There are two hw modes (5 GHz and 6 GHz) with HOSTAPD_MODE_IEEE80211A
and the current hw mode may be wrong after one channel switch to 6 GHz.
This will cause hostapd_set_freq_params() to return -1 when saving
previous state and the second channel switch to fail. Fix this by adding
hostapd_determine_mode() before every channel switch.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Define a struct to hold MLD station info and implement publishing of the
Basic Multi-Link element. Add it into Beacon and Probe Response frames.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
To simplify the handling of MLD stations, assume that all
interfaces/BSSs use the same IEEE 802.1X authenticator, the same RADIUS
server instance, and the same RADIUS client.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
- Include RNR element in Beacon frames of AP MLDs.
- Whenever a new interface is added to an AP MLD, reconfigure
the Beacon frame templates for all other interfaces, to allow
updating their RNR elements.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Add support for including multiple hostapd interfaces in the same AP
MLD, i.e., all using the same underlying driver network interface.
To do so, when a new hostapd interface is added, if there is already
another interface using the same underlying network interface, associate
the new interface with the same private data object, instead of creating
a new one.
As some of the BSSs are non-first BSSs, meaning that they reuse the
drv_priv of the initial BSS, make sure not to double free it.
Currently multiple BSS entries are not supported so always use bss[0]
for MLD.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
AP cannot come up in channels that are marked as NO_IR. If AP moves to
HAPD_IFACE_DISABLED state, it will deinitialize the nl80211 driver
interface and sockets.
Hence, introduce a new state called HAPD_IFACE_NO_IR, for 6 GHz APs to
handle NO_IR scenarios, such as AFC, where the channels not allowed by
AFC will have HOSTAPD_CHAN_NO_IR flag set. In this state, AP is still
kept in a non-operational state (stopped) without deinitializing the
nl80211 driver interface. wiphy reg change event can then update the
channels and bring up the AP in a valid channel.
Signed-off-by: Pooventhiran G <quic_pooventh@quicinc.com>
Prune-associations should be done only after the new station is
authorized. Otherwise any STA can cause denial of service to connected
stations in PMF case when more than a single interface is being
controlled by the same hostapd process.
Signed-off-by: Adil Saeed Musthafa <quic_adilm@quicinc.com>
Add validatation of center frequency, and filling of appropriate
bandwidth in the channel switch wrapper when the channel switch is done
to a 320 MHz channel.
Signed-off-by: Ramya Gnanasekar <quic_rgnanase@quicinc.com>
The execute bits were not supposed to be added to a source code file.
Fixes: 927dbfb453 ("Fix 40 MHz channel bringup with ACS on the 6 GHz band")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
When AP is brought up in HE40/EHT40 with ACS, the AP comes up with 20
MHz bandwidth. It is expected to come up with 40 MHz bandwidth.
conf->secondary_channel does not hold the correct value and it leads to
choosing 20 MHz in hostapd_set_freq_params(). conf->secondary_channel is
filled using the hostapd config he_oper_centr_freq_seg0_idx. When AP is
configured to use ACS, the hostapd config he_oper_centr_freq_seg0_idx is
not valid as the channel is not known during bring up. So using the
config he_oper_centr_freq_seg0_idx to fill the conf->secondary_channel
does not work with ACS.
Use op_class to determine the bandwidth and based on the bandwidth fill
the conf->secondary_channel to address this ACS case.
Signed-off-by: Hari Chandrakanthan <quic_haric@quicinc.com>
hostapd start AP flow changed in commit 931e5d4f9e. However, that
could cause a regression in a legacy AP driver where the set key
operation for GTK, IGTK, and BIGTK before AP start (set beacon) would
cause the driver to ignore the key set command. Restore the flow of the
set beacon and WPA key init operations to make sure drivers can receive
and set group keys correctly.
Fixes: 931e5d4f9e ("mbssid: Configure all BSSes before beacon setup")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
center_idx_to_bw_6ghz() does not return the bandwidth in MHz and as
such, the check here against 20 (MHz) is never true. The returned value
is greater than 0 for the over 20 MHz cases.
Fixes: 15742566fd ("6 GHz: Fix operating class in Supported Operating Classes element")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Once ACS picks a channel, iface->freq and iface->conf->channel are
updated. So, AP comes up in the last operating channel when 'ENABLED'
after 'DISABLED' though ACS is configured.
But this will fail for 6 GHz APs since configured_fixed_chan_to_freq()
checks if iface->conf->channel is filled or not irrespective of ACS
configuration, and the checks inside configured_fixed_chan_to_freq()
fail the AP setup. Fix this by clearing iface->freq and
iface->conf->channel in AP setup for ACS configuration.
Fixes: bb781c763f ("AP: Populate iface->freq before starting AP")
Signed-off-by: Pooventhiran G <quic_pooventh@quicinc.com>
"multicast_to_unicast" and "hairpin_mode" are usually set outside of
hostapd. However, DFS channel change events pull the BSS out of the
bridge causing these attributes to be lost. Make these settings tunable
within hostapd so they are retained after the BSS is brought up again.
