iio: __iio_update_buffers: Perform request_update() only for new buffers

We only have to call the request_update() callback for a newly inserted
buffer. The configuration of the already previously active buffers will not
have changed.

This also allows us to move the request_update() call to the beginning of
__iio_update_buffers(), before any currently active buffers are stopped.
This makes the error handling a lot easier since no changes were made to
the buffer list and no rollback needs to be performed.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: __iio_update_buffers: Slightly refactor scan mask memory management

Add a small helper function iio_free_scan_mask() that takes a mask and
frees its memory if the scan masks for the device are dynamically
allocated, otherwise does nothing. This means we don't have to open-code
the same check over and over again in __iio_update_buffers.

Also free compound_mask as soon a we are done using it. This constrains its
usage to a specific region of the function will make further refactoring
and splitting the function into smaller sub-parts more easier.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Replace printk in __iio_update_buffers with dev_dbg

While more verbose error messages are useful for debugging we should really
not put those error messages into the kernel log for normal errors that are
already reported to the application via the error code, when running in
non-debug mode.

Otherwise application authors might expect that this is part of the ABI and
to get the error they should scan the kernel log. Which would be rather
error prone itself since there is no direct mapping between a operation and
the error message so it is impossible to find out which error message
belongs to which error.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: add m62332 DAC driver

m62332 is a simple 2-channel DAC used on several Sharp Zaurus boards to
control LCD voltage, backlight and sound. The driver use regulators to
control the reference voltage and enabling/disabling the DAC.

Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: ltr501: Add light channel support

Added support to calculate lux value from visible
and IR spectrum adc count values. Also added IIO_LIGHT
channel to enable user read the lux value directly
from device using illuminance input ABI.

Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: pressure: hid-sensor-press: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: orientation: hid-sensor-rotation: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: orientation: hid-sensor-incl-3d: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: hid-sensor-als.c: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: gyro: hid-sensor-gyro-3d: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: accel: hid-sensor-accel-3d: Fix memory leak in probe()

'channels' is allocated via kmemdup and it is never freed in the
subsequent error paths.

Use 'indio_dev->channels' directly instead, so that we avoid such
memory leak problem.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: ltr501: Fix proximity threshold boundary check

Currently, proximity sensor boundary check is done
inside the switch block but outside the case
statement.Since this code will never get executed,
moved the check outside the switch case statement.

   867          case IIO_PROXIMITY:
   868                  switch (dir) {
   // Following line has been moved outside the switch block.
   869                  if (val > LTR501_PS_THRESH_MASK)
   870                          return -EINVAL;
   871                  case IIO_EV_DIR_RISING:

Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio:temp:mlx90614 trivial drop of unnecessary ret return from write_raw.

This is mostly part of an effort to clean out our current warnings
and make the autobuilder build reports more useful.

Still a worthwhile if trivial cleanup!

Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Cc: Vianney le Clément de Saint-Marcq <vianney.leclement@essensium.com>
Cc: Peter Meerwald <pmeerw@pmeerw.net>

iio: adc: ti_am335x_adc: make sample delay, open delay, averaging DT parameters

Add optional DT properties to set open delay, sample delay and number
of averages per sample for each adc step. Open delay, sample delay
and averaging are some of the parameters that affect the sampling rate
and accuracy of the sample. Making these parameters configurable via
DT will help in balancing speed vs accuracy.

Signed-off-by: Vignesh R <vigneshr@ti.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: acpi: Add support for ACPI0008 Ambient Light Sensor

This driver adds the initial support for the ACPI Ambient Light Sensor
as defined in Section 9.2 of the ACPI specification (Revision 5.0) [1].

Sensors complying with the standard are exposed as ACPI devices with
ACPI0008 as hardware ID and provide standard methods by which the OS
can query properties of the ambient light environment the system is
currently operating in.

This driver currently allows only to get the current ambient light
illuminance reading through the _ALI method, but is ready to be
extended extended to handle _ALC, _ALT and _ALP as well.

[1] http://www.acpi.info/DOWNLOADS/ACPIspec50.pdf

Signed-off-by: Martin Liska <marxin.liska@gmail.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Gabriele Mazzotta <gabriele.mzt@gmail.com>
Cc: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

imu: inv_mpu6050: adds device tree bindings

Signed-off-by: Varka Bhadram <varkab@cdac.in>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-press: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-rotation: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-incl-3d: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-magn-3d: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: hid-sensor-prox: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: hid-sensor-als: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-gyro-3d: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: adc: axp288: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensor-accel-3d: Constify platform_device_id

The platform_device_id is not modified by the driver and core uses it as
const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: add support for ROHM BH1710/BH1715/BH1721/BH1750/BH1751 ambient light sensors

Add support for ROHM BH1710/BH1715/BH1721/BH1750/BH1751 ambient light
sensors.

