This functionality was repeated for couple of times. Use a shared helper
function to avoid code duplication.
Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
Add initial_freq_list to wpa_supplicant configuration. This option
allows wpa_supplicant to scan a smaller list of frequencies when it
starts. This in turn allows for a faster connection to an already known
network. This limit applies only for the initial scan operation and does
not restrict other channels from being used in consecutive scans.
Tests have shown this to reduce the amount of time for connecting to a
network from roughly 3 seconds to roughly 0.1 second.
Signed-off-by: Frederik Juul <frederik.juul@3shape.com>
Use the common IE parsing routine in hope of trying to make the length
checks easier for static analyzers. In addition, try to make the
*_vendor_ie_multi() cases easier to analyze as well even though they
cannot use for_each_element_id().
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When using NetworkManager to set up an access point, there seems to be a
race condition which can lead to a new log message every second.
The following message appears in AP mode:
CTRL-EVENT-SCAN-FAILED ret=-95 retry=1
Normally, this log message only appears once. But then (and only then)
the race is triggered and they appear every second, the following log
messages are also present:
Reject scan trigger since one is already pending
Failed to initiate AP scan
This patch just disables the retry for requests where the operation is
not supported anyway.
Signed-off-by: Georg Müller <georgmueller@gmx.net>
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>
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>
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>
Specific BSSID scan was replacing wildcard SSID with the known SSID if
any BSS with the specified BSSID is available in the known BSSes list.
Add control interface support to force use of a wildcard SSID in a
specific BSSID scan by user with the new "wildcard_ssid=1" argument to
the SCAN command.
Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
Local variable should be used. This fixes an issue where IEs are
available only from a Beacon frame.
Fixes: ad06ac0b0 ("Move throughput estimation into a helper function")
Signed-off-by: Matthew Wang <matthewmwang@chromium.org>
The 5 GHz channels are stored in one hw_features set with mode
HOSTAPD_MODE_IEEE80211A while the 6 GHz channels will need to be stored
in a separate hw_features set (but with same mode
HOSTAPD_MODE_IEEE80211A) due to possibility of different HT/VHT/HE
capabilities being available between the 5 GHz and 6 GHz bands.
Iterate through all hw_features sets and check and match the band of
channel supported by the hw_features set while getting the hw_features
set in get_mode(). This allows both the 5 GHz and 6 GHz channels to be
found and correct capabilities to be used in cases where the driver
reports different capability values between 5 and 6 GHz channels.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This is a step towards allowing this functionality to update the scan
result -based values with the values from a signal poll for the current
BSS.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
wpas_dbus_handler_scan() constructs a set of 'params' each time, but it
doesn't acknowledge the existing randomization settings when doing so.
That means that any D-Bus initiated scans weren't going to follow the
configured settings.
Signed-off-by: Eric Caruso <ejcaruso@chromium.org>
Add D-Bus property:
* MACAddressRandomizationMask: a{say}
which configure random MAC address functionality in the Wi-Fi
driver via netlink.
Signed-off-by: Eric Caruso <ejcaruso@chromium.org>
This array can be freed either from the scan parameters or from
clearing the MAC address randomization parameters from the
wpa_supplicant struct. To make this ownership more clear, we have
each struct own its own copy of the parameters.
Signed-off-by: Eric Caruso <ejcaruso@chromium.org>
wpa_scan_result_compar() would return wb->est_throughput -
wa->est_throughput in case the comparison is done based on the
throughput estimates. While the return value from this function is a
signed integer, these est_throughput values are unsigned integers and
need to be explicitly typecast to avoid an UBSan warning.
scan.c:1996:30: runtime error: unsigned integer overflow: 54000 - 135000 cannot be represented in type 'unsigned int'
Signed-off-by: Jouni Malinen <j@w1.fi>
When setting scan with randomized MAC, check the requested scan type
against supported types, to ensure callers will not set an unsupported
type, since this can cause scan/connect failures later. It is better to
do this in wpas_mac_addr_rand_scan_set() instead of control interface
specific code to apply the constraint on all possible interfaces using
this setting.
Signed-off-by: Lior David <liord@codeaurora.org>
The HS 2.0 Indication element can be up to 9 octets in length, so add
two more octets to the minimum extra_ie buffer size for scanning.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Hotspot 2.0 tech spec mandates mobile device to not indicate a release
number that is greater than the release number advertised by the AP. Add
this constraint to the HS 2.0 Indication element when adding this into
(Re)Association Request frame. The element in the Probe Request frame
continues to show the station's latest supported release number.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
An OWE AP device that supports transition mode does not transmit the
SSID of the OWE AP in its Beacon frames and in addition the OWE AP does
not reply to broadcast Probe Request frames. Thus, the scan results
matching relies only on Beacon frames from the OWE open AP which can be
missed in case the AP's frequency is actively scanned.