Signed-off-by: Anthony Refuerzo <anthony96922@gmail.com>
When AP is started on a DFS channel and DFS is offloaded to the driver,
AP setup will be separated to two stages. In the first stage, hostapd
will set frequency and initialize BSS, then waits the driver CAC to
complete. Once CAC done, in the second stage,
hostapd_setup_interface_complete() will be called again from a callback
to continue AP/channel setup.
But the driver will fail to restart AP if it is disabled/reenabled
during a driver CAC procedure because some steps such as setting
freq/beacon in the first stage are skipped due to cac_started not
cleared when the AP is disabled.
Avoid this by clearing cac_started when the AP is disabled.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Add data per IEEE Std 802.11-2020, 9.4.2.240. Current implementation is
added for the security and extended supported rates only.
For the Extended rates element, add a new member 'xrates_supported'
which is set to 1 only if hostapd_eid_ext_supp_rates() returns success.
Without this change, there are cases where this function returns before
adding the element for the transmitting interface resulting in incorrect
addition of this element inside the MBSSID Non-Inheritance element.
Signed-off-by: Aloka Dixit <quic_alokad@quicinc.com>
Co-developed-by: John Crispin <john@phrozen.org>
Signed-off-by: John Crispin <john@phrozen.org>
Co-developed-by: Sowmiya Sree Elavalagan <quic_ssreeela@quicinc.com>
Signed-off-by: Sowmiya Sree Elavalagan <quic_ssreeela@quicinc.com>
When multiple BSSID advertisement feature is enabled in IEEE 802.11ax
mode or later, Beacon frames are not transmitted per interface, instead
only one of the interfaces transmits Beacon frames that include one or
more Multiple BSSID elements with configuration for the remaining
interfaces on the same radio.
Change the existing logic such that all configuration details for all
the interfaces are available while building the Beacon frame template
for the transmitting interface itself.
Do not change the flow for the cases where multiple BSSID advertisement
is not enabled.
Signed-off-by: Aloka Dixit <quic_alokad@quicinc.com>
When using multiple BSSes on a single radio, it is sometimes desirable
to reconfigure one BSS, without disconnecting the stations already
connected to other BSSes on the same radio.
When a BSS is reconfigured using the SET command, there is no "old"
configuration we can compare to (so we cannot compare a hash of the
configuration for example).
One possible solution would be to make the current RELOAD command
reload only the current BSS. However, that could break the workflow of
existing users. Instead, introduce a new RELOAD_BSS command, which
reloads only the current BSS.
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
In hostapd_clear_old() multiple steps are needed to clear a BSS.
There are some places where it would be desirable to clear only some
BSSes and not all.
To make it easier to clear only some BSSes, split hostapd_clear_old()
with hostapd_clear_old_bss(), which does the same actions but on a
single BSS.
Signed-off-by: Raphaël Mélotte <raphael.melotte@mind.be>
Move most of CHANWIDTH_* definitions from ieee80211_defs.h to defs.h as
the definitions are getting used mostly for internal purpose only. Also
change prefix of the definitions to CONF_OPER_CHWIDTH_* and update in
all the files accordingly.
Leave the couple of VHT-specific exceptions to use the old defines (the
reason why they were originally added as VHT values), to avoid use of
clearly marked configuration values in information elements. In
addition, use the defines instead of magic values where appropriate.
Signed-off-by: Aleti Nageshwar Reddy <quic_anageshw@quicinc.com>
BIT(r) is not sufficient here since it does not cover 64 bit values.
Write this out with 1ULL to be large enough for the shift operation.
Signed-off-by: Jouni Malinen <j@w1.fi>
It looks like fst_wpa_obj::get_hw_modes would have been left
uninitialized in hostapd. It is not obviously clear why this would not
have caused issues earlier, but in any case, better make this set
properly to allow unexpected behavior should that function pointer ever
be used.
Signed-off-by: Jouni Malinen <j@w1.fi>
Modify hostapd_set_freq_params() to include EHT parameters and update
the calling functions to match.
Signed-off-by: Muna Sinada <quic_msinada@quicinc.com>
Signed-off-by: Aloka Dixit <quic_alokad@quicinc.com>
Signed-off-by: Pradeep Kumar Chitrapu <quic_pradeepc@quicinc.com>
Add the core code for handling BSS color collision events and triggering
CCA inside the kernel. The caller of hostapd_switch_color() will be
added in the following commits.
Tested-by: Peter Chiu <chui-hao.chiu@mediatek.com>
Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
While these do not return error code within the current hostapd
implementation, matching functions in wpa_supplicant AP functionality
will have an error case and using consistent return type will make the
control interface code more consistent.
In addition, export hostapd_set_acl() in preparation for the
wpa_supplicant control interface implementation extension.
Signed-off-by: Chaoli Zhou <quic_zchaoli@quicinc.com>
hostapd should update Supported Channel Width Set of HT Capability
Information field after channel switching done. Otherwise, it would
continue to use the old setting.
Signed-off-by: peterhuang <peterhuang@realtek.com>
hostapd will build wrong beacon_after in hostapd_fill_csa_settings() if
it doesn't update ieee80211ac when channel switching.
