If the initial attempt to start a scan for p2p_find fails, an error
is reported. However, the P2P scan handler and search state was
left behind. That can result in unexpected behavior when the next
non-P2P scan results are indicated. Avoid this by clearing the
P2P search state on failure.
There is no need to request periodic bgscans when the driver claims
to have capability for roaming within ESS. Ignoring the bgscan
configuration allows the same configuration file to be used both
with drivers the handle roaming and with drivers that don't.
If the driver indicates that it supports BSS selection (including
roaming within an ESS) with WPA_DRIVER_FLAGS_BSS_SELECTION, modify
ap_scan=1 mode to behave like ap_scan=2 mode for BSS selection.
The initial scan is still done to avoid the need for strict
configuration of or security parameters (e.g., to figure out whether
TKIP or CCMP is being used as the group cipher). However, when
requesting the driver to connect, the bssid and freq parameters are
not provided to leave the driver in control of selecting which BSS
to use and to allow the driver to decide when to roam.
This new flag can be used to change wpa_supplicant behavior in the
default ap_scan=1 mode to move BSS selection into the driver (likely
firmware). This commit is only adding the flag; separate commits
will be used to change the actual connection/roaming behavior.
These protocols seem to be abandoned: latest IETF drafts have expired
years ago and it does not seem likely that EAP-TTLSv1 would be
deployed. The implementation in hostapd/wpa_supplicant was not complete
and not fully tested. In addition, the TLS/IA functionality was only
available when GnuTLS was used. Since GnuTLS removed this functionality
in 3.0.0, there is no available TLS/IA implementation in the latest
version of any supported TLS library.
Remove the EAP-TTLSv1 and TLS/IA implementation to clean up unwanted
complexity from hostapd and wpa_supplicant. In addition, this removes
any potential use of the GnuTLS extra library.
This fixes some build issues in GnuTLS wrapper to be compatible with
at least following GnuTLS versions: 2.2.5, 2.4.3, 2.6.6, 2.8.6,
2.10.5, 2.12.11, 3.0.3.
Instead of using separate bad_record_mac and decryption_failed alerts,
use only bad_record_mac alert regardless of how the CBC decryption
failed. This provides less information to attackers that could modify
packets. In addition, instead of returning immediately on error, run
through the MAC check to make timing attacks more difficult.
When the received data will be decrypted, there is no need to first
copy it and then handle decryption in-place when decryption step can
take care of both operations.
TLS v1.0 and v1.1 RFCs were not exactly clear on the use of the
protocol version in record later. As such, accept any {03,xx} value
to remain compatible with existing implementations and new protocol
versions.
The p2p->drv_in_listen variable is used to track Listen state operations
in the driver. This is cleared when the driver reports that the Listen
state has ended (p2p_listen_end() gets called). However, it is possible
that the driver does not indicate that if the Listen state is canceled.
This can apparently happen in some cases where p2p_connect command is
issues while the Listen state is in progress.
Work around this issue by clearing p2p->drv_in_listen when Listen state
is stopped as part of p2p_stop operation. This allows the P2P module to
process CONNECT_LISTEN timeout in p2p_timeout_connect_listen() to move
to CONNECT state, e.g., when starting GO Negotiation after Device
Discoverability mechanism.
The "too long" buffer (32+1 octets) prevented AP from starting up with
32 octet SSID with WE-21 and newer. Fix this by reducing the
SIOCGIWESSID buffer length.
wpa_supplicant returns the PIN in reply to the connect method. Treating
this value as an integer runs the risk of not returning the correct
value if the first digit(s) happens to be a zero(es). To return the
correct PIN it needs to be returned as a string.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Angie Chinchilla <angie.v.chinchilla@intel.com>
The default discovery type is to search for devices only on social
channels. Expand this to also support an initial full scan followed by the
scan of social channels as well as the progressive scan that scans through
all the channels in the Search state rounds. This does not change the
default of scanning only social channels although there is currently a
discrepancy wrt the default used by wpa_cli, which is the full scan first.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Angie Chinchilla <angie.v.chinchilla@intel.com>
When parsing a dict entry which is an array of an array of bytes the entry
representing the dict entry has DBUS_TYPE_ARRAY as its type and
WPAS_DBUS_TYPE_ARRAY as its array_type. The function freeing this parsed
data incorrectly tested the entry type for WPAS_DBUS_TYPE_ARRAY while doing
no testing of this value for array_type. This results in a memory leak
whenever a D-Bus message with this type of data is parsed.
Messages affected are:
fi.w1.wpa_supplicant1.Interface.P2PDevice
using RequestedDeviceTypes with Find method
using SecondaryDeviceTypes or VendorExtension with P2PDeviceProperties
fi.w1.wpa_supplicant1.Group
using WPSVendorExtensions with Properties property
All of the above messages are parsed with the same function,
wpa_dbus_dict_get_entry, so the assignment of the entry's type and
array_type is consistent. The parsed data is also consistently freed with
the same function, wpa_dbus_dict_entry_clear, so we can use the same checks
to free the data correctly.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Angie Chinchilla <angie.v.chinchilla@intel.com>
The action done handling needs to abort an off-channel period since one
might have been used for example for GO negotiation and after action
done the code assumes it can start a new off-channel period.
