hostapd configuration parameters fragm_threshold and rts_threshold were
documented to disable the threshold with value -1 and not change driver
configuration if the parameter is not included. However, -1 was mapped
into not changing the driver value, so the explicit disabling part did
not work.
Replace the default values for these to be -2 so that explicitly set
configuration value -1 can be distinguished from the case of not
including the parameter. Map the -1 value to a driver request to disable
the threshold. Ignore any error from this operation just in case to
avoid breaking functionality should some drivers not accept the (u32) -1
value as a threshold value request to disable the mechanism.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This allows better control of processing new SAE sessions so that other
operations can be given higher priority during bursts of SAE requests,
e.g., during a potential DoS attack. The receive commit messages are
queued (up to maximum of 15 entries) and processed from eloop callback.
If the queue has multiple pending entries, more wait time is used to go
through the each new entry to reduce heavy CPU load from SAE processing.
Enable anti-clogging token use also based on the pending commit message
queue and not only based on the already started sessions.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Avoid an unnecessary unsigned integer overflow warning due to loop index
j-- use.
hostapd.c:661:10: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'unsigned int'
Signed-off-by: Jouni Malinen <j@w1.fi>
Split the check and decrementation into separate steps to avoid an
unnecessary UBSan warning.
hostapd.c:1895:14: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'size_t' (aka 'unsigned long')
Signed-off-by: Jouni Malinen <j@w1.fi>
hapd->conf might be NULL in case initialized failed, so better be
prepared for that when debug printing interface name in the deinit path.
hostapd.c:312:54: runtime error: member access within null pointer of type 'struct hostapd_bss_config'
hostapd.c:351:29: runtime error: member access within null pointer of type 'struct hostapd_bss_config'
hostapd.c:2158:18: runtime error: member access within null pointer of type 'struct hostapd_bss_config'
Signed-off-by: Jouni Malinen <j@w1.fi>
iface->num_bss is unsigned integer, so need to explicit typecast it to
unsigned before decrementation by one even when the result is stored in
an unsigned integer.
../src/ap/hostapd.c:2185:26: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'unsigned long'
Signed-off-by: Jouni Malinen <j@w1.fi>
Move functions corresponding to neighbor report elements to
src/ap/neighbor_db.[c,h] in preparation to using them after channel
switch from src/ap/drv_callbacks.c.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
BSS additional/removal cases were not considered at all in the previous
implementation of hostapd configuration file reloading on SIGHUP. Such
changes resulted in num_bss values getting out of sync in runtime data
and configuration data and likely dereferencing of freed memory (e.g.,
when removing a BSS).
Fix this by forcing a full disable/enable sequence for the interface if
any BSS entry is added/removed or if an interface name changes between
the old and the new configuration.
Signed-off-by: Jouni Malinen <j@w1.fi>
Previously, when an AP interface was disabled through a control
interface DISABLE command during a channel switch window, the interface
could not be reenabled due to beacon setup failure (which validates if
CSA is in progress).
Fix this by clearing channel switch parameters while disabling the
hostapd interface.
Signed-off-by: Sriram R <srirrama@codeaurora.org>
This allows WEP mode AP to be re-enabled automatically after external
ifconfig down + up on netdev used by hostapd.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When removing an interface, hostapd_bss_deinit() frees all associated
STAs. If any of the stations is 40MHz intolerant, the cleanup invokes
ht40_intolerant_remove(), that in turn registers a 20->40MHz transition
timer for the last station (ap_ht2040_timeout() function). That timer is
never canceled; once it executes, the interface is gone, most likely
resulting in a segfault when referencing it.
While hostapd_interface_deinit() cancels the transition timer, it does
so before cleaning up STAs. Move the cancellation after STA cleanup to
cancel any timer that was registered during that operation.
Signed-off-by: Martin Willi <martin@strongswan.org>
Previously, hostapd CHAN_SWITCH command did not effect VHT configuration
for the following:
When VHT is currently disabled (ieee80211ac=0),
1. hostapd_cli -p /var/run/hostapd chan_switch 10 5180 \
sec_channel_offset=1 center_freq1=5190 bandwidth=40 ht
====> Comes up in HT40
2. hostapd_cli -p /var/run/hostapd chan_switch 10 5765 \
sec_channel_offset=-1 center_freq1=5775 bandwidth=40 vht
====> Comes up in HT40
3. hostapd_cli -p /var/run/hostapd chan_switch 10 5200 center_freq1=5200 \
bandwidth=20 vht
====> Comes up in HT20
When VHT is currently enabled (ieee80211ac=1),
1. hostapd_cli -p /var/run/hostapd chan_switch 10 5180 \
sec_channel_offset=1 center_freq1=5190 bandwidth=40 ht
====> Comes up in VHT40
2. hostapd_cli -p /var/run/hostapd chan_switch 10 5200 center_freq1=5200 \
bandwidth=20 ht
====> Comes up in VHT20
This is since VHT config from chan_switch is processed only for
bandwidths 80 and above (80P80, 160) and for VHT20, VHT40 cases, only
NLA chan type and chan width are updated.
