Process QCA_NL80211_VENDOR_SUBCMD_KEY_MGMT_ROAM_AUTH after NL80211_CMD_ROAM
NL80211_CMD_ROAM indication is scheduled via a kernel work queue, while
QCA_NL80211_VENDOR_SUBCMD_KEY_MGMT_ROAM_AUTH is a vendor event from the
driver. Thus, a race condition can exist wherein the vendor event is
received prior to the NL80211_CMD_ROAM indication.
The processing of this vendor event depends on the NL80211_CMD_ROAM
indication to update the roamed BSS/BSSID information and thus the out
of sequence processing of these events would result in not updating the
right BSS information.
This commit adds a workaround to hold the pending
QCA_NL80211_VENDOR_SUBCMD_KEY_MGMT_ROAM_AUTH event for up to 100 ms in
case NL80211_CMD_ROAM is not received first.
Vamsi Krishna [Fri, 18 Dec 2020 17:20:00 +0000 (22:50 +0530)]
nl80211: Do not ignore disconnection event after a connection event
After a disconnect command is issued, wpa_supplicant generates a
disconnection event to self and ignores the next disconnection event
coming from the driver. In a race condition in which the driver
generates a connected event due to roaming just before receiving the
disconnect command from userspace, wpa_supplicant processes the
connected event after processing the self-generated disconnection event
and enters WPA_COMPLETED state. The driver sends a disconnection event
after processing the disconnect command sent by wpa_supplicant but the
disconnection event is ignored by wpa_supplicant as the disconnection
event is considered to be a result of locally generated disconnect
command. Thus, wpa_supplicant continues to be in the connected
(WPA_COMPLETED) state though the driver is in disconnected state.
Fix this out-of-sync behavior between the driver and wpa_supplicant by
not ignoring the disconnection event from the driver because of the
locally generated disconnect command sent to the driver if there is a
connection event received after issuing the disconnect command to the
driver.
Jouni Malinen [Mon, 21 Dec 2020 13:50:33 +0000 (15:50 +0200)]
tests: Update ap_open_unexpected_assoc_event to match implementation
Now that wpa_supplicant ignores the unexpected association event in
disconnected-not-trying-to-connect state, this test cases needs to be
modified to avoid incorrect failures.
Vamsi Krishna [Thu, 3 Dec 2020 16:45:29 +0000 (22:15 +0530)]
Drop unexpected connection event while disconnected
If there is a disconnect command from wpa_supplicant immediately after
the driver sends a connection event to userspace but before that event
is received and processed by wpa_supplicant, wpa_supplicant processes
the disconnect command and a self-generated disconnected event first
followed by the connected event received from the driver. As a result
wpa_supplicant moves to the WPA_COMPLETED state. Whereas the driver
processes the disconnect command received from wpa_supplicant after it
sends the connected event and moves to the disconnected state. Due to
this race between the disconnect command from wpa_supplicant and the
connected event from the driver, wpa_supplicant is moving to the
connected state though the driver is moving to the disconnected state
which results in abnormal functionality.
Ignore the connection event coming from the driver when wpa_supplicant
is not trying to connect after a disconnect command is issued but before
the next connect command is issued to fix the above mentioned race
condition.
Sunil Dutt [Fri, 18 Dec 2020 04:36:19 +0000 (10:06 +0530)]
Vendor feature capability to notify TWT asynchronous response support
The response for the respective TWT operations can either be synchronous
or asynchronous (wherever specified). If synchronous, the response to
this operation is obtained in the corresponding vendor command reply to
the user space. For asynchronous case, the response is obtained as an
event with the same operation type.
Drivers shall support either of these modes but not both simultaneously.
The support for asynchronous mode is advertised through the new flag
QCA_WLAN_VENDOR_FEATURE_TWT_ASYNC_SUPPORT. If the driver does not
include this flag, it shall support synchronous mode.
Sunil Dutt [Tue, 24 Nov 2020 18:11:11 +0000 (23:41 +0530)]
New TWT operations and attributes to TWT Setup and Nudge
Define the following additional TWT operations:
QCA_WLAN_TWT_GET_STATS, QCA_WLAN_TWT_CLEAR_STATS,
QCA_WLAN_TWT_GET_CAPABILITIES, QCA_WLAN_TWT_SETUP_READY_NOTIFY.
Also define new attributes to qca_wlan_vendor_attr_twt_setup
and qca_wlan_vendor_attr_twt_nudge.
