mech_eap.git
9 years agomesh: Check mesh key management method
Masashi Honma [Sat, 15 Nov 2014 01:35:28 +0000 (10:35 +0900)]
mesh: Check mesh key management method

Signed-off-by: Natsuki Itaya <Natsuki.Itaya@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Add wpa_cli action script processing for mesh events
Masashi Honma [Sat, 15 Nov 2014 01:35:24 +0000 (10:35 +0900)]
mesh: Add wpa_cli action script processing for mesh events

Signed-off-by: Natsuki Itaya <Natsuki.Itaya@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Add scan result for mesh network
Masashi Honma [Mon, 1 Sep 2014 04:23:39 +0000 (00:23 -0400)]
mesh: Add scan result for mesh network

Android 4.4 uses "BSS" command instead of "SCAN_RESULT" command.
So this patch add the mesh scan result for BSS command.

Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Make BSSBasicRateSet configurable
Masashi Honma [Mon, 1 Sep 2014 04:23:38 +0000 (00:23 -0400)]
mesh: Make BSSBasicRateSet configurable

STAs that have different BSSBasicRateSet cannot connect to each other
as per IEEE 802.11s-2011 9.6.0c1:

"A mesh STA shall not establish a mesh peering with a mesh STA using a
different BSSBasicRateSet."

Make BSSBasicRateSet configurable to improve interoperability with other
stations.

Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agoAdd mesh_group_{add,remove} control interface commands
Javier Lopez [Mon, 1 Sep 2014 04:23:33 +0000 (00:23 -0400)]
Add mesh_group_{add,remove} control interface commands

Parse MESH_GROUP_ADD/REMOVE commands on ctrl interface and call
wpa_supplicant routines. These commands are used to start or
join and leave a mesh network.

The mesh id is given in the configuration file, therefore there is
no need to scan before joining a mesh network. We reuse the
connect_without_scan construct used by P2P for that same purpose.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Javier Lopez <jlopex@gmail.com>
9 years agomesh: Reduce none 11N Self-protected Action frame allocation size
Masashi Honma [Sat, 15 Nov 2014 01:35:27 +0000 (10:35 +0900)]
mesh: Reduce none 11N Self-protected Action frame allocation size

Signed-off-by: Natsuki Itaya <Natsuki.Itaya@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Fix 11N capability in Self-protected Action frame
Masashi Honma [Sat, 15 Nov 2014 01:35:26 +0000 (10:35 +0900)]
mesh: Fix 11N capability in Self-protected Action frame

mesh_ht_mode default value is CHAN_UNDEFINED.
So previous code set 11N capability even though 11N is not used.

Signed-off-by: Natsuki Itaya <Natsuki.Itaya@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Set driver capability flags to mesh interface
Masashi Honma [Sat, 15 Nov 2014 01:35:25 +0000 (10:35 +0900)]
mesh: Set driver capability flags to mesh interface

Signed-off-by: Kenzoh Nishikawa <Kenzoh.Nishikawa@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Fix segmentation fault by repeating MESH_GROUP_ADD/REMOVE
Masashi Honma [Sat, 15 Nov 2014 01:35:29 +0000 (10:35 +0900)]
mesh: Fix segmentation fault by repeating MESH_GROUP_ADD/REMOVE

Signed-off-by: Natsuki Itaya <Natsuki.Itaya@jp.sony.com>
Signed-off-by: Masashi Honma <masashi.honma@gmail.com>
9 years agomesh: Enable mesh HT mode
Jason Mobarak [Mon, 1 Sep 2014 04:23:36 +0000 (00:23 -0400)]
mesh: Enable mesh HT mode

Add a new option "mesh_ht_mode" that specifies the HT mode for the
mesh, with this option on, mesh beacons, actions frames, and probe
responses with include the appropriate HT information elements.

[original implementation by Chun-Yeow Yeoh <yeohchunyeow@gmail.com>]
[some fixes by Masashi Honma <masashi.honma@gmail.com>]
Signed-off-by: Ashok Nagarajan <ashok.dragon@gmail.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Jason Mobarak <x@jason.mobarak.name>
9 years agomesh: Add timer for SAE authentication in RSN mesh
Chun-Yeow Yeoh [Mon, 1 Sep 2014 04:23:31 +0000 (00:23 -0400)]
mesh: Add timer for SAE authentication in RSN mesh

Add timer to do SAE re-authentication with number of tries defined
by MESH_AUTH_RETRY and timeout defined by MESH_AUTH_TIMEOUT.

Ignoring the sending of reply message on "SAE confirm before commit"
to avoid "ping-pong" issues with other mesh nodes. This is obvious when
number of mesh nodes in MBSS reaching 6.

