This patch fixes checkpatch warning: missing a blank like after declarations.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patches fixes checkpatch warning: braces {} are not necessary for single
statement blocks. Remove some comments also.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch changes return value with linux error value, not 1 or 0.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch changes return value with linux error value, not 1 or 0.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes checkpatch warning braces{} are not necessary for single
statment blocks.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes checkpatch warning line over 80 characters.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes unused function pointer hif_sync and hif_clear_int, and
removes it's related functions sdio_clear_int, sdio_sync, wilc_spi_clear_int
and wilc_spi_sync.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes wilc_debug_func of hif_init and remove it's related
functions as well because it is not used anymore.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch use device print api instead of driver defined print. Remove
varialbe dPrint as well. String "[wilc sdio]" and "[wilc spi]" are also removed
from all the print statment if exist because it shows which device the message
is related to.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

linux_wlan.c and wilc_wlan.c was separated into two part at the beginning
to support various platforms. They are in charge of send/receive control and
packet data, so they will be merged into one file wlan.c later.
First of all, wilc_wlan_dev_t which is used as global variable of wilc_wlan.c
will be moved into struct wilc.
This patch moves all members of wilc_wlan_dev_t to struct wilc and use wilc
instead of g_wlan. Finally remove wilc_wlan_dev_t and g_wlan.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes unused variable tx_buffer_offset of wilc_wlan_dev_t.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

io_type of wilc_wlan_dev_t is unneeded, we can use io_type of struct wilc.
Remove io_type of wilc_wlan_dev_t and use io_type of wilc.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

hif_func of wilc_wlan_dev_t is duplicate because we have same struct
wilc_hif_func ops of struct wilc which is available in wilc_wlan.c.
Rename ops of struct wilc with hif_func and remove hif_func of wilc_wlan_dev_t,
and use wilc->hif_func instead of g_wlan.hif_func in all functions.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes define BIG_ENDIAN and use kernel define byte order macros
instead of swap itself. Remove unused BYTE_SWAP macro and __CHECK_ENDIAN__
in Makefile also.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch renames "index" of struct txq_entry_t to tcp_pending_ack_idx
since this name could be confused index of txq_entry_t. It is index of
tcp pending ack.

It fixes 8e55639d ("staging: wilc1000: rename tcp_PendingAck_index
of struct txq_entry_t")
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch remove memory allocation options in Kconfig. It was used a long time
ago to aquire memory, which we will not use this config anymore.
Remove it's config, related define and codes as well. We will take
PREALLOCATE_AT_LOADING_DRIVER as it is default.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes define COMPLEMENT_BOOT in Makefile. The feature was removed
by commit b46d6882 ('staging: wilc1000: remove COMPLEMENT_BOOT') but
the define was not removed. So remove completely.
Signed-off-by: NGlen Lee <glen.lee@atmel.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename these functions so they have namespace associated with the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename these functions so they have namespace associated with the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

An (*insn_read) can only happen if the subdevice is in a non-busy state,
i.e. an async command is not running. The board reset and subdevice
(*cancel) will ensure that the control bits (devpriv->ctrl) are already
cleared.

The (*insn_read) only needs to enable the software trigger before reading
samples. It should also disable the software trigger when done. Fix the
(*insn_read) to do this.

For aesthetics, rename the function so it has namespace associated with
the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There are a number of descrepencies in the various manuals for the boards
that this driver supports. Some show a 10 MHz clock for counters 1 and 2
others show a 1 MHz clock. Counter 0 can use either a div 10 of that clock
or an external clock (up to 10 MHz).

Currently this driver initializes counters 1 and 2 with a 10 MHz clock.
For consistency, return 1 MHz (10 MHz/10) for counter 0 when the user
queries the internal clock source with INSN_CONFIG_GET_CLOCK_SRC.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Change the return type to void, this function always succeeds and the
caller does not check the return value anyway.

Fix the initial programming of the control register. The SW bit enables
the software trigger and should not be set here. Setting CNT0 selects the
external clock source for counter 0 (the user counter). It makes more
sense to select the internal 1 MHz clock.