Extend the setband support for 6 GHz and band combinations
Support possible band combinations of 2.4 GHz, 5 GHz, and 6 GHz with
QCA_WLAN_VENDOR_ATTR_SETBAND_MASK attribute. Ensure backwards
compatibility with old drivers that are using
QCA_WLAN_VENDOR_ATTR_SETBAND_VALUE attribute and supporting only 2.4 GHz
and 5 GHz bands.
Jouni Malinen [Fri, 4 Dec 2020 11:45:03 +0000 (13:45 +0200)]
Fix gcc-10 build with -Werror=array-bounds and dl_list_for_each()
The earlier workaround for UBSAN issues in commit 3b6b3ae58133 ("Modify
dl_list_for_each() to not use unaligned access with WPA_TRACE") ended up
using a construction in which the type cast to the containing structure
was compared instead of the struct dl_list pointers. While that worked
around the UBSAN issue, it resulted in a comparison that gcc-10
interprets as being out of bounds for struct dl_list (which it obviously
is since this is to find the start of the containing structure).
Revert that workaround and instead, mark the struct dl_list used within
struct os_alloc_trace to have matching 16 octet alignment as the
containing structure. This is also restoring consistent design for
dl_list_for_each*().
Avraham Stern [Tue, 3 Nov 2020 07:54:16 +0000 (09:54 +0200)]
nl80211: Report invalid signal and noise when info is unavailable
When the driver sends a CQM RSSI threshold event, wpa_supplicant queries
the driver for the signal and noise values. However, it is possible that
by that time the station has already disconnected from the AP, so these
values are no longer valid. In this case, indicate that these values are
invalid by setting them to WPA_INVALID_NOISE.
Previously a value of 0 would be reported, which may be confusing as
this is a valid value.
Since nl80211_get_link_signal() and nl80211_get_link_noise() already set
invalid values for a case of failure, just use the value set by these
functions even if they fail.
Signed-off-by: Avraham Stern <avraham.stern@intel.com> Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com> Reviewed-by: Brian Norris <briannorris@chromium.org>
Thomas Pedersen [Thu, 22 Oct 2020 18:20:32 +0000 (11:20 -0700)]
wlantest: Allow missing RSNE in S1G beacon
S1G beacons save a few bytes by not requiring the RSNE in beacon if RSN
BSS is configured. Handle this in wlantest by only clearing RSNE from
the BSS info if frame is a Probe Response frame.
Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Frederik Juul [Thu, 23 Jul 2020 08:39:41 +0000 (08:39 +0000)]
wpa_supplicant: Initial connection speedup
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>
Shay Bar [Tue, 30 Jun 2020 14:23:46 +0000 (17:23 +0300)]
hostapd: Add ability to disable HT/VHT/HE per BSS
Add the ability to disable HT/VHT/HE for specific BSS from hostapd.conf.
- Add disable_11ax boolean to hostapd_bss_config.
- Change disable_11n and disable_11ac to bool in hostapd_bss_config.
- Add configuration option to set these disable_11* parameters
(which were previously used only automatically based on incompatible
security parameters to disable HT/VHT).
Markus Theil [Tue, 30 Jun 2020 12:19:04 +0000 (14:19 +0200)]
tests: DFS test for mesh
Add a test which uses a DFS channel, waits for CAC afterwards successful
mesh join and then triggers a radar event and checks if the mesh comes
up again on the same device.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Markus Theil [Tue, 30 Jun 2020 12:19:03 +0000 (14:19 +0200)]
mesh: Fix DFS deinit/init
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>
Peter Oh [Tue, 30 Jun 2020 12:19:00 +0000 (14:19 +0200)]
mesh: Fix channel switch error during CAC
Mesh interface would have used its channel parameters that were
configured during initialization even after channel switch due to DFS
radar detection during CAC which could result in a channel switch error.
Fix the error by updating the channel parameters when channel has been
changed from the initial one.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
Peter Oh [Tue, 30 Jun 2020 12:18:59 +0000 (14:18 +0200)]
nl80211: Do not set offchanok on DFS channels in non-ETSI for mesh
mac80211 does not allow mgmt tx to use off channel on
DFS channels in non-ETSI domain, because it will invalidate
CAC result on current operating channel.
(mac80211 commit: 34373d12f3cbb74960a73431138ef619d857996f)
Hence don't set offchanok for mgmt tx in case of DFS channels
in non-ETSI.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
Peter Oh [Tue, 30 Jun 2020 12:18:57 +0000 (14:18 +0200)]
mesh: Update ssid->frequency as pri/sec channels switch
ssid->frequency is one of the variables used to get the channel number
from a given frequency. Leaving it as unchanged when pri/sec channel
switch will cause picking up a wrong channel number after applying the
secondary channel offset for HT40 and leads to failing interface
bring-up.
