The previously used normal data TX depends on undefined driver behavior
after all keys have been removed. That may not be available, so do this
more properly with frame injection through a monitor interface.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
We have to delete PTK keys when either BIT(0) or BIT(15) are zero and
not only when both are zero.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Allow AP mode network profile in wpa_supplicant to be configured to
advertise Transition Disable DKE.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Check whether the Transition Disable KDE is received from an
authenticated AP and if so, whether it contains valid indication for
disabling a transition mode. If that is the case, update the local
network profile by removing the less secure options.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The new hostapd configuration parameter transition_disable can now be
used to configure the AP to advertise that use of a transition mode is
disabled. This allows stations to automatically disable transition mode
by disabling less secure network profile parameters.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Define the OUI Type and bitmap values for Transition Disable KDE. These
will be shared by both the AP and STA implementations.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Add QCA_WLAN_VENDOR_CFR_NDPA_NDP_ALL in enum
qca_wlan_vendor_cfr_capture_type. This capture type requests all NDPA
NDP frames to be filtered.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
hostapd will trigger EDMG auto channel selection by setting
QCA_WLAN_VENDOR_ATTR_ACS_EDMG_ENABLED. The 60 GHz driver will be
called to start an auto channel selection and will return the
primary channel and the EDMG channel.
Signed-off-by: Noam Shaked <nshaked@codeaurora.org>
Addition of chan_2ghz_or_5ghz_to_freq() broke 60 GHz ACS, because it
assumes reported ACS channel is on either 2.4 or 5 GHz band. Fix this
by converting chan_2ghz_or_5ghz_to_freq() to a more generic
chan_to_freq(). The new function uses hw_mode to support 60 GHz.
Fixes: 41cac481a8 ("ACS: Use frequency params in ACS (offload) completed event interface")
Signed-off-by: Noam Shaked <nshaked@codeaurora.org>
The parameters that need to be applied are symmetric to those of VHT,
however the validation code needs to be tweaked to check the HE
capabilities.
Signed-off-by: Shashidhar Lakkavalli <slakkavalli@datto.com>
Signed-off-by: John Crispin <john@phrozen.org>
This makes it easier to analyze certain encryption issues. Also print
out an error at the default INFO debug verbosity with the frame number.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Enhanced channel frequency response supports capturing of channel status
information based on RX. Define previous CFR as version 1 and enhanced
CFR as version 2. If target supports both versions, two versions can't
be enabled at same time. Extend attributes for enhanced CFR capture in
enum qca_wlan_vendor_peer_cfr_capture_attr.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
"GET last_tk_key_idx" can now be used in testing build to determine
which was the last configured Key ID for the pairwise key.
Signed-off-by: Jouni Malinen <j@w1.fi>
Support Extended Key ID in wpa_supplicant according to
IEEE Std 802.11-2016 for infrastructure (AP) associations.
Extended Key ID allows to rekey pairwise keys without the otherwise
unavoidable MPDU losses on a busy link. The standard is fully backward
compatible, allowing STAs to also connect to APs not supporting it.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Support Extended Key ID in hostapd according to IEEE Std 802.11-2016.
Extended Key ID allows to rekey pairwise keys without the otherwise
unavoidable MPDU losses on a busy link. The standard is fully backward
compatible, allowing an AP to serve STAs with and without Extended Key
ID support in the same BSS.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Stopping the AP from beaconing will also stop it from acknowledging
frames and that resulted in bgscan_learn_beacon_loss failing when
mac80211_hwsim is registering REPORTS_TX_ACK_STATUS. Work around this by
moving to using PMF so that the station ignores the unprotected
deauthentiation frames from the AP and also disabling SA Query. This
allows the AP to be stopped and restarted with large enough Beacon
interval to allow the station to detect beacon loss.
This is identical to the earlier design change for
bgscan_simple_beacon_loss (somehow this bgscan_learn_beacon_loss test
case managed to pass at that time).
Signed-off-by: Jouni Malinen <j@w1.fi>
There is no real use case for the scan to be requested on more than 100
channels individually. To avoid excessively long lists with invalid
configuration, use 100 entry limit for the list before dropping to the
fallback scan-all-channels option.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is more consistent with the other eloop registrations and avoids a
theoretical integer overflow with 16-bit int should more than 32767
sockets/signals/events be registered.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is more consistent with the other eloop registrations and avoids a
theoretical integer overflow with 16-bit int should more than 32767
sockets be registered (which is not really going to happen in practice).
Signed-off-by: Jouni Malinen <j@w1.fi>
This avoids a theoretical integer overflow with 16-bit unsigned int
should a certificate be encoded with more that 65535 friendly names or
icons.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is more consistent with the other eloop registrations and avoids a
theoretical integer overflow with 16-bit int (not that there would ever
be more that 32767 signal handlers getting registered).
Signed-off-by: Jouni Malinen <j@w1.fi>
There is no point in starting a huge number of pending SA Queries, so
limit the number of pending queries to 1000 to have an explicit limit
for how large sa_query_count can grow.
