This was done using the below semantic patch. There are a few more
places that were missed due to variable declarations or additional
checks in the for loop.
@@
iterator name for_each_link;
identifier max_links =~ "MAX_NUM_MLD_LINKS|MAX_NUM_MLO_LINKS";
expression links;
expression further_tests;
identifier i;
statement stmt;
@@
-for (i = 0; i < max_links; i++)
+for_each_link(links, i)
{
(
- if (!(links & BIT(i)))
- continue;
...
|
- if (!(links & BIT(i)) || further_tests)
+ if (further_tests)
continue;
...
|
- if (further_tests || !(links & BIT(i)))
+ if (further_tests)
continue;
...
|
- if (links & BIT(i))
stmt
|
- if (further_tests && (links & BIT(i)))
+ if (further_tests)
stmt
|
- if ((links & BIT(i)) && further_tests)
+ if (further_tests)
stmt
)
}
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
In some places the wrong index variable was used to access the link
configuration. Fix this by simply using link_id instead of i.
With this, the i loop variable is not needed anymore. We can simply
always pass 0 to nla_nest_start(). Also, the kernel does not care about
the order that the links are provides, so just remove the special
handling for the association link.
Fixes: a2c4c0b1b6 ("nl80211: Support MLD association request")
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
When a device supports both the 5 GHz band and the 6 GHz band,
these are reported as two separate modes, both with mode set to
HOSTAPD_MODE_IEEE80211A. However, as these are different modes,
each with its own characteristics, e.g., rates, capabilities etc.,
specifically differentiate between them by adding a flag to indicate
whether the mode describes a 6 GHz band capabilities or not.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
For AP MLD cases, the link id is required to determine the correct link
to stop in the stop_ap() driver op.
Signed-off-by: Chenming Huang <quic_chenhuan@quicinc.com>
Add wpa_supplicant support for interacting with the NAN discovery engine
to allow USD as Publisher or Subscriber.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
When ACS offloaded to the driver, the channel selected event carries
link id to specify the link if operating as AP MLD.
Find the specified link to handle this event.
Signed-off-by: Chenming Huang <quic_chenhuan@quicinc.com>
ACS is triggered per link, so link id is needed for the driver to handle
when the ACS operation is offloaded.
Signed-off-by: Chenming Huang <quic_chenhuan@quicinc.com>
If both 6 GHz and S1G channels are included, the previously used timeout
was not long enough at least with mac80211_hwsim. Increase the initial
timeout to allow such a scan to be completed.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
QCA_WLAN_VENDOR_ATTR_CONFIG_AP_ALLOWED_FREQ_LIST needs be to set per
link if operating as an AP MLD.
Signed-off-by: Chenming Huang <quic_chenhuan@quicinc.com>
Radar detection event could have resulted in the driver interface
instance getting deinitialized and the related memory freed in the
middle of the loop. This was not an issue when the event was passed only
into a single interface, but it became an issue when the loop tried to
send it to all interfaces. If the driver were removed, that loop check
would have used freed memory. Avoid this by explicitly checking that the
driver interface instance is still valid.
Fixes: f136837202 ("nl80211: Pass wiphy events to all affected interfaces")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
nl_socket_get_cb() increases cb_refcnf for the cb that is bound to a
socket and as such, nl_cb_put() needs to be used with the returned cb
after having cloned it to avoid leaking memory due to cb_refcnt never
getting back to 0.
Fixes: da0d51fee7 ("nl80211: Use socket cb instead of global->nl_cb in send_and_recv()")
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
NL80211_WPA_VERSION_3 was only added in kernel 5.2 so it should not be
set for older kernel versions. There is no direct way to check if the
value is supported. However, we can use the new infrastructure to check
whether the kernel has the NL80211_ATTR_SAE_PASSWORD attribute. It is
related and was added at the same time.
Fixes: 6cc78b3945 ("nl80211: Set NL80211_WPA_VERSION_2 vs. _3 based on AKM")
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
The kernel will reject commands if newer attributes are included even
though they should not be. Add appropriate checks in set_ap().
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
Older kernel versions may not support all attributes and may refuse
commands that include them. To avoid sending too new attributes query
the highest supported attribute. This allows adding appropriate checks
where needed.
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
Previously, we would only pass the event to the first interface that
matches. However, one wiphy can have multiple interfaces and each one
needs to get the event delivered. Without this, it could e.g. happen
that a radar detection event is forwarded to p2p-dev-wlan0 and not to
the wlan0 interface which actually needs it.
As such, keep iterating if we are processing a wiphy match and send the
event to all affected BSSs.
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
This was done with spatch using the following semantic patch and minor
manual edits to clean up coding style and avoid compiler warnings in
driver_wext.c:
@@
expression a,b;
@@
- os_memcmp(a, b, ETH_ALEN) == 0
+ ether_addr_equal(a, b)
@@
expression a,b;
@@
- os_memcmp(a, b, ETH_ALEN) != 0
+ !ether_addr_equal(a, b)
@@
expression a,b;
@@
- !os_memcmp(a, b, ETH_ALEN)
+ ether_addr_equal(a, b)
Signed-off-by: Jouni Malinen <j@w1.fi>
The AP MLD MAC address is meant to be preserved here, but since it's
reset to zeroes before attempting to authenticate in
nl80211_mark_disconnected(), we can't just point to the AP MLD MAC
address in drv->auth_ap_mld_addr. Fix it by using a copy.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
cfg80211 requires the link ID to be specified for requests to start a
remain-on-channel operation during an ML association. This feels wrong
since the ROC operation is in most cases unrelated to the
association. However, that requirement has been in place since kernel
commit 7b0a0e3c3a88 ("wifi: cfg80211: do some rework towards MLO link
APIs") from April 2022, and as such, it looks necessary to have
wpa_supplicant work around this by specifying the currently used link ID
that would seem to match the ROC channel most closely.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Previously only the SHA-1 -based AKM was supported. Extend that to cover
all PSK AKMs so that the PSK configuration to the driver happens for all
the possible cases during AP start.
