Jouni Malinen [Tue, 26 Feb 2013 16:07:17 +0000 (18:07 +0200)]
P2P: Only schedule a single p2p_go_neg_start timeout at a time
It is possible for the driver to indicate multiple Probe Request frames
that would be processed in a single loop. If those frames happen to be
from a peer which with we are trying to start GO Negotiation, multiple
timeouts to start GO Negotiation (p2p_go_neg_start) could end up being
scheduled. This would result in confusing burst of multiple GO
Negotiation Request frames being sent once the RX loop finally
concludes. Avoid this by scheduling only a single eloop timeout to
trigger GO Negotiation regardless of how many Probe Request frames from
the peer is received. In addition, make sure this timeout gets canceled
in p2p_deinit().
Jouni Malinen [Tue, 26 Feb 2013 16:02:51 +0000 (18:02 +0200)]
P2P: Do not start new GO Neg on Probe Req when waiting for Confirm
If we have already sent out GO Negotiation Response and are waiting for
the peer to reply with GO Negotiation Confirm, there is no point in
re-starting GO Negotiation based on Probe Request frame from the peer.
Doing that would just result in confusing GO Negotiation exchange with
multiple sessions running at the same time.
Jouni Malinen [Tue, 26 Feb 2013 15:22:11 +0000 (17:22 +0200)]
P2P: Fail GO Negotiation on missing Group ID
The device that is selected as the GO shall incode P2P Group ID
attribute in GO Negotiation Response/Confirm message. Previously we did
not reject a message without that attribute since it was possible to
continue operations even without knowing the SSID. However, this can
potentially result in confusing results since missing P2P Group ID
attribute can be a sign of conflicting GO role determination (both
devices assuming the peer is the GO). To get clearer end result for the
GO Negotiation, reject this as a fatal error. In addition, stop GO
Negotiation if GO Negotiation Confirm indicates non-zero status since
that is also a fatal error.
Jouni Malinen [Tue, 26 Feb 2013 14:56:48 +0000 (16:56 +0200)]
P2P: Assign GO tie breaker bit at the same time with dialog token
Commit 624b4d5a6469c92b039e0d39655b9e0eb26588aa changed GO Negotiation
to use the same Dialog Token value for all retransmissions of the GO
Negotiation Request within the same session. However, it did leave the
tie breaker bit changing for each frame. While this should not have
caused issues for most cases, it looks like there are possible sequences
where the peer may end up replying to two GO Negotiation Request frames
with different tie breaker values. If in such a case the different GO
Negotiation Response frames are used at each device, GO role
determination may result in conflicting results when same GO intent is
used.
Fix this by assigning the tie breaker value at the same time with the
dialog token (i.e., when processing the p2p_connect command instead of
for each transmitted GO Negotiation Request frame) to avoid issues with
GO selection.
Sunil Dutt [Mon, 25 Feb 2013 08:31:50 +0000 (10:31 +0200)]
TDLS: Pass peer's VHT Capability information during sta_add
The information of the peer's VHT capability is required for the
driver to establish a TDLS link in VHT mode with a compatible peer.
Pass this information to the driver when the peer station is
getting added.
Jouni Malinen [Sun, 24 Feb 2013 08:56:29 +0000 (10:56 +0200)]
WPS NFC: Add connection handover support for ER
wpa_supplicant can now generate the WPS carrier record for connection
handover response when acting as an ER. The AP whose configuration is
provided in this way is identified with an UUID as an argument to
wps-nfc.py.
Jouni Malinen [Sat, 16 Feb 2013 17:54:09 +0000 (19:54 +0200)]
Interworking: Select highest priority cred if multiple matches
Previously, the credential to use for a connection with a specific BSS
was picked arbitrary based on first found match of each matching
mechanism. While the credential priorities were used elsewhere, this did
not take into account that different mechanisms could find multiple
matching credentials. As such, the highest priority credential was not
always used in case more than one credential matched with the selected
BSS. Fix this by checking credential priorities again during connection
request.
