staging: ft1000-pcmcia: don't crash on load

Don't crash with NULL pointer dereference on load because of empty .name.

Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging/easycap: Fix bytesperline calculation

As described above fillin_formats()

"""
    /*
     *  THE 16-BIT easycap_format.mask HAS MEANING:
     *    (least significant) BIT  0:     0 => PAL, 25 FPS;   1 => NTSC, 30 FPS
     *                        BITS 2-4:   RESERVED FOR DIFFERENTIATING STANDARDS
     *                        BITS 5-7:   NUMBER OF BYTES PER PIXEL
     *                        BIT  8:     0 => NATIVE BYTE ORDER;  1 => SWAPPED
     *                        BITS 9-10:  RESERVED FOR OTHER BYTE PERMUTATIONS
     *                        BIT 11:     0 => UNDECIMATED;    1 => DECIMATED
     *                        BIT 12:     0 => OFFER FRAMES;   1 => OFFER FIELDS
     *                        BIT 13:     0 => FULL FRAMERATE; 1 => REDUCED
     *     (most significant) BITS 14-15: RESERVED FOR OTHER FIELD/FRAME OPTIONS
     *  IT FOLLOWS THAT:
     *     bytesperpixel IS         ((0x00E0 & easycap_format.mask) >> 5)
     *     byteswaporder IS true IF (0 != (0x0100 & easycap_format.mask))
     *
     *     decimatepixel IS true IF (0 != (0x0800 & easycap_format.mask))
     *
     *       offerfields IS true IF (0 != (0x1000 & easycap_format.mask))
     */
"""

bytes-per-pixel is stored in bits 5-7 of calculated mask.

But when calculating bytes-per-line we were extracting wrong value
instead of bytes-per-pixel, which was usually 2 times bigger -- e.g. for
PAL YUV 422 I was getting ((mask3 & 0x00F0) >> 4) = 4 bytes instead of 2.

The error here is that even in comments there is a line saying

     *     bytesperpixel IS         ((0x00E0 & easycap_format.mask) >> 5)

but we were using
                                    ((0x00F0 & easycap_format.mask) >> 4)

With 2 times bigger bytesperpixel and automatically bytesperline, the
video was shown halfheight'ed, which is understandable if we look at
video-memory layout:

    <------- bytesperline -------->
    <- real bpl ->

    x0----------y0  x1-----------y1
    x2----------y2  x3-----------y3

    xn----------yn  xn-----------yn
    <garbage>

for each line, we should display width pixels, then move to next line
with bytesperline, and oops, if bytesperline = 2*real-bytesperlin, we'll
skip one line and move to next-next line, and so only half lines will be
shown.

Initially I've debugged the problem with my video application[1], but
I've checked that after this patch both rawv (mine app) and tvtime work
correctly.

[1] http://repo.or.cz/w/rawv.git

P.S. why at all we use those mask/shifts? Why not use bitfields?

Cc: Mike Thomas <rmthomas@sciolus.org>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging/easycap: PAGE_SIZE is always defined

I'm not 100% sure, only 99.99% that PAGE_SIZE is always defined in
Linux. So there is no need to check for it.

Cc: Mike Thomas <rmthomas@sciolus.org>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging/easycap: There is no need to support V4L1 ioctls

Because V4L1 was completly removed from the kernel in 2.6.38. See e.g.

08af245d ([media] V4L: remove V4L1 compatibility mode).

Cc: Mike Thomas <rmthomas@sciolus.org>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging/easycap: Fix thinko in "bars" module option description

Both bars=0 and bars=1 were described as meaning to display bars on
signal lost. Actually bars=1 means "display bars", but bars=0 means
display raw source as is (usually black screen).

Instead of changing bars=0 to "_no_ testcard bars ..." as suggested by
Dan Carpenter reword the whole bars description for clarity.

