alexandru Costache [Mon, 12 May 2014 16:14:55 +0000 (19:14 +0300)]
Remove leftover timeouts on cleanup

Signed-off-by: Alexandru Costache <alexandru.costache.100@gmail.com>
10 years agotests: DUP_NETWORK
Jouni Malinen [Mon, 12 May 2014 15:50:05 +0000 (18:50 +0300)]
tests: DUP_NETWORK

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoAdd 'dup_network <id_s> <id_d> <name>' command
Dmitry Shmidt [Tue, 29 Apr 2014 20:41:47 +0000 (13:41 -0700)]
Add 'dup_network <id_s> <id_d> <name>' command

This command allows to copy network variable from one network to
another, e.g., to clone the psk field without having to extract it from
wpa_supplicant.

Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
10 years agodbus: Reorder deauthentication and cleanup calls when removing a network
Hannu Mallat [Mon, 12 May 2014 11:33:12 +0000 (14:33 +0300)]
dbus: Reorder deauthentication and cleanup calls when removing a network

Valgrind indicates reference to already freed memory if function
wpa_config_remove_network() is called prior to calling
wpa_supplicant_deauthenticate(), and this can lead to a crash.
Inverting the call order fixes the problem.

Signed-off-by: Hannu Mallat <hannu.mallat@jollamobile.com>
10 years agotests: Clear cfg80211 scan results before HT coex tests
Jouni Malinen [Sun, 11 May 2014 20:01:10 +0000 (23:01 +0300)]
tests: Clear cfg80211 scan results before HT coex tests

This is needed to avoid incorrect failures based on neighboring BSSes
from earlier test cases.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: UNAUTH-TLS
Jouni Malinen [Sun, 11 May 2014 17:59:07 +0000 (20:59 +0300)]
tests: UNAUTH-TLS

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-pwd with server fragmenting messages
Jouni Malinen [Sun, 11 May 2014 17:54:15 +0000 (20:54 +0300)]
tests: EAP-pwd with server fragmenting messages

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoEAP-pwd server: Allow fragment_size to be configured
Jouni Malinen [Sun, 11 May 2014 17:53:41 +0000 (20:53 +0300)]
EAP-pwd server: Allow fragment_size to be configured

Previously, the fragment_size parameter was ignored and the default
value of 1020 was hardcoded.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-IKEv2 fragmentation
Jouni Malinen [Sun, 11 May 2014 17:42:18 +0000 (20:42 +0300)]
tests: EAP-IKEv2 fragmentation

This adds a test case for the server fragmenting an EAP-IKEv2 message.
In addition, the fragmentation threshold is made shorter to trigger
fragmentation for all messages.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoEAP-IKEv2: Allow frag ack without integrity checksum
Jouni Malinen [Sun, 11 May 2014 17:40:44 +0000 (20:40 +0300)]
EAP-IKEv2: Allow frag ack without integrity checksum

RFC 5106 is not exactly clear on the requirements for the "no data"
packet that is used to acknowledge a fragmented message. Allow it to be
processed without the integrity checksum data field since it is possible
to interpret the RFC as this not being included. This fixes reassembly
of fragmented frames after keys have been derived.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-FAST and missing PAC configuration
Jouni Malinen [Sun, 11 May 2014 17:09:40 +0000 (20:09 +0300)]
tests: EAP-FAST and missing PAC configuration

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-FAST using binary PAC file format
Jouni Malinen [Sun, 11 May 2014 17:03:45 +0000 (20:03 +0300)]
tests: EAP-FAST using binary PAC file format

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-pwd with invalid group
Jouni Malinen [Sun, 11 May 2014 15:39:54 +0000 (18:39 +0300)]
tests: EAP-pwd with invalid group

This could result in wpa_supplicant segmentation fault prior to the
EAP-pwd fix for error case handling.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoEAP-pwd: Fix processing of group setup failure
Jouni Malinen [Sun, 11 May 2014 15:38:07 +0000 (18:38 +0300)]
EAP-pwd: Fix processing of group setup failure

If invalid group was negotiated, compute_password_element() left some of
the data->grp pointer uninitialized and this could result in
segmentation fault when deinitializing the EAP method. Fix this by
explicitly clearing all the pointer with eap_zalloc(). In addition,
speed up EAP failure reporting in this type of error case by indicating
that the EAP method execution cannot continue anymore on the peer side
instead of waiting for a timeout.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Fix scan_bss_operations
Jouni Malinen [Sun, 11 May 2014 15:18:58 +0000 (18:18 +0300)]
tests: Fix scan_bss_operations

The BSS id numbers were assumed to start from 0 at the beginning of this
test case, but that is only the case if this is run as the first test
after starting wpa_supplicant. Fix the test case to figure out the id
values dynamically to avoid false errors.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-SIM/AKA configuration parameters
Jouni Malinen [Sun, 11 May 2014 14:08:31 +0000 (17:08 +0300)]
tests: EAP-SIM/AKA configuration parameters