Jouni Malinen [Sat, 16 Feb 2013 17:15:05 +0000 (19:15 +0200)]
GAS server: Fix a regression in GAS server callback
Commit 2d9ffe1e855410839468a75cf6723ee339acdd7e broke GAS server
callback for receiving Public Action frames. The incorrect context
pointer was used in the public_action_cb2 case. Fix this to use the
correct context pointer.
Srinivasan B [Sat, 16 Feb 2013 09:15:13 +0000 (11:15 +0200)]
hostapd: Fix Max SP Length derivation from QoS Info
Hostapd provides QoS info of the STA (Service Period & AC mask) to the
kernel during wpa_driver_nl80211_sta_add call. Bit 5 and Bit 6 of QoS
info represents the Max SP length. Fix an issue in the code to fetch the
Max SP by shifting right the QoS info by value WMM_QOSINFO_STA_SP_SHIFT.
(operator ">" is replaced with ">>" operator).
Jouni Malinen [Fri, 15 Feb 2013 15:09:24 +0000 (17:09 +0200)]
WPS: Fix build without CONFIG_WPS_NFC
Commit bbf41865c90c2e16510712fa32290729c194422d added calls to functions
that are only defined if CONFIG_WPS_NFC=y. Protect those calls properly
to avoid build issues.
Jouni Malinen [Fri, 15 Feb 2013 10:34:55 +0000 (12:34 +0200)]
WPS: Add support for NFC handover select generation with wpa_supplicant
When wpa_supplicant is controlling an AP mode interface, it can generate
the alternative carrier record for NFC connection handover select
message similarly to the way this is done in hostapd.
Jouni Malinen [Fri, 15 Feb 2013 09:24:29 +0000 (11:24 +0200)]
WPS: Use pre-configured NFC password token instead of overriding it
"WPS_NFC_TOKEN <WPS/NDEF>" used to generate a new NFC password token
regardless of whether there was a pre-configured token in the
configuration. Change this to use the pre-configured value, if
available, instead. This allows the same command to be used to write the
password token to an NFC tag more conveniently.
Sunil Dutt [Tue, 5 Feb 2013 14:50:36 +0000 (16:50 +0200)]
TDLS: Pass peer's Capability and Ext Capability info during sta_add
The contents of the peer's capability and extended capability
information is required for the driver to perform TDLS P-UAPSD and Off
Channel operations. Pass this information to the driver when the peer
station is getting added.
Sunil Dutt [Tue, 5 Feb 2013 14:41:01 +0000 (16:41 +0200)]
TDLS: Pass peer's HT Capability and QOS information during sta_add
The information of the peer's HT capability and the QOS information is
required for the driver to perform TDLS operations. Pass this
information to the driver when the peer station is getting added.
Jouni Malinen [Thu, 14 Feb 2013 16:23:33 +0000 (18:23 +0200)]
TDLS: Fix add/set STA operation
Commit a9a1d0f08aaf7c96f40def0d7966399b89b2a7c0 added vht_capabilities
to struct hostapd_sta_add_params but forgot to update
wpa_supplicant_tdls_peer_addset() to initialize the variable to NULL.
This could result in uninitialized pointer being used in
driver_nl80211.c when adding a TDLS peer entry. Fix this by clearing the
hostapd_sta_add_params with memset.
Jouni Malinen [Thu, 14 Feb 2013 18:41:14 +0000 (20:41 +0200)]
WPS: Allow Device Password to be changed from M1 to M2
Registrar is allowed to propose another Device Password ID in M2. Make
Enrollee validate Device Password ID in M2 to check if this happened.
This commit adds support for changing from NFC password token to default
PIN for the case where the AP is the Enrollee and has both the NFC
password token and AP PIN enabled at the same time.