Signed-off-by: Chun-Yeow Yeoh <yeohchunyeow@gmail.com>
Signed-off-by: Bob Copeland <me@bobcopeland.com>
9 years agomesh: Start mesh peering after successful authentication
Bob Copeland [Mon, 1 Sep 2014 04:23:30 +0000 (00:23 -0400)]
mesh: Start mesh peering after successful authentication

[original patch by: Thomas Pedersen <thomas@noack.us>]
Signed-off-by: Bob Copeland <me@bobcopeland.com>
9 years agoSAE: Enhance AP implementation to handle auth for mesh interfaces
Bob Copeland [Mon, 1 Sep 2014 04:23:30 +0000 (00:23 -0400)]
SAE: Enhance AP implementation to handle auth for mesh interfaces

Add state transition logic to the SAE frame handling in order to more
fully implement the state machine from the IEEE 802.11 standard. Special
cases are needed for infrastructure BSS case to avoid unexpected
Authentication frame sequence by postponing transmission of the second
Authentication frame untile the STA sends its Confirm.

[original patch by: Thomas Pedersen <thomas@noack.us>]
Signed-off-by: Bob Copeland <me@bobcopeland.com>
9 years agomesh: Add mesh robust security network
Thomas Pedersen [Mon, 1 Sep 2014 04:23:29 +0000 (00:23 -0400)]
mesh: Add mesh robust security network

This implementation provides:

- Mesh SAE authentication mechanism
- Key management (set/get PSK)
- Cryptographic key establishment
- Enhanced protection mechanisms for robust management frames

Signed-off-by: Javier Lopez <jlopex@gmail.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Jason Mobarak <x@jason.mobarak.name>
Signed-off-by: Thomas Pedersen <thomas@noack.us>
9 years agohostapd: Add wowlan_triggers config param
Dmitry Shmidt [Wed, 3 Sep 2014 21:58:37 +0000 (14:58 -0700)]
hostapd: Add wowlan_triggers config param

New kernels in wiphy_suspend() will call cfg80211_leave_all() that will
eventually end up in cfg80211_stop_ap() unless wowlan_triggers were set.
For now, use the parameters from the station mode as-is. It may be
desirable to extend (or constraint) this in the future for specific AP
mode needs.

Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
9 years agonl80211: Fix Android compilation by adding include for fcntl.h
Ilan Peer [Wed, 5 Nov 2014 08:50:37 +0000 (03:50 -0500)]
nl80211: Fix Android compilation by adding include for fcntl.h

This is needed for fcntl() at least with Android KK.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Suite B
Jouni Malinen [Sun, 16 Nov 2014 13:42:29 +0000 (15:42 +0200)]
tests: Suite B

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoSuite B: Select EAPOL-Key integrity and key-wrap algorithms based on AKM
Jouni Malinen [Sun, 16 Nov 2014 13:40:02 +0000 (15:40 +0200)]
Suite B: Select EAPOL-Key integrity and key-wrap algorithms based on AKM

This adds support for AKM 00-0F-AC:11 to specify the integrity and
key-wrap algorithms for EAPOL-Key frames using the new design where
descriptor version is set to 0 and algorithms are determined based on
AKM.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoSuite B: PMKID derivation for AKM 00-0F-AC:11
Jouni Malinen [Sun, 16 Nov 2014 11:22:46 +0000 (13:22 +0200)]
Suite B: PMKID derivation for AKM 00-0F-AC:11

The new AKM uses a different mechanism of deriving the PMKID based on
KCK instead of PMK. hostapd was already doing this after the KCK had
been derived, but wpa_supplicant functionality needs to be moved from
processing of EAPOL-Key frame 1/4 to 3/4 to have the KCK available.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoSuite B: Add AKM 00-0F-AC:11
Jouni Malinen [Sun, 16 Nov 2014 11:20:51 +0000 (13:20 +0200)]
Suite B: Add AKM 00-0F-AC:11

This adds definitions for the 128-bit level Suite B AKM 00-0F-AC:11. The
functionality itself is not yet complete, i.e., this commit only
includes parts to negotiate the new AKM.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: Clear ignore_old_scan_res setting
Jouni Malinen [Sat, 15 Nov 2014 10:56:02 +0000 (12:56 +0200)]
tests: Clear ignore_old_scan_res setting

This parameter was used in some of the NFC test cases to make scanning
more robust in case of changing AP configuration. However, the parameter
was not cleared anywhere, so it could have been left in use for other
test cases as well. To get more consistent behavior, clear the value
back to its default between test cases.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoOpenSSL: Allow TLS v1.1 and v1.2 to be negotiated by default
Jouni Malinen [Sat, 15 Nov 2014 10:35:10 +0000 (12:35 +0200)]
OpenSSL: Allow TLS v1.1 and v1.2 to be negotiated by default

Use SSLv23_method() to enable TLS version negotiation for any version
equal to or newer than 1.0. If the old behavior is needed as a
workaround for some broken authentication servers, it can be configured
with phase1="tls_disable_tlsv1_1=1 tls_disable_tlsv1_2=1".

