Johannes Berg [Thu, 18 Apr 2013 21:42:19 +0000 (23:42 +0200)]
mac80211: fix P2P-Device management frame RX
There's an issue in receiving broadcast management frames
on P2P Device virtual interfaces, such frames have the RX
flag IEEE80211_RX_RA_MATCH cleared and are thus dropped
in ieee80211_rx_h_mgmt_check(). They should be let through
to make it to ieee80211_rx_h_userspace_mgmt() and then to
userspace.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Wed, 17 Apr 2013 11:43:22 +0000 (13:43 +0200)]
mac80211: optimize minstrel_ewma
Use powers of two in ewma of minstrel.
This changes :
- EWMA_DIV from 100 to 2^7
- EWMA_LEVEL from 75 (/EWMA_DIV=100) to 2^6 + 2^5 (/EWMA_DIV=128)
Note that this changes EWMA_DIV - EWMA_LEVEL from 25 to 2^5 and keeps
EWMA_LEVEL / EWMA_DIV == 0.75.
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Wed, 17 Apr 2013 12:08:26 +0000 (14:08 +0200)]
mac80211: cosmetics for minstrel_debugfs
This changes the minstrel stats ouput from:
rate throughput ewma prob this prob this succ/attempt success attempts
BCD 6 0.0 0.0 0.0 0( 0) 0 0
to:
rate throughput ewma prob this prob this succ/attempt success attempts
BCD 6 0.0 0.0 0.0 0( 0) 0 0
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 17 Apr 2013 09:26:40 +0000 (11:26 +0200)]
mac80211: fix station entry leak/warning while suspending
Since Stanislaw's patches, when suspending while connected,
cfg80211 will disconnect. This causes the AP station to be
removed, which uses call_rcu() to clean up. Due to needing
process context, this queues a work struct on the mac80211
workqueue. This will warn and fail when already suspended,
which can happen if the rcu call doesn't happen quickly.
To fix this, replace the synchronize_net() which is really
just synchronize_rcu_expedited() with rcu_barrier(), which
unlike synchronize_rcu() waits until RCU callback have run
and thus avoids this issue.
In theory, this can even happen without Stanislaw's change
to disconnect on suspend since userspace might disconnect
just before suspending, though then it's unlikely that the
call_rcu() will be delayed long enough.
Cc: stable@vger.kernel.org [3.7+] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Felix Fietkau [Tue, 16 Apr 2013 11:38:43 +0000 (13:38 +0200)]
mac80211: fix CTS protection handling
The rates[0] CTS and RTS flags are only set after rate control has been
called, so minstrel cannot use them to for setting the number of
retries. This patch adds two new flags to explicitly indicate RTS/CTS use.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Felix Fietkau [Tue, 16 Apr 2013 11:38:42 +0000 (13:38 +0200)]
mac80211: fix and optimize MCS mask handling
Currently the code always copies the configured MCS mask (even if it is
set to default), but only uses it if legacy rates were also masked out.
Fix this by adding a flag that tracks whether the configured MCS mask is
set to default or not.
Optimize the code further by storing a pointer to the configured rate
mask in txrc instead of using memcpy.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Mon, 15 Apr 2013 15:09:29 +0000 (17:09 +0200)]
mac80211: VHT off-by-one NSS
The number of VHT spatial streams (NSS) is found in:
- s8 ieee80211_tx_rate.rate.idx[6:4] (tx - filled by rate control)
- u8 ieee80211_rx_status.vht_nss (rx - filled by driver)
Tx discriminates valid rates indexes with the sign bit and encodes NSS
starting from 0 to 7 (note this matches some hw encodings e.g IWLMVM).
Rx does not have the same constraints, and encodes NSS starting from 1
to 8 (note this matches what wireshark expects in the radiotap header).
To handle ieee80211_tx_rate.rate.idx[6:4] ieee80211_rate_set_vht() and
ieee80211_rate_get_vht_nss() assume their nss parameter and return value
respectively runs from 0 to 7.
ATM, there are only 2 users of these: cfg.c:sta_set_rate_info_t() and
iwlwifi/mvm/tx.c:iwl_mvm_hwrate_to_tx_control(), but both assume nss
runs from 1 to 8.
This patch fixes this inconsistency by making ieee80211_rate_set_vht()
and ieee80211_rate_get_vht_nss() handle an nss running from 1 to 8.
