wlcore: Allow memory access when the FW crashes

When the no_recovery flag is used, the recovery work will not restart
the FW and the state will not be set to 'on'. To enable post-mortem
analysis, allow memory access in the 'restarting' state.

Also, since the FW might not be operational, don't fail the read/write
operations if elp_wakeup fails.

Reported-by: Arkady Miasnikov <a-miasnikov@ti.com>
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: use dynamic keep-alive template ids

Currently, all the (station) roles use the same
keep-alive template id (0). However, the klv
template ids shouldn't be shared by different
roles.

Implement a simple klv_templates bitmap, and let
each role allocate its own klv template id on
role initialization.

[Arik - remove invalidation of KLV template when getting into "idle".
This is already handled in unjoin]

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: invalidate keep-alive template on disconnection

Previously, invalidation of the keep-alive template was
done when going idle. However, while removing the
idle-handling we didn't move the keep-alive template
invalidation to another place.

This finally resulted in fw error when trying to use
the keep-alive template by another role.

(Note that we still have an error here - each role
should have its unique keep-alive template id, while
currently they all use CMD_TEMPL_KLV_IDX_NULL_DATA (0).
This only works now because we don't support concurrent
connected stations yet)

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx: default to siso40 in 2.4ghz with a single antenna

The driver used siso20 in this case for legacy reasons.

Reported-by: Ido Reis <idor@ti.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: Fix unbalanced interrupts enablement

The interrupt line is enabled by wl12xx_enable_interrupts and
wl18xx_enable_interrupts, but it will not be disabled in all failure
paths. Fix this.

Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: Don't recover during boot

While recursive recovery is avoided during shutdown, a new recovery may
be queued when the FW boots. The recovery work will then try to stop an
already stopped hardware, which will most likely result in a kernel
panic.

Fix this by verifying that wl->state is on before queueing a new
recovery.

Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore/wl18xx: add phy_fw_version_str to debugfs driver_state

add phy_fw_version_str to debugfs driver_state file.
information is taken during boot and stored in wl->chip.phy_fw_ver_str.

Signed-off-by: Yair Shapira <yair.shapira@ti.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: spi: use private max-buf-size limit

Limit SPI transmissions to the wl12xx max buffer size, as only 12xx was
tested with SPI.

This allows us to remove the global aggregation buffer constant.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx: number_of_assembled_ant5 indicates if A band is enabled

Use number_of_assembled_ant5 phy param to indicate if A band is enabled:
if number_of_assembled_ant5 != 0 then it is enabled otherwise it is
disabled.

This aligns with phy implementation that uses this param both to indicate
if band is active and the number of antennas.

This parameter replaces enable_11a module param that was removed.

User-Space applications can use wlconf and/or INI files to disable A band
using this parameter.

Signed-off-by: Yair Shapira <yair.shapira@ti.com>
Signed-off-by: Igal Chernobelsky <igalc@ti.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: allow up to 3 running STA interfaces in combinations

This allows us to have a p2p-management interface (in STA mode),
as we as a group dedicated interface.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: lazy-enable device roles

Enable device roles just before starting it.
This way, a single device role should be enough
for all vifs, as we can't use concurrent device
roles (which require ROC) anyway.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: always use wlvif->role_id for scans

enabled (but not-started) sta role should be good enough
for scanning (both normal and scheduled), so use it
instead of the device_role.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: make debug prints work without dynamic debug

Make debug prints operational when dynamic debug is not defined.
This allows better debugging in production environments.

Change the driver prefix to "wlcore" while were at it.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore/wl18xx/wl12xx: allow up to 3 mac addresses

Allow 3 native mac addresses on 18xx. On 12xx allow 2 native mac
addresses and set the LAA bit to create a third mac address. This
enabled operation with a separate group interface.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: resume() only if sta is associated

mac80211's resume() callback might get called even if
the sta is not associated (but only up). The
resume sequence in this case results in configuring
the wake-up conditions of a non-started role, which
causes fw assertion.