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoEAP-IKEv2: Fix compilation warning
Andrei Otcheretianski [Wed, 5 Nov 2014 08:50:35 +0000 (03:50 -0500)]
EAP-IKEv2: Fix compilation warning

Fix signed/unsigned comparison compilation warning introduced in
08ef442 "EAP-IKEv2: Fix the payload parser".

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
9 years agoAdd CTRL-EVENT-SCAN-FAILED notification in case of scan failure
Dmitry Shmidt [Wed, 5 Nov 2014 21:22:25 +0000 (13:22 -0800)]
Add CTRL-EVENT-SCAN-FAILED notification in case of scan failure

This is needed since the SCAN command with radio work returns before the
actual driver operation to trigger a scan has been executed and as such,
cannot return result of that operation.

Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
9 years agoWork around AP misbehavior on EAPOL-Key descriptor version
Jouni Malinen [Fri, 14 Nov 2014 18:57:05 +0000 (20:57 +0200)]
Work around AP misbehavior on EAPOL-Key descriptor version

It looks like some APs are incorrectly selecting descriptor version 3
(AES-128-CMAC) for EAPOL-Key frames when version 2 (HMAC-SHA1) was
expected to be used. This is likely triggered by an attempt to negotiate
PMF with SHA1-based AKM.

Since AES-128-CMAC is considered stronger than HMAC-SHA1, allow the
incorrect, but stronger, option to be used in these cases to avoid
interoperability issues with deployed APs.

This issue shows up with "WPA: CCMP is used, but EAPOL-Key descriptor
version (3) is not 2" in debug log. With the new workaround, this issue
is ignored and "WPA: Interoperability workaround: allow incorrect
(should have been HMAC-SHA1), but stronger (is AES-128-CMAC), descriptor
version to be used" is written to the log.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: PMF required with SHA1-based AKM
Jouni Malinen [Fri, 14 Nov 2014 18:56:43 +0000 (20:56 +0200)]
tests: PMF required with SHA1-based AKM

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agohostapd: Change drv_flags from unsigned int to u64
Yanbo Li [Sun, 2 Nov 2014 09:46:35 +0000 (17:46 +0800)]
hostapd: Change drv_flags from unsigned int to u64

Some flag already using a bit larger than 32, so extend the hostapd
drv_flags type similarly to the earlier wpa_supplicant change to get the
full flag content.

Signed-off-by: Yanbo Li <yanbol@qti.qualcomm.com>
9 years agoAssign QCA vendor specific nl80211 command id 52 for APFIND
Jouni Malinen [Thu, 13 Nov 2014 14:26:55 +0000 (16:26 +0200)]
Assign QCA vendor specific nl80211 command id 52 for APFIND

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: DFS with chanlist
Jouni Malinen [Sat, 1 Nov 2014 17:42:49 +0000 (19:42 +0200)]
tests: DFS with chanlist

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoRemove unused EVENT_MLME_RX
Jouni Malinen [Sat, 1 Nov 2014 17:31:23 +0000 (19:31 +0200)]
Remove unused EVENT_MLME_RX

This was used in driver_test.c, but that driver wrapper has been removed
and there are no remaining or expected users for EVENT_MLME_RX.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoRemove unused EVENT_FT_RRB_RX
Jouni Malinen [Sat, 1 Nov 2014 17:27:48 +0000 (19:27 +0200)]
Remove unused EVENT_FT_RRB_RX

This was used in hostapd driver_test.c, but that driver wrapper has been
removed and there are no remaining or expected users for
EVENT_FT_RRB_RX.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: Make vm-run.sh arguments non-positional
Ilan Peer [Mon, 27 Oct 2014 14:00:57 +0000 (10:00 -0400)]
tests: Make vm-run.sh arguments non-positional

This was currently breaking parallel-run.*, as it was passing
--split num/num parameters (intended for rnu-tests.py)
to vm-run.sh which broke the --codecov and --timewrap options.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Update vm README
Ilan Peer [Mon, 27 Oct 2014 14:00:56 +0000 (10:00 -0400)]
tests: Update vm README

Update the code coverage documentation to also specify the
source base directory for the code coverage generation.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Add usage() to run-all.sh
Ilan Peer [Mon, 27 Oct 2014 14:00:55 +0000 (10:00 -0400)]
tests: Add usage() to run-all.sh

The -h or --help command line arguments can now be used to request usage
information for run-all.sh.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Add option to build before running all tests
Ilan Peer [Mon, 27 Oct 2014 14:00:54 +0000 (10:00 -0400)]
tests: Add option to build before running all tests