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Mon, 15 Apr 2013 17:04:06 +0000 (19:04 +0200)]
mac80211: adjust initial chandefs assignments in ieee80211_register_hw
I noticed that monitor interfaces by default would start on 5GHz
while STA/AP ones would start 2GHZ - It stems from the fact that
ieee80211_register_hw unnecessarily adjusts the local->monitor_chandef
for each band.
This avoids this and while at it uses a single dflt_chandef to initialize
in one go local->{hw.conf.chandef,_oper_chandef,monitor_chandef}
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Mon, 15 Apr 2013 16:28:21 +0000 (18:28 +0200)]
mac80211: fix rate control tx handler for VHT rates
Handle VHT rates like HT ones, otherwise we easily trigger the pre-HT
rates WARN_ON(rc_rate->idx >= sband->n_bitrates) which will set
rc_rate->idx to -1.
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
mac80211: remove warning from ieee80211_beacon_loss
Currently, mac80211 assumes that connection monitor offload
for BSS station implies that the device:
- sends periodic keep alive packets to associated AP
- monitors missed beacons
- actively probes the AP in case of missed beacons
In case of poor connection conditions it expects the function
ieee80211_connection_loss() to be called by driver. However,
some devices implement connection monitor offload excluding
active AP probing.
To allow them to call ieee80211_beacon_loss() cleanly, remove
the warning there and thus allow them to use mac80211 for the
AP probing even if connection monitor offload is supported.
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 28 Mar 2013 09:44:18 +0000 (10:44 +0100)]
mac80211: handle wide bandwidth channel switch
Parse and react to the wide bandwidth channel switch element
in beacons/action frames. Finding the element was done in a
previous patch (it has different positions in beacons/action
frames), now handle it. If there's something wrong with it
simply disconnect.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Tue, 26 Mar 2013 13:54:16 +0000 (14:54 +0100)]
mac80211: parse VHT channel switch IEs
VHT introduces multiple IEs that need to be parsed for a
wide bandwidth channel switch. Two are (currently) needed
in mac80211:
* wide bandwidth channel switch element
* channel switch wrapper element
The former is contained in the latter for beacons and probe
responses, but not for the spectrum management action frames
so the IE parser needs a new argument to differentiate them.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Handle the (public) extended channel switch announcement
action frames. Parts of the data in these frames isn't
really in IEs, but put it into the elems struct anyway
to simplify the handling.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Mon, 25 Mar 2013 17:29:27 +0000 (18:29 +0100)]
mac80211: support secondary channel offset in CSA
Add support for the secondary channel offset IE in channel
switch announcements. This is necessary for proper handling
of CSA on HT access points.
For this to work it is also necessary to convert everything
here to use chandef structs instead of just channels. The
driver updates aren't really correct though. In particular,
the TI wl18xx driver update can't possibly be right since
it just ignores the new channel width for lack of firmware
API.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Tue, 26 Mar 2013 13:13:58 +0000 (14:13 +0100)]
mac80211: support extended channel switch
Support extended channel switch when the operating
class is one of the global operating classes as
defined in Annex E of 802.11-2012. If it isn't,
disconnect from the AP instead.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
CSA action frame content should be processed as variable IEs
rather than fixed to make it extensible. Unify the code and
process them just like CSA in beacons to make it easier to
extend for HT/VHT.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When a full scan 2.4 and 5 GHz scan is scheduled, but then the 2.4 GHz
part of the scan disables a 5.2 GHz channel due to, e.g. receiving
country or frequency information, that 5.2 GHz channel might already
be in the list of channels to scan next. Then, when the driver checks
if it should do a passive scan, that will return false and attempt an
active scan. This is not only wrong but can also lead to the iwlwifi
device firmware crashing since it checks regulatory as well.
Fix this by not setting the channel flags to just disabled but rather
OR'ing in the disabled flag. That way, even if the race happens, the
channel will be scanned passively which is still (mostly) correct.
Cc: stable@vger.kernel.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Fri, 5 Apr 2013 10:06:24 +0000 (12:06 +0200)]
mac80211: always pick a basic rate to tx RTS/CTS for pre-HT rates
When the 1st rate control entry is a pre-HT rate we want to set
rts_cts_rate_idx "as the fastest basic rate that is not faster than the
data rate"(code comments).
But in case some bss allowed rate indexes are lower than the lowest bss
basic rate, if the rate control selects a rate among the formers for its
1st rate control entry, rts_cts_rate_idx remains 0 and is not a basic
rate index.
