staging: brcm80211: fix rtnl_lock issue when bringing down brcmfmac

When bringing down the netdevice interface a deadlock occurred
sporadically due to the rtnl_lock being held by a task that was
waiting for another task trying to get the lock. This patch fixes
that issue.

Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix checkpatch warning in fullmac

Fix the __aligned checkpatch warning in wl_cfg80211.h

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: make use of cordic library function

The cordic function which calculates cosine and sine values for given
angle is now provided in library module. The phy code now uses this
module function.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: make use of crc8 library function

The crc8 function is now available as a library function provided
by kernel module crc8.ko. The function has been removed from brcmutil
module and srom.c now uses crc8 library function.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: only enable brcmsmac if bcma is not set

brcmsmac doesn't yet use bcma, but both drivers attempt to claim the same
device.  For now, turn of brcmsmac if bcma is enabled.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed unused MIPS specific SDRAM swap window

Certain Mips based, big endian Broadcom products contain a memory window
for endianness swap purposes. However, this is too system specific for
this driver as there are more big endian platforms to support. Thus, this
'window' functionality is not to be used and can be removed.

The swap window would be a performance improvement and could be provided
with a dedicated mapping function in the platform specific code.

Tested on BCM63281.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: restrict MIPS dma bug workaround to BCM47XX

The inline function dma_spin_for_len() was defined for MIPS platforms
but the problem only occurs with dma of the PCI core in bcm47xx chips.
This patch restricts the function further to BCM47XX platforms only.

Tested on BCM63281.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: restrict register access method for bcm47xx

The driver contained conditional code for resolving issue with
dma transaction reordering. This code was conditionalized using
__mips__ macro, but it actually is specific to bcm47xx chips.
This patch replaces it for the more speficic CONFIG_BCM47XX macro.

Tested on BCM63281.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove mips specific include from dma.c

The include <asm/addrspace.h> has been removed as it is not
needed. None of the definitions from it are used in the source
file.

Tested on BCM63281.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: Fix handling of firmware and inits on big-endian platforms

The firmware files are encoded as little-endian.  Do the appropriate swapping
for big-endian platforms.

Tested on Mac G5 PPC and BCM63281.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: Remove swaps in R_REG and W_REG macros

Swapping the addresses is unnecessary, since the swaps are handled by the
underlying platform code (i.e. readb() etc. handle any necessary swapping).

Tested on Mac G5 PPC and BCM63281.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix off-by-one error in swap functions

The original implementation iterated over the data in buf+1 through buf+size,
whereas it should have been over the data in buf through buf+size-1.

Tested on Mac G5 PPC and BCM63281.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Tested-by: Jonas Gorski <jonas.gorski@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replace BUS_SWAP32 macro with cpu_to_le32()

The macro BUS_SWAP32(a) expanded to (a) which was fine for a little
endian system. For big endian platform this should do as the name
implies. As the driver is intended for PCI which is little-endian
the macro cpu_to_le32() can be used instead.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove PHYSADDR macro from dma.c

Four macros with PHYSADDR name prefix are not doing a lot and
complicate code readability. These have been expanded in the source
code and consequently removed the definitions.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove W_SM and R_SM macros from dma.c

The macros W_SM and R_SM are not doing much conversion of the macro
parameters and complicate code readability without good cause. It
is more clear to remove usage of the macros expanding it in the source
code.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: void * cleanup in fullmac

Remove improper usage of void * in dhd_linux.c and dhd_sdio.c

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: got rid of several void pointers in softmac

Code cleanup. Replaced void pointers by less generic pointer types.

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced void pointers in otp functions

Code cleanup. Otp is 'One Time Programmable' functionality. Replaced void
pointers by less generic pointer types.

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced void pointers in dma functions

Code cleanup. Replaced by less generic pointer types.

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: got rid of several void pointers for softmac PCI

Code cleanup. Replace void * related to PCI functionality by less generic
pointer types.

