For aesthetics, add some whitespace to the Analog Output subdevice init.

Also, remove the '#ifdef PCIDMA' here and only hook up the async command
support if we have an irq and the board either has a fifo or DMA is
supported.

The ni_pcimio driver is the only place PCIDMA is defined. That driver
is also the only user of ni_mio_common that allocates devpriv->mite. Use
that test instead to determine if dma is supported.

Move 'is_67xx' check and init_ao_67xx() call into the Analog Output
subdevice init. The 67xx boards all have analog outputs and this check
does not need to be done for the boards that do not have analog outputs.
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>

For aesthetics, add some whitespace to the Analog Input subdevice init.

The callers of ni_E_init() do the request_irq() and set dev->irq if the
interrupt is available. Only hook up the async command support if we
have the irq.

Also, remove the '#ifdef PCIDMA' here. The ni_pcimio driver is the only
place PCIDMA is defined. That driver is also the only user of ni_mio_common
that allocates devpriv->mite. Use that test instead to determine if the
subdevice async_dma_dir member needs to be initialized.
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>

The comedi subdevice (*cancel) function can only be called by the core if the
subdevice supports async commands. Move the initialization of this callback
to it is only set if async commands are enabled.
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>

For aesthetics, change the 'aobits' in the boardinfo to 'ao_maxdata'
to remove the need for the calculation of the subdevice 'maxdata'.
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>

For aesthetics, change the 'adbits' in the boardinfo to 'ai_maxdata'
to remove the need for the calculation of the subdevice 'maxdata'.
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>

For aesthetics, use the subdevice 'maxdata' instead of the board 'adbits'.
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>

The comedi core expects (*insn_read) functions to return insn->n
data smaples. Fix this function to work like the core expects.
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>

This member of the private data can be determined by checking the
cmd->stop_src. Do that instead and remove the member.
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>

Currently the PCI6xxx boards use a different (*insn_write) function
for the analog output subdevice. Most of the code is identical to the
(*insn_write) used by all the other board types.

For aesthetics, merge the two (*insn_write) functions and handle the
'is_6xxx' differences in the common 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>

Currently ni_m_series_ao_config_chanlist() gets the boardinfo pointer in
order to get the number of analog output channels. Get this from the
comedi_subdevice instead.

The ni_old_ao_config_chanlist() function also gets the boardinfo pointer
to get the analog output resolution, 'aobits', in order to calculate the
value needed to convert samples from unsigned to two's complement for
bipolar ranges. Use the comedi_subdevice 'maxdata' instead.
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>

None of the boardinfo that uses this code sets the 'has_analog_trig' flag so
ni_ai_config_analog_trig() always returns -EINVAL.

Also, the private data 'atrig_low', 'atrig_high', and 'atrig_mode' values
that are set by ni_ai_config_analog_trig() are never used.

Remove the incomplete INSN_CONFIG_ANALOG_TRIG support along with the unused
boardinfo and private data members.
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>

Use the bytes_per_sample() helper instead of 'sizeof(short)' to clarify
the length calculation.

Use comedi_offset_munge() to handle the unsigned to two's complement
munge of the data for bipolar ranges.

Tidy up the local variables.
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>

The board->reg_type if checked quite often in the ni_mio_common.c code to
handle differences in the PCI devices supported by the ni_pcimio driver.
Simplify the code a bit by adding some 'is_{board type}' bit-field flags
to 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>

Comedi (*insn_write) functions are expected to write insn->n values
to the hardware. Fix this function to work like the core expects.

Also, don't rely on the return value of ni_ao_comfig_chanlist() to
determine if the values need converted to two's complement before
writing to the hardware. Use the comedi_range_is_bipolar() and
comedi_offset_munge() helpers 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>

This marco relies on a local variable having a specific name. Remove the
macro and just use ni_ao_win_outw() directly.
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>

Comedi (*insn_write) functions are expected to write insn->n values to
the hardware. Fix this function to work like the core expects.

Also, use the comedi_offset_munge() helper to convert the comedi unsigned
values into the two's complement values that the hardware needs.
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>

This member of the boardinfo for the NI MIO drivers is used to indicate if
the ranges for the analog output subdevice (ao_range_table in the boardinfo)
includes any unipolar ranges. If it's not set, the ao_range_table only has
bipolar ranges.

The 'ao_unipolar' flag is checked when munging the ao data values from the
user so that the values for bipolar ranges are converted to 2's complement
values before they are written to the hardware.

The flag is also used when programming the analog output configuration on
non-M series boards for bipolar/unipolar and external reference operation.

Simplify the driver a bit by removing this boardinfo flag and just using
the comedi_range_is_bipolar() and comedi_range_is_external() helpers to
check the range directly.
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>

The comedi_krange includes a flags member that currently identifies the
'units' of the range (RF_UNIT) and if the range is from an internal or
external source (RF_EXTERNAL).

