If a new netdev (e.g. an AP VLAN) is created while the driver
has queues stopped, the new netdev queues will be started even
though they shouldn't. This will lead to frames accumulating
on the internal mac80211 pending queues instead of properly
being held on the netdev queues.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Given the (nested) switch statements, this code can't
be reached, so make it warn instead of manipulating
the carrier state which seems purposeful.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When a device is unplugged while suspended, mac80211 is
de-initialized and all interfaces are removed while no
state is actually present in the driver. This can cause
warnings and driver confusion.

Fix this by reordering the do_stop code to not call the
driver when it is suspended, i.e. when there's no state
in the driver anyway.

The previous patches removed a few corner cases in ROC
and virtual monitor interfaces so that now this is safe
to do and no state should be left over.
Reported-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

It has to be removed from the driver, but completely
destroying it helps handle unplug of a device during
suspend since then the channel context handling etc.
doesn't have to happen later when it's removed.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

They can't really be executed while suspended and could
trigger work warnings, so abort all ROC items. When the
system resumes the notifications about this will be
delivered to userspace which can then act accordingly
(though it will assume they were canceled/finished.)
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The code now explicitly calls ieee80211_configure_filter()
anyway, so nothing needs to be explained.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The master interface no longer exists ... and hasn't for
a few years now, so remove this reference :-)
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

There are a number of situations in which mac80211 only
really needs to flush queues for one virtual interface,
and in fact during this frames might be transmitted on
other virtual interfaces. Calculate and pass a queue
bitmap to the driver so it knows which queues to flush.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This basically reverts commit b207cdb0.

Now is possible to use drv_{add,remove}_interface() and vif->debugfs_dir
to create/remove per interface debugfs files. Remove redundant
callbacks.
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Some devices have multicast filter capability for each individual
virtual interface rather than just a global one. Add an interface
specific driver callback allowing such drivers to configure this.
Signed-off-by: NAlexander Bondar <alexander.bondar@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

For AP interfaces, there's no need to flush stations
or keys again when the interface is stopped as already
happened when the BSS was stopped on the interface.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since hostapd will remove keys this isn't usually
an issue, but we shouldn't leak keys to the next
BSS started on the same interface. For VLANs this
also fixes a bug, keys that aren't removed would
otherwise be leaked.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

During roaming, the crypto_tx_tailroom_needed_cnt counter
will often take values 2,1,0,1,2 because first keys are
removed and then new keys are added. This is inefficient
because during the 0->1 transition, synchronize_net must
be called to avoid packet races, although typically no
packets would be flowing during that time.

To avoid that, defer the decrement (2->1, 1->0) when keys
are removed (by half a second). This means the counter
will really have the values 2,2,2,3,4 ... 2, thus never
reaching 0 and having to do the 0->1 transition.

Note that this patch entirely disregards the drivers for
which this optimisation was done to start with, for them
the key removal itself will be expensive because it has
to synchronize_net() after the counter is incremented to
remove the key from HW crypto. For them the sequence will
look like this: 0,1,0,1,0,1,0,1,0 (*) which is clearly a
lot more inefficient. This could be addressed separately,
during key removal the 0->1->0 sequence isn't necessary.

(*) it starts at 0 because HW crypto is on, then goes to
    1 when HW crypto is disabled for a key, then back to
    0 because the key is deleted; this happens for both
    keys in the example. When new keys are added, it goes
    to 1 first because they're added in software; when a
    key is moved to hardware it goes back to 0
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since two years no mac80211 driver implement support for NAPI. Looks
this feature is unneeded, so remove it from generic mac80211 code.
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When setting a monitor interface up or down, the idle state needs to be
recalculated, otherwise the hardware will just stay in its previous idle
state.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When the driver does not want a monitor mode VIF, no channel context is
allocated for it. This causes ieee80211_recalc_idle to put the hardware
into idle mode if only a monitor mode is active, breaking injection.

Fix this by checking local->monitors in addition to active channel
contexts.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If an interface is set down while authenticating or
associating, there's a station entry that will be
removed by the flushing in do_stop() and that will
cause a warning. It's otherwise harmless, but avoid
the warning by calling ieee80211_mgd_stop() first.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since mesh powersaving was added, pending bcast/mcast frames may go out the
CAB queue now.  Unfortunately, the queue was only set up for AP mode, so we
would try to tx on the IEEE80211_INVAL_HW_QUEUE.  Allow cab_queue for mesh
interfaces as well.