Jouni Malinen [Thu, 14 Feb 2013 17:44:54 +0000 (19:44 +0200)]
WPS: Fix wps_reg nfc-pw option
Commit ffdaa05a6b1b59c4b2e50f9b7fef82769fc2d3fe added support for using
NFC password token from an AP. However, it had a bug that prevented the
wpa_supplicant wps_reg command from being used with "nfc-pw" as the PIN
value. Fix string comparison to handle this correctly.
Sunil Dutt [Tue, 12 Feb 2013 21:25:21 +0000 (23:25 +0200)]
TDLS: Tear down peers when disconnecting from the AP
A TDLS Teardown frame with Reason Code 3 (Deauthenticated because
sending STA is leaving (or has left) IBSS or ESS) shall be transmitted
to all TDLS peer STAs (via the AP or via the direct path) prior to
transmitting a Disassociation frame or a Deauthentication frame to the
AP.
Jouni Malinen [Tue, 12 Feb 2013 17:25:18 +0000 (19:25 +0200)]
P2P: Do not use old scan result data for peer discovery
The driver may have cached (e.g., in cfg80211 BSS table) the scan
results for relatively long time. To avoid reporting stale information,
update P2P peers only based on results that have based on frames
received after the last p2p_find operation was started.
This helps especially in detecting when a previously operating GO stops
the group since the BSS entry for that could live for 30 seconds in the
cfg80211 cache. Running p2p_flush followed by p2p_find will now allow
wpa_supplicant to not add a P2P peer entry for that GO if the group had
been terminated just before that p2p_flush command. Previously, that GO
could have been indicated as a newly found device for up to 30 seconds
after it had stopped the group.
Jouni Malinen [Tue, 12 Feb 2013 17:14:32 +0000 (19:14 +0200)]
Use more accurate timestamps for scan results
For various P2P use cases, it is useful to have more accurate timestamp
for the peer information update. This commit improves scan result
handling by using a single timestamp that is taken immediately after
fetching the results from the driver and then using that value to
calculate the time when the driver last updated the BSS entry. In
addition, more debug information is added for P2P peer updates to be
able to clearly see how old information is being used here.
Jouni Malinen [Tue, 12 Feb 2013 16:24:56 +0000 (18:24 +0200)]
P2P: Postpone P2P-DEVICE-FOUND if config_methods not known
If we discover a P2P peer based on a Beacon frame from the GO role, we
do not get information about the supported configuration methods. This
can result in issues if the P2P managing entity above wpa_supplicant is
not prepared to handling config_methods=0x0. To avoid this, postpone
reporting of the P2P-DEVICE-FOUND event when this happens on one of the
social channels. It would be good to be able to this on all channels,
but that could result in issues of never indicating the event for a peer
that is operating a GO on a channel that requires passive scanning.
Jouni Malinen [Tue, 12 Feb 2013 16:14:48 +0000 (18:14 +0200)]
P2P: Do not allow peer update to clear config_methods
It could be possible for the scan results to include two entries for a
peer, one from the Listen state and the second one from the GO role. The
latter could be based on a Beason frame. If that happens and the entry
from GO is processed last, the P2P peer config_methods value could
potentially get cleared since Beacon frames do not include this
information in either WPS or P2P element. Avoid this by allowing the
config_methods value for P2P peers to be updated only if the new value
is non-zero.
Instead of reporting only one connection handover message, report
completion of NFC connection handover with carrier record from both the
request and select messages.
Jouni Malinen [Mon, 11 Feb 2013 10:21:03 +0000 (12:21 +0200)]
P2P: Avoid concurrent scans during all steps of group formation
Previously, only some P2P states were considered to postpone concurrent
station mode scans during group formation. Especially the WPS
provisioning step was skipped due to issues to scans run on the P2P
client interface (see commit fc6997b345d97c80965725b12c52b0d225f32a1a).
This is not ideal since a concurrent scan can slow down group formation
considerably and potentially make it time out. Enforce p2p-in-progress
through all steps of group formation on another interface to address
this.
