Occasionally, we would run into this warning:

  iwlwifi 0000:02:00.0: U iwl_mvm_protect_session extend 0x2601: only 200 ms left
  iwlwifi 0000:02:00.0: U iwl_mvm_remove_time_event Removing TE 0x2601
  iwlwifi 0000:02:00.0: I iwl_pcie_enqueue_hcmd Sending command TIME_EVENT_CMD (#29), seq: 0x0925, 60 bytes at 37[5]:9
  iwlwifi 0000:02:00.0: U iwl_pcie_send_hcmd_sync Attempting to send sync command TIME_EVENT_CMD
  iwlwifi 0000:02:00.0: U iwl_pcie_send_hcmd_sync Setting HCMD_ACTIVE for command TIME_EVENT_CMD
  iwlwifi 0000:02:00.0: I iwl_pcie_enqueue_hcmd Sending command TIME_EVENT_CMD (#29), seq: 0x0926, 60 bytes at 38[6]:9
  iwlwifi 0000:02:00.0: U iwl_mvm_time_event_response TIME_EVENT_CMD response - UID = 0x2601
  iwlwifi 0000:02:00.0: I iwl_pcie_hcmd_complete Clearing HCMD_ACTIVE for command TIME_EVENT_CMD
  iwlwifi 0000:02:00.0: U iwl_mvm_rx_time_event_notif Time event notification - UID = 0x2701 action 1
  wlan0: associate with 00:0a:b8:55:a8:30 (try 2/3)
  ------------[ cut here ]------------
  WARNING: at drivers/net/wireless/iwlwifi/mvm/time-event.c:269 iwl_mvm_time_event_send_add+0x163/0x1a0 [iwlmvm]()
  Modules linked in: [...]
  Call Trace:
   [<c1046e42>] warn_slowpath_common+0x72/0xa0
   [<c1046e92>] warn_slowpath_null+0x22/0x30
   [<f8cad913>] iwl_mvm_time_event_send_add+0x163/0x1a0 [iwlmvm]
   [<f8cadead>] iwl_mvm_protect_session+0xcd/0x1c0 [iwlmvm]
   [<f8ca2087>] iwl_mvm_mac_mgd_prepare_tx+0x67/0xa0 [iwlmvm]
   [<f882a130>] ieee80211_sta_work+0x8f0/0x1070 [mac80211]

The reason is a problem with asynchronous vs. synchronous
commands, what happens here is the following:
 * TE 0x2601 is removed, the TIME_EVENT_CMD for that is async
 * a new TE (will be 0x2701) is created, the TIME_EVENT_CMD
   for that is sync and also uses a notification wait for the
   response (to avoid another race condition)
 * the response for the TE 0x2601 removal comes from the
   firmware, and is handled by the notification wait handler
   that's really waiting for the second response, but can't
   tell the difference, we therefore see the message
   "TIME_EVENT_CMD response - UID = 0x2601" instead of
   "TIME_EVENT_CMD response - UID = 0x2701".

Fix this issue by making the TE removal synchronous as well,
this means that we wait for the response to that command
first, before there's any chance of sending a new one.

Also, to detect such issues more easily in the future, add
a warning to the notification handler that detects them.
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

This is helpful for debugging the time event warning,
but also in general to see what's going on.
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

All station commands must include a valid MAC ID,
the ID 0 is randomly valid in some cases, but we
must set the ID properly. Do that by passing the
right station and using its mac_id_n_color.
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

For the firmware to know when DTIM beacons arrive
we have to program the DTIM time in TSF and system
time in the MAC context. Since mac80211 now tracks
the different times (on demand), this becomes easy.
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The iwlwifi-next tree removed IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC
while the mac80211-next tree removed
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Otherwise memory corruption occurs when using channel contexts (ATM when
param 'channel' > 1).
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Even when connecting to an AP that doesn't support VHT,
and even when the local device doesn't support it either,
the downgrade message gets printed. Suppress the message
if HT and/or VHT is disabled.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Adding the flag to mac80211 already without testing was
clearly a mistake, one that we now pay for by having to
reserve bit 13 forever. The problem is cfg80211 doesn't
allow capability/rate changes for station entries that
were added unassociated, so the station entries cannot
be set up properly when marked associated.

Change the NL80211_FEATURE_FULL_AP_CLIENT_STATE value
to make it clear to userspace implementations that all
current kernels don't actually support it, even though
the previous bit is set, and of course also remove the
flag from mac80211 until we test and fix the issues.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The information of the peer's capabilities is required for the driver
to perform TDLS Peer UAPSD operations. This information of the peer is
passed by the supplicant using NL80211_CMD_SET_STATION command. This
commit enhances the function nl80211_set_station to pass this
information of the peer to the driver in case this command is used
with the TDLS peer STA.

In addition, make the HT/VHT capability configuration handled more
consistently for other STA cases (reject both instead of just HT).
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The information of the peer's capabilities and extended capabilities are
required for the driver to perform TDLS Peer UAPSD operations and off
channel operations. This information of the peer is passed from user space
using NL80211_CMD_SET_STATION command. This commit enhances
the function nl80211_set_station to pass the capability information of
the peer to the driver.