Fix it by bailing out if the STA is not connected
(like we do on suspend()).

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: configure wowlan regardless of wakeup conditions

wowlan filters should be configured in any case in suspend/resume.
This shouldn't be dependent on whether wakeup conditions are the
same for suspend and resume states. Only the FW command to
reconfigure wakeup conditions should be avoided in such a case.

Reported-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eyal Shapira <eyal@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore/wl18xx/wl12xx: aggregation buffer size set

Aggregation buffer size is set separately per 18xx/12xx chip family.
For 18xx aggragation buffer is set to 13 pages to utilize all
the available tx/rx descriptors for aggregation.

[Arik - remove redundant parts from the patch]

Signed-off-by: Igal Chernobelsky <igalc@ti.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx/wl12xx: defines for Tx/Rx descriptors num

Use defines for number of Tx/Rx descriptors.

Signed-off-by: Igal Chernobelsky <igalc@ti.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: tx_flush - optimize flow and force Tx during the flush

Force Tx during the flush if there are packets pending in the driver.
This actually solves a bug where we would get called from the mac80211
wq context, which would prevent tx_work from getting queued, even when
the mutex is unlocked.

Don't stop the queues needlessly if there's nothing to flush. Use a
larger delay when sleeping to give the driver a chance to flush and
avoid cpu busy looping. Re-arrange the loop so the last iteration is
not wasted.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: make Tx flush timings more verbose

Print how much time a flush took. This will help debug the time it takes
to switch between channels.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: allow only the lowest OFDM rate for p2p setup frames

The IEEE80211_TX_CTL_NO_CCK_RATE flag is only set for mgmt packets
transmitted during p2p connection setup. Make sure to use the lowest
OFDM rate guarantee the peer always hears us.
Change the p2p rate policy to contain only the 6mpbs rate to acheive
this effect.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: AP mode - send non-data packets with basic rates

This solves interoperability issues with peer that don't seem to "hear"
management packets transmitted in higher rates. Based on a previous
patch by Igal Chernobelsky.

Cc: Igal Chernobelsky <igalc@ti.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: cancel recovery_work on stop() instead of remove_interface()

recovery_work should be cancelled when stopping the device,
not when removing an interface (this is probably a leftover
from the single-role days)

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx: increase rx_ba_win_size to 32

The new FWs support a bigger BA window.

Signed-off-by: Ido Reis <idor@ti.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: Prevent interaction with HW after recovery is queued

When a function requests to recover, it would normally abort and will
not send any additional commands to the HW. However, other threads may
not be aware of the failure and could try to communicate with the HW
after a recovery was queued, but before the recovery work began.

Fix this by introducing an intermediate state which is set when recovery
is queued, and modify all state checks accordingly.

Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: consider single fw case

When a single fw is being used for both single-role
and multi-role cases (e.g. 18xx), wl->mr_fw_name is
NULL, which results in NULL dereference while trying
to load the multi-role fw.

In this case, always use the single-role fw, and avoid
redundant fw switch by checking for this case in
wl12xx_need_fw_change() as well.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx: update default phy configuration for pg2

default switch configuration set to pg2 chips (rdl 1/2/3/4).
removed hacks for specific board types.
pg1.x boards are now supported only using module params
or specific conf files.

Signed-off-by: Ido Reis <idor@ti.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl12xx: remove duplicated include from main.c

Remove duplicated include.

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wlcore: Declare MODULE_FIRMWARE usage

Declare any firmware that might be used by this driver.
If all drivers declare their firmware usage, then a sufficiently
complete list of firmware files can then be used to pare down
the external linux-firmware package to just the files in actual use.

Cc: Luciano Coelho <coelho@ti.com>
Cc: linux-wireless@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl12xx: use module_platform_driver

Use a module_platform_driver declaration instead of replicating the
init and exit functions in the driver.

Based on the patch for wl18xx by Devendra Naga.