Jouni Malinen [Sun, 10 Feb 2013 21:36:52 +0000 (23:36 +0200)]
P2P: Cancel group formation timeout on group removal (on client)
Group formation timeout is normally canceled when 4-way handshake is
completed (WPA_COMPLETED -> wpas_p2p_completed). However, it is possible
for the GO to stop the group before this happens (i.e., send
Deauthentication frame with reason code 3 before 4-way handshake is
completed). This resulted in the group getting removed, but the group
formation timeout left behind. The unexpected timeout may then result in
undesired termination of the following operation. Fix this by canceling
the group formation timeout in wpas_p2p_group_delete() similarly to how
group idle timeout was canceled there.
Jouni Malinen [Sun, 10 Feb 2013 16:42:43 +0000 (18:42 +0200)]
WPS: Report only the carrier record from NFC to wpa_supplicant
Since there could be multiple carrier records, it is cleaner to report
only the WPS carrier record instead of full NFC connection handover
select to wpa_supplicant.
Jouni Malinen [Sun, 10 Feb 2013 16:19:59 +0000 (18:19 +0200)]
WPS: Fetch only the carrier record from wpa_supplicant for NFC
Since there could be multiple carrier records, it is cleaner to build
only the WPS carrier record instead of full NFC connection handover
request within wpa_supplicant.
Johannes Berg [Sat, 9 Feb 2013 10:16:58 +0000 (12:16 +0200)]
hostapd: Do not change HT40 capability due to OBSS scan
The capability itself isn't really affected by an OBSS
scan, only the HT operation must then be restricted to
20 MHz. Change this, and therefore use the secondary
channel configuration to determine the setting of the
OP_MODE_20MHZ_HT_STA_ASSOCED flag.
This shouldn't really change anything functionally,
it just makes the code a little less confusing and
is also needed to implement more dynamic bandwidth
changes if ever desired.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
Peter Wu [Sat, 9 Feb 2013 10:13:54 +0000 (12:13 +0200)]
dbus: Add missing signal description for WPS (7)
Commit caff399250311abfea3376b38530a88bb78b957d added a property named
"WPS" to the dbus interface. It did not add the new
WPAS_DBUS_BSS_PROP_WPS property to the function responsible for marking
changed properties though. This cause the following messages to be
printed repeatedly to syslog:
dbus: wpas_dbus_bss_signal_prop_changed: Unknown Property value 7
Stevent Li [Wed, 6 Feb 2013 16:52:33 +0000 (18:52 +0200)]
EAP peer: Add Session-Id derivation
This adds a new getSessionId() callback for EAP peer methods to allow
EAP Session-Id to be derived. This commits implements this for EAP-FAST,
EAP-GPSK, EAP-IKEv2, EAP-PEAP, EAP-TLS, and EAP-TTLS.
Deepthi Gowri [Wed, 6 Feb 2013 22:55:06 +0000 (00:55 +0200)]
eap_proxy: Add mechanism for allowing EAP methods to be offloaded
In addition to the offload mechanism, the Android configuration and
makefiles are extended to allow this to be configured for the build by
dropping in platform specific configuration files and makefile without
having to modify any existing files.
Jouni Malinen [Thu, 7 Feb 2013 22:10:39 +0000 (00:10 +0200)]
P2P: Fix GO Probe Response IEs when Wi-Fi Display is enabled
Commit 1a9f6509b3a711071c12006ff7c3a9cb2a6691f2 added support for
fragmenting the P2P IE in Probe Response frames from a GO. However, it
did not take into account the possibility of Wi-Fi Display IE being
included in the same buffer and caused a regression for the cases where
Wi-Fi Display is enabled. Fix this by building the possibly fragmented
P2P IE first and then concatenating the separate IEs together.
Dmitry Shmidt [Thu, 7 Feb 2013 16:06:51 +0000 (18:06 +0200)]
Add 'SCAN TYPE=ONLY' functionality
Usual manual scan request may cause reassociation due to several
reasons. New command is intended to perform pure scan without taking any
automatic action based on the results.
