Added support by making a distinction between a struct and its little
endian relative.
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Reviewed-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Unnecessary endianess conversion and struct member setting was deleted.
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Reviewed-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This function is only called from within the driver. Because of that
certain simplifications (not required to lock down user mem pages,
not required to check user permissions) could be performed.
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Reviewed-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

net_device ioctl handler was called both by the OS as by the driver
itself. Split the ioctl handler into two functions to make code paths
more clear.
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Reviewed-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Substituted/moved/deleted macro's.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The source main.c provided an interface function that handled several
things demultiplexing on given identifier. This interface has been
replaced by separate functions to make it more straightforward.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The lookup function used for the devpath related functions did
not provide a variable table so the lookup would always fail. As
this makes the functions pretty useless these have been removed.
Reviewed-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Descriptor ring can only start at 8KB alignment, this requirement
is dictated by DMA hardware.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

since its variable assignment can be done in a more straightforward
manner.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Since structure member pub->_wme is set to AUTO early in init, macro
EDCF_ENAB is always 'true' after that. All code that uses ECDF_ENAB()
appears in the code flow after the init.
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Substituted macro's by functions. Renamed several struct members in dma.c
to get rid of a set of macro's.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Fullmac had a build error for CONFIG_PM_SLEEP=n
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The function ieee_set_channel() provides very little functionality
so it has been removed.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The ioctl layer in brcmsmac only provided an interface layer that
was internal to the driver. This is considered pointless and has
been removed.
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Reshuffle function order in dhd_sdio of fullmac to get rid of
static function declaration
Reported-by: NJohannes Berg <johannes@sipsolutions.net>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Reviewed-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NRoland Vossen <rvossen@broadcom.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This tabular comment is useful, but the original tabbed layout
is for a tab-wdth of 4. Reformat with spaces.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The HII and HII.5 configurations both use the mailboxes.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

CKIP is a Cisco proprietary alternative to TKIP. Support removed to
simplify code - the defines to interact with the firmware are left in
place.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Only used in one place, so just inline and get rid of the macro.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

SSN seems to be an old abbreviation for WPA. Remove all references.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

HCF_INT_OFF, _TCHAR_DEFINED, HCF_LITTLE_ENDIAN, HCF_TALLIES_EXTRA
HCF_ENTRY, HCF_EXIT, OUTPUTDEBUGMSG, ASSERTDEBUGMSG
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Encapsulation should always be enabled. Support is either via firmware
or in the driver, so we need to keep the macro. The driver can stop
exporting hcf_encap
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This is supposed to profile the code so that you can build a driver
that will only update the firmware. We don't need to support this
configuration.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Untabify with tab-width set to 4 (to match VI header). Then reindent
with tab-width reset to 8.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Semi-colon terminate various macros (HCFASSERT, IF_PROT_TIME, IF_DMA,
PROT_CNT_INI, HCFLOGENTRY, HCFLOGEXIT, DAWA_ACK, DAWA_ZERO_FID) so
auto-indent will work correctly. Add appropriate do {} while
protection to the macros.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Untabified with tab-width set to 4. Reindented with linux style.

Also remove BITN macros.
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

For now just remove from hcfcfg.h, we will slowly remove the unused
configurations as we go
Signed-off-by: NDavid Kilroy <kilroyd@googlemail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The recent reorganization of the sysfs attribute registration had
the side effect of moving iio_device_register after registration of
triggers etc.  The side effect of this is that the id hadn't been
allocated by the time of trigger registration. Thus all triggers
based on device got the name <dev_name>-dev0 instead of <dev_name>-devN
where N is the iio device id.

This should also fix the lack of device name for some error messages
that we have been seeing (and I'd been meaning to track down) as
that has now moved earlier as well.
Reported-by: NManuel Stahl <manuel.stahl@iis.fraunhofer.de>
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Tested-by: NManuel Stahl <manuel.stahl@iis.fraunhofer.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Cleaner and more consistent naming + makes one abi element we don't need
go away.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch set should bring all the attributes created outside
of chan_spec registration inline with the new abi.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Acked-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

There are no build warnings at the moment.
Signed-off-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Add a missing include to eliminate the following sparse warning:

drivers/staging/xgifb/vb_table.h:2572:28: warning: symbol 'XGI21_LCDCapList' was not declared. Should it be static?
Signed-off-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use readl() and writel() in FB memory test instead of direct pointer
access, and also add iomem annotations for the FB memory.

The patch eliminates the following sparse warnings:

drivers/staging/xgifb/XGI_main_26.c:2113:69: warning: incorrect type in assignment (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2113:69:    expected unsigned char *static [addressable] [toplevel] [assigned] pjVideoMemoryAddress
drivers/staging/xgifb/XGI_main_26.c:2113:69:    got void [noderef] <asn:2>*
drivers/staging/xgifb/XGI_main_26.c:2399:30: warning: incorrect type in assignment (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2399:30:    expected char [noderef] <asn:2>*screen_base
drivers/staging/xgifb/XGI_main_26.c:2399:30:    got char *[addressable] [toplevel] [assigned] video_vbase
drivers/staging/xgifb/XGI_main_26.c:2430:31: warning: incorrect type in argument 1 (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2430:31:    expected void volatile [noderef] <asn:2>*addr
drivers/staging/xgifb/XGI_main_26.c:2430:31:    got char *[addressable] [toplevel] [assigned] video_vbase
drivers/staging/xgifb/XGI_main_26.c:2454:31: warning: incorrect type in argument 1 (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2454:31:    expected void volatile [noderef] <asn:2>*addr
drivers/staging/xgifb/XGI_main_26.c:2454:31:    got char *[addressable] [toplev
Signed-off-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Include Linux headers before driver's own headers.
Signed-off-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use readb()/readw() instead of direct pointer access to read I/O mapped
memory and also add __iomem annotation.

The patch eliminates the following sparse warnings:

drivers/staging/xgifb/XGI_main_26.c:2125:35: warning: incorrect type in assignment (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2125:35:    expected char *[addressable] [toplevel] [assigned] mmio_vbase
drivers/staging/xgifb/XGI_main_26.c:2125:35:    got void [noderef] <asn:2>*
drivers/staging/xgifb/XGI_main_26.c:2439:31: warning: incorrect type in argument 1 (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2439:31:    expected void volatile [noderef] <asn:2>*addr
drivers/staging/xgifb/XGI_main_26.c:2439:31:    got char *[addressable] [toplevel] [assigned] mmio_vbase
drivers/staging/xgifb/XGI_main_26.c:2463:31: warning: incorrect type in argument 1 (different address spaces)
drivers/staging/xgifb/XGI_main_26.c:2463:31:    expected void volatile [noderef] <asn:2>*addr
drivers/staging/xgifb/XGI_main_26.c:2463:31:    got char *[addressable] [toplevel] [assigned] mmio_vbase
Signed-off-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Not much left out of this header file. All these typedefs can be found in stdint.h
Signed-off-by: NArvydas Sidorenko <asido4@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This flag is a NOOP and can be removed now.
Signed-off-by: NYong Zhang <yong.zhang0@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Consistently name the variable tracking the link status. Use a consistent
type for this variable and get rid of some unnecessary parentheses as well.
I would like to thank Joe Perches <joe@perches.com> for suggesting these
changes and patiently helping me get here!
Signed-off-by: NK. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: NHaiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>