Signed-off-by: Peter Oh <peter.oh@bowerswilkins.com>
Peter Oh [Tue, 30 Jun 2020 12:18:56 +0000 (14:18 +0200)]
mesh: Use setup completion callback to complete mesh join
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>
Sachin Ahuja [Mon, 23 Nov 2020 11:27:40 +0000 (16:57 +0530)]
QCA vendor attributes to configure TX and RX NSS
Define QCA vendor attributes to dynamically configure TX NSS and RX NSS
to be used with QCA_NL80211_VENDOR_SUBCMD_SET_WIFI_CONFIGURATION and
QCA_NL80211_VENDOR_SUBCMD_GET_WIFI_CONFIGURATION commands.
AP: Check driver support while auto-selecting bandwidth for AP/P2P GO
If the maximum operating channel width for AP/P2P GO is not specified,
it is auto-selected during configuration. While selecting the channel
width, if VHT is supported and 160 MHz channels are available, 160 MHz
channel width is preferred to 80 MHz.
During the selection of the channel width, the corresponding driver
capabilities were not checked. As a result, the AP/P2P GO configuration
was set to use the available 160 MHz channels even if the driver did not
have capability to support the 160 MHz channel width causing failure to
start the AP/P2P GO.
Fix this by checking the driver support for the 160 MHz channel width
while selecting the channel width for AP/P2P GO.
Jouni Malinen [Mon, 23 Nov 2020 18:32:26 +0000 (20:32 +0200)]
Indicate whether additional ANQP elements were protected
Store information on whether extra ANQP elements were received using the
protection alternative (protected GAS during an association using PMF)
and make this available through the control interface BSS command.
For example:
anqp[277]=<hexdump>
protected-anqp-info[277]=1
TWT nudge is a combination of suspend and resume in a single request.
Add TWT nudge operation and QCA vendor attributes to support
the TWT nudge request.
Jouni Malinen [Mon, 16 Nov 2020 14:25:17 +0000 (16:25 +0200)]
BSS: Use variable length array for IEs at the end of struct wpa_bss
Replace the previously used design "(u8 *) (bss + 1)" with a variable
length array at the end of struct wpa_bss bss->ies[] in hopes of making
this easier to understand for static analyzers.
Rohan Dutta [Sat, 7 Nov 2020 13:53:24 +0000 (19:23 +0530)]
nl80211: Send the sae_pwe value to the driver
Use NL80211_ATTR_SAE_PWE attribute to indicate the sae_pwe value
to the driver during the NL80211_CMD_START_AP and NL80211_CMD_CONNECT
in WPA3-Personal networks which are using SAE authentication.
Vamsi Krishna [Mon, 16 Nov 2020 07:01:34 +0000 (12:31 +0530)]
P2P: Disable P2P in the 6 GHz band for now
P2P usage in the 6 GHz band is not standardized yet by WFA. Disable P2P
operations in the 6 GHz band to avoid potential interop issues with
existing P2P devices in production. P2P operations in the 6 GHz band can
be reenabled later after defining standard ways to address potential
interop issues with existing P2P devices.
P2P: Include p2p_6ghz_disable in global configuration
Previously, the configuration to disable the 6 GHz band remained local
to the P2P interface. With this there is a possibility of 6 GHz channels
being included in the channel list when the channel list needs to be
updated if the state changes on one of the interfaces.
Include the configuration to disable the 6 GHz band for P2P as a global
configuration value to prevent the inclusion of 6 GHz channels in the
channel list for P2P when the channel list needs to be updated during
the state change in one of the interfaces.
Vinita S. Maloo [Thu, 15 Oct 2020 14:58:17 +0000 (20:28 +0530)]
Add connect fail reason code from the driver to assoc reject event
Add support to report a vendor specific connect fail reason code fetched
from the driver to users by adding the reason code to the event
CTRL-EVENT-ASSOC-REJECT. Fetch the connect fail reason code when the
driver sends a failure connection result and append the reason code, if
available, to assoc reject event.
Sunil Dutt [Wed, 4 Nov 2020 16:09:38 +0000 (21:39 +0530)]
Vendor feature capability to support concurrent sessions on Wi-Fi bands
Introduces a vendor specific feature capability
QCA_WLAN_VENDOR_FEATURE_CONCURRENT_BAND_SESSIONS to know if the device
supports concurrent network sessions on different Wi-Fi bands. This feature
capability is attributed to the hardware's capability to support the same
(e.g., DBS).