This commit sets rts_cts_rate_idx to the lowest bss basic rate index in
this situation.
Note that the code assumes that lowest indexes == lowest bitrates.
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Marek Puzyniak [Wed, 10 Apr 2013 11:19:13 +0000 (13:19 +0200)]
mac80211: provide SSID in IBSS mode
Some drivers need SSID in AP and IBSS mode. AP SSID is provided
through BSS_CHANGED_SSID notification. There was no easy way to
do the same for IBSS. In IBSS mode SSID is known but was not
stored in BSS configuration. Extend the AP-mode functionality
to also work in IBSS mode.
Signed-off-by: Marek Puzyniak <marek.puzyniak@tieto.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Sun, 7 Apr 2013 17:32:47 +0000 (19:32 +0200)]
mac80211: always advertise STBC/MCSes even if no AP support
Advertise STBC capabilities and MCS rates even if the AP
doesn't support them. This has always been the right thing
to do, but used to be problematic with some APs. Now WFA
testing requires this so re-enable it, problematic APs
would then presumably not pass the test and be fixed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Marek Puzyniak [Wed, 10 Apr 2013 11:47:45 +0000 (13:47 +0200)]
mac80211: clear SSID when stopping AP
When AP interface is stopped ssid_len in the BSS configuration
isn't cleared which can confuse drivers when switching modes.
Set the length to zero when stopping the AP interface.
Signed-off-by: Marek Puzyniak <marek.puzyniak@tieto.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Martin Pitt [Mon, 8 Apr 2013 09:30:01 +0000 (11:30 +0200)]
mac80211_hwsim: Register and bind to driver
Properly register our mac80211_hwsim_driver, attach it to the platform bus.
Bind newly created hwsim devices to that driver, so that our wlan devices get
a proper "driver" sysfs attribute.
This makes mac80211_hwsim interfaces work with NetworkManager.
Signed-off-by: Martin Pitt <martin.pitt@ubuntu.com>
[fix an old and a new message to not be line-broken,
also fix the driver_register error path] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Tue, 26 Mar 2013 21:23:20 +0000 (22:23 +0100)]
mac80211: don't start new netdev queues if driver stopped
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: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Tue, 26 Mar 2013 21:02:42 +0000 (22:02 +0100)]
mac80211: don't fiddle with netdev queues in MLME code
The netdev queues should always represent the state that
the driver gave them, so fiddling with them isn't really
appropriate in the mlme code. Also, since we stop queues
for flushing now, this really isn't necessary any more.
As the scan/offchannel code has also been modified to no
longer do this a while ago, remove the outdated smp_mb()
and comments about it.
While at it, also add a pair of braces that was missing.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
It's unlikely that an AP requires WMM mandatory admission control
for all access categories, and if it does then we still transmit
on the background AC without requesting admission. However, avoid
using uAPSD in this case since the implementation could run into
issues and might use other ACs etc.
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 27 Mar 2013 22:24:53 +0000 (23:24 +0100)]
mac80211: fix do_stop handling while suspended
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: Stanislaw Gruszka <sgruszka@redhat.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 27 Mar 2013 22:20:27 +0000 (23:20 +0100)]
mac80211: destroy virtual monitor interface across suspend
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: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 27 Mar 2013 21:49:19 +0000 (22:49 +0100)]
mac80211: purge remain-on-channel items when suspending
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: Johannes Berg <johannes.berg@intel.com>
Bob Copeland [Fri, 29 Mar 2013 13:38:39 +0000 (09:38 -0400)]
mac80211: return new mpath from mesh_path_add()
Most times that mesh_path_add() is called, it is followed by
a lookup to get the just-added mpath. We can instead just
return the new mpath in the case that we allocated one (or the
existing one if already there), so do that. Also, reorder the
code in mesh_path_add a bit so that we don't need to allocate
in the pre-existing case.
Signed-off-by: Bob Copeland <bob@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The mesh hwmp debug message is a bit confusing. The "sending PREP
to %p" should be the MAC address of mesh STA that has originated
the PREQ message and the "received PREP from %pM" should be the MAC
address of the mesh STA that has originated the PREP message.
Signed-off-by: Chun-Yeow Yeoh <yeohchunyeow@gmail.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Karl Beldan [Mon, 25 Mar 2013 15:26:57 +0000 (16:26 +0100)]
mac80211: Use a cfg80211_chan_def in ieee80211_hw_conf_chan
Drivers that don't use chanctxes cannot perform VHT association because
they still use a "backward compatibility" pair of {ieee80211_channel,
nl80211_channel_type} in ieee80211_conf and ieee80211_local.