Raja Mani [Thu, 7 Feb 2013 13:24:53 +0000 (15:24 +0200)]
nl80211: Add ctrl_iface message for AP mode connection rejection
When AP mode operation reject the client, nl80211 layer advertises the
connect failed event with the reason for failures (for example, max
client reached, etc.) using NL80211_CMD_CONN_FAILED.
This patch adds some debug messages whenever such an event is received
from the nl80211 layer and also the same event is posted to the upper
layer via wpa_msg().
Jouni Malinen [Thu, 7 Feb 2013 10:51:17 +0000 (12:51 +0200)]
P2P: Allow local configuration to use 5 GHz band 40 MHz channels
These channels were already enabled for P2P use, but the local
configuration parameter was not allowed to use the operating class in
which the 40 MHz channels are specified.
Amar Singhal [Thu, 7 Feb 2013 10:27:52 +0000 (12:27 +0200)]
Fix BSS RANGE command for no exact id match cases
The RANGE=N1-N2 command did not return any entries in some cases where
N1 does not match with any BSS entry. Fix this by allow entries to be
fetched even without knowing the exact id values.
Amar Singhal [Tue, 5 Feb 2013 22:28:48 +0000 (00:28 +0200)]
Fix the "BSS FIRST.." command
The "BSS FIRST.." command fails when additional parameters (e.g., MASK)
is used since the string comparsion does not take into account the
number of characters. Fix by comparing only 5 characters as in other
commands.
Jouni Malinen [Tue, 5 Feb 2013 15:40:09 +0000 (17:40 +0200)]
HS 2.0: Fix IE buffer length for extra scan IEs
The HS 2.0 Indication element is 7 (not 6) octets. The previous
implementation could result in wpabuf validation code stopping the
program if HS 2.0 was enabled without Interworking or P2P (which would
have created a large enough buffer to avoid hitting this) being enable.
Sunil Dutt [Tue, 5 Feb 2013 11:36:36 +0000 (13:36 +0200)]
TDLS: Fix MIC calculation for teardown frame to depend on reason code
The reason code used for calculating the MIC should correspond to the
reason code with which the teardown frame is sent, as the receiver shall
use the one obtained in the frame for validating the MIC.
Sunil Dutt [Tue, 5 Feb 2013 11:27:56 +0000 (13:27 +0200)]
TDLS: Remove link, if any, on an implicit set up request
If an implicit TDLS set up request is obtained on an existing link or an
to be established link, the previous link was not removed. This commit
disables the existing link on a new set up request. Also,
wpa_tdls_reneg() function was invoking wpa_tdls_start() on an already
existing peer for the case of internal setup, which is incorrect. Thus
the invocation of wpa_tdls_start() is removed in wpa_tdls_reneg() and
also this function is renamed to wps_tdls_remove() as it does not
renegotiation rather shall remove the link (if any) for the case of
external setup.
Sunil Dutt [Tue, 5 Feb 2013 11:10:34 +0000 (13:10 +0200)]
TDLS: Use existing peer entry if available when processing discovery
Peer entries were getting added on every discover request from the peer,
thus resulting in multiple entries with the same MAC address. Ensures
that a check is done for the presence of the peer entry and reuse the
existing entry instead of adding a new one.
Jouni Malinen [Mon, 4 Feb 2013 13:52:53 +0000 (15:52 +0200)]
P2P: Fix P2P-GROUP-STARTED event for p2p_connect-join
Commit 50178335bff450829729c91c8af798a1d3a7b408 introduced a regression
for P2P-GROUP-STARTED event indication during p2p_connect-join when
using a separate P2P group interface. wpa_s->global->p2p_group_formation
was already set in that case to point to the group interface and this
commit changed this to point to incorrect interface. Fix this by setting
p2p_group_formation here only in case a separate group interface is not
used.