There is no NL attribute for determining if it is HT or VHT for
bandwidths 20 & 40 and currently they are updated as HT20, HT40 (+ or -
depending on offset). Same is notified back via
NL80211_CMD_CH_SWITCH_NOTIFY.
Instead of adding new NL attribute for tracking HT/VHT enabled config,
we are adding new hostapd VHT config parameter to save the chan_switch
config and use only for chan_switch case of VHT20 and VHT40.
Tested with all combinations of chan_switch (noHT->20->40->80->) HT/VHT
and confirmed to be working.
Signed-off-by: Sathishkumar Muruganandam <murugana@codeaurora.org>
Extend RADIUS DAS to support CoA-Request packets for the case where the
HS 2.0 Terms And Conditions filtering VSA is used to remove filtering.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When AP initialization is completed in a callback (e.g., OBSS scan),
wpa_supplicant_deinit_ap() is not called in case of failure. Fix this by
calling setup_complete_cb in case of failure, too, which in turn calls
wpa_supplicant_deinit_ap() if needed.
Signed-off-by: Tova Mussai <tova.mussai@intel.com>
If there is no explicit wmm_enabled parameter in the configuration
(i.e., conf->wmm_enabled == -1), the configuration reload path needs to
initialize conf->wmm_enabled based on iconf->ieee80211n in
hostapd_reload_bss() similarly to what is done in the initial startup
case in hostapd_setup_bss().
This fixes issues with RSN capabilities being set incorrectly when WMM
is supposed to get enabled and unexpectedly enabling WMM when it is not
supposed to be enabled (HT disabled). Either of these issues could show
up when asking hostapd to reload the configuration file (and when that
file does not set wmm_enabled explicitly).
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
The previous implementation ended up starting a new EAPOL-Key 4-way
handshake if the STA were to attempt to perform another association.
This resulted in immediate disconnection since the PTK was not ready for
configuring FILS TK at the point when EAPOL-Key msg 1/4 is sent out.
This is better than alloing the association to continue with the same TK
reconfigured, but not really ideal.
Address this potential sequence by not starting a new 4-way handshake on
the additional association attempt. Instead, allow the association to
complete, but do so without reconfiguring the TK to avoid potential
issues with PN reuse with the same TK.
Signed-off-by: Jouni Malinen <j@w1.fi>
The owe_transition_bssid and owe_transition_ssid parameters can now be
replace with owe_transition_ifname to clone the BSSID/SSID information
automatically in case the same hostapd process manages both the OWE and
open BSS for transition mode.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This adds DPP bootstrapping, authentication, and configuration into
hostapd similarly to how the design was integrated in wpa_supplicant.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Replace the previously used extension of IEEE 802.11 managed Ethertype
89-0d (originally added for Remote Request/Response in IEEE 802.11r)
with Ethertype 88-b7 (OUI Extended EtherType) for FT inter-AP
communication. The new design uses a more properly assigned identifier
for the messages.
This assigns the OUI 00:13:74 vendor-specific subtype 0x0001 for the new
hostapd AP-to-AP communication purposes. Subtypes 1 (PULL), 2 (RESP),
and 3 (PUSH) are also assigned in this commit for the R0KH-R1KH
protocol.
This breaks backward compatibility, i.e., hostapd needs to be updated on
all APs at the same time to allow FT to remain functional.
Signed-off-by: Michael Braun <michael-dev@fami-braun.de>
With AP-AP communication, when hapd0 sends a packet, hapd1 can receive
it immediately and send a response. But hapd0 will only read and process
the response after it has returned from the sending context, that is
entered eloop again. So one does not need to consider the RX function of
the reply to run for the request sending hapd before the send calling
function has returned.
Previously, with intra-process communication, the packet is not
scheduled through eloop. Thus the RX handler of the reply might be run
while the sending context of the original request has not returned.
This might become problematic, e.g., when deferring a management frame
processing until an RRB response is received and then have the request
restarted and finished before the original request handling has been
stopped.
I'm not aware of any concrete bug this is currently triggering but came
across it while thinking of FT RRB AP-AP sequence numbering.
I think the non-eloop scheduling approach might be error-prone and thus
propose to model it more closely to the way the message would be
received from a socket. Additionally, this ensures that the tests model
AP-AP communication more closely to real world.
Solution: queue these packets through eloop.