Signed-off-by: Karl Beldan <karl.beldan@rivierawaves.com>
[fix kernel-doc] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Alexander Bondar [Tue, 22 Jan 2013 14:52:23 +0000 (16:52 +0200)]
mac80211: allow drivers to set default uAPSD parameters
mac80211 currently sets uAPSD parameters to have VO AC trigger-
and delivery-enabled, with maximum service period length.
Allow drivers to change these default settings since different
uAPSD client implementations may handle errors differently and
be able to recover from some errors.
Note: some APs may not function correctly if one or all ACs are
trigger- and delivery-enabled, see
http://thread.gmane.org/gmane.linux.kernel.wireless.general/93577.
We retested with this AP and later firmware doesn't have this
bug any more.
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 7 Mar 2013 19:54:29 +0000 (20:54 +0100)]
mac80211: ibss: use beacon_data struct for beacon and probe response
Instead of having an SKB all the time, use a beacon_data struct
with just the information required. This also allows removing a
synchronize_rcu() and using kfree_rcu() instead.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Ben Greear [Tue, 19 Mar 2013 21:19:56 +0000 (14:19 -0700)]
mac80211: make beacon-loss-count configurable
On loaded systems with lots of VIFs, I see lots of beacon
timeouts, even though the connection to the AP is very
good. Allow tuning the beacon-loss-count variable to
give the system longer to process beacons if the user
prefers.
Signed-off-by: Ben Greear <greearb@candelatech.com>
[add the number of beacons to the message] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
For the sake of speed of calculation and number accuracy,
mac80211 tracks the RSSI in dBm * 16. But it forgot to
divide back by 16 when the RSSI is asked by the driver.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 20 Mar 2013 13:05:52 +0000 (14:05 +0100)]
cfg80211: fix potential connection work crash
If wpa_supplicant and iw/iwconfig are used together, very
rarely the system crashes. It seems to be related to the
connection parameters not being set up, but it's not all
clear to me how this happens. In any case, checking that
the conn pointer exists here is probably a good idea.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 13 Feb 2013 11:25:28 +0000 (12:25 +0100)]
mac80211: stop queues temporarily for flushing
Sometimes queues are flushed in the middle of
operation, which can lead to driver issues.
Stop queues temporarily, while flushing, to
avoid transmitting new packets while they are
being flushed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 13 Feb 2013 11:11:00 +0000 (12:11 +0100)]
mac80211: pass queue bitmap to flush operation
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: Johannes Berg <johannes.berg@intel.com>
Felix Fietkau [Sat, 16 Mar 2013 16:00:27 +0000 (17:00 +0100)]
mac80211/minstrel_ht: do not sample actively used rates
max_tp_rate2 and max_prob_rate tend to get used occasionally during
retransmission, which is more useful for the statistics than probing
with individual probe packets.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Felix Fietkau [Sat, 16 Mar 2013 16:00:26 +0000 (17:00 +0100)]
mac80211/minstrel_ht: avoid useless sampling of high-probability slower rates
Slow rates that have >95% success probability do not need to be
monitored continuously. When the channel conditions change rapidly, the
slow sampling results are useless anyway. When conditions change slowly,
they will be monitored by gradual downgrading of the actively used
rates. This patch slightly improves throughput under good conditions.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Under load, otherwise stable rates can easily fluctuate because of
collisions. In my tests on a clean channel, the success probability of
the max throughput rate often stays somewhere between 90% and 100% under
load. This can cause some unnecessary switching to lower rates.
This patch improves stability by treating success probability values
between 90% and 100% the same.
In my tests on a 3x3 HT20 link with lots of TCP traffic, it improves the
average throughput by a few mbit/s.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
There is need create driver own per interface debugfs files. This is
currently done by drv_{add,remove}_interface_debugfs() callbacks. But it
is possible that after we remove interface from the driver (i.e.
on suspend) we call drv_remove_interface_debugfs() function. Fixing this
problem will require to add call drv_{add,remove}_interface_debugfs()
anytime we create and remove interface in mac80211. So it's better to
add debugfs dir dentry to vif structure to allow to create/remove
custom debugfs driver files on drv_{add,remove}_interface().
