staging: comedi: addi_apci_3120: introduce apci3120_ai_reset_fifo()

A dummy read of APCI3120_TIMER_MODE_REG resets the analog input FIFO. Introduce
a helper function to clarify this.

It's not necessary to do a dummy read of the FIFO (base + 0) before reseting it.
Remove the unnecessary dummy reads.

The APCI3120_TIMER_MODE_REG is a 16-bit register. This fixes a couple 8-bit reads.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: remove 'check' param from apci3120_setup_chan_list()

This parameter is always passed as '0'. Remove it and refactor the code.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up scan chanlist programming

Define the chanlist register and its bits and tidy up the programming
of the scan chanlist.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: remove unnecessary reset of the scan sequence

The scan sequence does not need to be reset when the driver is attached. Remove
the code in apci3120_reset() that does this.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_clr_timer2_interrupt()

A dummy read of APCI3120_CTR0_REG clears the timer 2 interrupt. Introduce a
helper function to clarify this.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: move apci3120_exttrig_enable() to driver source

Move this helper function from the included source file into the main driver
source file.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: move apci3120_timer_enable() to driver source

Move this helper function from the included source file into the main driver
source file.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename APCI3120_WR_ADDRESS

For aesthetics, rename this define.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: fix timer 2 disable in apci3120_write_insn_timer()

The wrong define is being used to disable the gate to stop timer 2 in this
function.

Use the apci3120_timer_enable() helper to properly disable the timer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_timer_enable()

Introduce a helper function to enable/disable a timer.

The disable of timers 0 and 1 in apci3120_interrupt() is probably not needed.
For now use the helper function to make sure the devpriv->ctrl bits are cleared
correctly to disable the timers.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up apci3120_exttrig_{enable,disable}()

Merge these two functions and use an 'enable' parameter to determine if the
external trigger needs to be enabled or disabled.

This function always succeeds and the callers don't check the return. Change
the return type to void.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename APCI3120_ENABLE_EXT_TRIGGER

For aesthetics, rename this define.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up APCI3120_ENABLE_TIMER[012]

For aesthetics, replace these defines with a macro that returns the
correct bit needed to set the gate bit to enable a timer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: remove APCI3120_DISABLE_TIMER[012]

For aesthetics, remove these defines and just use ~APCI3120_ENABLE_TIMER[012].

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up devpriv->ctrl use

The apci3120_setup_chan_list() function sets devpriv->ctrl to the scan length
(PR) and scan start (PA) values and writes the value to the register. There is
no need to mask the value first.

The apci3120_ai_insn_read() function calles apci3120_setup_chan_list(). There
is no need to clear devpriv->ctrl first or clear any additional bits and write
the register again. This also fixes an incorrect use of APCI3120_DISABLE_TIMER0
to disable the timer.

apci3120_cyclic_ai() also calls apci3120_setup_chan_list() so it does not need
to clear devpriv->ctrl or clear any addidional bits and write the register.

Update the comments in apci3120_reset() and apci3120_cancel().

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename private data 'us_OutputRegister'

Rename this CamelCase member of the private data and tidy up the mask/set
of its bits.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: move timer helpers to main driver source

Move the timer read/write and set mode helpers from the included source
file into the main driver source file.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_timer_set_mode()

Introduce a helper function to set the operation mode of a timer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up timer_mode masking

Define a macro that returns the mask of the timer_mode bits for a given
timer. Use the macro to remove the "magic" values used to clear the bits.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename private data 'b_TimerSelectMode'

Rename this CamelCase member of the private data and tidy up the mask/set
of its bits.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename APCI3120_TIMER_VALUE

For aesthetics, rename this register offset.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: fix counter and external interrupt disable

The APCI3120_DISABLE_ALL_TIMER define define is not needed, and it's used
incorrectly in apci3120_cancel(). The define is a mask of the bits needed
to disable the counters. Writing the value directly sets unintended bits.

Prior to writing the value, the 'devpriv->us_OutputRegister' is set to 0
and written to the register. This disables all the timers and the external
trigger.

Remove the unnecessary apci3120_exttrig_disable() call as well as the write
of APCI3120_DISABLE_ALL_TIMER.

In apci3120_interrupt(), remove the unnecessary mask and write to disable
all the timers. The COMEDI_CB_EOA event will cause the core to call the
(*cancel) operation which will disable the timers.