Signed-off-by: Tomasz Duszynski <tduszyns@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: make tools more cross-compilation friendly

When cross-compiling the IIO tools we need to opportunity to
specify a cross compiler prefix and some extra CFLAGS. This
patch enables this in the same way as for other stuff in
tools.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Daniel Baluta <daniel.baluta@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magnetometer: Add ACPI support for MMC35240

We assume that ACPI device tables use MMC35240 to
identify MEMSIC's 3 axis magnetic sensor.

Signed-off-by: Daniel Baluta <daniel.baluta@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magnetometer: mmc35240: Add PM sleep support

We rely on regmap to save the state of the registers at suspend,
and then we do an explicit sync at resume.

Signed-off-by: Daniel Baluta <daniel.baluta@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magnetometer: Add support for MEMSIC MMC35240 sensor

Minimal implementation for MMC35240 3-axis magnetometer
sensor. It provides processed readings and possiblity to change
the sampling frequency.

Signed-off-by: Daniel Baluta <daniel.baluta@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magn: bmc150_magn: add oversampling ratio

Export the oversampling ratio so that the user can change the
number of repetions for x/y/z axis.

The sampling frequency is limited by the oversampling ratio.
The available sampling frequencies might change depending
on the values of oversampling_ratio.

The specification can be downloaded from:
http://ae-bst.resource.bosch.com/media/products/dokumente/bmc150/BST-BMC150-DS000-04.pdf.

Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: core: Introduce IIO_CHAN_INFO_OVERSAMPLING_RATIO

Some magnetometers can perform a number of repetitions in HW
for each measurement to increase accuracy. One example is
Bosch BMC150:
http://ae-bst.resource.bosch.com/media/products/dokumente/bmc150/BST-BMC150-DS000-04.pdf.

Introduce an interface to set the oversampling ratio
for these devices.

Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magn: bmc150_magn: Add devicetree binding documentation

Add binding documentation for Bosch BMC150 magnetometer.

Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: magn: Add support for BMC150 magnetometer

Add support for the Bosh BMC150 Magnetometer.
The specification can be downloaded from:
http://ae-bst.resource.bosch.com/media/products/dokumente/bmc150/BST-BMC150-DS000-04.pdf.
The chip contains both an accelerometer and a magnetometer.
This patch adds support only for the magnetometer part.

The temperature compensation formulas are based on bmm050_api.c
authored by contact@bosch.sensortec.com.

Signed-off-by: Irina Tirdea <irina.tirdea@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Staging: iio: Allow compile test of GPIO consumers if !GPIOLIB

The GPIO subsystem provides dummy GPIO consumer functions if GPIOLIB is
not enabled. Hence drivers that depend on GPIOLIB, but use GPIO consumer
functionality only, can still be compiled if GPIOLIB is not enabled.

Relax the dependency on GPIOLIB if COMPILE_TEST is enabled, where
appropriate.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jonathan Cameron <jic23@kernel.org>
Cc: linux-iio@vger.kernel.org
Cc: devel@driverdev.osuosl.org
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: Allow compile test of GPIO consumers if !GPIOLIB

The GPIO subsystem provides dummy GPIO consumer functions if GPIOLIB is
not enabled. Hence drivers that depend on GPIOLIB, but use GPIO consumer
functionality only, can still be compiled if GPIOLIB is not enabled.

Relax the dependency on GPIOLIB if COMPILE_TEST is enabled, where
appropriate.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jonathan Cameron <jic23@kernel.org>
Cc: linux-iio@vger.kernel.org
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: Add threshold interrupt support for STK3310

Added interrupt support for proximity threshold events
to the stk3310 driver.

Signed-off-by: Tiberiu Breana <tiberiu.a.breana@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: light: Add support for Sensortek STK3310

Minimal implementation of an IIO driver for the Sensortek
STK3310 ambient light and proximity sensor. The STK3311
model is also supported.

Includes:
- ACPI support;
- read_raw and write_raw;
- reading and setting configuration parameters for gain/scale
  and integration time for both ALS and PS.
- power management

Signed-off-by: Tiberiu Breana <tiberiu.a.breana@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Merge tag 'iio-for-v4.2a' of git://git./linux/kernel/git/jic23/iio into staging-next

Jonathan writes:

First round of new drivers, functionality and cleanups for the 4.2 cycle

New drivers / device support
* st sensors driver, lsm303dlh magnetometer support.
* ltr501 - support ltr301 and ltr559 chips.

New functionality
* IIO_CHAN_INFO_CALIBEMISSIVITY for thermopile sensors.
* kxcjk1013 - make driver operational with external trigger.
* Add iio targets to the tools Makefile.