Fixes the following warning (or crash without MAC80211_VERBOSE_DEBUG):

WARNING: at net/mac80211/tx.c:1223 __ieee80211_tx+0x162/0x35f [mac80211]()
Modules linked in: mac80211_hwsim mac80211 cfg80211 [...]
Pid: 3085, comm: avahi-daemon Tainted: G        W    3.8.0-rc1+ #377
Call Trace:
 [<ffffffff81045c20>] warn_slowpath_common+0x83/0x9c
 [<ffffffff81045c53>] warn_slowpath_null+0x1a/0x1c
 [<ffffffffa083aef0>] __ieee80211_tx+0x162/0x35f [mac80211]
 [<ffffffffa083cb1d>] ieee80211_tx+0xd3/0xf9 [mac80211]
 [<ffffffffa083cc0f>] ieee80211_xmit+0xcc/0xd5 [mac80211]
 [<ffffffffa083db59>] ieee80211_subif_start_xmit+0xc53/0xcd8 [mac80211]
 [<ffffffff81319acd>] dev_hard_start_xmit+0x259/0x3ce
 [<ffffffff81333d6b>] sch_direct_xmit+0x74/0x17d
 [<ffffffff8131a0b1>] dev_queue_xmit+0x230/0x414
 [<ffffffff8134877a>] ip_finish_output2+0x348/0x3aa
 [<ffffffff81349029>] ip_finish_output+0x6c/0x71
 [<ffffffff81349046>] NF_HOOK_COND.constprop.44+0x18/0x58
 [<ffffffff8134a03a>] ip_mc_output+0x134/0x13c
 [<ffffffff8134835a>] dst_output+0x18/0x1c
 [<ffffffff81349a24>] ip_local_out+0x20/0x24
 [<ffffffff8134a8cf>] ip_send_skb+0x16/0x3c
 [<ffffffff8136bfba>] udp_send_skb+0x254/0x2b9
 [<ffffffff8136c85e>] udp_sendmsg+0x5a8/0x7d4
Signed-off-by: NBob Copeland <bob@cozybit.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When the vlan device is removed, ps->bc_buf processing can no longer
send its frames.
Signed-off-by: NMichael Braun <michael-dev@fami-braun.de>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add command to trigger radar detection in the driver/FW.
Once radar detection is started it should continuously
monitor for radars as long as the channel active.
If radar is detected usermode notified with 'radar
detected' event.

Scanning and remain on channel functionality must be disabled
while doing radar detection/scanning, and vice versa.

Based on original patch by Victor Goldenshtein <victorg@ti.com>
Signed-off-by: NSimon Wunderlich <siwu@hrz.tu-chemnitz.de>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Now that we have channel contexts, idle is (pretty
much) equivalent to not having a channel context.
Change the code to use this relation so that there
no longer is a need for a lot of idle recalculate
calls everywhere.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

There are only a few drivers that use HW scan, and
all of those don't need a non-idle transition before
starting the scan -- some don't even care about idle
at all. Remove the flag and code associated with it.

The only driver that really actually needed this is
wl1251 and it can just do it itself in the hw_scan
callback -- implement that.
Acked-by: NLuciano Coelho <coelho@ti.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Add debugfs driver callbacks so drivers can add
debugfs entries for interfaces. Note that they
_must_ remove the entries again as add/remove in
the driver doesn't correspond to add/remove in
debugfs; the former is up/down while the latter
is netdev create/destroy.
Signed-off-by: NAlexander Bondar <alexander.bondar@intel.com>
Reviewed-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Depending on the driver, having ARP filtering for
some addresses may be possible. Remove the logic
that tracks whether ARP filter is enabled or not
and give the driver the total number of addresses
instead of the length of the list so it can make
its own decision.
Reviewed-by: NLuciano Coelho <coelho@ti.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of checking every time bss_info_changed is called,
assign the pointer once depending on the interface type
and then leave it untouched until the interface type is
changed. This makes the ieee80211_bss_info_change_notify()
now a simple wrapper to call the driver only.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If there are VLANs, stopping an AP is inefficient as it
calls rcu_barrier() once for each interface (the VLANs
and the AP itself). Optimise this by moving rcu_barrier()
out of the station cleanups and calling it only once for
all interfaces combined.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The last fixes re-added the RCU synchronize penalty
on roaming to fix the races. Split up sta_info_flush()
now to get rid of that again, and let managed mode
(and only it) delay the actual destruction.
Tested-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When an interface is brought down it must have been
disconnected (or similar) in all modes other than WDS,
so warn if any stations were removed in other modes.
Tested-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

When all interfaces have been removed, there can't
be any stations left over, so there's no need to
flush again. Remove this, and all code associated
with it, which also simplifies the function.
Tested-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

The maximum MTU shouldn't take the headers into account,
the maximum MSDU size is exactly the maximum MTU.
Signed-off-by: NT Krishna Chaitanya <chaitanyatk@posedge.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Unfortunately, commit b22cfcfc, intended to speed up roaming
by avoiding the synchronize_rcu() broke AP/mesh modes as it moved
some code into that work item that will still call into the driver
at a time where it's no longer expected to handle this: after the
AP or mesh has been stopped.