Vinita S. Maloo [Tue, 27 Oct 2020 18:03:59 +0000 (23:33 +0530)]
Set NLA_F_NESTED flag with NL80211_ATTR_VENDOR_DATA conditionally
The newer kernel versions enforce strict netlink attribute policy
validation and will cause cfg80211 to reject vendor commands with
NL80211_ATTR_VENDOR_DATA if NLA_F_NESTED attribute is not set but
if the vendor command is expecting nested data within
NL80211_ATTR_VENDOR_DATA attribute.
Most of the earlier instances were addressed by adding NLA_F_NESTED
flag in nla_nest_start(). This commit addresses the remaining
instance in which NL80211_ATTR_VENDOR_DATA is populated using data
set by user through the control interface.
Enhance the control interface VENDOR command to indicate whether the
vendor subcommand uses nested attributes within NL80211_ATTR_VENDOR_DATA
attribute or not.
Set NLA_F_NESTED flag for existing QCA vendor commands which use nested
attributes within the NL80211_ATTR_VENDOR_DATA attributes so that the
old frameworks implementations for already existing commands work
without any issues.
Add test configuration attr to enable/disable full bandwidth UL MU-MIMO
Define a QCA vendor attribute
QCA_WLAN_VENDOR_ATTR_WIFI_TEST_CONFIG_FULL_BW_UL_MU_MIMO to
enable/disable full bandwidth UL MU-MIMO subfield in the HE PHY
capabilities information field for testing purposes.
Jouni Malinen [Wed, 4 Nov 2020 17:50:34 +0000 (19:50 +0200)]
BSS/scan: More conversions to for_each_element_id()
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().
Jouni Malinen [Tue, 3 Nov 2020 18:11:17 +0000 (20:11 +0200)]
EAP-IKEv2: Try to make transform parser simpler to understand
Use a local variable to try to make ikev2_parse_proposal() easier for
static analyzers to understand. Bounds checking in the loop is really
done by the ikev2_parse_transform() function, so the p->num_transforms
value itself is of no importance for that part and even that was already
implicitly limited in range.
Jouni Malinen [Tue, 3 Nov 2020 17:56:39 +0000 (19:56 +0200)]
HS 2.0: Clarify OSU Friendly Name length validation
This extends the changes in commit 0570a3ea7d87 ("HS 2.0: Clarify OSU
Provider list length validation") to cover the length field for the OSU
Friendly Name value to try to get this easier for static analyzers to
understand.
Jouni Malinen [Mon, 2 Nov 2020 16:46:35 +0000 (18:46 +0200)]
Avoid undefined behavior with memcpy PMK/PSK update
When SAE is used, the local pointer pmk may point to sm->PMK. Skip the
memcpy operation in such a case since it is not really needed and use of
overlapping memory buffers is undefined behavior for memcpy().
Jouni Malinen [Mon, 2 Nov 2020 16:38:41 +0000 (18:38 +0200)]
nl80211: Fix filtering of unsupported bands/modes
The loop for removing unsupported bands was assuming there is always
exactly one band/mode following the removed band. That was not at all
correct, so fix this by dynamically determining how many (if any) bands
need to be moved.
Jouni Malinen [Mon, 2 Nov 2020 15:26:04 +0000 (17:26 +0200)]
TDLS: Fix error path handling for TPK M1 send failures
Local allocation error or failure to get a random number could have
resulted in the peer entry getting freed and couple of the error path
cases in callers could have tried to reference or delete the peer after
that. Fix this by tracking the errors where the peer is freed.
Jouni Malinen [Mon, 2 Nov 2020 14:52:01 +0000 (16:52 +0200)]
DPP2: Fix error path handling in enterprise provisioning
The allocated memory pointed by the pem pointer was freed on an error
path without clearing the pointer to NULL before returning it from the
function. This could have resulted in use of freed memory in an error
case. Fix this by clearing the pointer so that the function returns NULL
properly in the case of this error.
Jouni Malinen [Mon, 2 Nov 2020 14:36:14 +0000 (16:36 +0200)]
Remove unused variable update
Commit e8b85c078e73 ("iface match: Unspecified matched interfaces should
not log driver fails") removed the only use of the added interface wpa_s
pointer, but left that pointer setting in place. Remove it to keep
static analyzers happy.
Jouni Malinen [Mon, 2 Nov 2020 14:31:02 +0000 (16:31 +0200)]
DPP2: Fix ppkey parsing
DPP_CONFIGURATOR_ADD processing of the new ppkey parameter had a
copy-paste error in determining the correct length of this parameter.
Fix that by referencing the correct pointer.