Signed-off-by: Luciano Coelho <luca@coelho.fi>

wl18xx: use module_platform_driver

the driver's init and exit routines can be implemented with the
module_platform_driver, as the init and exit code is same as
that of the module_platform_driver

Signed-off-by: Devendra Naga <devendra.aaru@gmail.com>
Signed-off-by: Luciano Coelho <luca@coelho.fi>

ath5k: add missing breaks in ath5k_hw_set_spur_mitigation_filter

https://bugzilla.kernel.org/show_bug.cgi?id=47681

Reported-by: David Binderman <dcb314@hotmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: fix BT sleep state on chip wakeup

Whenever both WLAN and BT in/out sleep mode, sometimes WLAN
is not able to take back the shared LNA control after resumes
from sleep mode. The idea is that for WLAN to check if BT owns
LNA control and BT is in sleep mode when WLAN just resumes from
sleep mode. If the condition is true, do a BTCOEX_RC_WARM_RESET
for WLAN to take back the control of shared LNA.
Now the issue is the BT sleep value read from MCI register is
overlooked by assigning u32 into u8. Hence the above condition never
be met so that WLAN used to report beacon losses and frequent
connection drops.

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Tested-by: Paul Stewart <pstew@chromium.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Replace open coded interface checking with interface combinations.

Mac80211 has formal infrastructure to specify which interface combinations
are supported. Make use of this facility in favor of open coding it
ourselves.
So far we only have to specify we can support multiple AP interfaces,
no other combinations are supported.

Inspired by an earlier patch from Paul Fertser.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Cc: Paul Fertser <fercerpav@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Deprecate max_sta_intf field of struct rt2x00_ops.

All drivers set this value to 1, so there is no need (currently) to let
drivers set this.
Therefor, remove the field; we can always add it back when it is needed.

Inspired by an earlier patch from Paul Fertser.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Cc: Paul Fertser <fercerpav@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwifiex: potential corruption in mwifiex_update_uap_custom_ie()

ap_custom_ie is a struct mwifiex_ie_list which is quite different and
also larger than struct mwifiex_ie.  It's a difference between 4196
bytes and 262.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Tested-by: Stone Piao <piaoyun@marvell.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Disable ASPM only for AR9285

Currently, ASPM is disabled for all WLAN+BT combo chipsets
when BTCOEX is enabled. This is incorrect since the workaround
is required only for WB195, which is a AR9285+AR3011 combo
solution. Fix this by checking for the HW version when enabling
the workaround.

Cc: stable@vger.kernel.org
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Tested-by: Paul Stewart <pstew@chromium.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Merge branch 'for-john' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

Merge branch 'for-john' of git://git./linux/kernel/git/jberg/mac80211-next

Merge tag 'nfc-next-3.7-1' of git://git./linux/kernel/git/sameo/nfc-3.0

So says Samuel Ortiz <sameo@linux.intel.com>:

This is the first NFC pull request for the 3.7 merge window.

With this one we get:

- HCI and LLC layers separation. We now can support various LLC
  protocols for HCI drivers, SHDLC being one of them. This will be needed as
  we're planning to support raw HCI chipsets that do the SHDLC encapsulation
  in firmware. So for now we have an SHDLC and a NOP LLC layers.

- pn533 command queueing implementation. This simplifies the pn533 locking
  logic and fixes a kernel warning.

- NCI p2p initiator mode implementation.

- Replace custom workqueues with system ones, for HCI and LLCP.

- Raw pn544 driver removal, as scheduled on the features-removal.txt file.

- A few HCI, SHDLC and LLCP fixes.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

cfg80211: Fix regulatory check for 60GHz band frequencies

The current regulatory code on cfg80211 performs a check to
see if a regulatory rule belongs to an IEEE band so that if
a Country IE is received and no rules are specified for a
band (which is allowed by IEEE) those bands are left intact.
The current band check assumes a rule is bound to a band
if the rule's start or end frequency is less than 2 GHz
apart from the center of frequency being inspected.