This fixes a bug I introduced when adding support for
in-kernel off-channel transmissions.
Reported-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Refer to kerneldoc2doxygen.pl as doc/kerneldoc2doxygen.pl since doxygen
is run in the top-level source directory.
Signed-off-by: Pavel Roskin <proski@gnu.org>
This adds support for RSN pre-authentication with nl80211 interface and
drivers that handle roaming internally (i.e., wpa_supplicant is not
responsible for requesting scan results that provide the needed
information without these events).
The Timeout Interval element needs to use Association Comeback time
for the case when (Re)Association Response frame uses Status Code 30.
Verify this before incrementing the (re)assocresp_comeback counters.
cfg80211 does not currently allow NL80211_CMD_CONNECT when there
is already an association. This can result in wpa_supplicant being
unable to request a connection if the interface is already in
connected state. Work around this by requesting disconnection if
the connect command fails with EALREADY error. This allows
wpa_supplicant to recover from the unexpected state and complete
the connect request on next attempt.
eapol_test command line argument -o<file> can now be used to request
the received server certificate chain to be written to the specified
file. The certificates will be written in PEM format. [Bug 391]
The proactive_key_caching parameter was missed in mapping the
config_ssid data into the WPA state machine configuration. This
prevented addition of PMKSA cache entries based on PMKSA caching
candidate events.
Make the WPS processing of (Re)Association Request frame IEs conditional
on WPS support. This fixes a build issue with wps_is_20() not being
defined and makes the non-WPS build a bit smaller.
os_*() wrappers should be used instead of functions from time.h.
Removing the header from includes.h enforces this. os_unix.c can
include this its uses are valid wrapper calls. wps_upnp.c uses
gmtime() for which there is no os_*() wrapper available yet, so
allow it to use time.h, too. Similarly, allow dump_state.c to
use time.h for ctime().
Currently, the following can happen:
1) P2P state machine requests R-O-C
2) user changes their mind and aborts
3) P2P state machine aborts R-O-C
4) driver_nl80211 rejects abort since there
was no notification about the start yet
5) R-O-C period start notification from kernel
6) P2P state machine requests new R-O-C
7) this overlaps with old R-O-C -- iwlwifi driver
can't handle that and returns -EBUSY
8) state machine stops dead in its tracks
The reason is that the abort isn't going through properly. Instead of
tracking whether a R-O-C is active in driver_nl80211, track whether one
was requested to avoid this scenario.
Reported-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
This allows drivers to disable CCK rates from Probe Request frames.
For nl80211, this is currently applying only to the supported rates
element(s), but this mechanism could be extended to address TX rate
control masking, too, to lessen need for global rate disabling.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Since the kernel can now advertise P2P concurrent support by advertising
interface combinations, we can take advantage of that and automatically
use P2P_CONCURRENT / P2P_MGMT_AND_NON_P2P for drivers that advertise
support.
Keep driver_param=use_p2p_group_interface=1 for anyone not advertising
interface combinations in their drivers yet.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The default config methods was hardcoded to claim support for
PushButton, Display, and Keypad. While these are supported by
most P2P devices, there may be some cases where it is convenient
to be able to disable a specific config method. Use config_methods
configuration parameter to set the default values for Config Methods
in the P2P Device Info attribute.
The P2P specification (3.1.4.3) disallows use of the Label configuration
method between two P2P devices. This was previously enforced at upper
level, but the obsolete code can be removed from wpa_supplicant. This
adds a bit more strict enforcement of the policy, but should not result
in practical differences since no known P2P implementation uses Label
config method.
If a network configuration block is removed or modified, flush
all PMKSA cache entries that were created using that network
configuration. Similarly, invalidate EAP state (fast re-auth).
The special case for OKC on wpa_supplicant reconfiguration
(network_ctx pointer change) is now addressed as part of the
PMKSA cache flushing, so it does not need a separate mechanism
for clearing the network_ctx values in the PMKSA cache.
If the driver maintains its own copy of the PMKSA cache, we need to
clear an entry from the driver whenever wpa_supplicant is dropping
an old PMKSA cache entry.
The Deauth/Disassoc TX status callbacks were ending up kicking the
station entry from kernel driver when test functionality was used to
inject Deauth/Disassoc frames from the AP with the purpose of leaving
the local association in place. Fix this by using STA flags to figure
out whether there was a pending callback for the frame that we need
to act on.
In addition, add forgotten functionality for the Disassoc TX status
callback to match the behavior with Deauth.
The previous code was trying to figure out which WPA version is
used based on the extra IEs requested for Association Request. That
did not work properly in cases where non-WPA networks are used with
some extra IEs. Fix this by using more robust mechanism for passing
the WPA versions from core wpa_supplicant to the driver_ops
associate().