To improve the discovery of transition mode APs, include their SSID in
the scan command to perform an active scan for the SSIDs learned from
the open mode BSSs.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
If the device supports OCE features and OCE is enabled, set the relevant
scan parameters and FILS Request Parameters element with Max Channel
Time.
Signed-off-by: Roee Zamir <roee.zamir@intel.com>
Add a flag to scan parameters that enables OCE scan features. If this
flag is set the device should enable the following features as defined
in the Optimized Connectivity Experience Technical Specification v1.0:
- Overwrite FILS request Max Channel Time with actual value (clause 3.8)
- Send Probe Request frame in high rate (at least 5.5 Mbps) (clause 3.12)
- Probe Request frame Transmission Deferral and Suppression (clause 3.5)
- Accept broadcast Probe Response frame (clause 3.6)
Signed-off-by: Roee Zamir <roee.zamir@intel.com>
When failing to trigger scan for beacon report (e.g., when the
requested duration is not supported by the driver), send a
Radio Measurement response with the mode set to refused and don't
retry the scan.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
In ap_scan=2 mode, wpa_supplicant_assoc_try() did not check whether the
SSID is temporarily disabled before trying to associate and this may
result in an infinite connect/disconnect loop. If the association
succeeds while the SSID is temporarily disabled, wpa_supplicant will
request to deauthenticate and that in turn will cause the SSID to be
temporarily disabled again. Fix that by postponing the association until
the SSID is no longer temporarily disabled.
Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
The userspace may want to delay the the first scheduled scan.
This enhances sched_scan to add initial delay (in seconds) before
starting first scan cycle. The driver may optionally choose to
ignore this parameter and start immediately (or at any other time).
This uses NL80211_ATTR_SCHED_SCAN_DELAY to add this via user
global configurable option: sched_scan_start_delay.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This leads to cleaner code overall, and also reduces the size
of the hostapd and wpa_supplicant binaries (in hwsim test build
on x86_64) by about 2.5 and 3.5KiB respectively.
The mechanical conversions all over the code were done with
the following spatch:
@@
expression SIZE, SRC;
expression a;
@@
-a = os_malloc(SIZE);
+a = os_memdup(SRC, SIZE);
<...
if (!a) {...}
...>
-os_memcpy(a, SRC, SIZE);
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Previously it was possible for wpa_s->scan_res_handler to remain set to
its old value in case wpa_drv_scan() failed and no retry for the scan
trigger was scheduled (i.e., when last_scan_req == MANUAL_SCAN_REQ).
This could result in getting stuck with the next connection attempt
after a failed "SCAN TYPE=ONLY" operation when wpa_s->scan_res_handler
was set to scan_only_handler().
Fix this by clearing wpa_s->scan_res_handler if wpa_drv_scan() fails and
no retry is scheduled.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This functionality was originally added in commit
204c9ac4ee ('Extend select_network command
with freq= to reduce scan time') re-using wpa_s->manual_scan_freqs and
MANUAL_SCAN_REQ. That got broken when commit
35d403096e ('Set NORMAL_SCAN_REQ on
SELECT_NETWORK/ENABLE_NETWORK') started overriding wpa_s->scan_req for
SELECT_NETWORK.
Fix this by adding a new scan frequency list specifically for
SELECT_NETWORK so that this does not need to depend on any specific
wpa_s->scan_req value.
Signed-off-by: Jouni Malinen <j@w1.fi>
Previously, the est_throughput comparison was done only when SNR
difference was less than 5 dB. Since the throughput estimation take into
account SNR, this can be done in more cases. For now, add a conservative
2 dB more to the difference so that any SNR difference below 7 dB
results in BSS selection based on throughput estimates.
In addition, the throughput estimates require SNR values to be
available, so separate this from the 5 GHz preference that can be done
based on either SNR or qual values.
Signed-off-by: Jouni Malinen <j@w1.fi>
This allows throughput estimates and 5 GHz preference over 2.4 GHz band
to be used in more cases. The previously used value of 30 was
significantly more conservative than the SNR limits used for the highest
rate in scan_est_throughput() and this resulted in cases where 5 GHz AP
was ignored while SNR with it would have been close to reaching the
maximum TX rate.
Signed-off-by: Jouni Malinen <j@w1.fi>
Add support to set sched scan relative RSSI parameters so that the
drivers can report BSSs after relative comparision with the current
connected BSS. This feature is applicable only when in connected mode.