Signed-off-by: Jouni Malinen <j@w1.fi>
While RADIUS messages are limited to 4 kB, use size_t to avoid even a
theoretical overflow issue with 16-bit int.
Signed-off-by: Jouni Malinen <j@w1.fi>
These are theoretical cases with 32-bit integers, but cases that could
potentially hit an integer overflow with 16-bit int.
Signed-off-by: Jouni Malinen <j@w1.fi>
While int and unsigned int are not going overflow in practice as 32-bit
values, these could at least in theory hit an integer overflow with
16-bit int. Use size_t to avoid such potential issue cases.
Signed-off-by: Jouni Malinen <j@w1.fi>
This avoids a theoretical unsigned integer overflow case with 32-bit
integers, but something that could potentially be hit with 16-bit int
(though, even that part looks pretty theoretical in this particular case
of number of BSSs in scan results).
Signed-off-by: Jouni Malinen <j@w1.fi>
It looks ike the authentication timeouts may continue a bit longer after
some kernel changes and that could result in temporarily disabling the
network profile. Give this test case more time to complete the
connection to avoid reporting failures unnecessarily.
Signed-off-by: Jouni Malinen <j@w1.fi>
Make max_*_rate() functions and rate calculation at the beginning of
wpas_get_est_tpt() more continuous. In wpa_supplicant_need_to_roam(), we
compare these values to make a roaming decision. However, at certain
SNRs, we see unrealistically large jumps in estimated throughput
according to these functions, leading us to make incorrect roaming
decisions. Perform linear interpolation where applicable to more
accurately reflect actual throughput.
Example:
wlan0: Current BSS: 88:3d:24:b4:95:d2 freq=2412 level=-69 snr=20 est_throughput=54000
wlan0: Selected BSS: 88:3d:24:b4:89:9e freq=2417 level=-67 snr=22 est_throughput=63500
wlan0: Using signal poll values for the current BSS: level=-69 snr=20 est_throughput=54000
wlan0: Allow reassociation - selected BSS has better estimated throughput
2 dB increase in RSSI likely isn't responsible for a 17% increase in
throughput.
Signed-off-by: Matthew Wang <matthewmwang@chromium.org>
Stopping the AP from beaconing will also stop it from acknowledging
frames and that resulted in bgscan_simple_beacon_loss failing when
mac80211_hwsim is registering REPORTS_TX_ACK_STATUS. Work around this by
moving to using PMF so that the station ignores the unprotected
deauthentiation frames from the AP and also disabling SA Query. This
allows the AP to be stopped and restarted with large enough Beacon
interval to allow the station to detect beacon loss.
Signed-off-by: Jouni Malinen <j@w1.fi>
The new wpa_supplicant control interface SET parameter disable_sa_query
can now be used to disable SA Query on receiving unprotected
disconnection event.
Signed-off-by: Jouni Malinen <j@w1.fi>
If the kernel rtm_newlink or rtm_dellink send the maximum length of
ifname (IFNAMSIZ), the event handlers in
wpa_driver_nl80211_event_rtm_addlink() and
wpa_driver_nl80211_event_rtm_dellink() did not copy the IFLA_IFNAME
value. Because the RTA_PAYLOAD (IFLA_IFNAME) length already includes the
NULL termination, that equals the IFNAMSIZ.
Fix the condition when IFNAME reach maximum size.
Signed-off-by: Ouden <Ouden.Biz@gmail.com>
Add more hang reason codes for the hang reason in the
QCA_NL80211_VENDOR_SUBCMD_HANG events. This also introduces the
attribute QCA_WLAN_VENDOR_ATTR_HANG_REASON_DATA to carry the required
data for the respective hang reason. This data is expected to contain
the required dump to analyze the reason for the hang.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Some APs have managed to add two SSID elements into Beacon frames and
that used to result in picking the last one which had incorrect data in
the known examples of this misbehavior. Pick the first one to get the
correct SSID.
Signed-off-by: Jouni Malinen <j@w1.fi>
This was not supposed to be conditional on CONFIG_FILS.
Fixes: ecbf59e693 ("wpa_supplicant configuration for Beacon protection")
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
int_array_concat() and int_array_add_unique() could potentially end up
overflowing the int type variable used to calculate their length. While
this is mostly theoretical for platforms that use 32-bit int, there
might be cases where a 16-bit int overflow could be hit. This could
result in accessing memory outside buffer bounds and potentially a
double free when realloc() ends up freeing the buffer.
All current uses of int_array_add_unique() and most uses of
int_array_concat() are currently limited by the buffer limits for the
local configuration parameter or frame length and as such, cannot hit
this overflow cases. The only case where a long enough int_array could
be generated is the combination of scan_freq values for a scan. The
memory and CPU resource needs for generating an int_array with 2^31
entries would not be realistic to hit in practice, but a device using
LP32 data model with 16-bit int could hit this case.
It is better to have more robust checks even if this could not be
reached in practice, so handle cases where more than INT_MAX entries
would be added to an int_array as memory allocation failures instead of
allowing the overflow case to proceed.
Signed-off-by: Jouni Malinen <j@w1.fi>
