It doesn't seem problematic to change the weight for the average
beacon signal from 3 to 4, so use DECLARE_EWMA. This also makes
the code easier to maintain since bugs like the one fixed in the
previous patch can't happen as easily.

With a fix from Avraham Stern to invert the sign since EMWA uses
unsigned values only.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

These file aren't really useful:
 - if per beacon data is required then you need to use
   radiotap or similar anyway, debugfs won't help much
 - average beacon signal is reported in station info in
   nl80211 and can be looked up with iw
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Sometimes we are interested in testing TDLS performance in a specific
width setting. Add the ability to disable the wider-band feature, thereby
allowing the TDLS channel width to be controlled by the BSS width.
Signed-off-by: NArik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Define rc_rateidx_vht_mcs_mask array and rate_idx_match_vht_mcs_mask()
method in order to apply mcs mask for vht rates
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Move TDLS channel-switch Rx handling into an RTNL locked work. This is
required to add proper regulatory checking to incoming channel-switch
requests.
Queue incoming requests in a dedicated skb queue and handle the request
in a device-specific work to avoid deadlocking on interface removal.
Signed-off-by: NArik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add support for declaring MU-MIMO beamformee capability for
relevant hardware.
When sending association request, the capability is included if both
hardware and the AP support it, and no other virtual interface
is using it.
This is in order to avoid multiple interfaces using MU-MIMO in parallel
which might lead to contradictions in the group-id mechanism.
Signed-off-by: NSara Sharon <sara.sharon@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Most of the fields in this struct use too wide types, change
that to shrink the struct from 64 to 48 bytes (on 64-bit.)

This results in a total saving of 64 bytes for each interface.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The IEEE802.11-2012 specification is vague regarding SMPS operation during
TDLS. It does not define a clear way to transition between SMPS states.

To avoid interop issues, set SMPS to off when TDLS peers are connected.
Accomplish this by extending the definition of the AUTOMATIC state. If the
driver forces a state other than OFF, disconnect all TDLS peers.

While at it, avoid changing the SMPS state of the peer STA. We have no
way to control it, so try and behave correctly towards it.

Move the TDLS peer-teardown function to where the rest of the TDLS code
resides.
Signed-off-by: NArik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Define a station chandef, to be used for wider-bw TDLS peers. When both
peers support the feature, upgrade the channel bandwidth to the maximum
allowed by both peers and regulatory. Currently widths up to 80MHz are
supported in the 5GHz band.

When a TDLS peer connects/disconnects recalculate the channel type of the
current chanctx.
Make the chanctx width calculation consider wider-bw TDLS peers and
similarly fix the max_required_bw calculation for the chanctx min_def.
Since the sta->bandwidth is calculated only later on, take
bss_conf.chandef.width as the minimal width for station interface.

Set the upgraded channel width in the VHT-operation set during TDLS setup.
Signed-off-by: NArik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

There's a long-standing TODO item to use this flag in the cooked
monitor RX, but clearly it was never needed and now this hasn't
been used by userspace for a long time, so no userspace changes
could require it now.

Remove the unused flag.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When introducing multiple RX queues, a single NAPI struct will not
be sufficient. Instead of trying to store multiple, simply change
the API to have the NAPI struct passed to the RX function. This of
course means that drivers using rx_irqsafe() cannot use NAPI, but
that seems a reasonable trade-off, particularly since only two of
all drivers are currently using it at all.

While at it, we can now remove the IEEE80211_RX_REORDER_TIMER flag
again since this code path cannot have a napi struct anyway.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This function is only used in the RX code, so moving it into
that file gives the compiler better optimisation possibilities
and also allows us to remove the check for short frames (which
in the RX path cannot happen, but as a generic utility needed
to be checked.)
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Short frames less than 16 octets are already blocked in the monitor
code by the should_drop_frame() function, and cannot get into the
regular RX path. Therefore, this check can never trigger and the
counter invariably stays zero. Remove the useless code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Due to remain-on-channel scheduling delays, when we split an ROC
while coalescing, we'll usually get a picture like this:

existing ROC:  |------------------|
current time:              ^
new ROC:                   |------|              |-------|

If the expected response frames are then transmitted by the peer
in the hole between the two fragments of the new ROC, we miss
them and the process (e.g. ANQP query) fails.

mac80211 expects that the window to miss something is small:

existing ROC:  |------------------|
new ROC:                   |------||-------|

but that's normally not the case.

To avoid this problem, coalesce only if the new ROC's duration
is <= the remaining time on the existing one:

existing ROC:  |------------------|
new ROC:                   |-----|

and never split a new one but schedule it afterwards instead:

existing ROC:  |------------------|
new ROC:                                       |-------------|

type=bugfix
bug=not-tracked
fixes=unknown
Reported-by: NMatti Gottlieb <matti.gottlieb@intel.com>
Reviewed-by: NEliadX Peller <eliad@wizery.com>
Reviewed-by: NMatti Gottlieb <matti.gottlieb@intel.com>
Tested-by: NMatti Gottlieb <matti.gottlieb@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

No matter how the driver manages its NAPI context, there's no way
sending frames to it from a timer can be correct, since it would
corrupt the internal GRO lists.