Add an option --build to run-all.sh to build before starting to run all
the tests. In addition, add an option --codecov to extract the code
coverage data at the end of the run.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Make run-all.sh arguments non-positional
Ilan Peer [Mon, 27 Oct 2014 14:00:53 +0000 (10:00 -0400)]
tests: Make run-all.sh arguments non-positional

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agoRemove completed to-do items
Jouni Malinen [Sat, 1 Nov 2014 14:24:36 +0000 (16:24 +0200)]
Remove completed to-do items

While this file has not really been updated in years, some of the
completed items can easily be removed.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoRemove references to madwifi from various files
Jouni Malinen [Sat, 1 Nov 2014 14:15:53 +0000 (16:15 +0200)]
Remove references to madwifi from various files

Number of documentation and configuration files had references to the
madwifi driver interface that was removed in the previous commit. Remove
these references as well.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agomadwifi: Remove obsolete madwifi driver interface
Jouni Malinen [Sat, 1 Nov 2014 14:13:41 +0000 (16:13 +0200)]
madwifi: Remove obsolete madwifi driver interface

The MadWifi project is not active anymore and the last release happened
in early 2008. As such, there is no remaining justification for
maintaining the madwifi-specific driver interface for hostapd either.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agodbus: Fix memory leak in wpas_dbus_getter_bss_wps
Slava Monich [Tue, 28 Oct 2014 16:20:37 +0000 (18:20 +0200)]
dbus: Fix memory leak in wpas_dbus_getter_bss_wps

Signed-off-by: Slava Monich <slava.monich@jolla.com>
9 years agowpa_gui: Use dialog window type
Martin Kletzander [Wed, 29 Oct 2014 10:12:36 +0000 (11:12 +0100)]
wpa_gui: Use dialog window type

This helps window managers treat the window properly. Mostly tiling WMs
are affected by this. All other windows inherit this option from QDialog
already.

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
9 years agotests: Concurrent station mode scan and p2p_find with P2P Device
Jouni Malinen [Thu, 30 Oct 2014 22:48:12 +0000 (00:48 +0200)]
tests: Concurrent station mode scan and p2p_find with P2P Device

This verifies that station mode interface SCAN command gets executed if
P2P Device instance is running p2p_find.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: P2P_STOP_FIND/P2P_FLUSH to ignore new results
Jouni Malinen [Thu, 30 Oct 2014 18:01:21 +0000 (20:01 +0200)]
tests: P2P_STOP_FIND/P2P_FLUSH to ignore new results

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Do not consider the p2p_scan results after p2p_stop_find
Sunil Dutt [Thu, 30 Oct 2014 15:21:52 +0000 (20:51 +0530)]
P2P: Do not consider the p2p_scan results after p2p_stop_find

If p2p_stop_find is issued after the p2p_scan request is triggered to
the host driver, the obtained scan results are also considered to update
the P2P peer entries. This is not always desired behavior, i.e., it can
be clearer if no P2P-DEVICE-FOUND events are generated based of that
final pending scan.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Document p2p_in_progress() return value 2
Jouni Malinen [Thu, 30 Oct 2014 18:13:16 +0000 (20:13 +0200)]
P2P: Document p2p_in_progress() return value 2

Function documentation was not in sync with the implementation.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Delay scan operation only when P2P is not in search state
Sunil Dutt [Thu, 30 Oct 2014 12:07:46 +0000 (17:37 +0530)]
P2P: Delay scan operation only when P2P is not in search state

With the radio work interface in place, station interface SCAN command
was not scheduled (i.e., it got continously delayed with "Delay station
mode scan while P2P operation is in progress") when a p2p_find was
operational. Fix this be delaying station mode scan only when a P2P
operation is in progress, but not in search state.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Set p2p_scan_running based on driver scan request result
Sunil Dutt [Thu, 30 Oct 2014 10:50:22 +0000 (16:20 +0530)]
P2P: Set p2p_scan_running based on driver scan request result

With the radio work interface, the actual request to start p2p_scan
operation is scheduled from a radio work and hence the initial return
value cannot provide the real result of the driver operation to trigger
a scan. Introduce a new notification API to indicate the scan trigger
status based on which the p2p_scan_running instance can be set using the
real return value from the driver operation.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoMACsec: Fix policy configuration
Jouni Malinen [Thu, 30 Oct 2014 09:43:47 +0000 (11:43 +0200)]
MACsec: Fix policy configuration

macsec_validate variable was set incorrectly to FALSE(0) or TRUE(1)
instead of the enum validate_frames values (Disabled(0), Checked(1),
Strict(2). This ended up policy == SHOULD_SECURE to be mapped to
macsec_validate == Checked instead of Strict. This could have resulted
in unintended SecY forwarding of invalid packets rather than dropping
them.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Clean up p2p_go_neg_failed() calls
Jouni Malinen [Wed, 29 Oct 2014 11:04:09 +0000 (13:04 +0200)]
P2P: Clean up p2p_go_neg_failed() calls