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Alexander Bondar [Mon, 11 Feb 2013 12:56:29 +0000 (14:56 +0200)]
mac80211: add driver callback for per-interface multicast filter
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: Alexander Bondar <alexander.bondar@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 6 Mar 2013 22:09:11 +0000 (23:09 +0100)]
mac80211: batch key free synchronize_net()
Instead of calling synchronize_net() for every key
on an interface or when a station is removed, do it
only once for all keys in both of these cases.
As a side-effect, removing station keys now always
calls synchronize_net() even if there are no keys,
which fixes an issue with station removal happening
in the driver while the station could still be used
for TX.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 6 Mar 2013 21:53:52 +0000 (22:53 +0100)]
mac80211: clean up key freeing a bit
When a key is allocated but not really added, there's no
need to go through the entire teardown process. Also, if
adding a key fails, ieee80211_key_link() can take care of
freeing it instead of the (only) caller.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Ilan Peer [Tue, 5 Mar 2013 13:27:20 +0000 (15:27 +0200)]
mac80211: Call drv_set_tim only if there is a change
It is possible that sta_info_recalc_tim() is called consecutively
without changing the station's tim bit. In such cases there is no
need to call the driver's set_tim() callback.
Signed-off-by: Ilan Peer <ilan.peer@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 15 Feb 2013 20:38:08 +0000 (21:38 +0100)]
mac80211: provide ieee80211_sta_eosp()
The irqsafe version ieee80211_sta_eosp_irqsafe() exists, but
drivers must not mix calls to any irqsafe/non-irqsafe function.
Both ath9k and iwlwifi, the likely first users of this interface,
use non-irqsafe RX/TX/TX status so must also use a non-irqsafe
version of this function. Since no driver uses the _irqsafe()
version, remove that.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Tue, 5 Feb 2013 09:55:21 +0000 (10:55 +0100)]
mac80211: provide race-free 64-bit traffic counters
Make the TX bytes/packets counters race-free by keeping
them per AC so concurrent TX on queues can't cause lost
or wrong updates. This works since each station belongs
to a single interface. While at it also make the bytes
counters 64-bit.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Pedersen [Mon, 4 Mar 2013 21:06:11 +0000 (13:06 -0800)]
cfg80211: rename mesh station types
The mesh station types used to refer to whether the
station was secure or nonsecure. Really the salient
information is whether it is managed by the kernel or
userspace
Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Pedersen [Mon, 4 Mar 2013 21:06:10 +0000 (13:06 -0800)]
nl80211: explicit userspace MPM
Secure mesh had the implicit requirement that the Mesh
Peering Management entity be in userspace. However
userspace might want to implement an open MPM as well, so
specify a mesh setup parameter to indicate this.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:07 +0000 (23:30 +0100)]
mac80211: improve minstrels rate sorting by means of throughput & probability
This patch improves the way minstrel sorts rates according to throughput
and success probability. 3 FOR-loops across the entire rate set in function
minstrel_update_stats() which where used to determine the fastest, second
fastest and most robust rate are reduced to 1 FOR-loop.
The sorted list of rates according throughput is extended to the best four
rates as we need them in upcoming joint rate and power control. The sorting
is done via the new function minstrel_sort_best_tp_rates().
The most robust rate selection is aligned with minstrel_ht's approach.
Once any success probability is above 95% the one with the highest
throughput is chosen as most robust rate. If success probabilities of all
rates are below 95%, the rate with the highest succ. prob. is elected as
most robust one
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:06 +0000 (23:30 +0100)]
mac80211: treat minstrel success probabilities below 10% as implausible
Based on minstrel_ht this patch treats success probabilities below 10% as
implausible values for throughput calculation in minstrel's statistics.
Current throughput per rate with such a low success probability is reset
to 0 MBit/s.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:05 +0000 (23:30 +0100)]
mac80211: add lowest rate into minstrel's random rate sampling table
While minstrel bootstraps and fills the success probabilities of each
rate the lowest rate has typically a very high success probability
(often 100% in our tests).
Its statistics are never updated but considered to setup the mrr chain.
In our tests we see that especially the 3rd mrr stage (which is that
rate providing highest success probability) is filled with the lowest rate
because its initial high sucess probability is never updated. By design
the 4th mrr stage is filled with the lowest rate so often 3rd and 4th
mrr stage are equal.