Remove the unused define.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: tidy up CTR0 register defines

Define the CTR0 register in the main driver source file and remove all
the old defines in hwrdv_apci3120.c.

For aesthetics, save the raw digital output state (devpriv->do_bits) in
the digital output (*insn_bits) function and use a macro to set them when
reading/writing a timer.

Use the CTR0 register define in the digital output (*insn_bits) function
and remove the current register define.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_timer_read()

Introduce a helper function to select a timer and read a value from it.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_timer_write()

Introduce a helper function to select a timer and write a value to it.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: rename private data 'b_DigitalOutputRegister'

Rename this CamelCase member of the private data.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: introduce apci3120_ns_to_timer()

The timer divisor calculations in this driver are over complicated.

There are three timers on the board. They all use the same base clock
with a fixed prescaler for each timer. The base clock used depends on
the board version and type:

  APCI-3120 Rev A boards OSC = 14.29MHz base clock (~70ns)
  APCI-3120 Rev B boards OSC = 20MHz base clock (50ns)
  APCI-3001 boards OSC = 20MHz base clock (50ns)

The prescalers for each timer are:

  Timer 0 CLK = OSC/10
  Timer 1 CLK = OSC/1000
  Timer 2 CLK = OSC/1000

Add a new member to the private data, 'osc_base', to hold the base clock
time. Set this member during the board attach.

Introduce a helper function to calculate the divisor needed to generate
a nanosecond time with a given timer.

Use the new helper function in the driver to clarify the code.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgap: introducing find_board_by_major()

use find_board_by_major function instead of getting a brd from static
arrary. Becasue tty's major number doesn't start zero and we can
find a brd from dgap_board[].

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgap: remove useless variables for saving tty's major

The board_t has a tty_struct(serial_driver and print_driver)
that has a major number. When major number is compared in
dgap_tty_open(), just use brd->serial_driver{print_driver}->major.

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgap: fix memory leak caused by double allocation of tty_structs

The tty_struct of serial_driver and print_driver were getting
allocated twice. One is allocated in tty_alloc_driver(), the other
is in dgap_tty_register(). So remove these in dgap_tty_register().

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgap: set tty's flags by tty_alloc_driver()

tty's flags can be set by calling tty_alloc_driver().

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: das1800: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
add to the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: quatech_daqp_cs: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: cb_pcidas64: use comedi_async 'scans_done' to detect AO EOA

Remove the private data member 'ao_count' and use the comedi_async 'scans_done'
member to detect the analog output end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
read from the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: cb_pcidas64: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'ai_count' and use the comedi_async 'scans_done'
member to detect the analog output end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
add to the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: cb_pcidas: use comedi_async 'scans_done' to detect AO EOA

Remove the private data member 'ao_count' and use the comedi_async 'scans_done'
member to detect the analog output end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
read from the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: cb_pcidas: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'count' and use the comedi_async 'scans_done'
member to detect the analog input end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
add to the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: usbduxfast: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'ai_sample_count' and use the comedi_async
'scans_done' member to detect the analog input end-of-acquisition.

Use the comedi_nsamples_left() helper to get the number of samples to actually
add to the async buffer.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: amplc_pci230: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'ai_scan_count' and use the comedi_async 'scans_done'
member to detect the analog input end-of-acquisition.

Use the comedi_nsamples_left() helper to work out the number of 'wake' samples in
pci230_ai_update_fifo_trigger_level() and the number of 'todo' samples actually
added to the async buffer in pci230_handle_ai().

Remove the unnecessary COMEDI_CB_OVERFLOW event for the hardware FIFO overflow
error. The COMEDI_CB_ERROR event will terminate the command.

comedi_buf_write_samples() can fail if the async buffer does not have room for
the sample. The it will set the COMEDI_CB_OVERFLOW event and return 0. Detect
this and quit trying to read and add more samples. The event will terminate the
command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: drivers: introduce comedi_nsamples_left()

Introduce a helper function to calculate the number of samples remaining
when the cmd->stop_src is TRIG_COUNT.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: amplc_pci230: use comedi_async 'scans_done' to detect AO EOA

Remove the private data member 'ai_count' and use the comedi_async 'scans_done'
member to detect the analog output end-of-acquisition.