To fix this problem remove the per-station work struct, maintain a
station cleanup list instead and flush this list when stations are
flushed. To keep this patch smaller for stable, do this when the
stations are flushed (sta_info_flush()). This unfortunately brings
back the original roaming delay; I'll fix that again in a separate
patch.

Also, Ben reported that the original commit could sometimes (with
many interfaces) cause long delays when an interface is set down,
due to blocking on flush_workqueue(). Since we now maintain the
cleanup list, this particular change of the original patch can be
reverted.

Cc: stable@vger.kernel.org [3.7]
Reported-by: NBen Greear <greearb@candelatech.com>
Tested-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Make AP_VLAN type interfaces track the AP master channel
context so they have one assigned for the various lookups.
Don't give them their own refcount etc. since they're just
slaves to the AP master.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If the sdata work is pending while the interface is stopped,
we currently flush it. If it's not running this means waiting
for it to run, which could take a while if the workqueue is
backlogged. However, the work exits right away if it starts
to run while the interface is already stopping. There's no
point in waiting for that, so use cancel_work_sync() instead.
Reported-by: NBen Greear <greearb@candelatech.com>
Tested-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

If a driver registers an address mask we should ensure that no
interface gets an address assigned that isn't covered by the
registered address mask. This prevents invalid configurations
from reaching the device and causing problems.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
[change function flow to reduce indentation, fix locking]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Convert mac80211 (and where necessary, some drivers a
little bit) to the new channel definition struct.

This will allow extending mac80211 for VHT, which is
currently restricted to channel contexts since there
are no drivers using that which makes it easier. As
I also don't care about VHT for drivers not using the
channel context API, I won't convert the previous API
to VHT support.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Even before channel contexts/multi-channel, having a
single global TX power limit was already problematic,
in particular if two managed interfaces connected to
two APs with different power constraints. The channel
context introduction completely broke this though and
in fact I had disabled TX power configuration there
for drivers using channel contexts.

Change everything to track TX power per interface so
that different user settings and different channel
maxima are treated correctly. Also continue tracking
the global TX power though for compatibility with
applications that attempt to configure the wiphy's
TX power globally.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Since cfg80211 will now call the explicit stop_ap
operation when an AP interface goes down, move all
teardown code there and remove it from interface
handling. The only thing that needs to stay is the
code to dev_close() all dependent VLANs.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

This patch prepares mac80211 for a later implementation of mesh or
ad-hoc powersave clients.
The structures related to powersave (buffer, TIM map, counters) are
moved from the AP-specific interface structure to a generic structure
that can be embedded into any interface type.
The functions related to powersave are prepared to allow easy
extension with different interface types. For example with:

+ } else if (sta->sdata->vif.type == NL80211_IFTYPE_MESH_POINT) {
+         ps = &sdata->u.mesh.ps;

Some references to the AP's beacon structure are removed where they
were obviously not used.

The patch compiles without warning and has been briefly tested as AP
interface with one client in PS mode.
Signed-off-by: NMarco Porsch <marco.porsch@etit.tu-chemnitz.de>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

On each channel that the device is operating on, it
may need to listen using one or more chains depending
on the SMPS settings of the interfaces using it. The
previous channel context changes completely removed
this ability (before, it was available as the SMPS
mode).

Add per-context tracking of the required static and
dynamic RX chains and notify the driver on changes.
To achieve this, track the chains and SMPS mode used
on each virtual interface and update the channel
context whenever this changes.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Instead of operating on a single channel only,
use the new channel context infrastructure in
all mac80211 code.

This enables drivers that want to use the new
channel context infrastructure to use multiple
channels, while nothing should change for all
the other drivers that don't support it.

Right now this disables both TX power settings
and spatial multiplexing powersave. Both need
to be re-enabled on a channel context basis.

Additionally, when channel contexts are used
drop the connection when channel switch is
received rather than trying to handle it. This
will have to be improved later.

[With fixes from Eliad and Emmanuel incorporated]
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NEmmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>