To avoid that, always use the non-NAPI path when releasing frames
from the timer.

Cc: stable@vger.kernel.org
Reported-by: NJean Trivelly <jean.trivelly@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

As I was testing with hwsim, I missed that my previous commit to
make LED work depend on activation broke the code because I missed
removing the old trigger struct and some code was still using it,
now erroneously, causing crashes.

Fix this by always using the correct struct.
Reported-by: NFelix Fietkau <nbd@openwrt.org>
Tested-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When LED triggers are compiled in, but not used, mac80211 will still
call them to update the status. This isn't really a problem for the
assoc and radio ones, but the TX/RX (and to a certain extend TPT)
ones can be called very frequently (for every packet.)

In order to avoid that when they're not used, track their activation
and call the corresponding trigger (and in the TPT case, account for
throughput) only when the trigger is actually used by an LED.

Additionally, make those trigger functions inlines since theyre only
used once in the remaining code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since these counters can only be read through debugfs, there's
very little point in maintaining them all the time. However,
even just making them depend on debugfs is pointless - they're
not normally used. Additionally a number of them aren't even
concurrency safe.

Move them under MAC80211_DEBUG_COUNTERS so they're normally
not even compiled in.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The debugfs statistics macros are pointlessly verbose, so change
that macro to just have a single argument. While at it, remove
the unused counters and rename rx_expand_skb_head2 to the better
rx_expand_skb_head_defrag.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

With promisc support gone, only AP and P2P-Device type interfaces
still clear IEEE80211_RX_RA_MATCH. In both cases this isn't really
necessary though, so we can remove that flag and the code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This support is essentially useless as typically networks are encrypted,
frames will be filtered by hardware, and rate scaling will be done with
the intended recipient in mind. For real monitoring of the network, the
monitor mode support should be used instead.

Removing it removes a lot of corner cases.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

In order to speed up mac80211's TX path, add the "fast-xmit" cache
that will cache the data frame 802.11 header and other data to be
able to build the frame more quickly. This cache is rebuilt when
external triggers imply changes, but a lot of the checks done per
packet today are simplified away to the check for the cache.

There's also a more detailed description in the code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

As the next patch will require the IE splitting utility functions
in cfg80211, move them there from mac80211.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This allows drivers to request per-vif and per-sta-tid queues from which
they can pull frames. This makes it easier to keep the hardware queues
short, and to improve fairness between clients and vifs.

The task of scheduling packet transmission is left up to the driver -
queueing is controlled by mac80211. Drivers can only dequeue packets by
calling ieee80211_tx_dequeue. This makes it possible to add active queue
management later without changing drivers using this code.

This can also be used as a starting point to implement A-MSDU
aggregation in a way that does not add artificially induced latency.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
[resolved minor context conflict, minor changes, endian annotations]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

We currently have a hand-rolled table with 256 entries and are
using the last byte of the MAC address as the hash. This hash
is obviously very fast, but collisions are easily created and
we waste a lot of space in the common case of just connecting
as a client to an AP where we just have a single station. The
other common case of an AP is also suboptimal due to the size
of the hash table and the ease of causing collisions.

Convert all of this to use rhashtable with jhash, which gives
us the advantage of a far better hash function (with random
perturbation to avoid hash collision attacks) and of course
that the hash table grows and shrinks dynamically with chain
length, improving both cases above.

Use a specialised hash function (using jhash, but with fixed
length) to achieve better compiler optimisation as suggested
by Sergey Ryazanov.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add VHT support for IBSS. Drivers could activate
this feature by setting NL80211_EXT_FEATURE_VHT_IBSS
flag.
Signed-off-by: NJanusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

In case of wide bandwidth (wider than 20MHz) used by IBSS,
scan all channels in chandef to be able to find neighboring
IBSS netwqworks that use the same overall channels but a different
control channel.
Signed-off-by: NJanusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This will expose in /sys whether the ifname of a device is set by
userspace or generated by the kernel. The latter kind (wlanX, etc)
is not deterministic, so userspace needs to rename these devices
to names that are guaranteed to stay the same between reboots. The
former, however should never be renamed, so userspace needs to be
able to reliably tell the difference.

Similar functionality was introduced for the rtnetlink core in
commit 5517750f ("net: rtnetlink - make create_link take name_assign_type")
Signed-off-by: NTom Gundersen <teg@jklm.no>
Cc: Kalle Valo <kvalo@qca.qualcomm.com>
Cc: Brett Rudley <brudley@broadcom.com>
Cc: Arend van Spriel <arend@broadcom.com>
Cc: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Cc: Hante Meuleman <meuleman@broadcom.com>
Cc: Johannes Berg <johannes@sipsolutions.net>
[reformat changelog to fit 72 cols]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of looking up the destination station twice in the TX path
(first to build the header, and then for control processing), save
it when building the header and use it later in the TX path.