Signed-off-by: Michael Braun <michael-dev@fami-braun.de>
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>
The new broadcast_deauth parameter can be used to disable sending of the
Deauthentication frame whenever AP is started or stopped. The default
behavior remains identical to the past behavior (broadcast_deauth=1).
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Commit 01e2231fdc ('hostapd: Skip some
configuration steps for mesh cases') removed some operations based on
hapd->iface->mconf being NULL. This was within #ifdef CONFIG_MESH, so it
should not impact hostapd, but it can impact AP mode with
wpa_supplicant. That does not sound intentional, so make these
conditional on hapd->conf->mesh being enabled.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The local custom version of allocating and initializing struct
hostapd_data within wpa_supplicant_mesh_init() is problematic. This has
already missed couple of initialization steps that are required. Instead
of trying to remember to keep this up to date, use
hostapd_alloc_bss_data() so that there is only one place for this
initialization.
This is fixing a recent issue where FILS HLP started using
hapd->dhcp_server and expected that to be initialized to -1. For the
mesh case, that did not happen and when removing the interface, the FILS
HLP implementation ended up unregistering eloop socket for
hapd->dhcp_server (= 0). This could result in missing socket callbacks
for an arbitrary socket.
Signed-off-by: Jouni Malinen <j@w1.fi>
The new dhcp_server configuration parameter can now be used to configure
hostapd to act as a DHCP relay for DHCPDISCOVER messages received as
FILS HLP requests. The dhcp_rapid_commit_proxy=1 parameter can be used
to configure hostapd to convert 4 message DHCP exchange into a 2 message
exchange in case the DHCP server does not support DHCP rapid commit
option.
The fils_hlp_wait_time parameter can be used to set the time hostapd
waits for an HLP response. This matches the dot11HLPWaitTime in IEEE Std
802.11ai-2016.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
It was possible to try to do driver operations before the driver
interface had been initialized when processing a SIGHUP signal. This
would result in NULL pointer dereference. Fix this by skipping the steps
when SIGHUP is issued before the interface is enabled.
Signed-off-by: Jouni Malinen <j@w1.fi>
hostapd didn't send "TERMINATING" event when stopped by
SIGTERM. Android handles this event to stop monitor thread.
This commit adds "TERMINATING" event same as with wpa_supplicant.
Signed-off-by: Tomoharu Hatano <tomoharu.hatano@sonymobile.com>
We should unconditionally remove inactivity timers for wired network
cases. This commit checks for this after a new station association:
hapd->iface->drv_flags & WPA_DRIVER_FLAGS_WIRED
and then cancels the timeout and does not register a new one.
It prints out a debug message like this:
1476740180.276286: IEEE 802.1X: 00:02:00:00:00:07 CTRL_DIR entering
state FORCE_BOTH
1476740180.276295: hostapd_new_assoc_sta: canceled wired ap_handle_timer
timeout for 00:02:00:00:00:07
This was tested on a debian jessie amd64 system with a configured 120
second inactivity timer and the session did not timeout.
Signed-off-by: Sam Tannous <stannous@cumulusnetworks.com>
Previously, CONFIG_IEEE80211R enabled build that supports FT for both
station mode and AP mode. However, in most wpa_supplicant cases only
station mode FT is required and there is no need for AP mode FT.
Add support to differentiate between station mode FT and AP mode FT in
wpa_supplicant builds by adding CONFIG_IEEE80211R_AP that should be used
when AP mode FT support is required in addition to station mode FT. This
allows binary size to be reduced for builds that require only the
station side FT functionality.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
The Channel Center Frequency Segment subfields use the channel index
instead of frequency in MHz.
Signed-off-by: Beni Lev <beni.lev@intel.com>
Signed-off-by: David Spinadel <david.spinadel@intel.com>
Add a configuration option in hostapd.conf and in neighbor report that
sets an AP as stationary. To enable this option on the current AP set
the config option stationary_ap to 1. To set a neighbor entry to be
marked as stationary add the word stat to the SET_NEIGHBOR command. This
option tells hostapd to send LCI data even if it is older than requested
by max age subelement in RRM request.
Signed-off-by: David Spinadel <david.spinadel@intel.com>
A weakness in the initial client taxonomy mechanism is from storing both
the Probe and Associate in struct sta_info. struct sta_info is created
after a client associates (or starts authentication frame exchange),
which means that any Probe Request frames sent prior to association are
not retained. The Associate Request frame has to be seen, and then
another Probe Request frame after association, before we have a
signature for the client.
Most clients send lots of Probe Request frames (lots and lots and lots
of Probes, actually), but a few do not. ChromeOS is notably sparing in
sending Probe Request frames, it can take a long time before a signature
for a ChromeOS device is available.
Store the most recent Probe Request frame in struct hostapd_sta_info
tracking list. When a struct sta_info is created, move the Probe Request
frame information from struct hostapd_sta_info to struct sta_info.
