Bluetooth: Add new BT_SNDMTU and BT_RCVMTU socket options

This patch adds new socket options for LE sockets since the existing
L2CAP_OPTIONS socket option is not usable for LE. For now, the new
socket options also require LE CoC support to be explicitly enabled to
leave some playroom in case something needs to be changed in a backwards
incompatible way.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Limit L2CAP_OPTIONS socket option usage with LE

Most of the values in L2CAP_OPTIONS are not applicable for LE and those
that are have different semantics. It makes therefore sense to
completely block this socket option for LE and add (in a separate patch)
a new socket option for tweaking the values that do make sense (mainly
the send and receive MTU). Legacy user space ATT code still depends on
getsockopt for L2CAP_OPTIONS though so we need to plug a hole for that
for backwards compatibility.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Track LE L2CAP credits in l2cap_chan

This patch adds tracking of L2CAP connection oriented channel local and
remote credits to struct l2cap_chan and ensures that connect requests
and responses contain the right values.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add LE L2CAP flow control mode

The LE connection oriented channels have their own mode with its own
data transfer rules. In order to implement this properly we need to
distinguish L2CAP channels operating in this mode from other modes.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Make l2cap_le_sig_cmd logic consistent

This patch makes the error handling and return logic of l2cap_le_sig_cmd
consistent with its BR/EDR counterpart.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add L2CAP Disconnect suppport for LE

The normal L2CAP Disconnect request and response are also used for LE
connection oriented channels. Therefore, we can simply use the existing
handler functions for terminating LE based L2CAP channels.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Fix L2CAP channel closing for LE connections

Sending of the L2CAP Disconnect request should also be performed for LE
based channels. The proper thing to do is therefore to look at whether
there's a PSM specified for the channel instead of looking at the link
type.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add basic LE L2CAP connect request receiving support

This patch adds the necessary boiler plate code to handle receiving
L2CAP connect requests over LE and respond to them with a proper connect
response.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Refactor L2CAP connect rejection to its own function

We'll need to have a separate code path for LE based connection
rejection so it's cleaner to move out the response construction code
into its own function (and later a second one for LE).

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add smp_sufficient_security helper function

This function is needed both by the smp_conn_security function as well
as upcoming code to check for the security requirements when receiving
an L2CAP connect request over LE.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add initial code for LE L2CAP Connect Request

This patch adds the necessary code to send an LE L2CAP Connect Request
and handle its response when user space has provided us with an LE
socket with a PSM instead of a fixed CID.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add definitions for LE connection oriented channels

This patch adds the necessary defines and structs for LE connection
oriented channels.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Move LE L2CAP initiator procedure to its own function

Once connection oriented L2CAP channels over LE are supported they will
need a completely separate handling from BR/EDR channels.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Pass command length to LE signaling channel handlers

The LE signaling PDU length is already calculated in the
l2cap_le_sig_channel function so we can just pass the value to the
various handler functions to avoid unnecessary recalculations (byte
order conversions). Right now the only user is the connection parameter
update procedure, but as new LE signaling operations become available
(for connection oriented channels) they will also be able to make use of
the value.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Allow l2cap_chan_check_security() to be used for LE links.

With connection oriented L2CAP channels some code paths will be shared
with BR/EDR links. It is therefore necessary to allow the
l2cap_chan_check_security function to be usable also for LE links in
addition to BR/EDR ones. This means that smp_conn_security() needs to be
called instead of hci_conn_security() in the case of an LE link.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Update l2cap_global_chan_by_psm() to take a link type

Once connection oriented L2CAP channels become possible for LE we need
to be able to specify the link type we're interested in when looking up
L2CAP channels. Therefore, add a link_type parameter to the
l2cap_global_chan_by_psm() function which gets compared to the address
type associated with each l2cap_chan.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add module parameter to enable LE CoC support

Along with the L2CAP Connection Oriented Channels features it is now
allowed to use both custom fixed CIDs as well as PSM based (connection
oriented connections). Since the support for this (with the subsequent
patches) is still on an experimental stage, add a module parameter to
enable it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Remove unnecessary braces from one-line if-statement

This patch is just a trivial coding style fix to remove unnecessary
braces from a one-line if-statement.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Remove dead code from SMP encryption function

The AES cipher is used in ECB mode by SMP and therefore doesn't use an
IV (crypto_blkcipher_ivsize returns 0) so the code trying to set the IV
was never getting called. Simply remove this code to avoid anyone from
thinking it actually makes some difference.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Remove useless smp_rand function

This function was always just making a single get_random_bytes() call
and always returning the value 0. It's simpler to just call
get_random_bytes() directly where needed.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Enable autosuspend for Intel Bluetooth device

This patch enables autosuspend for Intel Bluetooth device.