In order to support 60 GHz for 802.11ad we need to increase
this to account for the channel spacing of 2160 MHz whereby
a channel somewhere in the middle of a regulatory rule may
be more than 2 GHz apart from either the beginning or
end of the frequency rule.

Without a fix for this even though channels 1-3 are allowed world
wide on the rule (57240 - 63720 @ 2160), channel 2 at 60480 MHz
will end up getting disabled given that it is 3240 MHz from
both the frequency rule start and end frequency. Fix this by
using 2 GHz separation assumption for the 2.4 and 5 GHz bands
but for 60 GHz use a 10 GHz separation before assuming a rule
is not part of the band.

Since we have no 802.11ad drivers yet merged this change has
no impact to existing Linux upstream device drivers.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Acked-by: Luis R. Rodriguez <mcgrof@do-not-panic.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

NFC: Fix typo negociating -> negotiating

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Don't handle consequent RSET frames after UA

During processing incoming RSET frame chip, possibly due to
its internal timout, can retrnasmit an another RSET which
is next queued for processing in shdlc layer.

In case when we accept processed RSET skip those remaining on
the rcv queue until chip will send it's first S or I frame.
This will mean the chip completed connection as well.

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Handle RSET in SHDLC_CONNECTING state

As queue_work() does not guarantee immediate execution of sm_work it
can happen in crossover RSET usecase that connect timer will constantly
change the shdlc state from NEGOTIATING to CONNECTING before shdlc has
chance to handle incoming frame.

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Acked-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: xmit from hci ops must return 0 or negative

xmit callback provided by a driver encapsulates upper layers
data and sends it to the hardware. So, HCI does not know the
exact amount of data being sent and thus can't handle partially
sent frames properly.

Therefore, the driver must return 0 for completely sent frame or
negative for failure.

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Acked-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

MAINTAINERS: Add NFC specific mailing list

linux-nfc@lists.01.org is where all the Linux NFC related discussions take
place, and one can also send kernel patches there.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Remove the pn544 raw driver

This was scheduled for 3.6, we're late.

Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Add HCI module description

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Fix LLC registration definitions for ANSI compliance

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Move the pn544_hci dereference below the NULL test

The dereference should be moved below the NULL test.

spatch with a semantic match is used to found this.
(http://coccinelle.lip6.fr/)

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Move the nfcwilink dereference below the NULL test

spatch with a semantic match is used to found this.
(http://coccinelle.lip6.fr/)

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Remove unneeded LLC symbols export

After fixing the LLC Makefile, we no longer need those exports.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Changed HCI and PN544 HCI driver to use the new HCI LLC Core

The previous shdlc HCI driver and its header are removed from the tree.
PN544 now registers directly with HCI and passes the name of the llc it
requires (shdlc).
HCI instantiation now allocates the required llc instance. The llc is
started when the HCI device is brought up.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Add an shdlc llc module to llc core

This is used by HCI drivers such as the one for the pn544 which require
communications between HCI and the chip to use shdlc.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Add a nop (passthrough) llc module to llc core

This is a passthrough llc. It can be used by HCI drivers that don't
need link layer control. HCI will then write directly to the driver, and
driver will deliver incoming frames directly to HCI without any
processing.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Add an LLC Core layer to HCI

The LLC layer manages modules that control the link layer protocol (such
as shdlc) between HCI and an HCI driver. The driver must simply specify
the required llc when it registers with HCI.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Modified hci_transceive to become an asynchronous operation

This enables the completion callback to be called from a different
context, preventing a possible deadlock if the callback resulted in the
invocation of a nested call to the currently locked nfc_dev.
This is also more in line with the im_transceive nfc_ops for NFC Core or
NCI drivers which already behave asynchronously.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Add a public nfc_hci_send_cmd_async method

This method initiates execution of an HCI cmd. Result will be delivered
through an asynchronous callback.