Signed-off-by: dgentry@google.com (Denton Gentry)
Signed-off-by: denny@geekhold.com (Denton Gentry)
Signed-off-by: rofrankel@google.com (Richard Frankel)
Signed-off-by: richard@frankel.tv (Richard Frankel)
Previously, struct hostapd_iface sta_seen list head was initialized only
when completing interface setup. This left a window for operation that
could potentially iterate through the list before the list head has been
initialized. While the existing code checked iface->num_sta_seen to
avoid this case, it is much cleaner to initialize the list when struct
hostapd_iface is allocated to avoid any accidental missing of the extra
checks before list iteration.
Signed-off-by: Jouni Malinen <j@w1.fi>
Previously, this was initialized in hostapd_setup_bss() which made it
possible for a REMOVE_NEIGHBOR control interface command to be issued
prior to the list head pointers having been set. That resulted in a NULL
pointer dereference. Fix this by initializing the list head at the time
the data structure gets allocated.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This can be used to mandate the presence of the Message-Authenticator
attribute on CoA/Disconnect-Request packets.
Signed-off-by: Nick Lowe <nick.lowe@lugatech.com>
hostapd deinit functions were not ready to handle a case where the data
structures were not fully initialized. Make these more robust to allow
wpa_supplicant mesh implementation to use the current deinit design in
OOM error cases without causing NULL pointer dereferences.
Signed-off-by: Jouni Malinen <j@w1.fi>
Add a hostapd control interface command REQ_LCI to request LCI from an
associated station using radio measurement.
Signed-off-by: David Spinadel <david.spinadel@intel.com>
Add a configurable neighbor database that includes the content of
Nighbor Report element, LCI and Location Civic subelements and SSID.
All parameters for a neighbor must be updated at once; Neighbor Report
element and SSID are mandatory, LCI and civic are optional. The age of
LCI is set to the time of neighbor update.
The control interface API is:
SET_NEIGHBOR <BSSID> <ssid=SSID> <nr=data> [lci=<data>] [civic=<data>]
To delete a neighbor use:
REMOVE_NEIGHBOR <BSSID> <SSID>
Signed-off-by: David Spinadel <david.spinadel@intel.com>
Interface additions/removals are not guaranteed to be for the driver
listening to the kernel events. As such, send the events to
wpa_supplicant_event_global() which can then pick the correct interface
registered with wpa_supplicant to send the event to.
Signed-off-by: Roy Marples <roy@marples.name>
Add a new 'use_driver_iface_addr' configuration parameter to allow use
of the default interface address generated by the driver on interface
creation. This can be useful when specific MAC addresses were allocated
to the device and we want to use them for multi-BSS operation.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Add UDP support for ctrl_iface:
New config option could be set:
CONFIG_CTRL_IFACE=udp
CONFIG_CTRL_IFACE=udp-remote
CONFIG_CTRL_IFACE=udp6
CONFIG_CTRL_IFACE=udp6-remote
And hostapd_cli usage:
hostapd_cli -i localhost:8877
Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
There is no need to maintain two implementations of the functionality.
is_zero_ether_addr() is easier to understand, so use it.
Signed-off-by: Jouni Malinen <j@w1.fi>
r1_key_holder is an identifier that was always set to zero if unless
configured before.
See 11.6.1.7.4 of IEEE Std 802.11-2012 which reads
"R1KH-ID is a MAC address of the holder of the PMK-R1 in the
Authenticator of the AP"
See 12.2.2 of IEEE Std 802.11-2012 which reads
"Each R0KH-ID and R1KH-ID is assumed to be expressed as a unique
identifier within the mobility domain."
"The R1KH-ID shall be set to a MAC address of the physical entity
that stores the PMK-R1 ..."
Defaulting this to BSSID is a more reasonable value since we have not
rejected the missing r1_key_holder as invalid configuration.
Signed-off-by: Michael Braun <michael-dev@fami-braun.de>
Rework the Acct-Session-Id and Acct-Multi-Session-Id implementation to
give better global and temporal uniqueness. Previously, only 32-bits of
the Acct-Session-Id would contain random data, the other 32-bits would
be incremented. Previously, the Acct-Multi-Session-Id would not use
random data. Switch from two u32 variables to a single u64 for the
Acct-Session-Id and Acct-Multi-Session-Id. Do not increment, this serves
no legitimate purpose. Exclusively use os_get_random() to get quality
random numbers, do not use or mix in the time. Inherently take a
dependency on /dev/urandom working properly therefore. Remove the global
Acct-Session-Id and Acct-Multi-Session-Id values that serve no
legitimate purpose.
Signed-off-by: Nick Lowe <nick.lowe@lugatech.com>