Remove the unnecessary initialization of the private data members. This
function is only called during the (*auto_attach) after the private
data was kzalloc'ed.

Remove the redundant clearing of the A/D FIFO and pending interrupts.
Just do it once.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The (*cancel) operation should do just that. Remove the setting of the SW bit
which enables the software trigger.

For aesthetics, rename the function so it has namespace associated with the
driver and add a couple comments.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

For aesthetics, post-increment the 'subdev' index when used to get a
comedi_subdevice pointer instead of incrementing it after the subdevice
is initialized.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the SDF_COMMON flag, the analog reference is not programmable and
the default aref (AREF_GROUND -> SDF_GROUND) provides adequate information
about the reference.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the SDF_COMMON flag, the analog reference is not programmable and
the default aref (AREF_GROUND -> SDF_GROUND) provides adequate information
about the reference.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The analog outputs can use an external reference to create the D/A output
range. Add an entry to the comedi_lrange table for it and modify the
(*insn_write) to support it.

Note that the D/A output range is 0 to +Vref with a -Vref. The comedi_lrange
does not include the sign of the range. It simmply allows the user to convert
between the 12-bit samples values (0x0000 - 0x0fff) and a physical value (0.0
to 1.0) using the comedilib comedi_to_phys() and comedi_from_phys() functions.
A physical value of 0.0 would actually be 0V with a -Vref and -V with a +Vref
and 1.0 would be +V with a -Vref and 0V with a -Vref. Ths user will need to
work this out but at least they can now use the external reference.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename this function so it has namespace associated with the driver.

For aesthetics, move the function so it is located in the middle of
the analog input support functions.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename these functions so they have namespace associated with the driver.

For aesthetics, move the functions so they are not located in the middle
of the analog input/output support functions.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename this function so it has namespace associated with the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename this function so it has namespace associated with the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the unnecessary comments and rename the 'rangelist_ai' member for
aesthetics.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The pci1711/31 boards are the only ones that have a smaller FIFO (1K vs 4K) and
single-ended analog inputs (no differential).

Replace the 'has_large_fifo' and 'has_diff_ai' members of the boardinfo with
'is_pci1711' and use that to determine how to initialize the analog input
subdev_flags as well as the private data 'max_samples'.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This member of the boardinfo isn't really necessary. All the boards
except the pci1713 have 16 digital inputs and 16 digital outputs.

There is already a 'is_pci1713' member in the boardinfo so that can
be used to determine the subdevices for the digital inputs and outputs
need to be allocated and initialized.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This member of the boardinfo isn't really necessary. All the boards
have analog inputs, the pci1713 has 32 channels the rest have 16
channels.

There is already a 'is_pci1713' member in the boardinfo so that can
be used to determine the number of channels for the analog input
subdevice.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

For aesthetics, define some macros to set the bits in the mux control
register. Also, rename the 'mux_ext' member of the private data.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The gain codes used to program the analog output range are currently
stored in const char arrays. The values look a bit "magic" and it's
not clear how they associate with the comedi_lrange without looking
through user manuals.

Refactor the ai range programming to clarify the driver and remove
the magic numbers. Also, refine the bits in the range register that
set the differential and unipolar modes.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Currently this driver calls pci1710_reset() during the (*detach) of
the driver. That function does the following:

  1) program the control register to stop any operations
  2) clears the analog input FIFO
  3) clears any pending interrupts
  4) sets all the analog output channels to unipolar 5V range and 0V output
  5) sets all the digital outputs to 0V

Before detaching the comedi core will (*cancel) any running async commands.
This will handle 1-3 above.

Depending on the application, it might not be safe to reset the analog and
digital outputs when the driver is detached.

Remove the board reset when detaching and just use comedi_pci_detach()
directly for the driver (*detach).
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

All the boards supported by this driver have a 8254 counter. Channels
1 and 2 are used to create the cascaded 32-bit analog input pacer.
Counter 0 is available for the user on all the boards except the PCI-1713.

Remove the 'has_counter' boardinfo and use the 'is_pci1713' boardinfo to
determine if the user counter subdevice needs to be allocated and
initialized.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

All the boards supported by this driver can use an interrupt. Remove
the unnecessary boardinfo.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>