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove unused func parameter in wl_cfg80211 of fullmac

Get rid of some unused void *data parameter

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove dead SDTEST code from fullmac

Remove legacy dead code of SDTEST

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove unused parameter from data interface in fullmac

Remove redundant complete function pointer in send/recv buffer interface.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove wrapper functions in bcmsdh_sdmmc.c

Use brcmf_sdioh_request_byte directly instead of calling
brcmf_sdioh_cfg_read and brcmf_sdioh_cfg_write in brcmfmac

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove dead code from dhd_sdio.c

Code cleanup for brcmfmac

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdioh_abort into brcmf_sdcard_abort

brcmfmac wrapper function cleanup

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdioh_interrupt_deregister into brcmf_sdcard_intr_dereg

brcmfmac wrapper function cleanup

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdioh_interrupt_register into brcmf_sdcard_intr_reg

brcmfmac wrapper function cleanup

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdioh_iovar_op into brcmf_sdcard_iovar_op

fullmac wrapper function cleanup.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove wext traces in fullmac

wext support code have been removed a while ago. This patch removes the
traces left behind.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove global var gInstance from brcmfmac

Use sdio function pointer stored in struct brcmf_sdio_dev instead
of those in gInstance.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove struct sdioh_info from brcmfmac

Use brcmf_sdio_dev as the unified structure to store information
of device.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove iovar IOV_BLOCKSIZE in brcmfmac

Iovar IOV_BLOCKSIZE code is not actually setting the sdio
function's block size. Use cur_blksize provided by mmc core where
necessary.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: use wait queues instead of semaphores in wl_cfg80211.c

In wl_cfg80211.c two semaphores were used to trigger a task to process
an event. The wait queues are better suited for that purpose. This also
removes checkpatch warning on sema_init() calls in this source file.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: updated TODO file

The only remaining TODO item has been tackled so it was removed. No
other TODO items left.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove target platform limitations for drivers

Several issue have been reported building the driver on big endian
systems. Following commits were done on driver Kconfig:

d462039 Staging: brcm80211: disable drivers except for X86 or MIPS platforms
15e5201 Staging: brcm80211: disable drivers for PPC platforms

These problems have been resolved so the added 'depends' lines in
the Kconfig have been removed.

Compile tested it for following architectures:
little endian
x86
x86_64
arm
mips
big endian
ppc64
sparc64
mips

Reported-by: David S. Miller <davem@davemloft.net>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reported-by: Greg Kroah-Hartman <gregkh@suse.de>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix 'uninitialized usage' compiler warning

Building for the ARM resulted in a compiler warning about usage of
a possibly uninitialized variable. The warning is valid and the
code has been fixed accordingly.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: SPARC build error fix

Due to missing memset function declaration.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed watchdog function from softmac

Code cleanup. Watchdog function served no purpose since it had an empty body.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed brcms_c_module_unregister() call in ampdu.c

The AMPDU module was never registered, so it does not have to be unregistered.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed void * from ai_ functions

Code cleanup. Replaced void * by less generic pointer types.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced void *btparam into struct pci_dev *btparam

Code cleanup. Since the softmac driver only supports a PCI(e) bus, the void *
could be replaced with a less generic type.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed unused bus code from softmac

Code cleanup. For the softmac, the 'bustype' in use is always PCI_BUS. Hence
code related to dealing with different bus types (eg: PCI_BUS, JTAG_BUS,
SI_BUS) could be removed.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: cleanup to get rid of 'over 80 character' line

One checkpatch warning was remaining in the brcmsmac driver. The
code has been restructured to get rid of this last checkpatch
warning.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove function pointer of interrupt isr in brcmfmac

The use of function pointer of bus interrupt isr is no longer needed
in fullmac as there is only one available isr.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove dead client interrupt code from brcmfmac

In fullmac, brcmf_sdcard_intr_enable/disable is a legacy layer and
doesn't really enable/disable any interrupt. This patch removes the
corresponding code.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove dead code from brcmfmac

brcmf_sdcrad_cfg_read_word and brcmf_sdcard_cfg_write_word are
not used by anyone.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove struct brcmf_sdio_card from brcmfmac

Use brcmf_sdio_dev as the unified structure to store information
of wifi dongle

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove vendor and device id check from brcmfmac

mmc core is handling the vendor/device id check and matching drivers
with devices. No need to do it again in driver.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove struct brcmf_sdioh_driver from brcmfmac

brcmf_sdioh_driver is a legacy bus driver interface we dont need
any more.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: placed suspend flag in gInstance in brcmfmac

One of the steps to remove global variable. gInstance will be
replaced by brcmf_sdio_dev in the upcoming patch.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: simplified register access macro's in softmac

Code cleanup. Removed MIPS specific 'sync' instruction since this is not
required for the chips that this driver supports. MIPS specific macro's
were now the same as non-MIPS register access macro's and thus have been
deleted. Also added comment that makes clearer what the benefit of these
macro's is. Unified big and little end register access macro's.

Reported-by: Dan Carpenter <error27@gmail.com>
Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed void * from softmac phy

Code cleanup. Replacing void * by other pointer types improves code
readability and enforces stronger type checking.