After btusb is loaded for Intel Bluetooth device, the power/control
attribute contains "on" value by default which disables the autosuspend.
Based on the USB PM document(Documentation/usb/power-management.txt),
kernel disabled the autosuspend for all devices other than hub by default.

"The USB specification states that all USB devices must support power
management.  Nevertheless, the sad fact is that many devices do not
support it very well.  You can suspend them all right, but when you
try to resume them they disconnect themselves from the USB bus or
they stop working entirely.  This seems to be especially prevalent
among printers and scanners, but plenty of other types of device have
the same deficiency.

For this reason, by default the kernel disables autosuspend (the
power/control attribute is initialized to "on") for all devices other
than hubs.  Hubs, at least, appear to be reasonably well-behaved in
this regard."

This document also described how the driver can enables the autosuspend
by using an USB api.

"Drivers can enable autosuspend for their devices by calling

usb_enable_autosuspend(struct usb_device *udev);

in their probe() routine, if they know that the device is capable of
suspending and resuming correctly.  This is exactly equivalent to
writing "auto" to the device's power/control attribute."

For Intel Bluetooth device, the autosuspend needs to be enabled so the
device can transit to LPM(Low Power Mode) and ULPM(Ultra LPM) states after
receiving suspend message from the host.

Signed-off-by: Tedd Ho-Jeong An <tedd.an@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add support for Intel Bluetooth device [8087:0a2a]

This patch adds support for new Intel Bluetooth device.

T:  Bus=02 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  4 Spd=12   MxCh= 0
D:  Ver= 2.01 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=8087 ProdID=0a2a Rev= 0.01
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms

Signed-off-by: Tedd Ho-Jeong An <tedd.an@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Refactor hci_disconn_complete_evt

hci_disconn_complete_evt() logic is more complicated than what it
should be, making it hard to follow and add new features.

So this patch does some code refactoring by handling the error cases
in the beginning of the function and by moving the main flow into the
first level of function scope. No change is done in the event handling
logic itself.

Besides organizing this messy code, this patch makes easier to add
code for handling LE auto connection (which will be added in a further
patch).

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: Remove unneeded check in hci_disconn_complete_evt()

According to b644ba336 (patch that introduced HCI_CONN_MGMT_CONNECTED
flag), the HCI_CONN_MGMT_CONNECTED flag tracks when mgmt has been
notified about the connection.

That being said, there is no point in calling mgmt_disconnect_failed()
conditionally based on this flag. mgmt_disconnect_failed() removes
pending MGMT_OP_DISCONNECT commands, it doesn't matter if that
connection was notified or not.

Moreover, if the Disconnection Complete event has status then we have
nothing else to do but call mgmt_disconnect_failed() and return.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: ath3k: Add support for a new AR3012 device

T:  Bus=02 Lev=01 Prnt=01 Port=04 Cnt=01 Dev#=  9 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=0489 ProdID=e05f Rev= 0.02
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 2 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb

Reported-by: Joshua Richenhagen <richenhagen@gmail.com>
Signed-off-by: Sujith Manoharan <sujith@msujith.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: ath3k: Add support for another AR3012 card

T:  Bus=03 Lev=01 Prnt=01 Port=02 Cnt=01 Dev#=  2 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=04ca ProdID=300b Rev= 0.01
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 2 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb

Reported-by: Face <falazemi@gmail.com>
Signed-off-by: Sujith Manoharan <sujith@msujith.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Remove unnecessary 'send' parameter from smp_failure()

The send parameter has only been used for determining whether to send a
Pairing Failed PDU or not. However, the function can equally well use
the already existing reason parameter to make this choice and send the
PDU whenever a non-zero value was passed.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Remove link type check in hci_disconn_complete_evt()

We can safely remove the link type check from hci_disconn_complete_
evt() since this check in not required for mgmt_disconnect_failed()
and mgmt_device_disconnected() does it internally.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Add an extra check in mgmt_device_disconnected()

This patch adds an extra check in mgmt_device_disconnected() so we only
send the "Device Disconnected" event if it is ACL_LINK or LE_LINK link
type.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Check address in mgmt_disconnect_failed()

Check the address and address type in mgmt_disconnect_failed() otherwise
we may wrongly fail the MGMT_OP_DISCONNECT command.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: btmrvl: remove cal-data byte swapping and redundant mem copy

The device tree property can define the cal-data in proper order.
There is no need to swap the bytes in driver.
Also remove the redundant cal-data memory copy after removing the
byte swapping.

