The NLMSG_PUT() macro contains a hidden goto that jumps to the
nlmsg_failure label. Since the sk_buff was allocated before the macro,
jumping to the nlmsg_failure label leaks the memory allocated for it.

Calling kfree() before returning would fix it, but is better to avoid
using this error prone macro and use nlmsg_put() instead.

Also, use nlmsg_data() instead of NLMSG_DATA() to check type.
Signed-off-by: NJavier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The default (return -EINVAL) case is not needed when working
out the correct value to set the analog input range. As pointed
out by Ian Abbott, the comedi core checks the range in
comedi_check_chanlist() before calling the insn_read() function.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reported-by: NIan Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The original comment about the PCMCIA support code was wrong in that
it said this driver was for the "pcm-das08". When it was moved the
comment was reworded badly.

Fix the comment so it makes sense.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Reported-by: NIan Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix a multi-line comment to follow the kernel CodingStyle.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The das16cs_ao_winsn() function has a couple commented out
debug messages. Just remove them.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The BIP_RANGE(a) macro can be used instead of the RANGE(a,b)
macro when -a == b. And it's a bit clearer that this is a bipolar
range.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The 'link' variable in the private data struct is not used by the
driver. Remove it.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Instead of open coding the 'dev->iobase + n' for each io register
in the device, use the provided #define's for the offsets.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Cleanup to analog input read function.

1) Initialize the chan, range, and aref locale variables when
   they are declared.
2) Remove need for the static local variable.
3) Remove the unnecessary cast of inw()'s return value.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Add some whitespace to the subdev initialization to improve
readability.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The comedi core expects a < 0 value during the attach to indicate
an error. The normal 'success' return to the kernel is '0' so use
that here.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Consolidate all the attach messages into one dev_info() output
at the end of the successful attach.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

For aesthetic reasons, reorder the boardinfo to match the
MODULE_DEVICE_TABLE. Also reorder the data to match the
struct definition and add some whitespace to improve
readability.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

During the attach, all of the device i/o registers are read and
the data is output as a dev_dbg() "fingerprint". This just adds
a bunch of noise during the loading of the driver. Remove the
output.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

During the attach of the comedi_driver, the device type probe
can fail. We should do the probe before requesting the irq
for the driver.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This driver is not for a pci device. Remove the unneeded include
of the pci.h header.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The analog output subdevice is only available on the -AO version
of the DAS16/16 device and the number of channels is provided in
the boardinfo. Make sure the subdevice is marked as unused for
devices that do not support the analog out.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The digital i/o subdevice is always initialized due to the 'if (1)'.
Simplify the attach by removing the test.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The timer subdevice is never initialized due to the 'if (0)'. The
comedi callbacks also don't do anything and just return -EINVAL.

Remove the subdevice and associated code.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

There are a number of comments in this driver that are cut-and-paste
from the skel driver. They are not needed in "real" drivers.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The 'n_boards' define is only used one place in the driver.
Just put the ARRAY_SIZE() where used and remove the define.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The 'thisboard' and 'devpriv' macros rely on a local variable having
a specific name and yield pointers derived from that local variable.
Replace the macro with local variables where used. Use the
comedi_board() helper to get the 'thisboard' pointer.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the comedi_driver attach/detach functions, as well as the
probe function used during the attach. This removes the need
for all the forward declarations.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the comedi_driver variable to remove the need for the
forward declarations. Add some whitespace to the declaration
for aesthetic reasons.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The pcmcia device-specific data is not longer needed by this
driver. Remove it.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The pcmcia suspend/resume callbacks don't do anything. Remove
them.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The pcmcia support code in this driver appears to be cut-and-
paste from some other driver. It has code in it to stop the
device during suspend but nothing in the main comedi_driver
uses it.

Remove the 'stop' variable from the pcmcia private data and
all the logic that deals with it.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the pcmcia_disable_device() call where needed in the
pcmcia attach/detach and delete the das16cs_pcmcia_release()
function.

Move the logic of das16cs_pcmcia_config() directly into the
attach function and properly return an error code when the
config fails.

Only set the cur_dev, used by the comedi_driver, if the pcmcia
attach is successful. Also, make sure to NULL it in the detach.