Reported-by: Julian Calaby <julian.calaby@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: cleaned up softmac DMA layer

Code cleanup. Since this driver only needs to support 64 bit DMA hardware,
an unnecessary layer of abstraction could be removed. Also DMA functions that
were not called have been removed.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: revert removal of atomic initialization

The commit "staging: brcm80211: remove code for unsupported chip"
unintentionally got rid of initialization of the atomic variable
brcmf_mmc_suspend. The patch restore that particular piece of code.

Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove dma_addrwidth() function

The function dma_addrwidth() always returns the same value so
it is redundant. As such it has been removed.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove wl_alloc_dma_resources() function

The function wl_alloc_dma_resources() does not provide any value
for the brcmsmac driver as it only returns true. It has been
removed from the driver.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: reformat long lines in brcmsmac to 80 columns

Linux coding style strongly suggest to limit length of source lines
to 80 characters. This commit correct this for the brcmsmac sources.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove unused rx status definitions

The file pub.h contained a number of definitions that are not
used in the brcmsmac driver and consequently have been removed.

Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: use PCI_DEVICE() macro in device table

The macro PCI_DEVICE() fills in the entry in abbreviated manner. Using
this removes the "lines over 80 characters" checkpatch warning on these
entries.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fixed checkpatch warnings for 'include' dir

All of them being 'line longer than 80 chars' type of warning.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fixed checkpatch warnings for brcmutil dir

Most of them being 'line longer than 80 chars' type of warning.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fixed checkpatch warnings for fullmac

Most of them being 'line longer than 80 chars' type of warning.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replace simple_strtoul usage in brcmfmac

The usage of simple_strtoul is not preferred. Instead kstrtoul
should be used. This patch fixes this for the brcmfmac driver.

Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdcard_detach into brcmf_sdio_remove

To increase code readability of brcmfmac

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: absorb brcmf_sdcard_attach into brcmf_sdio_probe

Increase readability of brcmfmac

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove BRCMF_SD_* debug macros from brcmfmac

Use unified debug macros BRCMF_* in fullmac.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove SDLX_MSG from brcmfmac

Use unified debug macros in fullmac

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove structure sdio_hc in brcmfmac

Most members in sdio_hc are no longer needed anymore. And fullmac
is keeping a global pointer of this structure. This patch deletes
the structure and places the useful member to a new structure
brcmf_sdio_dev. The pointer of brcmf_sdio_dev will be save in the
private driver data during sdio_probe. Therefore, we don't need to
keep the global pointer.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: get rid of sd debug message macro in fullmac

Use BRCMF_* instead of sd_* as debug message interface

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove code for unsupported chip

Currently fullmac only support bcm4329. Remove dead code for
unsupported chip.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove global wait queue head sdioh_spinwait_sleep

Remove global wait queue head sdioh_spinwait_sleep in fullmac.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move brcmf_mmc_suspend to sdio layer in fullmac

brcmf_mmc_suspend is used for sdio suspend/resume function. Hence it
should be placed in sdio interface layer.

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove private timeout functions in fullmac

Use kernel timer macros to replace current private timeout functions
used in dhd_sdio.c

Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replace simple_strtoul usage in brcmsmac

The usage of simple_strtoul is not preferred. Instead kstrtoul
should be used. This patch fixes this for the brcmsmac driver.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed global variable from sdio fullmac

Code cleanup. bus->card is assigned in brcmf_sdbrcm_probe (before
brcmf_sdbrcm_probe_attach()). Since w_sdreg32() and r_sdreg32() are called
only after that assignment, they can safely use bus->card. Thus there
is no instance left where brcmf_sdcard_reg_read() or brcmf_sdcard_reg_write()
is called with a NULL parameter, so the mechanism in bcmsdh.c that deals with
a NULL pointer could be deleted.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed R_REG and OR_REG macro's from fullmac

Code cleanup. Macro invocations have been substituted with macro expansion.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove MIPS specific 'sync' instruction in fullmac

This instruction was required for the bcm4716/bcm4322, but since the
fullmac driver only supports bcm4329, it could be removed. After that,
the R_REG macro's were identical and thus were reduced to just 1 R_REG
macro.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Franky (Zhenhui) Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fixed build issue for big endian platforms

Driver now builds for big endian mips platform, possibly also for other
big endian platforms. A change was made to the R_REG and W_REG macro's.
These macro's perform an xor (^) operation for endianess swap purposes.
Gcc complained because an xor operation is not allowed on a pointer type.
Fixed this by casting the pointer to an unsigned long.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: resolved checkpatch warnings in N phy