Cc: Mike Frysinger <vapier@chromium.org>
Cc: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Hyuckjoo Lee <hyuckjoo.lee@samsung.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: btmrvl: use cal-data from device-tree instead of conf file

Some ARM versions of Chromebook need to download a new calibration
data from host driver to firmware. They do have EEPROM but still
need a piece of new calibration data in test mode.

The cal-data is platform dependent. It's simpler and more feasible
to use device tree based cal-data instead of configuration file
based cal-data.

This patch remove configuration file based cal-data downloading
and replace it using cal-data from device tree.

When CONFIG_OF is not selected, or the specific property is not
present in the device tree, the calibration downloading will not
happen.

Cc: Mike Frysinger <vapier@chromium.org>
Cc: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Hyuckjoo Lee <hyuckjoo.lee@samsung.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: btmrvl: operate on 16-bit opcodes instead of ogf/ocf

Replace ogf/ocf and its packing with 16-bit opcodes.

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Bluetooth: Store supported commands only during setup procedure

The list of supported commands of a controller can not change during
its lifetime. So store the list just once during the setup procedure
and not every time the HCI command is executed.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: Remove debug statement for features complete event

The complete list of local features are available through debugfs and
so there is no need to add a debug print here.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: Set default own address type only during controller setup

The default own address type is currently set at every power on of
a controller. This overwrites the value set via debugfs. To avoid
this issue, set the default own address type only during controller
setup.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

Bluetooth: Fix limited discoverable mode for Zeevo modules

There is an old Panasonic module with a Zeevo chip in there that is
not really operating according to Bluetooth core specification when
it comes to setting the IAC LAP for limited discoverable mode.

For reference, this is the vendor information about this module:

  < HCI Command: Read Local Version Information (0x04|0x0001) plen 0
  > HCI Event: Command Complete (0x0e) plen 12
        Read Local Version Information (0x04|0x0001) ncmd 1
          Status: Success (0x00)
          HCI version: Bluetooth 1.2 (0x02) - Revision 196 (0x00c4)
          LMP version: Bluetooth 1.2 (0x02) - Subversion 61 (0x003d)
          Manufacturer: Zeevo, Inc. (18)

The module reports only the support for one IAC at a time. And that
is totally acceptable according to the Bluetooth core specification
since the minimum supported IAC is only one.

  < HCI Command: Read Number of Supported IAC (0x03|0x0038) plen 0
  > HCI Event: Command Complete (0x0e) plen 5
        Read Number of Supported IAC (0x03|0x0038) ncmd 1
          Status: Success (0x00)
          Number of IAC: 1

The problem arises when trying to program two IAC into the module
on a controller that only supports one.

  < HCI Command: Write Current IAC LAP (0x03|0x003a) plen 7
          Number of IAC: 2
          Access code: 0x9e8b00 (Limited Inquiry)
          Access code: 0x9e8b33 (General Inquiry)
  > HCI Event: Command Status (0x0f) plen 4
        Write Current IAC LAP (0x03|0x003a) ncmd 1
          Status: Unknown HCI Command (0x01)

While this looks strange, but according to the Bluetooth core
specification it is a legal operation. The controller has to
ignore the other values and only program as many as it supports.

  This command shall clear any existing IACs and stores Num_Current_IAC
  and the IAC_LAPs in to the controller. If Num_Current_IAC is greater
  than Num_Support_IAC then only the first Num_Support_IAC shall be
  stored in the controller, and a Command Complete event with error
  code Success (0x00) shall be generated.

This specific controller has a bug here and just returns an error. So
in case the number of supported IAC is less than two and the limited
discoverable mode is requested, now only the LIAC is written to
the controller.

  < HCI Command: Write Current IAC LAP (0x03|0x003a) plen 4
          Number of IAC: 1
          Access code: 0x9e8b00 (Limited Inquiry)
  > HCI Event: Command Complete (0x0e) plen 4
        Write Current IAC LAP (0x03|0x003a) ncmd 1
          Status: Success (0x00)

All other controllers that only support one IAC seem to handle this
perfectly fine, but this fix will only write the LIAC for these
controllers as well.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>

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

brcmfmac: expose chip information through debugfs

Adding the debugfs file <debugfs_mnt>/brcmfmac/<devid>/chipinfo
which contains the chip number and revision.