This function is always called with the peer argument equal to
p2p->go_neg_peer, so there is no need for that argument to be there. In
addition, p2p->go_neg_peer is not NULL in cases where there is an
ongoing GO Negotiation, so the function can be simplified to just check
once whether the peer pointer is set and if not, skip all processing.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: GO Negotiation and timeout while waiting for peer
Jouni Malinen [Tue, 28 Oct 2014 14:39:33 +0000 (16:39 +0200)]
tests: GO Negotiation and timeout while waiting for peer

grpform_cred_ready_timeout2 is similar to the grpform_cred_ready_timeout
test case with the difference being in initiating a P2P_FIND operation
during the wait.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoP2P: Use timer to fail GO Negotation while waiting for peer
Rashmi Ramanna [Tue, 28 Oct 2014 11:26:43 +0000 (16:56 +0530)]
P2P: Use timer to fail GO Negotation while waiting for peer

The timeout check while waiting for the peer to accept the GO
Negotiation depended on the WAIT_PEER_IDLE or WAIT_PEER_CONNECT states
being in use. Any P2P command to alter such states would have resulted
in the failure to time out GO Negotiation and thus ended up in not
indicating GO Negotiation failure or left the selected peer available
for new GO negotiation after the expected two minute timeout.

Fix this by using a separate timer to time out GO Negotiation
irrespective of the P2P state.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: Remove ap_ht40_csa and ap_ht40_csa3 workarounds
Jouni Malinen [Tue, 28 Oct 2014 12:01:16 +0000 (14:01 +0200)]
tests: Remove ap_ht40_csa and ap_ht40_csa3 workarounds

The issue in mac80211 was identified and proposed fix is now available,
so remove the workaround here and start reporting failures from
unexpected disconnection during CSA.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoP2P: Do not call wpas_p2p_disconnect_safely() if no match is found
Jouni Malinen [Mon, 27 Oct 2014 23:47:17 +0000 (01:47 +0200)]
P2P: Do not call wpas_p2p_disconnect_safely() if no match is found

In theory, wpas_p2p_disconnect_safely() could have dereferenced the
wpa_s == NULL argument, but in practice, it won't due to the
calling_wpa_s == wpa_s check and wpas_p2p_disconnect() accepting NULL.
Anyway, it is cleaner to add an explicit check for this. (CID 74492)

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Fix a copy-paste error in key offload management
Jouni Malinen [Mon, 27 Oct 2014 23:41:57 +0000 (01:41 +0200)]
nl80211: Fix a copy-paste error in key offload management

Commit b41f26845aaa7cf8aed6e4889e7041debc476ef9 ('Add support for
offloading key management operations to the driver') used incorrect
variable for determining the KCK length. This does not get triggered in
normal use cases since KCK and KEK are always included and of the same
length (in currently supported key management cases). Anyway, this needs
to be fixed to check the correct attribute. (CID 74491)

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoAP: Add support for Proxy ARP, DHCP snooping mechanism
Kyeyoon Park [Fri, 26 Sep 2014 05:32:55 +0000 (22:32 -0700)]
AP: Add support for Proxy ARP, DHCP snooping mechanism

Proxy ARP allows the AP devices to keep track of the hardware address to
IP address mapping of the STA devices within the BSS. When a request for
such information is made (i.e., ARP request, Neighbor Solicitation), the
AP will respond on behalf of the STA device within the BSS. Such
requests could originate from a device within the BSS or also from the
bridge. In the process of the AP replying to the request (i.e., ARP
reply, Neighbor Advertisement), the AP will drop the original request
frame. The relevant STA will not even know that such information was
ever requested.

This feature is a requirement for Hotspot 2.0, and is defined in IEEE
Std 802.11-2012, 10.23.13. While the Proxy ARP support code mainly
resides in the kernel bridge code, in order to optimize the performance
and simplify kernel implementation, the DHCP snooping code was added to
the hostapd.

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agocommon: Add definition for ETH_HLEN
Kyeyoon Park [Thu, 25 Sep 2014 06:17:29 +0000 (23:17 -0700)]
common: Add definition for ETH_HLEN

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agoAP: Add support for setting bridge network parameter
Kyeyoon Park [Sat, 27 Sep 2014 04:09:54 +0000 (21:09 -0700)]
AP: Add support for setting bridge network parameter

This allows setting a network parameter on the bridge that the BSS
belongs to.

This commit adds the needed functionality in driver_nl80211.c for the
Linux bridge implementation. In theory, this could be shared with
multiple Linux driver interfaces, but for now, only the main nl80211
interface is supported.