Signed-off-by: Vinayak Yadawad <vinayak.yadawad@broadcom.com>
An affiliated AP of an AP MLD can temporarily be disabled. Other
affiliated APs of the AP MLD indicate this in the Reduced Neighbor
Report (RNR) elements added to their Beacon and Probe Response frames.
When an affiliated AP is disabled, it should be included in the
association exchange, but can be activated only after it is enabled.
Add support identifying disabled APs and propagate the information
to the driver within the associate() callback.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
If the interface initialization fails, no links might be set when
calling the deinit functions. Those functions need to be prepared for
bss->n_links being 0.
Fixes: 859cbc396f ("nl80211: Remove links when stopping AP MLD in hostapd")
Signed-off-by: Jouni Malinen <j@w1.fi>
If the interface initialization fails, no links might be set when
calling the deinit functions. Those functions need to be prepared for
bss->flink being NULL.
Fixes: 47269be36e ("nl80211: Refactor i802_bss to support multiple links")
Signed-off-by: Jouni Malinen <j@w1.fi>
This helper is of no real use anymore, so get rid of it. This completes
send_and_recv*() cleanup. What remains is the most generic
send_and_recv() and two wrappers for it to cover the most common simpler
cases.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is a variant for the second most common case of send_and_recv() needs:
send a command with a special response handling.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is a variant for the most common case of send_and_recv() needs:
send a command without needing a special response handling. In addition,
move the helper functions into driver_nl80211.h since these are now
simple wrappers for the more flexible send_and_recv().
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is not really that helpful as a separate helper function, so get
rid of one of the many send_and_recv() variants.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is not helpful as a separate helper function anymore, so get rid of
one of the many send_and_recv() variants.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is independent of the NL80211_ATTR_SOCKET_OWNER use, so add these
attributes from a separate helper function that is called only from
locations that actually start an operation that uses EAPOL frames.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
Doing this based on driver support for control port RX or SAE seems like
undesired extra complexity. Just use this in all cases where the special
handle for a longer term operation, like connection or AP mode
operation, is needed.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is a step towards cleaning up all the workarounds that have showed
up over the years and made the design difficult to understand. This
removes use of the separate registration of process_bss_event() as
NL_CB_VALID for the individual commands since the use of the appropriate
cb (i.e., the one from nl80211_init_bss() in these cases) will already
point to the correct handler.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
This is a workaround for some race conditions where the hardcoded use of
global->nl_cb in send_and_recv() ends up getting events delivered
through unexpected context when they happen to arrive at the time when a
command is being processed.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
If ioctl() returns EBUSY on the command SIOCBRADDIF, the interface might
have already been added to the bridge by an external operation (e.g.,
netifd in OpenWrt), and linux_br_add_if() should not indicate an error.
Check whether the interface is correctly brigded when ioctl()
returns EBUSY and if so, report success.
Signed-off-by: Michael-CY Lee <michael-cy.lee@mediatek.com>
Link ID needs to be specified for MLD case when doing channel switch.
Add it to the driver command.
Signed-off-by: Chenming Huang <quic_chenhuan@quicinc.com>
With changes to optimize scan for specific BSSID, there arises a
scenario where in nl80211_trigger_scan() is called with a scan
randomization enabled. A combination of NL80211_ATTR_MAC for BSSID and
scan randomization, which uses NL80211_ATTR_MAC for a different purpose,
results in invalid error for the scan request. To fix the issue use
attribute NL80211_ATTR_BSSID instead of NL80211_ATTR_MAC.
NL80211_ATTR_BSSID was introduced in kernel commit 2fa436b3a2a7
("nl80211: Use different attrs for BSSID and random MAC addr in scan
req") in 2016. Prior to that, only NL80211_ATTR_MAC could be used for
specifying the target BSSID. For backwards compatibility, add the
NL80211_ATTR_MAC attribute as well when not using a random MAC address.
Signed-off-by: Vinayak Yadawad <vinayak.yadawad@broadcom.com>
The underlying driver can reject only specific links in an association
request. In that case we will only ignore the corresponding BSS. Add a
test for this flow by simulating an association failure on the second
link specifically.
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
Beacon hinting is a feature that can temporarily change the regulatory
rule flags on the channel where the radio hears the beacon. Add a new
event CTRL-EVENT-REGDOM-BEACON-HINT to notify the wpa_supplicant user
about an important update to the regulatory rules including which
frequencies are impacted, new power limit, and new rule flags.
Signed-off-by: Jun Yu <junyuu@chromium.org>
Do not allow offchannel operation for action frame transmission if
no frequency is specified, as this doesn't make sense.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
USe the link ID information to determine the specific affiliated link
when processing channel switch events on an AP MLD.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
There is an issue when starting a non-MLD AP on the same interface that
previous operated as an AP MLD. When the previous AP MLD got stopped,
links were not removed when using hostapd. Next non-MLD AP will fail to
start because some nl80211 operations still require link id (e.g., set
freq).
Remove links when AP MLD is stop to avoid such issue. This was already
done in the deinit_ap() handler when using wpa_supplicant, but hostapd
needs to do same.
Signed-off-by: Jouni Malinen <quic_jouni@quicinc.com>
If a link specific error occured, mark the offending link within the
association parameters.
Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