Use the helper function comedi_nscans_left() to get the number of scans in the
async buffer of left in the command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: comedi_test: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Use the helper function comedi_nscans_left() to check if the number of scans
left in the command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: amplc_pci224: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ao_stop_count' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Use the helper function comedi_nscans_left() to determine the number of scans
available in the async buffer or left in the command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: drivers: introduce comedi_nscans_left()

Introduce a helper function to determine the number of scans left in
the async command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: split out PCI support into new module

Setting the `CONFIG_COMEDI_PCI_DRIVERS` kernel configuration option
makes the main "comedi" module depend on the PCI support in the kernel.
That's not that big a deal since PCI support in the kernel is either
built into the kernel or is absent, and is not in a separate module.
Still, not all low-level Comedi drivers need PCI support, so we could
save a bit of space by not including it.  The Comedi PCI support
functions are all in "comedi_pci.c".  Turn it into a separate module so
the support code doesn't have to be loaded unnecessarily.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: split out USB support into new module

Setting the `CONFIG_COMEDI_USB_DRIVERS` kernel configuration option
makes the main "comedi" module depend on the "usbcore" module.  But
perhaps the machine has no USB controllers (or they have been disabled),
in which case the "usbcore" module may have been pulled in
unnecessarily.  Only a few low-level Comedi drivers require USB support.
The Comedi USB support functions are all in "comedi_usb.c".  Turn it
into a separate module so we don't have to pull in the "usbcore" and
"usb_common" modules unnecessarily.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: split out PCMCIA support into new module

Setting the `CONFIG_COMEDI_PCMCIA_DRIVERS` kernel configuration option
makes the main "comedi" module depend on the "pcmcia" module, but many
machines don't have PCMCIA slots and only a few low-level Comedi drivers
need PCMCIA support.  The Comedi PCMCIA support functions are all in
"comedi_pcmcia.c".  Turn it into a separate module so we don't have to
pull in the other PCMCIA support modules unnecessarily.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: comedidev.h: remove some #ifdefs

Some declarations and macro definitions in "comedidev.h" are protected
by various `#ifdef`s for kernel configuration options, but the header
file compiles fine without the `#ifdef`s regardless of whether those
config option macros are defined or not.  Remove the `#ifdef`s and
compile the affected code unconditionally.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: comedidev.h: remove dummy PCI support functions

Some low-level Comedi driver modules used to handle PCI devices and ISA
devices in the same module with some conditional compilation.  To reduce
the amount of conditional compilation, some dummy inline versions of
`comedi_to_pci_dev()`, `comedi_pci_enable()`, `comedi_pci_disable()`,
and `comedi_pci_detach()` are defined in "comedidev.h" if
`CONFIG_COMEDI_PCI_DRIVERS` is undefined.  Since those Comedi low-level
driver modules have since had PCI support split out into separate
modules, there is no need to keep the dummy inline functions around any
more, so remove them.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: Fix issue of byPacketType is not set at all.

set by calling CARDbyGetPktType.

change CARDbIsOFDMinBasicRate to check basic_rates.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: BBbReadEmbedded replace __iomem with vnt_private

Changing callers to point to private

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: BBbWriteEmbedded replace __iomem with vnt_private for callers

Referencing dwIoBase inside function.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: baseband/rf s_bAL7230Init change __iomem for baseband functions

Replace __iomem with vnt_private providing pointer for dwIoBase

Change BBvPowerSaveModeON and BBvPowerSaveModeOFF to vnt_private and
reference __iomem there, to be removed later.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: baseband replace __iomem where caller is priv dereferenced.

Replace with stucture vnt_private *priv

in functions
BBvSoftwareReset
BBvSetTxAntennaMode
BBvSetRxAntennaMode
BBvSetDeepSleep
BBvExitDeepSleep

__iomem *dwIoBase will be moved into BBbWriteEmbedded and BBbReadEmbedded
later.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: baseband.h remove pDevice from struct vnt_private pointer

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: dereferencing NULL pointer

we can reach the label Away in total five situation and in four of
of them pDevInfo is NULL. so we were basically dereferencing a NULL
pointer.

Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: unneeded NULL check

the NULL check for memregion is not required as it has already been
checked for NULL after kzalloc. so we can reach this part of the code
only if memregion is not NULL.

Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Add braces to else arm of if/else statement

Add braces to the else arm of an if/else block in visorchannel_funcs.c

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Rename goto label Away

Rename the goto label used throughout visorchannel_funcs.c from Away to cleanup.