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Changed the HCI cmd execution callback prototype

Make it match the data_exchange_cb_t so that it can be used directly in
the implementation of an asynchronous hci_transceive

Signed-off-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Correct outgoing frame before requeueing

Driver must handle its data added to the frame, so at this point
removeing control field of shdlc frame is enough.

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Acked-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Remove crc generation from shdlc layer

Checksum is specific for a chip spcification and it varies
(in size and type) between different hardware. It should be
handled in the driver then.

Moreover, shdlc spec doesn't mention crc as a part of the frame.

Update pn544_hci driver as well.

Signed-off-by: Waldemar Rymarkiewicz <waldemar.rymarkiewicz@tieto.com>
Acked-by: Eric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Fix possible LLCP memory leak

nfc_llcp_build_tlv() malloced the memory and should be free in
nfc_llcp_build_gb() after used, and the same in the error handling
case, otherwise it will cause memory leak.

spatch with a semantic match is used to found this problem.
(http://coccinelle.lip6.fr/)

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Remove pointless conditional before HCI kfree_skb()

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Set the IRQF_ONESHOT flag from the pn544_hci IRQ handler request

As we don't have a primary handler but only a threaded one, __setup_irq()
ends up failing if we don't set this flag.

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Don't use WQ_MEM_RECLAIM for pn533

NFC driver doesn't sit in memory reclaim path and has no reason to use
WQ_MEM_RECLAIM.  Drop WQ_MEM_RECLAIM from pn533->wq and use
alloc_ordered_workqueue() instead of WQ_UNBOUND w/ max_active == 1.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Use system_nrt_wq instead of custom ones

NFC is using a number of custom ordered workqueues w/ WQ_MEM_RECLAIM.
WQ_MEM_RECLAIM is unnecessary unless NFC is gonna be used as transport
for storage device, and all use cases match one work item to one
ordered workqueue - IOW, there's no actual ordering going on at all
and using system_nrt_wq gives the same behavior.

There's nothing to be gained by using custom workqueues.  Use
system_nrt_wq instead and drop all the custom ones.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Remove repeated code for NULL check

This patch remove the repeated code for checking llcp_sock &
llcp_sock->dev against NULL.

Signed-off-by: Syam Sidhardhan <s.syam@samsung.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Use module_platform_driver macro for nfcwilink.c

Simplify the code by make use of module_platform_driver macro.

Signed-off-by: Syam Sidhardhan <s.syam@samsung.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Implement NCI dep_link_up and dep_link_down

During NFC-DEP target activation, store the remote
general bytes to be used later in dep_link_up.
When dep_link_up is called, activate the NFC-DEP target,
and forward the remote general bytes.
When dep_link_down is called, deactivate the target.

Signed-off-by: Ilan Elias <ilane@ti.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Parse NCI NFC-DEP activation params

Signed-off-by: Ilan Elias <ilane@ti.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Set local general bytes in nci_start_poll

If initiator protocol is NFC-DEP, set the local general bytes
in nci_start_poll.

Signed-off-by: Ilan Elias <ilane@ti.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

NFC: Queue pn533 commands

Instead of returning EBUSY when getting a command while another one is
running, we queue them. Upon completion of the pending command, the next
one is processed.
Besides the fact that it simplifies the pn533 locking scheme, it also
comes with the nice side effect of fixing the following warning:

[   82.274297] =====================================
[   82.274297] [ BUG: bad unlock balance detected! ]
[   82.274298] 3.5.0-rc1+ #1 Not tainted
[   82.274299] -------------------------------------
[   82.274300] kworker/u:1/16 is trying to release lock (&dev->cmd_lock) at:
[   82.274305] [<ffffffff8144f246>] mutex_unlock+0x9/0xb
[   82.274305] but there are no more locks to release!
[   82.274306]
[   82.274306] other info that might help us debug this:
[   82.274306] 2 locks held by kworker/u:1/16:
[   82.274311]  #0:  (pn533){.+.+..}, at: [<ffffffff8103a67d>]
+process_one_work+0x145/0x2e2
[   82.274314]  #1:  ((&dev->cmd_work)){+.+...}, at: [<ffffffff8103a67d>]
+process_one_work+0x145/0x2e2
[   82.274314]
[   82.274314] stack backtrace:
[   82.274315] Pid: 16, comm: kworker/u:1 Not tainted 3.5.0-rc1+ #1
[   82.274315] Call Trace:
[   82.274317]  [<ffffffff8144f246>] ? mutex_unlock+0x9/0xb
[   82.274321]  [<ffffffff81059841>] print_unlock_inbalance_bug+0xda/0xe4
[   82.274323]  [<ffffffff8105c74c>] lock_release_non_nested+0xb2/0x232
[   82.274325]  [<ffffffff8105a61e>] ? mark_held_locks+0x6d/0x95
[   82.274326]  [<ffffffff8144f246>] ? mutex_unlock+0x9/0xb
[   82.274328]  [<ffffffff81451105>] ? _raw_spin_unlock_irqrestore+0x40/0x5c
[   82.274329]  [<ffffffff8144f246>] ? mutex_unlock+0x9/0xb
[   82.274330]  [<ffffffff8105ca42>] lock_release+0x176/0x1ac
[   82.274333]  [<ffffffff8123de14>] ? pn533_send_complete+0xa8/0xa8
[   82.274334]  [<ffffffff8144f1d6>] __mutex_unlock_slowpath+0xb0/0x117
[   82.274336]  [<ffffffff8144f246>] mutex_unlock+0x9/0xb
[   82.274337]  [<ffffffff8123de65>] pn533_wq_cmd_complete+0x51/0x55
[   82.274338]  [<ffffffff8103a6db>] process_one_work+0x1a3/0x2e2
[   82.274340]  [<ffffffff8103a67d>] ? process_one_work+0x145/0x2e2
[   82.274341]  [<ffffffff8103b119>] worker_thread+0xcf/0x153
[   82.274343]  [<ffffffff8103b04a>] ? manage_workers.isra.22+0x16b/0x16b
[   82.274344]  [<ffffffff8103b04a>] ? manage_workers.isra.22+0x16b/0x16b
[   82.274346]  [<ffffffff8103eb11>] kthread+0x95/0x9d
[   82.274348]  [<ffffffff81452ef4>] kernel_thread_helper+0x4/0x10
[   82.274351]  [<ffffffff81046561>] ? finish_task_switch+0x45/0xc3
[   82.274352]  [<ffffffff814514f0>] ? retint_restore_args+0x13/0x13
[   82.274353]  [<ffffffff8103ea7c>] ? __init_kthread_worker+0x55/0x55
[   82.274354]  [<ffffffff81452ef0>] ? gs_change+0x13/0x13

Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>

ath5k: disable HW crypto in management frame

Hardware support for MFP is not available in ath5k. Thus,
this implementation allows the mac80211 stack to do the
actuall crypto operation.

Signed-off-by: Chun-Yeow Yeoh <yeohchunyeow@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

carl9170: connect to 11w protected networks

Previously, it was not possible to connect to
networks which requires 11w to be supported by
the stations.

While the documentation hints that there's some
hardware support for offloading MFP "decryption",
this simple implementation relies on the mac80211
stack to do the actual crypto operations.

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

p54: connect to 11w protected networks

Previously, it was not possible to connect to
networks which requires 11w to be supported by
the stations.

Note:

As all current (and old) firmwares corrupt
incoming, protected management frames, the
decryption offloading needs to be disabled.
This will be done automatically if needed.

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlegacy: use eth_broadcast_addr

Instead of copying from a constant array
(which is still needed for other purposes)

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

ipw2x00: silence GCC warning for unused variable 'dev'

Building the libipw component without CONFIG_LIBIPW_DEBUG set triggers this GCC
warning:
    drivers/net/wireless/ipw2x00/libipw_wx.c:526:21: warning: unused variable 'dev' [-Wunused-variable]

The cause of this warning is that, without CONFIG_LIBIPW_DEBUG set,
LIBIPW_DEBUG_WX compiles away. Fix it by substituting ieee->dev for (its
equivalent) dev.

Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

bcma: handle BCM43227

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Jack <x6719620@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwifiex: block scan request during heavy Tx traffic

Currently scan operation is delayed/aborted based on Tx traffic
consistency. This decision is taken after receiving scan
response of first scan command from FW. But when heavy traffic
is running, we can not even afford to send first scan command
and go off channel for 30msec. We will block scan request in
this case.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwifiex: disconnect the device before entering suspend state

By default, device is disconnected before entering suspend state.
User can keep the connection alive by using module parameter
"disconect_on_suspend=0".

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: get rid of void pointer in struct brcmf_cfg80211_priv

Field 'pub' in struct brcmf_cfg80211_priv was types as void pointer
and filled with parameter passed in brcmf_cfg80211_attach(). This
patch makes the type specific, ie. struct brcmf_pub.

Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: change struct brcmf_cfg80211_priv comments to kernel-doc

Small step to fix structure commenting using kernel-doc syntax.

Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: clear control lock on usb error.

On a usb error the lock bit should be cleared.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmsmac: don't start device when RfKill is engaged

This patch fixes a bug when device is being started
while RfKill switch is engaged, leading to hang
due to partial initialization of hardware.

Tested-by: <dragonn@op.pl>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Piotr Haber <phaber@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: remove unused usb bmac model code.

clean up code.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: fix bug in determining phy bands.

This patch fixes a bug in routine where phy bands are determined.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: remove unused function.

clean up code.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: streamline SDIO read frame routine

SDIO read non-glomming frame routine handles first frame and
follow up frame read separately. But they share a lot of common
code. This patch abstracts a brcmf_sdio_hdparser function and
optimize the code flow for better readability and future
optimization.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: store usb fw images in local linked list.

For suspend/resume it is necessary to store firmware in memory.
In order to support multiple usb dongles at the same time a linked
list of firmwares was created.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: fix debug printout of event data.

Some events result in printing of the buffer when debug is
enabled. This printing was not very efficient. Changed to
macro so it comes out nice and clean without filling log buffer.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: introduce scheduled scan support

This change add support for NL80211 scheduled scan. This may be used to
offload scanning to the device, which may give the host opportunity to
sleep. The newer versions of wpa_supplicant have support for this
functionality.

Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: extend brcmf_term_iscan() to abort e-scan

With the introduction of E-Scan there are two scan mechanisms
in the driver. I-Scan was aborted on suspend and bringing down
the device using brcmf_term_iscan(). The function has been
renamed to brcmf_abort_scanning() and covers e-scan abort as
well.

Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: fix sparse warnings in e-scan related code

With the introduction of e-scan mechanism in brcmfmac
(e756af5 brcmfmac: add e-scan support.) a couple of sparse
warnings were introduced. This patch resolves those.

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwifiex: fix issue in resumed scan operation

When delayed scan operation is resumed, we just add next scan
command in queue but don't wakeup main thread to process the
command. Hence the command is downloaded to firmware only after
waking up the main thread by any other means.

This bug which was introduced after "mwifiex: improve scan delay
logic.." patch is fixed here.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Clean up RFCSR1 programming in rt2800_config_channel_rf3xxx.

Setting of the individual fields of the RF register can be simplified.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Code style cleanup in rt2800lib.c

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: rt2800lib - code cleanup.

Move RT3290 BBP initialization sequence (part) to the right place in the code.

This is just a code style change, no functional changes.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: rt2800 - Fix default vgc values for RT3572

Align with the values used by the RT3572 Ralink vendor driver v2.5.0.0.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Code clean up in rt2800lib.

Make the code in rt2800_get_default_vgc more understandable and
readable, especially for the 5GHz band values.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: Enable WLAN RX diversity for AR9565

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>