Similarly, there may be need for capability information for other modes,
so allow this to be provided with both add_station and change_station.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Use the new extended capabilities advertising to advertise
the fact that operating mode notification is supported.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

In many cases, userspace may need to know which of the
802.11 extended capabilities ("Extended Capabilities
element") are implemented in the driver or device, to
include them e.g. in beacons, assoc request/response
or other frames. Add a new nl80211 attribute to hold
the extended capabilities bitmap for this.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Instead of modifying the HT SMPS capability field
for stations, track the SMPS mode explicitly in a
new field in the station struct and use it in the
drivers that care about it. This simplifies the
code using it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Mesh interfaces will now respond to any broadcast (or
matching directed mesh) probe requests with a probe
response.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Previously, the entire mesh beacon would be generated each
time the beacon timer fired. Instead generate a beacon
head and tail (so the TIM can easily be inserted when mesh
power save is on) when starting a mesh or the MBSS
parameters change.

Also add a mutex for protecting beacon updates and
preventing leaks.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Make all the parsed IE pointers const, and propagate
the change to all the users etc.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

When drivers or regulatory have limitations on
40, 80 or 160 MHz channels, advertise these to
userspace via nl80211. Also add a new feature
flag to let userspace know this is supported.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Some drivers might support 80 or 160 MHz only on some
channels for whatever reason, so allow them to disable
these channel widths. Also maintain the new flags when
regulatory bandwidth limitations would disable these
wide channels.
Reviewed-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

A while ago, I made the mac80211 station code never change
the channel type after association. This solved a number of
issues but is ultimately wrong, we should react if the AP
changes the HT operation IE and switches bandwidth. One of
the issues is that we associate as HT40 capable, but if the
AP ever switches to 40 MHz we won't be able to receive such
frames because we never set our channel to 40 MHz.

This addresses this and VHT operation changes. If there's a
change that is incompatible with our setup, e.g. if the AP
decides to change the channel entirely (and for some reason
we still hear the beacon) we'll just disconnect.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The next patch will need it further up in the file, so
move it unchanged now.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

For HT and VHT the current bandwidth can change,
add the function ieee80211_vif_change_bandwidth()
to take care of this. It returns a failure if the
new bandwidth isn't compatible with the existing
channel context, the caller has to handle that.
When it happens, also inform the driver that the
bandwidth changed for this virtual interface (no
drivers would actually care today though.)

Changing to/from HT/VHT isn't allowed though.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The channel use is confusing, some uses the channel
context and some the bss_conf.chandef. The latter is
fine, so get rid of the channel context part.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Having HT/VHT operation IEs but not capability IEs
leads to a strange situation where we configure the
channel to an HT or VHT bandwidth and then can't
actually use it. Prevent this by checking that the
HT and VHT capability IEs are present as well as
the operation IEs; if not, disable HT and/or VHT.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

In beacons and association response frames an AP may include an
operating mode notification element to advertise changes in the
number of spatial streams it can receive. Handle this using the
existing function that handles the action frame, but only handle
NSS changes, not bandwidth changes which aren't allowed here.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

This should be called ieee80211_change_chanctx() since
it changes the channel context, not a chandef.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

The code to disable HT and VHT if VHT was advertised
without VHT is wrong -- it accidentally uses the wrong
flags. Fix that.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

In case of connection, the station data is initialised from
the beacon/probe response first and then updated from the
association response. If the latter is different we update
the rate control algorithm and driver. Instead of doing it
this way, set the station data properly with data from the
association response before initializing rate control.

Also simplify the code by passing the station pointer.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Handle the operating mode notification action frame.
When the supported streams or the bandwidth change
let the driver and rate control algorithm know.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

With VHT, a station can change the number of spatial
streams it can receive on the fly, not unlike spatial
multiplexing in HT. Prepare for that by tracking the
maximum number of spatial streams it can receive when
the connection is established.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Define the action frame format, the VHT category
and its action types and the field format and EID
for operating mode notifications. The frame may
be used outside of VHT context as well, so don't
include "VHT" in the names.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

For VHT, many more bandwidth changes are possible. As a first
step, stop toggling the IEEE80211_HT_CAP_SUP_WIDTH_20_40 flag
in the HT capabilities and instead introduce a bandwidth field
indicating the currently usable bandwidth to transmit to the
station. Of course, make all drivers use it.

To achieve this, make ieee80211_ht_cap_ie_to_sta_ht_cap() get
the station as an argument, rather than the new capabilities,
so it can set up the new bandwidth field.

If the station is a VHT station and VHT bandwidth is in use,
also set the bandwidth accordingly.

Doing this allows us to get rid of the supports_40mhz flag as
the HT capabilities now reflect the true capability instead of
the current setting.

While at it, also fix ieee80211_ht_cap_ie_to_sta_ht_cap() to not
ignore HT cap overrides when MCS TX isn't supported (not that it
really happens...)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Like with HT, make things a bit simpler in future patches by
passing the station to ieee80211_vht_cap_ie_to_sta_vht_cap()
instead of the vht_cap pointer. Also disable VHT here if HT
isn't supported.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

Since no driver calls the TKIP functions from interrupt
context, there's no need to use spin_lock_irqsave().
Just use spin_lock_bh() (and spin_lock() in the TX path
where we're in a BH or they're already disabled.)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

There's no need to use _irqsave() as the lock
is never used in interrupt context.

This also fixes a problem in the iwlwifi MVM
driver that calls spin_unlock_bh() within its
set_tim() callback.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

There's no use for it, WPA is entirely handled in
wpa_supplicant in userspace, so don't pick the IE.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

In some cases when disconnecting after (or during?) CSA
the queues might not recover, and then the only way to
recover is reloading the module.

Fix this by always unblocking the queue CSA reason when
disconnecting.

Cc: stable@vger.kernel.org
Reported-by: Jan-Michael Brummer <jan.brummer@tabos.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>