Fixes: 9c1fbff07404 ("DPP2: Generate a privacy protection key for Configurator") Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
6 GHz: Fix opclasses mapping in ieee80211_freq_to_channel_ext()
Previously only primary channel number used to calculate 6GHz operating
class in ieee80211_freq_to_channel_ext() and it is always giving 131
operating class. Fix this by mapping operating class using chanwidth and
sec_channel also.
This is needed to avoid OCV failures on the 6 GHz band when the channel
width is larger than 20 MHz.
Hai Shalom [Fri, 16 Oct 2020 19:47:28 +0000 (12:47 -0700)]
EAP-SIM peer: Send AT_IDENTITY first
For EAP-SIM connections, reorder the order of the attributes in
EAP-Response/SIM/Start message: Send AT_IDENTITY first, then
AT_NONCE and AT_VERSION instead of AT_IDENTITY last. Even though there
is no order requirements in the RFC, some implementations expect the
order of the attributes to be exactly as described in the RFC figures.
Peer Authenticator
| |
| +------------------------------+
| | Server does not have a |
| | Subscriber identity available|
| | When starting EAP-SIM |
| +------------------------------+
| |
| EAP-Request/SIM/Start |
| (AT_ANY_ID_REQ, AT_VERSION_LIST) |
|<------------------------------------------------|
| |
| |
| EAP-Response/SIM/Start |
| (AT_IDENTITY, AT_NONCE_MT, |
| AT_SELECTED_VERSION) |
|------------------------------------------------>|
| |
Pooventhiran G [Tue, 27 Oct 2020 06:24:21 +0000 (11:54 +0530)]
nl80211: Check for proper nlmsg allocation in send_and_recv_msgs_owner()
When nlmsg allocation fails, nl80211_drv_msg() returns NULL and the call
to send_and_recv_msgs_owner() from nl80211_leave_ibss() could have ended
up dereferencing a NULL pointer. Fix this by make
send_and_recv_msgs_owner() more consistent with other send_and_recv*()
cases that check msg == NULL internally.
Fixes: 12ea7dee311b ("nl80211: Use nl80211 control port for receiving EAPOL frames") Signed-off-by: Pooventhiran G <pooventh@codeaurora.org>
P2P: Consider BSS entry pending for P2P joining as a known BSS
Consider the BSS entry that is pending for the P2P group join operation
also as a known network along with the existing configured networks.
This prevents removal of the BSS entry that is still in the process of
P2P join operation from the BSS table when the number of entries exceed
bss_max_count.
Jouni Malinen [Mon, 26 Oct 2020 20:34:07 +0000 (22:34 +0200)]
nl80211: Filter out unsupported bands
If the driver indicates capability for a band that
hostapd/wpa_supplicant does not support, the struct hostapd_hw_modes
array of bands got an empty entry for that with NUM_HOSTAPD_MODES as the
mode. This resulted in various issues, e.g., with fst_hw_mode_to_band()
hitting a WPA_ASSERT(0).
Fix this by filtering out unsupported bands from the internal data
structures.
P2P: Include p2p_add_cli_chan parameter while cloning the configuration
The dynamically created P2P group interface did not consider the
channels that can be used by the P2P client during the P2P group
formation. Copy the p2p_add_cli_chan parameter while cloning the
configuration to the P2P group interface. This allows the dynamically
created group interface case to form the group in the specific
client-only channels when the device is a P2P client in the group.
If SAE_CONFIG_PK is not defined and sae->pk isn't zero (which is
possible as it is controlled by the commit message status code),
sae_derive_keys() may end up deriving PMK and KCK from an
uninitialized array. Fix that.
Fixes: 6b9e99e571ee ("SAE-PK: Extend SAE functionality for AP validation") Fixes: 20ccf97b3dc1 ("SAE-PK: AP functionality") Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
wpa_supplicant: Remove unfeasible conditions in config parsing
pos can't be NULL in wpa_global_config_parse_str(), so there is no point
checking this, especially when pos was already dereferenced earlier.
Remove the redundant conditions.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Pooventhiran G [Sun, 18 Oct 2020 15:30:09 +0000 (21:00 +0530)]
OpenSSL: Make openssl_debug_dump_certificate() more robust
SSL_CTX_get0_certificate() returns NULL if no certificate is installed.
While this should not be the case here due to the loop in
openssl_debug_dump_certificate_chains() proceeding only if the
SSL_CTX_set_current_cert() returns success, it is safer to make
openssl_debug_dump_certificate() explicitly check against NULL before
trying to dump details about the certificate.
Signed-off-by: Pooventhiran G <pooventh@codeaurora.org>
projects / thirdparty / hostap.git / log