To avoid having to look up the station in the many callers, allow
those to pass %NULL which keeps the existing lookup.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This mechanism was historic, and only ever used by IBSS, which
also doesn't need to have it as it properly manages station's
802.1X PAE state (or, with WEP, always has a key.)

Remove the mechanism to clean up the code.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If the AP is confused and starts doing a CSA to the same channel,
just ignore that request instead of trying to act it out since it
was likely sent in error anyway.

In the case of the bug I was investigating the GO was misbehaving
and sending out a beacon with CSA IEs still included after having
actually done the channel switch.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add the AID and VHT-cap/operation IEs during TDLS setup. Remove the
block of TDLS peers when setting HT-caps of the peer station.
Signed-off-by: NArik Nemtsov <arikx.nemtsov@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Some APs experience problems when working with
U-APSD. Decreasing the probability of that
happening by using legacy mode for all ACs but VO
isn't enough.

Cisco 4410N originally forced us to enable VO by
default only because it treated non-VO ACs as
legacy.

However some APs (notably Netgear R7000) silently
reclassify packets to different ACs. Since u-APSD
ACs require trigger frames for frame retrieval
clients would never see some frames (e.g. ARP
responses) or would fetch them accidentally after
a long time.

It makes little sense to enable u-APSD queues by
default because it needs userspace applications to
be aware of it to actually take advantage of the
possible additional powersavings. Implicitly
depending on driver autotrigger frame support
doesn't make much sense.

Cc: stable@vger.kernel.org
Signed-off-by: NMichal Kazior <michal.kazior@tieto.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Revert commit ad38bfc9 ("mac80211: Tx frame latency statistics")
(along with some follow-up fixes).

This code turned out not to be as useful in the current form as we
thought, and we've internally hacked it up more, but that's not
very suitable for upstream (for now), and we might just do that
with tracing instead.

Therefore, for now at least, remove this code. We might also need
to use the skb->tstamp field for the TCP performance issue, which
is more important than the debugging.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If suspend starts while ieee80211_scan_completed() is running, between
the point where SCAN_COMPLETED is set and the work is queued,
ieee80211_scan_cancel() will not catch the work and we may finish
suspending before the work is actually executed, leaving the scan
running while suspended.

To fix this race, queue the scan work during resume if the
SCAN_COMPLETED flag is set and flush it immediately.
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When we go to suspend, there is complex set of states that
avoids races. The quiescing variable is set whlie
__ieee80211_suspend is running. Then suspended is set.
The code makes sure there is no window without any of these
flags.

The problem is that workers can still be enqueued while we
are quiescing. This leads to situations where the driver is
already suspending and other flows like disassociation are
handled by a worker.

To fix this, we need to check quiescing and suspended flags
in the worker itself and not only before enqueueing it.
I also add here extensive documentation to ease the
understanding of these complex issues.
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Control per packet Transmit Power Control (TPC) in lower drivers
according to TX power settings configured by the user. In particular TPC is
enabled if value passed in enum nl80211_tx_power_setting is
NL80211_TX_POWER_LIMITED (allow using less than specified from userspace),
whereas TPC is disabled if nl80211_tx_power_setting is set to
NL80211_TX_POWER_FIXED (use value configured from userspace)
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

local->radar_detect_enabled should tell whether
radar_detect is enabled on any interface belonging
to local.

However, it's not getting updated correctly
in many cases (actually, when testing with hwsim
it's never been set, even when the dfs master
is beaconing).

Instead of handling all the corner cases
(e.g. channel switch), simply check whether
radar detection is enabled only when needed,
instead of caching the result.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When roaming / suspending, it makes no sense to wait until
the transmit queues of the device are empty. In extreme
condition they can be starved (VO saturating the air), but
even in regular cases, it is pointless to delay the roaming
because the low level driver is trying to send packets to
an AP which is far away. We'd rather drop these packets and
let TCP retransmit if needed. This will allow to speed up
the roaming.

For suspend, the explanation is even more trivial.
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Commit e1a0c6b3 ("mac80211: stop toggling IEEE80211_HT_CAP_SUP_WIDTH_20_40")
mistakenly removed the actual update of sta->sta.bandwidth.

Refactor ieee80211_sta_cur_vht_bw() into multiple functions
(calculate caps-bw and chandef-bw separately, and min them
with cur_max_bandwidth).

On ht chanwidth action frame set only cur_max_bandwidth
(according to the sta capabilities) and recalc the sta bw.
Signed-off-by: NEliad Peller <eliadx.peller@intel.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>