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agoAP: Add support for setting bridge port attributes
Kyeyoon Park [Fri, 26 Sep 2014 04:04:43 +0000 (21:04 -0700)]
AP: Add support for setting bridge port attributes

This allows setting a bridge port attribute. Specifically, the bridge
port in this context is the port to which the BSS belongs.

This commit adds the needed functionality in driver_nl80211.c for the
Linux bridge implementation. In theory, this could be shared with
multiple Linux driver interfaces, but for now, only the main nl80211
interface is supported.

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agoAP: Add support for IPv4 neighbor entry management to the BSS bridge
Kyeyoon Park [Wed, 24 Sep 2014 06:02:04 +0000 (23:02 -0700)]
AP: Add support for IPv4 neighbor entry management to the BSS bridge

This allows adding/deleting an IPv4 neighbor entry to/from the bridge,
to which the BSS belongs. This commit adds the needed functionality in
driver_nl80211.c for the Linux bridge implementation. In theory, this
could be shared with multiple Linux driver interfaces, but for now, only
the main nl80211 interface is supported.

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agonl80211: Make br_ifindex available in i802_bss
Kyeyoon Park [Wed, 10 Sep 2014 05:25:16 +0000 (22:25 -0700)]
nl80211: Make br_ifindex available in i802_bss

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agol2_packet: Add support for DHCP packet filter in l2_packet_linux
Kyeyoon Park [Mon, 8 Sep 2014 21:27:11 +0000 (14:27 -0700)]
l2_packet: Add support for DHCP packet filter in l2_packet_linux

Signed-off-by: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
9 years agotests: Scan result update changing last_scan_res and connect_work
Jouni Malinen [Mon, 27 Oct 2014 22:24:11 +0000 (00:24 +0200)]
tests: Scan result update changing last_scan_res and connect_work

This verifies that last_scan_res and connect work pointers to BSS
entries get updated when scan result update ends up removing or
reallocating a BSS entry.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoUpdate pending connect radio work BSS pointer on scan update
Jouni Malinen [Mon, 27 Oct 2014 22:19:24 +0000 (00:19 +0200)]
Update pending connect radio work BSS pointer on scan update

It is possible for scan result processing or BSS entry removal to occur
while there is a pending connect or sme-connect radio work with a
previously selected BSS entry. The BSS pointer was previously verified
to be valid, i.e., still point to a BSS entry, at the time the actual
connection operation is started. However, that BSS entry could have
changed to point to another BSS if the old BSS entry was either removed
or reallocated and a new BSS entry was added at the same location in
memory. This could result in the connection attempt failing to configure
parameters properly due to different BSS information (e.g., different
BSSID).

Fix this by updated the pending connect radio work data on BSS entry
updates similarly to how the last_scan_res array was updated. If the
selected BSS entry is removed, this will still result in a failed
connection, but reallocated BSS entry is now followed properly and used
when the connection work starts.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: WPS with ap_scan=2
Jouni Malinen [Mon, 27 Oct 2014 18:19:45 +0000 (20:19 +0200)]
tests: WPS with ap_scan=2

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoInvoke connect work done for all the connection failure cases
Sunil Dutt [Mon, 20 Oct 2014 04:48:47 +0000 (10:18 +0530)]
Invoke connect work done for all the connection failure cases

This commit fixes couple of connection failure paths where
wpas_connect_work_done() was not called, thus enabling the radio work
interface to proceed with the other queued actions.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agoWPS: Fix ap_scan=2 special association case
Jouni Malinen [Mon, 27 Oct 2014 18:13:39 +0000 (20:13 +0200)]
WPS: Fix ap_scan=2 special association case

Commit 22628eca3440976bf51846da0554099f7429b206 ('Support driver-based
BSS selection in ap_scan=1 mode') ended up disabling the special
ap_scan=2 WPS case where ap_scan=1 like scan followed by association is
used to find the WPS AP using wildcard SSID. Fix this by allowing
association request even with wpas_driver_bss_selection() when searching
for a WPS AP.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agodbus: Redirect P2P request to the managment device if present
Tomasz Bursztyka [Thu, 23 Oct 2014 08:31:21 +0000 (11:31 +0300)]
dbus: Redirect P2P request to the managment device if present

In case of a P2P management device being present, it will be up to that
interface to handle P2P requests.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
9 years agonotify: Do not raise any signal from a P2P management interface
Tomasz Bursztyka [Thu, 23 Oct 2014 08:31:20 +0000 (11:31 +0300)]
notify: Do not raise any signal from a P2P management interface

Such interface is not registered on DBus, thus there is no need to raise
any signal from it.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
9 years agowpa_supplicant: Do not register a P2P management interface on DBus
Tomasz Bursztyka [Thu, 23 Oct 2014 08:31:19 +0000 (11:31 +0300)]
wpa_supplicant: Do not register a P2P management interface on DBus

DBus client should always request the proper netdev interface. This will
be necessary to get a consistent behavior whatever driver is in use:
iwlwifi (which requires a P2P mgmt dev), ath9/10k (which does not),
etc...

