This patch fixes 802.11n stability and performance regression we have
since 2.6.35. It boost performance on my 5GHz N-only network from about
5MB/s to 8MB/s. Similar percentage boost can be observed on 2.4 GHz.

These are test results of 5x downloading of approximately 700MB iso
image:

vanilla: 5.27 5.22 4.94 4.47 5.31 ; avr 5.0420 std 0.35110
patched: 8.07 7.95 8.06 7.99 7.96 ; avr 8.0060 std 0.055946

This was achieved with NetworkManager configured to do not perform
periodical scans, by configuring constant BSSID. With periodical scans,
after some time, performance downgrade to unpatched driver level, like
in example below:

patched: 7.40 7.61 4.28 4.37 4.80 avr 5.6920 std 1.6683

However patch still make better here, since similar test on unpatched
driver make link disconnects with below messages after some time:

wlan1: authenticate with 00:23:69:35:d1:3f (try 1)
wlan1: authenticate with 00:23:69:35:d1:3f (try 2)
wlan1: authenticate with 00:23:69:35:d1:3f (try 3)
wlan1: authentication with 00:23:69:35:d1:3f timed out

On 2.6.35 kernel patch helps against connection hangs with messages:

iwlagn 0000:20:00.0: queue 10 stuck 3 time. Fw reload.
iwlagn 0000:20:00.0: On demand firmware reload
iwlagn 0000:20:00.0: Stopping AGG while state not ON or starting

Cc: stable@kernel.org # 2.6.35+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This reverts commit aac6af55.

Conflicts:

	net/mac80211/key.c

That commit has a race that causes a warning, as documented in the thread
here:

	http://marc.info/?l=linux-wireless&m=130717684914101&w=2Signed-off-by: John W. Linville <linville@tuxdriver.com>

Commit 06e8935f adds an IRQ handling
optimization for single-function SDIO cards like this one, but at the
same time exposes a small hardware bug.

During hardware init, an interrupt is generated with (apparently) no
source. Previously, mmc threw this interrupt away, but now (due to the
optimization), the mmc layer passes this onto libertas, before it is ready
(and before it has enabled interrupts), causing a crash.

Work around this hardware bug by registering the IRQ handler later and
making it capable of handling interrupts with no cause. The change that
makes the IRQ handler registration happen later actually eliminates
the spurious interrupt as well.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This partially reverts 1c5cae81, because
the netdev name is copied into sdata->name, which is used for debugging
messages, for example. Otherwise, we get messages like this:

wlan%d: authenticated
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
Cc: Jiri Pirko <jpirko@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Johannes Berg <johannes@sipsolutions.net>
Cc: "John W. Linville" <linville@tuxdriver.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

We use priv->mutex to avoid race conditions between iwl_chswitch_done()
and iwlagn_mac_channel_switch(), when marking channel switch in
progress. But iwl_chswitch_done() can be called in atomic context
from iwl_rx_csa() or with mutex already taken from iwlagn_commit_rxon().

These bugs were introduced by:

commit 79d07325
Author: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Date:   Thu May 6 08:54:11 2010 -0700

    iwlwifi: support channel switch offload in driver

To fix remove mutex from iwl_chswitch_done() and use atomic bitops for
marking channel switch pending.

Also remove iwl2030_hw_channel_switch() since 2000 series adapters are
2.4GHz only devices.

Cc: stable@kernel.org # 2.6.36+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Disable fast channel change by default on AR2413/AR5413 due to
some bug reports (it still works for me but it's better to be safe).
Add a module parameter "fastchanswitch" in case anyone wants to enable
it and play with it.
Signed-off-by: Nick Kossifidis <mickflemm@gmail.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

We were incorrectly executing PCIe specific workarounds on PCI cards.
This resulted in:
Machine check in kernel mode.
Caused by (from SRR1=149030): Transfer error ack signal
Oops: Machine check, sig: 7 [#1]
Reported-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

In some cases we can read wrong temperature value. If after that
temperature value will not be updated to good one, we badly configure
tx power parameters and device is unable to send a data.

Resolves:
https://bugzilla.kernel.org/show_bug.cgi?id=35932

Cc: stable@kernel.org # 2.6.39+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

read_lock() ... read_unlock_bh() is clearly bogus.
This was broken by

commit 23691d75
Author: Gustavo F. Padovan <padovan@profusion.mobi>
Date:   Wed Apr 27 18:26:32 2011 -0300

    Bluetooth: Remove l2cap_sk_list
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This is the same fix as

   commit 84105160
   Author: Matteo Croce <technoboy85@gmail.com>
   Date:   Fri Dec 3 02:25:08 2010 +0100

   The ath9k driver subtracts 3 dBm to the txpower as with two radios the
   signal power is doubled.
   The resulting value is assigned in an u16 which overflows and makes
   the card work at full power.

in two more places. I grepped the ath tree and didn't find any others.