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Adjust lines to contain a maximum of 80 characters

Split multiple logging lines to get them closer to 80 characters per line.

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Place logical continuation at the end of a line

Move the || logical continuation from the start of the second line of an if statement to the end of the first line.

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Remove unnecessary parentheses

Remove extraneous parentheses around a variable in visorchannel_funcs.c

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Remove space between cast and variable

Remove the whitespace between a cast and the variable in visorchannel_funcs.c

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Fix alignment issues

Use the appropriate whitespace for multiline statements in visorchannel_funcs.c

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Remove multiple blank lines

Remove the instances of multiple blank lines in the visorchannel files.

Signed-off-by: Bryan Thompson <bryan.thompson@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: das800: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_3120: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_AiActualScan' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

The function v_APCI3120_InterruptDmaMoveBlock16bit() is then just a wrapper
for comedi_buf_write_samples(). Remove it.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: usbduxsigma: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'ai_sample_count' and use the comedi_async
'scans_done' member to detect the analog input end-of-acquisition.

Move the EOA check so it happens after adding the samples from the current
urb to the async buffer. This prevents the unnecessary resubmit of the urb
when the EOA occurs.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: usbduxsigma: use comedi_async 'scans_done' to detect AO EOA

Remove the private data member 'ao_sample_count' and use the comedi_async
'scans_done' member to detect the analog output end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: usbdux: use comedi_async 'scans_done' to detect AI EOA

Remove the private data member 'ai_sample_count' and use the comedi_async
'scans_done' member to detect the analog input end-of-acquisition.

Move the EOA check so it happens after adding the samples from the current
urb to the async buffer. This prevents the unnecessary resubmit of the urb
when the EOA occurs.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: usbdux: use comedi_async 'scans_done' to detect AO EOA

Remove the private data member 'ao_sample_count' and use the comedi_async
'scans_done' member to detect the analog output end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_sample_count' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Also, remove the unnecessary COMEDI_CB_EOS event. The core automatically
detects and adds that event.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcmuio: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'stop_count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcmmio: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: adv_pci1710: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_act_scan' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl818: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_act_scan' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl818: remove private data member 'ai_act_chan'

This member of the private data is set to '0' but never used. Remove it.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl816: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_act_scan' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl812: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ai_act_scan' and use the comedi_async
'scans_done' member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl711: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'ntrig' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: amplc_dio200_common: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'stopcount' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_2032: use comedi_async 'scans_done' to detect EOA

Remove the private data member 'stop_count' and use the comedi_async 'scans_done'
member to detect the end-of-acquisition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: comedidev.h: add 'scans_done' member to comedi_async

Introduce a new member to comedi_async to count the number of scans completed.
This member is cleared by comedi_buf_reset() along with the other comedi_async
members. It is incremented in comedi_inc_scan_progress() when the end of scan
is detected.

This member will be used to clean up the scan counting in the comedi drivers.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: panel: Fix single-open policy race condition

Fix the implementation of a single-open policy for both
devices (lcd and keypad) by using atomic_t instead of plain ints.

Signed-off-by: Mariusz Gorski <marius.gorski@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: change tx wake queue

Wake queue in the dwIsr loop of device_intr instead of device_tx_srv.

This fixes an issue when ISR_TXDMA0 or ISR_AC0DMA does not occur
device_tx_srv is not called and the queue gets stuck in stopped
condition.

On test if the queue is stuck another MACvTransmitAC0 or MACvTransmit0
in vnt_tx_packet will clear it.

Check on vif that both buffers are available and the queue is stopped.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: imx-drm: remove old FSF address from license text

Linux already includes a copy of the GPL, checkpatch compains about the address.
Remove it from the license text.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: add ioctls to set per-file read and write subdevice

Now that Comedi has the structures in place to support setting the
current "read" and/or "write" subdevice on a per-file object basis, add
new ioctls to set them.  The newly chosen "read" ("write") subdevice
needs to support "read" ("write") commands, and the file cannot be busy
handling a "read" ("write") command on the previous subdevice (if any).

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: prepare support for per-file read and write subdevices

Comedi devices may have several subdevices that support "read" and/or
"write" asynchronous commands that use the "read" or "write" file
operations for data transfer.  The low-level Comedi drivers may nominate
a default "read" subdevice and/or a default "write" subdevice, but it
may have other subdevices that support asynchronous commands.