The below commands can be used to configure relative RSSI parameters
SET relative_rssi <disable|rssi_value>
disable - to disable the feature
rssi_value - amount of relative RSSI in dB
SET relative_band_adjust <band:adjust_value>
band - "2G" or "5G" for 2.4 GHz or 5 GHz respectively
adjust_value - amount of RSSI to be adjusted in dB
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Sched scan is supported by the kernel also in the connected state, so
allow PNO scan to be issued in the connected state from wpa_supplicant
as well.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Add the following parameters to scan request:
1. Dwell time on each channel.
2. Whether the specified dwell time is mandatory.
In addition, add to scan results info the time that the scan actually
started, and to each scan result the time the beacon/probe was received,
both in terms of TSF of the BSS that the interface that requested the
scan is connected to (if available).
Add flags to indicate whether the driver supports dwell time
configuration and scan information reporting.
This scan configuration and information is required to support beacon
report radio measurement.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
This makes wpa_supplicant behavior more consistent with FLUSH command to
clear all state. Previously, it was possible for an ongoing scan to be
aborted when the FLUSH command is issued and the scan results from that
aborted scan would still be processed and that would update the BSS
table which was supposed to cleared by the FLUSH command.
This could result in hwsim test case failures due to unexpected BSS
table entries being present after the FLUSH command.
Signed-off-by: Jouni Malinen <j@w1.fi>
Skipping the scan altogether will hurt auto-reconnect. Also move the PNO
check down since the scan might be canceled for other reasons before we
defer it.
Signed-off-by: Arik Nemtsov <arikx.nemtsov@intel.com>
This commit enhances the existing implementation of abort scan to also
abort concurrent active vendor scans. This is achieved by passing the
the scan_cookie to the driver interface with the intention to abort
the specific scan request. This scan_cookie is returned from the driver
interface when the scan request is scheduled.
This scan_cookie is 0 if the scan is triggered through the upstream
cfg80211 interface. Thus, the scan_cookie is used to determine whether
to abort the cfg80211 or vendor scan request.
Also, the previous implementation of relying on scan_work/p2p_scan_work
for the active work to trigger the abort scan is enhanced to check for
the started state of either of these work operations. This should also
help to abort the concurrent active scan/p2p-scan operations.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This commit enhances the abort scan implementation to also abort the
vendor scan, if one was used to trigger the scan.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
cfg80211 rejects the scans issued with random MAC address if the STA is
in connected state. This resulted in failures when using MAC_RAND_SCAN
while connected (CTRL-EVENT-SCAN-FAILED ret=-95). Enable random MAC
address functionality only if the STA is not in connected state to avoid
this. The real MAC address of the STA is already revealed in the
association, so this is an acceptable fallback mechanism for now.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
There are a couple of places in wpa_supplicant/hostapd where qsort() can
be called with a NULL base pointer. This results in undefined behavior
according to the C standard and with some standard C libraries (ARM RVCT
2.2) results in a data abort/memory exception. Fix this by skipping such
calls since there is nothing needing to be sorted.
Signed-off-by: Joel Cunningham <joel.cunningham@me.com>
PNO is sometimes restarted due to changes in scan parameters
(e.g., selected network changed or MAC randomization being
enabled/disabled). Restart is done by stopping PNO and immediately
starting it again. This may result in the SCHED_SCAN_STOPPED event being
received after the request for new PNO, which will make wpa_supplicant
believe PNO is not active although it is actually is. As a result, the
next request to start PNO will fail because PNO is active and should be
stopped first.
Fix this by deferring the request to start PNO until the
SCHED_SCAN_STOPPED event is received in case sched_scan is being
stopped.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
When scheduled scan stops without the interface request (for example,
driver stopped it unexpectedly), start a regular scan to continue
scanning for networks and avoid being left with no scan at all.
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
This code sequence was already used at two different places (and an
additional one has been proposed), so add a common helper function to
avoid having to copy-paste this functionality in multiple locations.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
If a scheduled scan is running on select network command,
cancel and reset it before kicking off a regular scan request.
Signed-off-by: Max Stepanov <Max.Stepanov@intel.com>
This makes wpa_supplicant set default scan IEs to the driver (if the
vendor command is supported). The driver can use these IEs in the scan
requests initiated by the driver itself. Also the driver can merge these
IEs into further scan requests that it receives, in case if the scan
request doesn't carry any of the IEs sent in this command.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, wpa_set_scan_ssids() fully exhausted
wpa_driver_scan_params.ssid list when hidden network IDs are provided
via the control interface. This results in us exceeding the max size for
the list advertised by the driver when we add the "wildcard" scan SSID
entry. So, ensure that we leave space for one more scan SSID entry in
the list when we exit out of wpa_set_scan_ssids().
Signed-off-by: Roshan Pius <rpius@google.com>