Cc: Mike Thomas <rmthomas@sciolus.org>
Cc: Dan Carpenter <error27@gmail.com>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging/easycap: Kill leftover build options in readme

The following options were removed from in-tree driver, but were left
in README:

- EASYCAP_IS_VIDEODEV_CLIENT was removed in cb81fa07 (staging/easycap:
  kill EASYCAP_IS_VIDEODEV_CLIENT compilation conditional);

- EASYCAP_NEEDS_V4L2_DEVICE_H/EASYCAP_NEEDS_V4L2_FOPS were removed in
  30516058 (staging/easycap: kill EASYCAP_NEEDS_V4L2_DEVICE_H and
  EASYCAP_NEEDS_V4L2_FOPS);

- EASYCAP_NEEDS_UNLOCKED_IOCTL was removed in f2b3c685 (staging/easycap:
  kill EASYCAP_NEEDS_UNLOCKED_IOCTL).

Remove them.

Cc: Mike Thomas <rmthomas@sciolus.org>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: rtl8712: Add missing whitespace in struct definition

Fixes scripts/checkpatch.pl warning.

Signed-off-by: Michael Fiedler <michael.fiedler@informatik.stud.uni-erlangen.de>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: rtl8712: Use linux/atomic.h instead of asm/atomic.h

Fixes warning from scripts/checkpatch.pl.

Signed-off-by: Michael Fiedler <michael.fiedler@informatik.stud.uni-erlangen.de>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: rtl8712: Use __packed instead of __attribute__ ((packed))

According to scripts/checkpatch.pl, the __packed macro
from linux/compiler.h should be used.

Signed-off-by: Michael Fiedler <michael.fiedler@informatik.stud.uni-erlangen.de>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: rtl8712: Remove duplicate #define for __attribute__ ((packed))

There is already a macro defined in linux/compiler.h.

Signed-off-by: Michael Fiedler <michael.fiedler@informatik.stud.uni-erlangen.de>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: easycap: use after free in easycap_delete()

The JOM() macro dereferences peasycap, so I moved the free down some
lines.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: usbip: Fix out-of-tree build for userspace

This fixes out-of-tree builds ($builddir != $srcdir). In such cases,
libusbip.la cannot be found because it is generated in the build
directory and not the source directory.

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Acked-by: matt mooney <mfm@muteddisk.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: keucr: usb: fixed a brace coding style warning

Fixed a brace coding style warning.

Signed-off-by: Zac Storer <zac.3.14159@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: r8192e_pci: Handle duplicate PCI ID 0x10ec:0x8192 conflict with rtl8192se

There are two devices with PCI ID 0x10ec:0x8192, namely RTL8192E and
RTL8192SE. The method of distinguishing them is by the revision ID
at offset 0x8 of the PCI configuration space. If the value is 0x10,
then the device uses rtl8192se for a driver.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers:rtl8192u:remove the use of obsolete create_workqueue in ieee80211_softmac.c

Remove the use of obsolete create_workqueue(name, 0) interface
in ieee80211_softmac.c, since it has changed.

Signed-off-by: Wanlong Gao <wanlong.gao@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers:rtl8192u:remove the use of obsolete create_workqueue in r8192U_core.c

Remove the use of obsolete create_workqueue(name, 0) interface
in r8192U_core.c, since it has changed.

Signed-off-by: Wanlong Gao <wanlong.gao@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

drivers:rtl8187se:remove the use of obsolete create_workqueue in ieee80211_softmac.c

Remove the use of obsolete create_workqueue(name, 0) in
ieee80211_softmac.c, since the create_workqueue interface has changed.

Signed-off-by: Wanlong Gao <wanlong.gao@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: transciever/transceiver spelling fixes

Just tyops.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: altera-stapl: Remove superfluous comment from Kconfig

Same text as for the config ALTERA_STAPL entry, so remove it.

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: zram: Replace mutex lock by a R/W semaphore

Currently, nothing protects zram table from concurrent access.
For instance, ZRAM_UNCOMPRESSED bit can be cleared by zram_free_page()
called from a concurrent write between the time ZRAM_UNCOMPRESSED has
been set and the time it is tested to unmap KM_USER0 in
zram_bvec_write(). This ultimately leads to kernel panic.

Also, a read request can occurs when the page has been freed by a
running write request and before it has been updated, leading to
zero filled block being incorrectly read and "Read before write"
error message.

This patch replace the current mutex by a rw_semaphore. It extends
the protection to zram table (currently, only compression buffers are
protected) and read requests (currently, only write requests are
protected).

Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: zram: allow partial page operations

Commit 7b19b8d45b216ff3186f066b31937bdbde066f08 (zram: Prevent overflow
in logical block size) introduced ZRAM_LOGICAL_BLOCK_SIZE constant to
prevent overflow of logical block size on 64k page kernel.
However, the current implementation of zram only allow operation on block
of the same size as a page. That makes theorically legit 4k requests fail
on 64k page kernel.

This patch makes zram allow operation on partial pages. Basically, it
means we still do operations on full pages internally, but only copy the
relevent segments from/to the user memory.

Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: zram: Refactor zram_read/write() functions

This patch refactor the code of zram_read/write() functions. It does
not removes a lot of duplicate code alone, but is mostly a helper for
the third patch of this series (Staging: zram: allow partial page
operations).

Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Staging: zram: Remove useless offset calculation in handle_uncompressed_page()

The offset of uncompressed page is always zero: handle_uncompressed_page()
doesn't have to care about it.

Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brmc80211: cleaned up sdio related error codes

Code cleanup. Replaced Broadcom proprietary error codes with Linux native
ones.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: updated TODO file

Updated TODO file to reflect current driver state, results of recent
cleanup and fixes.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: fix for reported log spam problem

Every few minutes, this message would appear in syslog:

ieee80211 ph0: wl_ops_bss_info_changed: BSS idle: true (implement)

The message has been deleted, the driver requires no special action on this
particular event (). See: https://bugzilla.kernel.org/show_bug.cgi?id=38162

Reported-by: David Hill <hilld@binarystorm.net>
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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: rename static functions in wl_cfg80211.c

The static functions in wl_cfg80211.c starting with wl_ prefix have
been renamed starting with brcmf_ prefix to be consistent.

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 unused macros from dhd_dbg.h

cleanup of some unused macros.

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: removed last global function declaration in .c file

Code cleanup.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed global function declarations from .c files

Code cleanup. Function declarations were moved to .h files.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: rename struct dhd_pub to struct brcmf_pub

The structure dhd_pub has been renamed to have more consistent naming
in the fullmac driver. This also remove the type definition for this
structure.

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: rename structure definitions in dhd_linux.c

renamed structures to consitent naming convention and removed the
structure type definitions.

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: structure renaming in dhd.h

structure have been renamed for consistency and the structure typedefs
have been removed.

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 unused macro definitions from dhd.h

Several unused macro definitions have been removed.

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: rename macro definitions in dhd.h

The definitions in dhd.h have been renamed with consistent prefix
BRCMF_.

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: rename structure definitions in dhd.h

Structure definitions in dhd.h have been prefixed with brcmf_ for
consistent naming in fullmac driver and typedef of these structures
has been removed.

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 unused definitions from dhd.h

Several defintions from dhd.h are not used in the fullmac driver
and have been removed.

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: rename function variables

The fullmac driver uses dhd as variable name for the driver instance
in a number of functions. This has been changed to a more generic name
as the dhd is not meaningful (broadcom internal name).

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: clean up firmware download code in fullmac

In brcmfmac the firmware download related code scatter all around.
And some old dead code which is not following the standard firmware
routine is still present. This patch limits the firmware download
related code only in dhd_sdio.c and removes the rest.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move sdio resource lock to dhd_sdio.c

In fullmac the lock is for race condition for sdio interface. Move
the 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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move dpc code to dhd_sdio.c

Dpc thread handles data transaction which should be placed in
bus interface layer. Move related code 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>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move debug console related code to dhd_sdio.c

Debug console functionality is provided by dhd_sdio.c. So the related
variables and iovars should go there as well.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move fullmac watchdog timer code to dhd_sdio.c

The watchdog timer is used in bus interface layer in fullmac. Move
related code 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>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move waitqueue code to dhd_sdio.c

In fullmac waitqueue for tx control is used by bus interface layer
only. So move the related 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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: move sdio related suspend/resume code to bus interface layer

In fullmac some SDIO configurations should be done in suspend/resume
routine. It was placed under pm ops in wl_cfg80211.c which is
inappropriate. This patchs move them to sdio layer.