The Comedi core provides a somewhat clunky mechanism to provide access
to the asynchronous command support of the non-default subdevices.  When
a low-level device is "attached" to a core Comedi device, it dynamically
allocates a minor device number for each of the subdevices that support
asynchrounous commands and associates them with files created in SysFS
named "comediX_subdY", where "X" is the minor device number of the main
comedi device, and "Y" is the subdevice number.  An application can open
these subdevice-specific files and they behave like the regular
"comediX" files except that the "read" and/or "write" subdevice may be
different to the default chosen by the low-level driver.

This patch adds a layer of indirection between the file object and the
comedi device object to allow the current "read" and/or "write"
subdevice to be altered after opening the Comedi device, on a per-file
object basis.  The advantage is that an application only needs to open
the main Comedi device file and can then choose which subdevice it wants
to "read" or "write".  The main Comedi device file can be opened more
than once, and each file object can choose the "read" and "write"
subdevices independently.

The new `struct comedi_file` is created on "open" and freed on
"release".  It includes pointers to the main Comedi device structure,
and to the current "read" and "write" subdevice structures (which may be
NULL).  It also has information to keep track of when a low-level device
has been attached or detached since the previous time the file object
was used.  In that case, the current "read" and "write" subdevices in
the `struct comedi_file` will be changed to the new defaults (or set to
NULL).  (The change to new defaults is done by `comedi_file_reset()`.
The checking for attach/detach is done by `comedi_file_check()` which
will call `comedi_file_reset()` if there have been any attach/detach
operations since the previous call.)

A subsequent patch will add the ioctls to change the current "read" and
"write" subdevices.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: media: lirc: modify print calls

This patches replaces one pr_debug call by dev_dbg and
changes the device used by one of the dev_err calls.

Signed-off-by: Aya Mahfouz <mahfouz.saif.elyazal@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgap: remove unnecessary function

The dgap_init_global() initialize the dgap_board
that is a global variable as static and dgap_poll_timer.
But init_timer() is called twice in dgap_start() and dgap_board
doesn't need to be initialized to NULL.

Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655 baseband.c camel case replace pDevice -> priv

for struct vnt_private

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: device.h remove dead member wstats

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: device_set_options remove dead variables

uConnectionRate
wRTSThreshold
byOpMode
b11hEnable
uChannel

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: baseband.c/h remove dead functions

These functions are not used so remove them
BBbIsRegBitsOn
BBbIsRegBitsOff
BBvReadAllRegs
BBvLoopbackOn
BBvLoopbackOff

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: device_init_registers remove uConnectionRate

The device starts up with a default rate of 54M

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: device_set_options remove unused ethernet addresses

Removing these variables
abyBroadcastAddr
abySNAP_RFC1042
abySNAP_Bridgetunnel

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: device_get_options remove unused device parameters

IP_byte_align
Channel
PreambleType
RTSThreshold
ConnectionRate
OPMode
b80211hEnable

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: remove fragmentation from driver

fragmentation is now handled by mac80211.

Remove functions
device_alloc_frag_buf
device_init_defrag_cb
device_free_frag_buf

Removing
typedef struct tagSDeFragControlBlock
frag_thresh
sRxDFCB
cbDFCB;
cbFreeDFCB;
uCurrentDFCBIdx;

macros
FRAG_THRESH_MIN
FRAG_THRESH_MAX

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: mac.c and mac.h remove dead functions.

MACvReadAllRegs
MACbyReadMultiAddr
MACvWriteMultiAddr
MACvSetMultiAddrByHash
MACvResetMultiAddrByHash
MACvSetRxThreshold
MACvGetRxThreshold
MACvSetTxThreshold
MACvGetTxThreshold
MACvSetDmaLength
MACvGetDmaLength
MACvGetLongRetryLimit
MACbIsInLoopbackMode
MACbCompareContext
MACvOneShotTimer0MicroSec
MACbTxDMAOff
MACvClearBusSusInd
MACvEnableBusSusEn
MACbFlushSYNCFifo
MACvSetDefaultKeyEntry
MACvEnableDefaultKey
MACvDisableDefaultKey
MACvSetDefaultTKIPKeyEntry
MACvSetDefaultKeyCtl

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: mac.c/h remove member type comments

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>