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

brcmfmac: move firmware event related defines to fweh.h

The dhd.h file contained a number of definitions that are
related to events received from the firmware. Those are
processed and dispatched in the driver by fweh. Hence the
definitions are moved to its include file.

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

brcmfmac: remove some unused definitions

Removing WLC_PHY_TYPE and some BRCMF_E_.* definitions as these
are not used in the driver sources.

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

brcmfmac: Use consistent naming for BCDC.

The BCDC protocol layer is using a mix of naming of CDC, BDC and
BCDC. Use the name BCDC consistenly over all functions, defines
and variables. This patch does not change code functionality.

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: move firmware command code definitions

Move the command codes to the firmware interface module as
that makes a bit more sense.

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

brcmfmac: move firmware related structures to fwil_types.h

The firmware control interface is provided by fwil source file, but
a number of structures used to communicate with the firmware still
resided in dhd.h. The patch moves them to fwil_types.h.

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

brcmfmac: remove unnecessary EXPORT_SYMBOL() usage

In bcmsdh.c the functions brcmf_sdio_probe() and brcmf_sdio_remove()
were exported, but that is not needed. The functions are linked into
the driver module, which is the only one needing to call these.

Reviewed-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Daniel (Deognyoun) Kim <dekim@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: provide firmware version in ethtool driver info

Knowing the firmware version is pretty useful information when
looking at issues. It is retrieved during initialization so
store it in driver data structure to fill the ethtool driver
info when requested.

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

brcmfmac: Dynamically register a protocol layer.

BCDC is the default protocol layer and being called directly. This
patch installs the functions for this layer dynamically. This allows
new protocols to be added and selected dynamically depending on the
hw capabilties. As currently only BCDC is supported this is always
the installed protocol.

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: remove redundant ioctl handlers

The ioctl() entry points were empty except for handling SIOC_ETHTOOL
but that has been obsoleted in favor of struct ethtool_ops. Cleaning
up removing the ioctl() handlers.

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

brcmfmac: use platform specific alignment in SDIO

The SDIO part of the brcmfmac driver uses a static define BRCMF_SDALIGN
to align buffers used for SDIO transfers. This patch replaces it by
using alignment derived from the platform specific data.

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

brcmfmac: determine alignment values during probe

The alignment values were being determined for each transmit
and receive depending on platform data. Instead determine
these once during the probe.

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

brcmfmac: removed dhd_proto.h.

dhd_proto.h was cleaned up and prototypes were moved to dhd.h.
dhd_proto.h was removed. This is a step in cleaning and
restucturing protocol layer.

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: rename dhd_cdc to bcdc

dhd_cdc is renamed to bcdc. This is a step in cleaning and
restructuring protocol layer. This is done so new protocols can
be added in the future. This step only renames the source files.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Hante Meuleman <meuleman@broadcom.com>
[arend@broadcom.com: use 'git mv' to do the rename]
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmfmac: add host tx glomming support

New WiFi full dongle supports receiving chained packets in one command
through the SDIO bus. This patch adds the support on the host side to
send chained packets.

Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@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: add firmware and nvram file name for bcm4339

Add firmware/nvram file name for bcm4339 so fmac can actually be functional with
the chip.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@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: fix driver build issue when CONFIG_BRCMDBG is not set

When CONFIG_BRCMDBG is not set we get the following build issue:
  CC [M]  drivers/net/wireless/brcm80211/brcmfmac/fwsignal.o
drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c:
    In function ‘brcmf_fws_hdrpush’:
drivers/net/wireless/brcm80211/brcmfmac/fwsignal.c:852:18:
error: ‘BRCMF_FWS_TYPE_SEQ_LEN’ undeclared

The define BRCMF_FWS_TYPE_SEQ_LEN was introduced by:

    commit 6918f38e4ed4e0493a90a4331e0033bdfc806e00
    Author: Hante Meuleman <meuleman@broadcom.com>
    Date:   Wed Oct 23 14:58:51 2013 +0200

        brcmfmac: Update fwsignal to fix out of order tx.

Unfortunately, it was put in conditional part of the source
file under #ifdef DEBUG.

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

brcmfmac: determine sd host controller related variable earlier

The commit "eb9c174 brcmfmac: determine host controller related
variables during probe" was not implemented correctly as the
information is already needed in brcmf_sdbrcm_probe(). This patch
moves it to brcmf_sdioh_attach() instead.