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 linux native ethertype iso ETH_P_BRCM

The linux native ethertype ETH_P_LINK_CTL is matching ETH_P_BRCM so
lets use that definition and get rid of ETH_P_BRCM.

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: rename EPI_VERSION_STR macro

The fullmac version macro has been renamed and moved to dhd.h to
get rid of having it defined in two separate c source files.

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 conditional code fragments from brcmfmac

The fullmac sources contained a lot of conditional code sections that
are never to be enabled for the fullmac driver. These have been removed.

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 custom gpio support code

The custom gpio code is intended for hardware platform for specific
customers and require supporting functionality not included in the
fullmac driver. The fullmac driver does not need this so it has been
removed.

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 unnecessary macro defintions

The common makefile and the fullmac makefile contained a few
macro definitions that are not or no longer used. These are
removed.

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 DHD_DEBUG macro with BCMDBG

The common Makefile for both brcm80211 drivers defines the macro
BCMDBG upon select Kconfig item BRCMDBG. The fullmac source add
another macro DHD_DEBUG with the same meaning. This is removed
and all conditional code now checks for BCMDBG definition.

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: define SHOW_EVENTS macro only for debugging

The macro SHOW_EVENTS will only be defined when Kconfig setting
BRCMDBG is set.

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 TOE macro definition

The code conditional by TOE macro is always needed so TOE macro
usage has been removed from the fullmac sources.

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 MMC_SDIO_ABORT macro definition

The fullmac driver is always needing the code conditionalized with
MMC_SDIO_ABORT macro so the macro has been removed and the code
is always compiled in.

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 EMBEDDED_PLATFORM macro definition

The macro is always intended to be defined so no need for 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: remove DHD_SCHED macro definition

The fullmac driver now always uses scheduler functionality.

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: always enable code for PLATFORM_BUS functionality

Some code is conditionally compiled in upon defintion of PLATFORM_BUS
macro. This patch will have it compiled in unconditionally.

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 BCMLXSDMMC macro definition

The conditional code depending on BCMLXSDMMC macro definition is
always compiled into the fullmac driver.

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 PKT_FILTER_SUPPORT macro definition

Packet filtering support is now always compiled into the fullmac
driver.

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 ARP_OFFLOAD_SUPPORT macro definition

The conditional code for ARP offloading is now always enabled
in the fullmac driver.

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: further renaming in fullmac sources

couple of items needed renaming to more consistent naming convention.

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: replaced more sdio related typedefs by structs

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced typedef SDIOH_API_RC by int

Code cleanup.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced typedef by struct for several sdio types

Code cleanup.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced typedef bcmsdh_info_t by struct brcmf_sdio

Code cleanup.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: replaced typedef sdioh_info_t by struct sdioh_info

Code cleanup.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: rename more brcmfmac module parameters

Renamed remaining brcmfmac module parameters to more consistent
naming convention.

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: rename static functions in dhd_linux.c

functions in dhd_linux.c are renamed according more consistent
naming convention.

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: rename netdevice related functions in fullmac

Change functions related to netdevice in fullmac to more consistent
naming convention.

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 conditional SIMPLE_ISCAN code

Code between #ifdef SIMPLE_ISCAN is not enabled and unnecessary.
It has been removed from the fullmac driver.

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: rename dhd_prot and get rid of dhd_prot_t typedef

The structure dhd_prot has been renamed to brcmf_proto to align with
fullmac naming convention and removed structure typedef dhd_prot_t.

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: get rid of PKT[TO/FROM]NATIVE macros

These macros are not needed in a linux-only driver and have been
removed.

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: rename fullmac protocol related functions

The fullmac functions for firmware communication protocol handling
have been renamed to make naming more consistent.

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: removed BDC defines

Code cleanup. BDC (Broadcom Device Class) protocol extension is used on top
of the USB related CDC (Communication Device Class) definition. BDC is always
used, so the #ifdefs were redundant.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove unused prototypes from dhd.h

A couple of prototypes are not defined or not used and have been
removed from the driver.

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: rename fullmac functions

The function provided by dhd_linux.c have been renamed to make
the naming throughout the driver more consistent and remove
Broadcom specific acronyms which are not meaningful.