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoEAP-pwd peer: Export Session-Id through getSessionId callback
Jouni Malinen [Sun, 11 May 2014 18:22:55 +0000 (21:22 +0300)]
EAP-pwd peer: Export Session-Id through getSessionId callback

EAP-pwd was already deriving the EAP Session-Id, but it was not yet
exposed through the EAP method API.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoeapol_test: Check EAP-Key-Name
Jouni Malinen [Sun, 11 May 2014 09:43:09 +0000 (12:43 +0300)]
eapol_test: Check EAP-Key-Name

The new command line argument -e can be used to request the server to
send EAP-Key-Name in Access-Accept. If both the local EAP peer
implementation and server provide the EAP Session-Id, compare those
values and indicate in debug log whether a match was seen.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoRADIUS: Define EAP-Key-Name
Jouni Malinen [Sun, 11 May 2014 09:42:55 +0000 (12:42 +0300)]
RADIUS: Define EAP-Key-Name

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP-SIM/AKA/AKA' with SQLite
Jouni Malinen [Sat, 10 May 2014 22:06:00 +0000 (01:06 +0300)]
tests: EAP-SIM/AKA/AKA' with SQLite

Extend EAP-SIM/AKA/AKA' test coverage by setting up another
authentication server instance to store dynamic SIM/AKA/AKA' information
into an SQLite database. This allows the stored reauth/pseudonym data to
be modified on the server side and by doing so, allows testing fallback
from reauth to pseudonym/permanent identity.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoEAP-SIM peer: Fix counter-too-small message building
Jouni Malinen [Sun, 11 May 2014 14:54:59 +0000 (17:54 +0300)]
EAP-SIM peer: Fix counter-too-small message building

The extra data (nonce_s) used in this message was pointing to the
parsed, decrypted data and that buffer was previously freed just before
building the new message. This resulted in use of freed data and
possibly incorrect extra data value that caused the authentication
attempt to fail. Fix this by reordering the code to free the decrypted
data only after the new message has been generated. This was already the
case for EAP-AKA/AKA', but somehow missing from EAP-SIM.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP with expanded Nak
Jouni Malinen [Sat, 10 May 2014 18:33:49 +0000 (21:33 +0300)]
tests: EAP with expanded Nak

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: EAP roundtrip limit
Jouni Malinen [Sat, 10 May 2014 18:20:54 +0000 (21:20 +0300)]
tests: EAP roundtrip limit

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: HS 2.0 with FT
Jouni Malinen [Sat, 10 May 2014 13:17:41 +0000 (16:17 +0300)]
tests: HS 2.0 with FT

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoInterworking: Allow FT to be used for connection
Jouni Malinen [Sat, 10 May 2014 13:15:20 +0000 (16:15 +0300)]
Interworking: Allow FT to be used for connection

This extends Interworking network selection to enable FT-EAP as an
optional key_mgmt value to allow FT to be used instead of hardcoding
WPA2-Enterprise without FT.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoRemove duplicated ibss_rsn_deinit() call
Eduardo Abinader [Wed, 7 May 2014 22:42:52 +0000 (18:42 -0400)]
Remove duplicated ibss_rsn_deinit() call

No need to call ibss_rsn_deinit() again since it is already being called
by wpa_supplicant_mark_disassoc().

Signed-off-by: Eduardo Abinader <eduardo.abinader@openbossa.org>
10 years agoX.509: Fix v3 parsing with issuerUniqueID/subjectUniqueID present
Jouni Malinen [Sat, 10 May 2014 09:36:19 +0000 (12:36 +0300)]
X.509: Fix v3 parsing with issuerUniqueID/subjectUniqueID present

The current position pointer was not updated when issuerUniqueID or
subjectUniqueID were present. This could result in extensions being
ignored.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Fix couple of compiler warnings
Jouni Malinen [Sat, 10 May 2014 09:15:36 +0000 (12:15 +0300)]
tests: Fix couple of compiler warnings

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agotests: Add 4.16.1 and 4.16.2 from NIST PKITS
Jouni Malinen [Fri, 9 May 2014 19:49:49 +0000 (22:49 +0300)]
tests: Add 4.16.1 and 4.16.2 from NIST PKITS

This adds two more X.509 path validation test cases.

Signed-off-by: Jouni Malinen <j@w1.fi>
10 years agoP2P: Refrain from performing extended listen during P2P connection
Sunil Dutt [Tue, 6 May 2014 16:34:37 +0000 (22:04 +0530)]
P2P: Refrain from performing extended listen during P2P connection

Do not perform extended listen period operations when either a P2P
connection is in progress. This makes the connection more robust should
an extended listen timer trigger during such an operation.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
10 years agoAdd macsec_qca driver wrapper
Hu Wang [Tue, 25 Mar 2014 19:39:02 +0000 (21:39 +0200)]
Add macsec_qca driver wrapper

This is based on driver_wired.c and provides driver interface for the
QCA MACsec driver.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>