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
9 years agowpa_supplicant: Parent interface should know about its dedicated P2P dev
Tomasz Bursztyka [Thu, 23 Oct 2014 08:31:18 +0000 (11:31 +0300)]
wpa_supplicant: Parent interface should know about its dedicated P2P dev

This will be useful to hand-off P2P commands from the parent interface
to its dedicated P2P device if that one is present, in DBus interface.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
9 years agoP2P: Support GCMP as part of 60 GHz support
Ahmad Masri [Sun, 26 Oct 2014 12:04:29 +0000 (12:04 +0000)]
P2P: Support GCMP as part of 60 GHz support

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: Allow GCMP to be added into ap_wps_mixed_cred network profile
Jouni Malinen [Mon, 27 Oct 2014 14:51:04 +0000 (16:51 +0200)]
tests: Allow GCMP to be added into ap_wps_mixed_cred network profile

This is needed in preparation for WPS AES being mapped to enabling both
CCMP and GCMP if the driver supports both ciphers.

Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
9 years agotests: Add support for reading modules list from a file
Ilan Peer [Sun, 26 Oct 2014 07:06:33 +0000 (03:06 -0400)]
tests: Add support for reading modules list from a file

The new run-tests.py argument "-l <file>" can now be used to specify the
test modules using a text file.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Add usage() to build.sh
Ilan Peer [Sun, 26 Oct 2014 07:06:30 +0000 (03:06 -0400)]
tests: Add usage() to build.sh

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Add an option to force copy config files
Ilan Peer [Sun, 26 Oct 2014 07:06:29 +0000 (03:06 -0400)]
tests: Add an option to force copy config files

"./build.sh --force-config" can be used to force copying configuration
files before building. This will overwrite any changes in .config files.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Add an option to build with code coverage
Ilan Peer [Sun, 26 Oct 2014 07:06:28 +0000 (03:06 -0400)]
tests: Add an option to build with code coverage

"./build.sh --codecov" can now be used to request CONFIG_CODE_COVERAGE=y
to be added to hostapd and wpa_supplicant .config files.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agonl80211: Move more event processing into driver_nl80211_event.c
Jouni Malinen [Sun, 26 Oct 2014 14:23:46 +0000 (16:23 +0200)]
nl80211: Move more event processing into driver_nl80211_event.c

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move event handling into a separate file
Jouni Malinen [Sun, 26 Oct 2014 14:13:27 +0000 (16:13 +0200)]
nl80211: Move event handling into a separate file

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move capability determination into a separate file
Jouni Malinen [Sun, 26 Oct 2014 11:28:39 +0000 (13:28 +0200)]
nl80211: Move capability determination into a separate file

This moves significant amount of code away from large driver_nl80211.c.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move QCA DFS offload capability determination to init time
Jouni Malinen [Sun, 26 Oct 2014 11:23:30 +0000 (13:23 +0200)]
nl80211: Move QCA DFS offload capability determination to init time

There is no need to fetch this capability for each get_capa() call.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move AP monitor interface handling to a separate file
Jouni Malinen [Sun, 26 Oct 2014 10:40:10 +0000 (12:40 +0200)]
nl80211: Move AP monitor interface handling to a separate file

This moves the old monitor interface design to driver_nl80211_monitor.c.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Make some helper functions non-static
Jouni Malinen [Sun, 26 Oct 2014 15:20:37 +0000 (17:20 +0200)]
nl80211: Make some helper functions non-static

This allows more functionality to be moved to separate files.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move most of the Android code into a separate file
Jouni Malinen [Sun, 26 Oct 2014 10:27:35 +0000 (12:27 +0200)]
nl80211: Move most of the Android code into a separate file

This moves most of the Android specific implementation from
driver_nl80211.c to driver_nl80211_android.c.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Move definitions into separate header files
Jouni Malinen [Sun, 26 Oct 2014 10:02:36 +0000 (12:02 +0200)]
nl80211: Move definitions into separate header files

This is an initial step in starting to split the large driver_nl80211.c
implementation into somewhat smaller parts.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: cfg80211 connect command for WEP connection
Jouni Malinen [Sun, 26 Oct 2014 13:48:57 +0000 (15:48 +0200)]
tests: cfg80211 connect command for WEP connection

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoAvoid unsigned vs. signed comparison warning
Jouni Malinen [Sun, 26 Oct 2014 10:16:53 +0000 (12:16 +0200)]
Avoid unsigned vs. signed comparison warning