Cc: stable@kernel.org
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Commit 0a35d36d ("cfg80211: Use capability info to detect mesh beacons")
assumed that probe response with both ESS and IBSS bits cleared
means that the frame was sent by a mesh sta.

However, these capabilities are also being used in the p2p_find phase,
and the mesh-validation broke it.

Rename the WLAN_CAPABILITY_IS_MBSS macro, and verify that mesh ies
exist before assuming this frame was sent by a mesh sta.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

For 6150 devices, modify the supported PCI subsystem ID.

Cc: stable@kernel.org
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

zd1211 devices register 'EP 4 OUT' endpoint as Interrupt type on USB 2.0:

      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x04  EP 4 OUT
        bmAttributes            3
          Transfer Type            Interrupt
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               1

However on USB 1.1 endpoint becomes Bulk:

      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x04  EP 4 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0040  1x 64 bytes
        bInterval               0

Commit 37939810 assumed that endpoint is
always interrupt type and changed usb_bulk_msg() calls to usb_interrupt_msg().

Problem here is that usb_bulk_msg() on interrupt endpoint selfcorrects the
call and changes requested pipe to interrupt type (see usb_bulk_msg).
However with usb_interrupt_msg() on bulk endpoint does not correct the
pipe type to bulk, but instead URB is submitted with interrupt type pipe.

So pre-2.6.39 used usb_bulk_msg() and therefore worked with both endpoint
types, however in 2.6.39 usb_interrupt_msg() with bulk endpoint causes
ohci_hcd to fail submitted URB instantly with -ENOSPC and preventing zd1211rw
from working with OHCI.

Fix this by detecting endpoint type and using correct endpoint/pipe types
for URB. Also fix asynchronous zd_usb_iowrite16v_async() to use right
URB type on 'EP 4 OUT'.

Cc: stable@kernel.org
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Fix kernel oops when trying to use passive scheduled scans.  The
reason was that in passive scans there are no SSIDs, so there was a
NULL pointer dereference.

To solve the problem, we now check the number of SSIDs provided in the
sched_scan request and only access the list if there's one or more
(ie. passive scan is not forced).  We also force all the channels to
be passive by adding the IEEE80211_CHAN_PASSIVE_SCAN flag locally
before the checks in the wl1271_scan_get_sched_scan_channels()
function.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Use a different value for DFS dwell time when performing a scheduled
scan.  Previously we were using the same value as for normal passive
scans.  This adds some flexibility between these two different types
of passive scan.

For now we use 150 TUs for DFS channel dwell time.  This may need to
be fine-tuned in the future.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

DFS channels were never getting included in the scheduled scans,
because they always contain the passive flag as well and the call was
asking for DFS and active channels.

Fix this by ignoring the passive flag when collecting DFS channels.
Also, move the DFS channels in the channel list before the 5GHz active
channels (this was implemented in the FW differently than specified).
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

We were comparing bitwise AND results with a boolean, so when the
boolean was set to true, it was not matching as it should.

Fix this by booleanizing the bitwise AND results with !!.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Before this patch, the command sequence number is being set before
lbs_queue_cmd() adds the command to the queue. However, lbs_queue_cmd()
sometimes forces commands to queue-jump (e.g. CMD_802_11_WAKEUP_CONFIRM).
It currently does this without considering that sequence numbers might need
adjusting to keep things running in order.

Fix this by setting the sequence number at a later stage, just before
we're actually submitting the command to the hardware. Also fixes a
possible race where seqnum was being modified outside of the driver
lock.
Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Ingo Molnar noticed that we have this unnecessary ratelimit.h
dependency in linux/net.h, which hid compilation problems from
people doing builds only with CONFIG_NET enabled.

Move this stuff out to a seperate net/net_ratelimit.h file and
include that in the only two places where this thing is needed.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Ingo Molnar <mingo@elte.hu>

Several networking headers were depending upon the implicit
linux/sysctl.h include they get when including linux/net.h

Add explicit includes.
Signed-off-by: David S. Miller <davem@davemloft.net>

This got missed back in 2006 when Jes Sorensen deleted
net/ethernet/sysctl_net_ether.c
Signed-off-by: David S. Miller <davem@davemloft.net>

Several crashes in cleanup_once() were reported in recent kernels.

Commit d6cc1d64 (inetpeer: various changes) added a race in
unlink_from_unused().

One way to avoid taking unused_peers.lock before doing the list_empty()
test is to catch 0->1 refcnt transitions, using full barrier atomic
operations variants (atomic_cmpxchg() and atomic_inc_return()) instead
of previous atomic_inc() and atomic_add_unless() variants.

We then call unlink_from_unused() only for the owner of the 0->1
transition.

Add a new atomic_add_unless_return() static helper

With help from Arun Sharma.