Signed-off-by: peterhuang <peterhuang@realtek.com>
Previously, the secondary channel was set only in presence of HT
capabilities based on HT40+ or HT40-. As HT capabilities and
secondary_channel are not present for the 6 GHz bamd, this causes
incorrect operating class indication in the Supported Operating Classes
element.
Fix this by assigning the secondary channel for bandwidths greater than
20 MHz in the 6 GHz band.
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
If WMM is not set explicitly in the configuration, it can be set based
on HT/HE config. As HE can be used without HT/VHT (which was introduced
as a special behavior for the 6 GHz band), add a similar automatic
enabling of WMM for HE without HT.
Signed-off-by: Lavanya Suresh <lavaks@codeaurora.org>
While it does not look like the stale pointer could have been
dereferenced in practice, it is better not to leave the stale pointer to
freed memory in place to avoid accidental uses.
Signed-off-by: Jouni Malinen <j@w1.fi>
hapd->tmp_eap_user needs to be cleared on interface deinit to avoid
leaving stale pointers to freed memory.
Fixes: ee431d77a5 ("Add preliminary support for using SQLite for eap_user database")
Signed-off-by: Jouni Malinen <j@w1.fi>
When an interface is disabled, e.g. due to radar detected,
hapd->time_adv is freed by hostapd_free_hapd_data(), but later
used by ieee802_11_build_ap_params() calling hostapd_eid_time_adv().
Thus hapd->time_adv needs to be cleared as well.
Fixes: 39b97072b2 ("Add support for Time Advertisement")
Signed-off-by: Michael Braun <michael-dev@fami-braun.de>
Commit f1df4fbfc7 ("mesh: Use setup completion callback to complete
mesh join") added a check for iface->conf being NULL into a debug print.
However, it is not clear how that could be NULL here. In any case,
setup_interface() could end up dereferencing iface->conf in the call to
hostapd_validate_bssid_configuration(), so better be consistent with the
checks and not get warnings from static analyzers regardless of whether
this can happen in practice.
Signed-off-by: Jouni Malinen <j@w1.fi>
Enable user to configure Maximum MPDU Length, Maximum A-MPDU Length
Exponent, Rx Antenna Pattern Consistency, and Tx Antenna Pattern
Consistency of 6 GHz capability through config file.
Signed-off-by: Rajkumar Manoharan <rmanohar@codeaurora.org>
The hostapd DFS code deinitializes and initializes the AP interface, if
a clean channel switch is not possible. In this case the AP code paths
would deinit the driver, for example nl80211, without wpa_supplicant
code paths getting notice of this.
Therefore add callbacks for wpa_supplicant mesh methods, which are
called on init/deinit of the AP BSS. These callbacks are then used to
handle the reset in the mesh code.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Mesh join function is the last function to be called during mesh join
process, but it's been called a bit earlier than it's supposed to be, so
that some mesh parameter values such as VHT capabilities were not
applied correct when mesh join is in process. Moreover, the current
design of mesh join that is called directly after mesh initialization
isn't suitable for DFS channels to use, since mesh join process should
be paused until DFS CAC is done and resumed after it's done.
The callback will be called by hostapd_setup_interface_complete_sync().
There is a possibility that completing mesh init fails, so add error
handling codes for that.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
When a single hostapd process manages both the OWE and open BSS for
transition mode, owe_transition_ifname can be used to clone the
transition mode information (i.e., BSSID/SSID) automatically. When both
BSSs use ACS, the completion of ACS on the 1st BSS sets state to
HAPD_IFACE_ENABLED and the OWE transition mode information is updated
for all the other BSSs. However, the 2nd BSS is still in the ACS phase
and the beacon update messes up the state for AP startup and prevents
proper ACS competion.
If 2nd BSS is not yet enabled (e.g., in ACS), skip beacon update and
defer OWE transition information cloning until the BSS is enabled.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Verify that all associated STAs that claim support for OCV initiate an
SA Query after CSA. If no SA Query is seen within 15 seconds,
deauthenticate the STA.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This allows nl80211-based drivers to get the frame out. The old earlier
location resulted in the driver operation getting rejected before the
kernel was not ready to transmit the frame in the BSS context of the AP
interface that has not yet been started.
While getting this broadcast Deauthentication frame transmitted at the
BSS start is not critical, it is one more chance of getting any
previously associated station notified of their previous association not
being valid anymore had they missed previous notifications in cases
where the AP is stopped and restarted.
Signed-off-by: Jouni Malinen <j@w1.fi>
cfg80211 takes care of key removal when link/association is lost, so
there is no need to explicitly clear old keys when starting AP.
Signed-off-by: Jouni Malinen <j@w1.fi>
Clean struct hapd_interfaces pointers and interface count during
deinitialization at the end of theh ostapd process termination so that a
call to hostapd_for_each_interface() after this does not end up
dereferencing freed memory. Such cases do not exist before this commit,
but can be added after this, e.g., for DPP needs.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
hostapd_broadcast_wep_set() can be called without a WEP key set.
Don't try to install a default key in that case.
This patch is not critical for the new API. With key_flag we just would
report an (ignored) error and do nothing. With the patch we simply do
nothing.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