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoWrite human readable version of channel width to CSA event debug log
Jouni Malinen [Sun, 26 Oct 2014 09:36:14 +0000 (11:36 +0200)]
Write human readable version of channel width to CSA event debug log

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: External bridge control for hostapd interface
Jouni Malinen [Sun, 26 Oct 2014 09:29:16 +0000 (11:29 +0200)]
tests: External bridge control for hostapd interface

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agonl80211: Store externally managed bridge name in driver status
Jouni Malinen [Sun, 26 Oct 2014 09:20:01 +0000 (11:20 +0200)]
nl80211: Store externally managed bridge name in driver status

This makes it easier to figure out from hostapd control interface
whether an interface had been added to a bridge externally at the time
hostapd interface was enabled or if the interface gets added during
hostapd operations.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: hostapd rfkill
Jouni Malinen [Sun, 26 Oct 2014 08:53:37 +0000 (10:53 +0200)]
tests: hostapd rfkill

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agohostapd: Add INTERFACE-ENABLED and INTERFACE-DISABLED events
Jouni Malinen [Sun, 26 Oct 2014 08:52:54 +0000 (10:52 +0200)]
hostapd: Add INTERFACE-ENABLED and INTERFACE-DISABLED events

These can be convenient for upper layer programs to determine if the
hostapd interface gets disabled/re-enabled, e.g., due to rfkill
block/unblock.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: Print failure exeception if logger output goes to file
Jouni Malinen [Sun, 26 Oct 2014 08:39:32 +0000 (10:39 +0200)]
tests: Print failure exeception if logger output goes to file

This makes it more convenient to determine failure reasons during test
case development and VM testing.

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agotests: IP address in STATUS
Jouni Malinen [Sun, 26 Oct 2014 08:21:48 +0000 (10:21 +0200)]
tests: IP address in STATUS

Signed-off-by: Jouni Malinen <j@w1.fi>
9 years agoP2P: Save group common frequencies in invitation result
Ilan Peer [Mon, 7 Jul 2014 11:21:01 +0000 (14:21 +0300)]
P2P: Save group common frequencies in invitation result

Save the group common frequencies when starting a GO due to
an invitation signaling requesting to re-invoke a persistent GO.

To do so, move the code that handles the translation of p2p_channels to
frequency list into a public function so it can be re-used both when GO
Negotiation is done and invitation signaling is done.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agoP2P: Save group common frequencies
Ilan Peer [Mon, 7 Jul 2014 11:20:59 +0000 (14:20 +0300)]
P2P: Save group common frequencies

Once a P2P GO interface is configured, save the group common
frequencies, as this can be useful later for channel selection
considerations during channel switch, etc.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agoWPS: Add missing device types
Ilan Peer [Mon, 7 Jul 2014 11:21:06 +0000 (14:21 +0300)]
WPS: Add missing device types

Add missing device category and device sub category definitions.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agonl80211: Clear beacon_set when deleting a beacon from deinit_ap()
Ilan Peer [Mon, 7 Jul 2014 11:20:58 +0000 (14:20 +0300)]
nl80211: Clear beacon_set when deleting a beacon from deinit_ap()

When a beacon was deleted from the kernel in
wpa_driver_nl80211_deinit_ap(), bss->beacon_set was not cleared so
restarting the AP again was not possible.

Fix this by clearing the variable once the beacon was deleted.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agonl80211: Add indoor only and GO concurrent flags
Ilan Peer [Mon, 7 Jul 2014 11:20:55 +0000 (14:20 +0300)]
nl80211: Add indoor only and GO concurrent flags

Add the following channel flags:

* INDOOR_ONLY: The channel can be used if and only if there is a clear
  assessment that the device is operating in an indoor environment, i.e.,
  it is AC power.

* CONCURRENT_GO: The channel can be used for instantiating a GO if and
  only if there is an additional station interface that is currently
  connected to an AP on the same channel or on the same U-NII band
  (assuming that the AP is an authorized master).

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agowpa_supplicant: Use the 'no_ir' notation
Ilan Peer [Mon, 7 Jul 2014 11:20:54 +0000 (14:20 +0300)]
wpa_supplicant: Use the 'no_ir' notation

Use the 'no_ir' notation instead of the 'passive scan' and
'no_ibss' notations to match the earlier change in nl80211.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
9 years agotests: Make ap_ht40_csa and ap_ht40_csa3 less likely to fail
Jouni Malinen [Sat, 25 Oct 2014 19:27:24 +0000 (22:27 +0300)]
tests: Make ap_ht40_csa and ap_ht40_csa3 less likely to fail

It looks like cfg80211 can trigger disconnection even without disabling
HT, so the same issue may be hit with both of the channel switches in
these test cases.

Signed-off-by: Jouni Malinen <j@w1.fi>