Refs: https://bugzilla.kernel.org/show_bug.cgi?id=32772Reported-by: Arun Sharma <asharma@fb.com>
Reported-by: Maximilian Engelhardt <maxi@daemonizer.de>
Reported-by: Yann Dupont <Yann.Dupont@univ-nantes.fr>
Reported-by: Denys Fedoryshchenko <denys@visp.net.lb>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

It's currently exposed only through /proc which, besides requiring
screen-scraping, doesn't allow userspace to distinguish between two
identical ATM adapters with different ATM indexes.  The ATM device index
is required when using PPPoATM on a system with multiple ATM adapters.
Signed-off-by: Dan Williams <dcbw@redhat.com>
Reviewed-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: David Woodhouse <dwmw2@infradead.org>
Cc: stable@kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

The AR9287 calibration code was not being called because of an
incorrect MAC revision check.
This forced the AR9287 to use the AR9285 initial calibration code and
bypass the AR9287 code entirely.
Signed-off-by: Adrian Chadd <adrian@freebsd.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Commit 79f460ca add a duplicate
linux/slab.h include to net/mac80211/scan.c - remove it.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Acked-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

local->ps_data wasn't cleared on disassociation, which
(in some corner cases) caused reconnections to enter
psm before association completed.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

We make oldconfig every time when a new kernel arrives, but
if we don't have such a device(I guess this is the most common
case for a new device), the default value should be 'n' so
that the kernel size we build doesn't grow up too much quickly.
For anyone who has the device, it is OK for them to turn it on
by themselves.

Cc: "John W. Linville" <linville@tuxdriver.com>
Cc: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rx_status.band is used uninitialized, what disallow to work on 5GHz .

Cc: stable@kernel.org # 2.6.39+
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

When ip_vs was adapted to netns the ftp application was not adapted
in a correct way.
However this is a fix to avoid kernel errors. In the long term another solution
might be chosen.  I.e the ports that the ftp appl, uses should be per netns.
Signed-off-by: Hans Schillstrom <hans.schillstrom@ericsson.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

Stop tx queues before updating rate control to ensure
proper rate selection. Otherwise packets can be transmitted
in 40 Mhz whereas hw is configured in HT20.
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Whenever there is a channel width change from 40 Mhz to 20 Mhz,
the hardware is reconfigured to ht20. Meantime before doing
the rate control updation, the packets are being transmitted are
selected rate with IEEE80211_TX_RC_40_MHZ_WIDTH.

While transmitting ht40 rate packets in ht20 mode is causing
baseband panic with AR9003 based chips.

==== BB update: BB status=0x02001109 ====
ath: ** BB state: wd=1 det=1 rdar=0 rOFDM=1 rCCK=1 tOFDM=0 tCCK=0 agc=2
src=0 **
ath: ** BB WD cntl: cntl1=0xffff0085 cntl2=0x00000004 **
ath: ** BB mode: BB_gen_controls=0x000033c0 **
ath: ** BB busy times: rx_clear=99%, rx_frame=0%, tx_frame=0% **
ath: ==== BB update: done ====

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

While receiving unsupported rate frame rx state machine
gets into a state 0xb and if phy_restart happens in that
state, BB would go hang. If RXSM is in 0xb state after
first bb panic, ensure to disable the phy_restart.
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Resetting hardware helps to recover from baseband
hang/panic for AR9003 based chips.

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Reported-by: Larry Finger <larry.finger@lwfinger.net>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Although a previous fix handles the kernel panics that result from
failure to allocate a new RX buffer, memory fragmentation can be
reduced if the amsdu_8k capability is disabled as new buffers need only
be of O(0), not O(2).
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

To handle amsdu_8k capability, the PCI routine of this driver must
allocate receive buffers of order 2. Under heavy load, this causes
fragmentation of memory. The present code releases the current buffer
before checking to see if a new one is availble. Recovery from
allocation failures is not possible, which results in kernel panics.

The fix is to reorder the code to check that a new buffer can be
allocated before the old one is released. If not possible, the
received frame is dropped and the old one is reused. Without this
change, it is impossible to transfer a 2 GB file without a kernel panic.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>              [2.6.{37,38,39}]
Signed-off-by: John W. Linville <linville@tuxdriver.com>

In both trigger_scan and sched_scan operations, we were checking for
the SSID length before assigning the value correctly.  Since the
memory was just kzalloc'ed, the check was always failing and SSID with
over 32 characters were allowed to go through.

This was causing a buffer overflow when copying the actual SSID to the
proper place.

This bug has been there since 2.6.29-rc4.

Cc: stable@kernel.org
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

While decoding received event packet from firmware, 4 bytes
of interface header are already removed unconditionally.
So for handling event only 4 more bytes needs to be pulled.
This is achieved by changing event header length to 4.

Almost all the events, except BA stream related and AMSDU
aggregation control events, do not have the payload in their
event skb. Such events handling depends only on the event ID.
This event ID is the first four bytes of the event skb, which
is copied to a separate variable before pulling the skb header.
Hence event handling worked only for those events that didn't
have payload in event skb.

This patch fixes the broken event path of the events with
payload in their event skb without harming existing working
event path for the events without payload.
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Kiran Divekar <dkiran@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>