Remove all the kernel messages in the pcmcia support code. They
are just added noise.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Refactor the pcmcia support code to remove the need for the
forward declarations.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove some useless comments and whitespace.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

For aesthetic reasons, add some whitespace to the declaration
of the pcmcia_driver and reorder it a bit.

Also, the symbol 'das16cs_driver' is only referenced in this file,
make it static. This quiets the following sparse warning:

warning: symbol 'das16cs_driver' was not declared. Should it be static?
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

As done with the other comedi driver types, register the
comedi_driver first then the pcmcia_driver. Also, make
sure the pcmcia_driver registration succeeds and unregister
the comedi_driver it it fails.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The register/unregister of the pcmcia driver is done is separate
functions that are called by the module_{init,exit} routines.
Simplify the code a bit by moving the register/unregister into
the module_{init,exit} routines.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The init and exit functions for this module were being declared
as global symbols. Add the module_{init,exit} declarations and
make the functions static. Also, rename the functions so they
have namespace associated with the module.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the MODULE_* declarations to the end of the file.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The Kconfig ensures that this driver can only be build it PCMCIA is
enabled. Remove the unneeded '#if defined (CONFIG_PCMCIA) ,,,' and
the '#else' code.
Signed-off-by: NH Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Frank Mori Hess <fmhess@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the CONFIG_COMEDI_MITE config options into the `if
COMEDI_PCI_DRIVERS` block so that the `depends on PCI` condition can be
omitted.  Move the CONFIG_COMEDI_NI_TIOCMD config option to the same
block as it selects COMEDI_MITE.  Move the CONFIG_COEMDI_NI_TIO config
option slightly for aesthetic reasons without changing its conditions.
Signed-off-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

At the moment, the options to select drivers for various National
Instruments boards in the "Comedi PCI drivers", "Comedi ISA and PC/104
drivers" and "Comedi PCMCIA drivers" menus do not appear unless you
first enable and enter the "Comedi National Instruments card support"
menu and possibly select some of the options in there.  This is pretty
confusing for the user.

Get rid of the "Comedi National Instruments card support" menu
(CONFIG_COMEDI_NI_COMMON) and make the CONFIG_COMEDI_MITE,
CONFIG_COMEDI_NI_TIO and CONFIG_COMEDI_NI_TIOCMD options selectable by
other options instead of dependencies of those other options.

We lose the ability to build the 'mite', 'ni_tio' and 'ni_tiocmd'
modules independently of other modules, but they are not useful on their
own anyway.
Signed-off-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Move the CONFIG_COMEDI_NI_LABPC option ("NI Lab-PC and compatibles ISA
and PCI support") from the CONFIG_COMEDI_NI_COMMON menu ("Comedi
National Instruments card support  --->") to the
CONFIG_COMEDI_PCI_DRIVERS menu ("Comedi PCI drivers  --->").

It currently depends on PCI but should be split up into separate ISA and
PCI config options eventually as was done for the "das08" driver.  For
now, the PCI card menu seems the best place for it and I plan to get rid
of the CONFIG_COMEDI_NI_COMMON menu altogether.
Signed-off-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Selecting the CONFIG_COMEDI_NI_TIO config option causes the 'ni_tio' and
'ni_tiocmd' modules to be built. CONFIG_COMEDI_NI_TIO depends on
CONFIG_COMEDI_MITE which in turn depends on CONFIG_PCI.  However, not
all the drivers that need the 'ni_tio' module also need the 'ni_tiocmd'
module.  Specifically, the ISA and PCMCIA drivers do not need the
'ni_tiocmd' module.

Add a new config option CONFIG_COMEDI_NI_TIOCMD to control building of
the 'ni_tiocmd' module.  It depends on CONFIG_COMEDI_NI_TIO and
CONFIG_COMEDI_MITE.  The existing CONFIG_COMEDI_NI_TIO option no longer
needs to depend on CONFIG_COMEDI_MITE.

Make CONFIG_COMEDI_NI_660X ('ni_660x' module) and
CONFIG_COMEDI_NI_PCIMIO ('ni_pcimio' module) depend on
CONFIG_COMEDI_NI_TIOCMD instead of CONFIG_COMEDI_NI_TIO.
Signed-off-by: NIan Abbott <abbotti@mev.co.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>