Jouni Malinen [Mon, 4 Feb 2013 13:38:35 +0000 (15:38 +0200)]
P2P: Fix Action frame processing if Interworking is enabled
GAS server used the same public_action_cb mechanism as P2P to process
Action frames. This ended up overriding P2P processing of Action frames
while running an AP/GO interface with a build that enables Interworking
(e.g., for Hotspot 2.0) and a driver that uses hostapd for AP mode
SME/MLME. Fix this by adding a separate callback registration for the
GAS server. This should really be cleaned up by supporting arbitrary
number of callback handlers, but for now, this addresses the regression
with a minimal change.
Mark Kettenis [Sun, 3 Feb 2013 19:16:29 +0000 (21:16 +0200)]
Add driver for OpenBSD net80211 layer
Very basic support for OpenBSD. No support for scanning yet, so this needs
ap_scan=0 and expects that the user has configured the interface manually
using ifconfig(8).
Signed-hostap: Mark Kettenis <kettenis@openbsd.org>
Paul Stewart [Sun, 3 Feb 2013 19:08:31 +0000 (21:08 +0200)]
wpa_supplicant: Implement fast-associate on SelectNetwork
If scan results are available when we perform a SelectNetwork, use
them to make an associate decision. This can save an entire scan
interval-worth of time in situations where something external to
wpa_supplicant (like a connection manager) has just previously
requested a scan before calling SelectNetwork.
Pontus Fuchs [Sun, 3 Feb 2013 16:14:05 +0000 (18:14 +0200)]
Update scan interval gracefully
When the scan interval is changed the new interval is effective
after the old interval timer fires off one last time. This can cause
an unacceptable long delay when updating the interval.
Change this behaviour to use MIN(left of old interval, new interval)
for the scan interval following the interval change.
Jouni Malinen [Sun, 3 Feb 2013 14:23:13 +0000 (16:23 +0200)]
Fix compilation with PMKSA caching support disabled
Commit 6aaac006af7fd39d618c6546939bed9f0f0cea37 modified the
pmksa_cache_init() prototype, but forgot to update the empty wrapper
function which is used when PMKSA caching is not included in the build.
Ben Greear [Sun, 3 Feb 2013 11:51:54 +0000 (13:51 +0200)]
hostapd: Fix crash when scan fails
When scan failed, the wpa_driver_nl80211_scan method tried
to recursively call itself, but it passed in the wrong argument
for the void*, and so then it crashed accessing bad memory.
With this fix, hostapd still will not retry the scan later, but
at least it will exit cleanly and won't polute the file system
with core files.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Sunil Dutt [Tue, 22 Jan 2013 12:09:54 +0000 (14:09 +0200)]
TDLS: Disable the link also on driver request for teardown
The link was not disabled for the case of implicit trigger from the
driver unlike in the case of explicit trigger fromc ctrl_iface. Make the
tear down sequences match in these cases by adding the TDLS_DISABLE_LINK
tdls_oper to the driver when processing the TDLS_REQUEST_TEARDOWN event.
Jouni Malinen [Thu, 17 Jan 2013 14:22:41 +0000 (16:22 +0200)]
wlantest: Add radiotap header when re-writing DLT_IEEE802_11 file
When using DLT_IEEE802_11 datalink type in a pcap file, wlantest will now
add a radiotap header to the re-written pcap file to make sure all frames
in the output file will include the radiotap header.
Jouni Malinen [Thu, 17 Jan 2013 10:55:30 +0000 (12:55 +0200)]
wlantest: Add -F option for assuming FCS is included
When using DLT_IEEE802_11 datalink type in a pcap file, wlantest can now
be instructed to assume there is an FCS included in the frame by adding
the new -F command line argument. This will make wlantest validate the
FCS and strip it from the frame before processing.
Pavan Kumar [Tue, 15 Jan 2013 09:52:20 +0000 (11:52 +0200)]
P2P: Send P2P-FIND-STOPPED event in the new continue-search states
The P2P-FIND-STOPPED event was sent only in the P2P_SEARCH state, but
this needs to be send also in the new continue-search-when-ready states
P2P_CONTINUE_SEARCH_WHEN_READY and P2P_SEARCH_WHEN_READY for consistent
behavior.