Introduce some helper functions to check if a given range is from an
external source.
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>

Convert the inline MITE_* functions, used to calculate the mite channel
register offsets, into simple macros.
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>

For aesthetics, factor the code that disables the input filters out of
ni_65xx_auto_attach().
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>

During the attach of this driver, the digital output ports are all
initialized to a known state. Some of the boards supported by this
driver have output ports that are inverted from the comedi view of
the output state. For these boards the values written to the ports
needs to be inverted.

Currently, only bit 0 of each port is inverted when the boardinfo
indicates that the outputs are inverted. This results in channels
0, 8, 16, etc. being set to '0' and all other channels being set
to '1'. If the boardinfo does not indicate that the outputs are
inverted, all the channels are set to '0'.

This initialization is unnecessary for the input only ports. The
input/output ports also do not need to be initialized since they
are configured as inputs during the attach.

Move the output port initialization so it occurs when the digital
output subdevice is setup. Use the 's->io_bits' value to initialize
the ports so that the correct inverted/non-inverted state is used
for the comedi '0' value.
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>

Some of the boards supported by this driver have output ports that are
inverted from the comedi view of the output state. For these boards the
read values from the output ports needs to be inverted before being
modified and inverted again before being written back in the (*insn_bits)
operation.

Currently the subdevice type and the boardinfo is checked in the (*insn_bits)
to determine if the inverted outputs need to be handled.

Since thise driver does not use the subdevice 'io_bits', simplify the driver
a bit by initializing the 'io_bits' during the attach to handle the inversion.
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>

Remove the need for the 'output_bits' in the private data by just
reading the current state of the data port when updating the output
channels in the (*insn_bits) function.

Add a local variable to handle the inverting of the hardware values
when the boardinfo indicates that the outputs are inverted.

Rename the local variable 'port_read_bits' to 'bits' so that we can
use it for updating the output channels and reading back the actual
state to return to the user.
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>

For aesthetics, tidy up the whitespace of this function declarations to
follow the form in the rest of the driver.
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>

Update the MODULE_DESCRIPTION to better describe the driver.
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>

Clean up the multi-line comments at the beginning of the file so they
follow the kernel CodingStyle.
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>

For aesthetics, add some whitespace to the comedi_driver declaration.
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>

For aesthetics, add some whitespace to the subdevice init.
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>

Introduce some convienence macros to handle the port to channel,
channel to port, and channel mask calculations based on the 8
channels per port of the hardware.
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>

Refactor this function to follow the standard (*insn_config) form.

Add a sanity check of the number of data parameters (insn->n). Currently
the core does not check INSN_CONFIG_CHANGE_NOTIFY.

Fix the writes to the rise/fall edge enable registers. The macro expects
a "port" value not the port offset value.
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>

Subdevice 3 is used in this driver to provide edge detection of the input
channels.

Move the reset/disable of the interrupts and the request_irq() so that
when subdevice 3 is setup we can conditionally hookup the async command
support only if the irq is available.

Also, remove the noise when the irq is not available.
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>

There is only one member in the subdevice private data, an unsigned value
that is the 'base_port' that the subdevice uses to access the port registers.

Just cast the appropriate value into s->private instead of allocating the
private data for each subdevice. The casts are a bit of a nusance but it
removes the unnecessary allocations.
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>

The IO Select registers are readable. Remove the need for the
'dio_direction' member in the private data by just checking the
register value for the INSN_CONFIG_DIO_QUERY instruction.

Also, refactor the switch statement to return -EINVAL for unhandled
instructions and have the (*insn_config) return insn->n normally.
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>

The INSN_CONFIG_FILTER instruction is used to set the deglitch filter
interval used to debounce the input channels.

Absorb the helper function into the (*insn_config) function and refactor
the code to not require the 'filter_interval' and 'filter_enable' members
in 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>

Convert the inline functions used to calculate the offsets to the
recurring port registers and rename them to remove the CamelCase.

Define all the recurring registers.
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>

Rename the CamelCase defines used for the non-recurring registers.

Define all the non-recurring registers and 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>

According to the register programming manual, the filter interval
register is 32-bit. Fix the writes to this register.
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>

This patch adds the virtpci debugfs directory and the info entry
inside of it.
Signed-off-by: Erik Arfvidson <erik.arfvidson@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The mite module provides the DMA interface for the PCI MITE ASIC used on
many National Instruments DAQ boards. This driver does not use DMA and only
depends on the mite module to initialize the MITE ASIC.

Handle the initialization localy and remove the unnecessary dependancies.
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>

The mite module provides the DMA interface for the PCI MITE ASIC used on
many National Instruments DAQ boards. This driver does not use DMA and only
depends on the mite module to initialize the MITE ASIC.

Handle the initialization localy and remove the unnecessary dependancies.
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>