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

brcmfmac: support hardware extension header in trace_brcmf_sdpcm_hdr()

The SDPCM header can be traced, but it used a fixed header size. With
txglom feature the SDPCM header will have additional 8 bytes of hardware
extension header so SDIO core can properly handle the txglom packet.

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

brcmsmac: select CONFIG_BCMA when possible

The brcmsmac relies BCMA functionality to access the device. This
patch selects CONFIG_BCMA when CONFIG_BCMA_POSSIBLE is set. This
way the user does not need to be select BCMA to make the brcmsmac
driver show up in his menuconfig.

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

brcmfmac: reduce logging noise accessing SDIO SleepCSR register

The SleepCSR register is accessed to wakeup the device from the
host side. Depending on the state of the device this may take
multiple attempts. The failed attempt are not real failures so
reduce the log level specifically for this register. The calling
function will scream when the multiple attempts all failed.

Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Reviewed-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: remove empty brcmf_proto_stop

remove empty brcmf_proto_stop from protocol layer

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
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: start netif queues only when setup is completed successful

Moving the call to netif_start_queue() after brcmf_cfg80211_up() is
completed successful. If not return -EIO instead of -1 as that results
in 'Operation not permitted' which can put user on wrong track.

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: replace dongle command list with .preinit() callback

The bus-specific interface allowed a list of dongle commands to be
provided to the common driver part. However, upcoming functionality
requires a more dynamic behaviour. Hence the list is replaced
by a new callback function so the bus-specific driver part can
implement this behaviour.

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

brcmfmac: add separate function for passing bus tx overhead

The common driver needs the packet overhead for the bus in order
to reserve headroom for sk_buffs. For the SDIO driver this depends
on firmware features so it is not possible to provide it in the
brcmf_attach() call.

Reviewed-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@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: Update fwsignal to fix out of order tx.

When using fwsignal it is possible that tx packets get delivered out
of order. This patch fixes that by reordering suppressed packets and
tracking generation bit and sequence number per packet.

Reviewed-by: Arend Van Spriel <arend@broadcom.com>
Reviewed-by: Franky (Zhenhui) Lin <frankyl@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>

ath9k: Fix initvals for freq 2484

This is missing for AR9300, AR9580 and AR9340.

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

ath9k: Update AR9340 initvals

* Baseband updates
* Remove ar9340Common_rx_gain_table_1p0 since it is a duplicate.

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

ath9k: Update initvals for AR9580 v1.0

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

ath9k: Update initvals for AR9300 v2.2

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

ath9k: Remove duplicate initvals for AR9462 v2.1

The initvals for AR9462 v2.1 are very similar to v2.0.
Identify duplicate arrays and reuse the values from v2.0
to reduce module size.

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

ath9k: Attach INI arrays for AR9565 v1.1

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

ath9k: Add version macros for AR9565 1.1

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

ath9k: Add initvals for AR9565 1.1

The initialization arrays for v1.1 AR9565
are mostly the same as v1.0/v1.0.1 except for
radio_postamble.

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

ath9k: Cleanup IQ calibration for PCOEM chips

Since IQ calibration is done as part of AGC calibration for
AR9485 and above, remove the seperate IQ calibration code.

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

ath9k_hw: allow fast channel change when only CHANNEL_HT changes

The CHANNEL_HT flag is insignificant for fast channel change conditions,
since it does not affect any important part of the hardware reset /
channel setup.
Scanning usually runs with HT disabled, so this change will slightly
improve scan time on many chipsets.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: fix TSF save/restore around chip reset

A cold reset can be triggered because of DMA stop issues, and this leads
to TSF being cleared on all chipsets. To properly deal with this, always
save the TSF.
Additionally, account for the time it takes to do the actual chip reset,
which can be quite significant. On AR9344 it takes around 4.5 ms.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: optimize ath9k_flush

Instead of checking the queues in a loop with hardcoded sleep times
inbetween, use a wait queue to trigger queue checks after the tx
processing tasklet has run.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: optimize ath_drain_all_txq

If the software has processed all packets, checking the hardware queue
is unnecessary.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: use a timer to put hardware into full sleep

When operating in client mode, the short period of time between scanning
and associating is often enough to put the hardware through several
FULL-SLEEP <-> AWAKE transitions, each wakeup requiring a reset to fully
recover the hardware.
This is completely unnecessary and can easily be avoided by deferring
the switch to full sleep.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Fix TX IQ calibration for SoC chips

Since calibration data reuse is not enabled in
SoC chips, simplify the IQ calibration code.