Code that exceeded the 80 char limit has been placed in separate
functions. Checkpatch warnings for the phy dir are now reduced to 1.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: resolved checkpatch warnings in LCN phy

Refactored code to not exceed the 80 char limit.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix for checkpatch warnings in phy directory

Most of them being 'line exceeds 80 chars'. Still checkpatch warnings
for the phy dir left, these will be resolved in the subsequent commits.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove volatile keyword used in struct rte_console

Two members vcons_in and vcons_out were declared using the volatile
keyword. However, there is no reason for doing so. The member is
used only to calculate the backplane address to access using offsetof.
This address is passed to subsequent read function. Use of volatile
is not warranted.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replace semaphore by wait_queue for sysioc thread

The sysioc thread was triggered using a semaphore. Now it waits for
a wake_up() on its wait queue and the semaphore has been removed as
the semaphore serves another purpose. This removes a checkpatch
warning for dhd_linux.c.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move ioctl response wait code to dhd_sdio.c

Ioctl response wait is used only by bus layer in fullmac. Move the
corresponding code to dhd_sdio.c

Signed-off-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move sdio related variables to dhd_sdio.c

In fullmac some variables are used by sdio layer declared in
dhd_linux.c. Move them to dhd_sdio.c for clean up.

Signed-off-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: brcmfmac: Roamed channel info passed to cfg80211

With this patch, brcmfmac driver will pass roamed channel information to
cfg80211 via cfg80211_roamed() API.

Signed-off-by: Sukesh Srikakula <sukeshs@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: brcmfmac: Fixed issues with iscan

Disabling the interface doesn't terminate undergoing
iscan in FW. When the interface is enabled immediately,
first scan request is returning the stale scan results,
which are populated during the earlier iscan. These stale
scan results are causing random failures with chromium
auto test cases.

With this patch, iscan will be terminated in FW whenever
iscan thread is terminated by the host driver.

Signed-off-by: Sukesh Srikakula <sukeshs@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix for 'remove unnecessary braces' checkpatch warning

Removed unnecessary braces in single-statement blocks. Used the tools
'uncrustify' and 'coccinelle' to accomplish this.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: use mac_pton() instead of own implementation

The function brcmu_ether_atoe() does exactly the same as mac_pton().
The driver now uses the latter and brcmu_ether_atoe() has been removed
as it is not used anymore.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: brcmfmac: Connect request made robust

Connect request made robust by passing broadcast bssid to the
FW in SET_SSID request. This fix helps STA to connect
to any available AP in ESS instead of sticking to one
unresponsive AP.

Signed-off-by: Sukesh Srikakula <sukeshs@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: brcmfmac: Enabling FW roaming by default

With this patch, FW roaming will be enabled by default in
brcmfmac driver

Signed-off-by: Sukesh Srikakula <sukeshs@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: power save issue fixed in brcmfmac driver

brcmfmac driver is getting request for enabling power save
before the interface is up. With the earlier implementation,
this request used to fail resulting out of sync power save
state between cfg80211 & brcmfmac driver. With this fix,
power save state is stored in configuration parameters and
will be used while initializing the adapter.

Signed-off-by: Sukesh Srikakula <sukeshs@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix for checkpatch 'avoid externs in c file' warning

Signed-off-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: use native error code in brcmf_c_pattern_atoh()

The function brcmf_c_pattern_atoh() returned meaningless -1 value.
These have been replaced by linux error code -EINVAL.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove volatile keyword from driver sources

As checkpatch.pl rightfully indicated the volatile keyword was used
but not necessary. It has been removed from the code.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove duplicated code from brcmf_init_iscan

Part of the code in brcmf_init_scan was identical to the code in
brcmf_invoke_iscan and has been replaced by a call to it.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: use mutex instead of semaphore in dhd_linux.c

The semaphore proto_sem has been replaced with mutex proto_block
which lock certain code paths for one thread.

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: modify the FOREACH_BSS macro

The FOREACH_BSS macro was changed to remove a checkpatch error. However,
the patch was considered ugly. This patch is another attempt to make it
more clear. END_FOREACH_BSS is defined with braces to avoid an invalid
checkpatch.pl warning. Ideally, the checkpatch.pl script should be adapted
but lacking good perl knowledge to do so.

Reported-by: Dan Carpenter <error27@gmail.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced all volatile typedefs

Volatile keyword is not needed, hardware is accessed using native Linux
calls that provide synchronization.

Signed-off-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>