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 unnecessary abstraction for scheduler

The source dhd_linux_sched.c provided function setScheduler, but it
simply calls sched_setscheduler() linux function. The setScheduler()
function has been removed.

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: rename module parameters

Fullmac source is renamed to be consistent throughout the driver. This
commit renames the modules parameters for module loading.

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: rename event handling definition

Renaming fullmac to have more consistent naming throughout the
driver. This commit renames identifiers related to firmware event
handling.

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 macro BCM_MSG_IFNAME_MAX with IFNAMSIZ

In dhd.h the macro BCM_MSG_IFNAME_MAX was defined and used. As this
is used to carry the interface name we might as well use IFNAMSIZ
which results in the same size.

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 unused definitions from dhd.h

The header file dhd.h contained a number of macro definitions that
are not used within the driver and have been removed.

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: rename ioctl command codes in dhd.h

The naming convention in the driver is to be aligned. This commit
changes the prefix used for the ioctl command codes.

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: rename functions and variables in dhd_common.c

The naming convention of the drivers is to be made consistent. This
commit fixes the function in dhd_common.c and use the new names from
the calling code.

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 extern function prototypes from c files

It is not recommended to specify function prototypes in c sources
using the extern keyword. Several occurrences have been removed or
placed in a header file or simply removed as it was already present
in a header file.

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 unused globals from dhd_common.c

Three global variable were defined but never used and have been
removed.

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 dhd_set_timer function prototype

This function prototype is not implemented nor referenced in the
brcmfmac driver so it has been removed.

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: replaced prefix of SDIO related functions

Code cleanup. Prefixed functions with brcmf_ for unique namespace. In addition
to that, prefix has been changed to categorize functions depending on their
place in the drivers SDIO stack.

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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: remove wireless extensions support from brcmfmac

The brcmfmac driver contained functions for wireless extensions
support. However, brcmfmac is intended to be a cfg80211 driver
so all the entire source file is obsolete.

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: correct bcmsdh_recv_buf() calls

The calls in dhd_sdio.c to bcmsdh_recv_buf() function passed a wrong
handle. The parameter in the function prototype was typed as void
pointer so compiler could not detect. This patch makes the type
explicit and fixes the calls to the function.

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: reorg brcms_c_validboardtype for clarity

The structure of the function made it difficult to understand, so reorganize it
to make it clearer.

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

staging: brcm80211: remove unnecessary null check

kfree(NULL) is safe, and the rest of the cleanup should be done regardless
to make sure the state is consistent.

Signed-off-by: Henry Ptasinski <henryp@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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: flatten wlc_phy_shared_detach()

wlc_phy_shared_detach() is an unnecessary level of indirection to just
call kfree(), so replace the function with kfree().

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

staging: brcm80211: add missing KERN_ facility level to printk()

Debug printk() call was missing KERN_DEBUG facility level.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: cleanup whitespace errors

Cleanup remaining whitespace errors:
- space before tabs
- spaces at the start of a line
- space between function name and open parenthesis '('
- unnecessary whitespace before a quoted newline

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: run scripts/cleanfile to remove whitespace errors

Used scripts/cleanfile to remove assorted whitespace errors.

Signed-off-by: Henry Ptasinski <henryp@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: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: delete empty brcmfmac/README file

brcm80211/brcmfmac/README contents were move to brcm80211/README previously,
but this empty file is still around.

Signed-off-by: Henry Ptasinski <henryp@broadcom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

staging: brcm80211: removed unneeded 'if' statement

The 'if' is unnecessary, since no action is taken as a result of the test.

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

staging: brcm80211: Changed comments

Code cleanup. BCM4321 is not supported by this driver. Removed 'XXX' in
comments since they do not signal a critical issue to be patched. Removed
FIXME in comment since it signalled a problem in the Minstrel algorithm,
not in our driver.

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

staging: brcm80211: corrected tx status feedback for non AMPDU packets

Code still contained a 'FIXME' label that said tx status feedback (consumed by
the rate control algorithm) was not yet taking RTS/CTS into account.

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

staging: brcm80211: taking max AMPDU length advertized by peer into account

AP advertizes max AMPDU rx length in beacon/probe response frame. Code now
uses this information to limit the length of AMPDU frames it transmits.

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