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

ath9k: Fix Carrier Leak calibration for SoC chips

CL calibration is applicable for all chips and the
enable/disable knob comes via the INI file. For PCOEM
chips, the calibration data is reused when Fast Channel Change
is used. Caldata reuse is not enabled for SoC chips, so remove
the CL post processing code.

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

ath9k: Remove unnecessary check

TX IQ calibration is always enabled for SoC chips.

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

ath9k: Remove RTT/MCI code from SoC calibration

RTT is enabled only for AR9462 and MCI for AR9462/AR9565.
Also, manual peak calibration is not done for any of the
SoC chips.

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

ath9k: Separate routines for PCOEM and SoC calibration

Though there is some overlap between the calibration mechanisms
of PC-OEM cards and SoC chip families, dumping both of them
into a single function makes things hard to understand.

ar9003_hw_init_cal() is unreadable with chip-specific segments
scattered around. To make the logic understandable, use
different functions for client cards and SoC chips. Some
code is duplicated, but in the long run, it makes the code
more maintanable.

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

ath9k: fix SC_OP_INVALID test in ath9k_tx99_init()

SC_OP_INVALID is zero so the test is always false.  We're supposed to be
testing the lowest bit instead.

Fixes: 89f927af7f33 ('ath9k: add TX99 support')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Update MAINTAINERS

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

ath9k: Apply CUS227 specific TX gain values

CUS227, which is an AR9340 based card used in Qualcomm's
Allplay platforms requires a custom TX gain array, based
on the index 7. Add suport for this.

Cc: Michael Larson <mlarson@qce.qualcomm.com>
Cc: Stephen Collmeyer <scollmey@qce.qualcomm.com>
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Remove pcieSerDesWrite

This HW config option is always set to true and is not needed.

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

ath9k: Remove unused AR9462 2.0 initvals

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

ath9k: Add SERDES initvals for AR9462 2.1

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

ath9k: Add an initialization routine for WoW

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

ath9k: Use CONFIG_ATH9K_WOW

Move the WoW code to wow.c and compile it conditionally
based on CONFIG_ATH9K_WOW.

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

ath9k: Add a config option for WoW

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

ath9k: Fix wow.c compilation

The HW routines to set various WoW registers are present
in wow.c. For some reason, it has been compiled as part
of the main ath9k.ko module all this time, when it should
really be part of ath9k_hw.ko. This patch renames the file to
ar9003_wow.ko and adds it to ath9k_hw.ko.

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

ath9k: Fix TX99 config option usage

Use CONFIG_ATH9K_TX99 to properly enclose the tx99 code
and make sure that it is not compiled as part of the driver
when it is not selected. Move the tx99 code to a new file tx99.c
and also add ATH9K_DEBUGFS as a dependency in Kconfig.

This reduces the module size on platforms like OpenWrt where
ATH9K_DEBUGFS is selected, but TX99 might be disabled.

Cc: Luis R. Rodriguez <mcgrof@do-not-panic.com>
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Identify Killer Wireless cards

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

ath9k: Add WB335 PCI IDs

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

mac80211: fix rx_nss calculation for drivers with hw rc

Drivers with hardware rate control were given
sta->rx_nss set to 0. This was because rx_nss
calculation procedure was protected by hw/sw rate
control check.

Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

mac80211: don't leak next beacon when csa is aborted

Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

mac80211: don't cancel csa finalize work within stop_ap

The current channel switch code has a potential deadlock:
1) * cfg80211_stop_ap acquires wdev-lock
   * ieee80211_stop_ap calls cancel_work_sync for the csa_finalize_work,
     which acquires the associated worker-lock
2) * ieee80211_csa_finalize_work holds the worker-lock when run
   * it calls cfg80211_ch_switch_notify which will claim the wdev-lock,
     and also needs to claim the sdata-lock (which is the same as the
     wdev-lock) to modify the beacons.

It is sufficient to just set the channel switch active to false. If the
worker is running later, it will find the channel switch to not be
active anymore and returns immediately without changing anything.

Canceling the worker is done anyway when the interface goes down
(ieee80211_do_stop).

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

cfg80211/mac80211/ath6kl: acquire wdev lock outside ch_switch_notify

The channel switch notification should be sent under the
wdev/sdata-lock, preferably in the same moment as the channel change
happens, to avoid races by other callers (e.g. start/stop_ap).
This also adds the previously missing sdata_lock protection in
csa_finalize_work.

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>