This patch follows minstrels general approach of assuming as little
as possible about rate dependencies. Consequently we include the
lowest rate into the random sampling table to get balanced up-to-date
statistics of all rates and therefore balanced decisions.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:04 +0000 (23:30 +0100)]
mac80211: extend minstrel's rate sampling to avoid unsampled rates
Minstrel's decision which rate should be directly sampled within the
1st mrr stage is limited to such rates faster than the current max
throughput rate. All rates below the current max. throughput rate
are indirectly sampled via the 2nd mrr stage.
This approach leads to deprecated per rate statistics and therfore
a deprecated mrr chain setup.
This patch uses the sampling approach from minstrel_ht. A counter is
added to sum all indirect sample attempts per rate. After 20 indirect
sampling attempts the rate is directly sampled within the 1st mrr stage.
Therefore more up-to-date statistics for all rates are maintained and
used to setup the mrr chain.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:03 +0000 (23:30 +0100)]
mac80211: add documentation and verbose variable names in
Add documentation and more verbose variable names to minstrel's
multi-rate-retry setup within function minstrel_get_rate() to
increase the readability of the algorithm.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:02 +0000 (23:30 +0100)]
mac80211: merge value scaling macros of minstrel_ht and minstrel
Both minstrel versions use individual ways to scale up integer values
to perform calculations. Merge minstrel_ht's scaling macros into
minstrels header file and use them in both minstrel versions.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Thomas Huehn [Mon, 4 Mar 2013 22:30:01 +0000 (23:30 +0100)]
mac80211: merge EWMA calculation of minstrel_ht and minstrel
Both rate control algorithms (minstrel and minstrel_ht) calculate
averages based on EWMA. Shift function minstrel_ewma() into
rc80211_minstrel.h and make use of it in both minstrel version.
Also shift the default EWMA level (75%) definition to the header file
and clean up variable usage.
Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The last minstrel_ht changes increased the sampling frequency for
potentially useful rates to decrease the response time to rate
fluctuations. This caused an increase in sampling frequency that can
slightly reduce throughput, so this patch limits the sampling attempts
to one per rate instead of two.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Mon, 3 Dec 2012 21:09:22 +0000 (22:09 +0100)]
regulatory: allow VHT channels in world roaming
For VHT, the wider bandwidths (up to 160 MHz) need
to be allowed. Since world roaming only covers the
case of connecting to an AP, it can be opened up
there, we will rely on the AP to know the local
regulations.
Acked-by: Luis R. Rodriguez <mcgrof@do-not-panic.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 1 Mar 2013 12:07:48 +0000 (13:07 +0100)]
mac80211: restrict peer's VHT capabilities to own
Implement restricting peer VHT capabilities to the device's own
capabilities. This is useful when a single driver supports more
than one device and the devices have different capabilities
(often they will differ in the number of spatial streams), but
in particular is also necessary for VHT capability overrides to
work correctly -- otherwise it'd be possible to e.g. advertise,
due to overrides, that TX-STBC is not supported, but then still
use it to TX to the AP because it supports RX-STBC.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 1 Mar 2013 10:54:43 +0000 (11:54 +0100)]
mac80211: fix HT capability overrides for AP station
HT capabilites are asymmetric -- e.g. beamforming is both an
RX and TX capability. If, for example, we support RX but not
TX, the RX capability of the AP station is masked out (if it
supports it). This works correctly if it's really the driver
capability.
If, on the other hand, the reason for not supporting TX BF
is that it was removed by HT capability overrides then the
wrong thing happens: the AP's TX capability will be removed
rather than its RX capability, because the override function
works on own capabilities, not remote ones, and doesn't take
the asymmetry into account.
To fix this make a copy of our own capabilities, apply the
overrides to them (where needed) and then use that to set up
the peer's capabilities.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 1 Mar 2013 10:43:30 +0000 (11:43 +0100)]
mac80211: don't apply HT overrides to TDLS peers
The HT overrides are intended only for the connection
to the AP, not for any other purpose. Therefore, don't
apply them to TDLS peers that are also stations added
to a managed station interface.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Sat, 23 Feb 2013 00:17:56 +0000 (01:17 +0100)]
mac80211: simplify AP interface stop
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: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Sat, 23 Feb 2013 00:14:20 +0000 (01:14 +0100)]
mac80211: flush keys when stopping AP
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: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 22 Feb 2013 23:59:03 +0000 (00:59 +0100)]
mac80211: defer tailroom counter manipulation when roaming
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: Johannes Berg <johannes.berg@intel.com>
projects / linux-beck.git / log