Cleanups
* st sensors - more helpful error message if device id wrong or irq request
  fails, explicitly make the Block Data Update optional rather
  than relying on writes to address 0 not doing anything, make interrupt
  support optional (Not always wired, and not all devices actually have
  an interrupt line.)
* kxcjk-1013 white space additions for readability, add the KXCJ9000 ACPI
  id as seen in the wild.
* sx9500 - GPIO reset support, refactor the GPIO interrupt code, add power
  management, optimize power usage by powering down when possible, rename
  the gpio interrupt pin to be more useful, trivial return path simplification,
  trivial formatting fixes.
* isl29018 -  move towards ABI compliance with a view to moving this driver
  out of staging, add some brackets to ensure code works as expected.  Note
  there is no actual bug as the condition being tested is always true
  (with current devices).
* ltr501 - add regmap support to get caching etc for later patches,
  fix a parameter sanity check that always fails (bug introduced
  earlier in this series), ACPI enumeration support,
  interrupt rate control support, interrupt support in general and
  integration time control support, code alignment cleanups.
* mma9553 - a number of little cleanups following a review from Hartmut
  after I'd already applied the original driver patch.
* tmp006 - prefix some defines with TMP006 for consistency.
* tsl4531 - cleanup some wrong prefixes, presumably from copy and paste.
* mlx90614 - check for errors in read values, add power management,
  add emissivity setting, add device tree binding documentation,
  fix a duplicate const warning.
* ti_am335x_adc - refactor the DT parsing into a separate function.

staging: unisys: uislib.h: Remove unused cache allocation prototypes

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchipset: Remove unused cache allocator

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: Remove wrapper around parser_init_guts()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: parser_init_guts(): standard_payload_header is always false

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: parser: Remove unused functions and mark others static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchipset: Declare parser_init_byte_stream() static

In addition remove unused parser_init()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: Make visorchannel_create take a gfp_t

This allows the caller to specify an appropriate GFP flag instead of
hardcoding the lowest common denominator.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: remove visorutil from top level Makefile

The visorutil directory is still mentioned in the top level makefile for
the Unisys drivers, so remove it.

Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: fix visorbus Kconfig

Removing visorutil made it impossible to build visorbus. Remove the config
setting from the Kconfig so the module can be enabled in the build.

Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: move periodic_work.c into the visorbus directory

This patch removes visorutil directory, move periodic_work.c into
the visorbus directory.

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>

staging: unisys: remove timskmod.h and procobjecttree.h

This patch move the needed linux include files from timskmod.h
to the files that calls those include. Also procobjecttree.h is
removed since it is dead code.

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>

staging: unisys: remove charqueue.c

This patch removes charqueue.[ch] since it no longer called

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>

staging: unisys: move timskmod.h functionality

This patch removes all timksmod.h pound defines. It also removes
visorkmodutils.c since it no longer has any use by itself. Since
visorkmodutils.c is no longer needed the module_init for
visorchipset.c is modified to call visorutil_spar_detect directly
instead of the extern variable in timksmod.h.

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>

staging: unisys: add ACPI and PCI requirement to Kconfig

Later patches will require PCI and ACPI support in the kernel, so add these
features to the Kconfig.

Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: add acpi pnp driver

According to Unisys, another OS detects the PNP0A07 as the auto load
device.  We can also do this in the linux kernel by simply converting the
driver over to the ACPI driver model.

Notes: This changes the usage of __init and it had to be removed from some
functions to avoid a !__init function calling an __init function.
Additionally I also cleaned up the headers in visorchipset.c since I was
adding a header file.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: remove BOOL,TRUE,FALSE definitions

These shouldn't be defined in the code and can be replaced with the
standard bool, true, and false usage that the kernel uses.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: Finally remove the last remnants of memregion

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: Eliminate visor_memregion_write()

visorchannel's signal code should call visorchannel_write() directly.
This is more consistent and cleaner, and allows us to remove the last
memregion call.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: Eliminate visor_memregion_read()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_create_guts(): Use visorchannel_read()

There is no benefit to calling visor_memregion_read() at this point.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_write(): Handle partial channel_header writes

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_read(): Use memcpy_fromio() directly

Note, this changes the behavior of visorchannel_read(). The old code
would return the channel header, if the offset argument was 0, and the
caller tried to read beyond the size of the visorchannel. Note this
only worked for offset == 0, but not for
(offset > 0) && (offset < header_size), which was inconsistent.

The new implementation returns an error if someone tries to read
beyond the visorchannel size.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_write(): Use memcpy_toio() directly

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: {un, }mapit() are no longer used

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Eliminate visor_memregion_get_*() functions

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Eliminate visor_memregion_resize()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Eliminate visor_memregion_destroy()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: Eliminate visor_memregion_create()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: get rid of serialloopbacktest

serialloopbacktest is no longer supported. Get rid of it.

Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: cleanup and align iochannel.h comments

This patch reorganizes, aligns, and corrects grammar mistakes on
comments.

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>

staging: unisys: remove unused #define in iochannel.h

This patch simply remove all unused #define's in iochannel.h

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>

staging: unisys: remove unused chipset feature PCIVBUS

This patch removes unused chipset feature
ULTRA_CHIPSET_FEATURE_PCIVBUS

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>

staging: unisys: remove unused CONTROLVM_CHIPSET_SHUTDOWN

this patch removes unused chipset command on the guest
CONTROLVM_CHIPSET_SHUTDOWN

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>

staging: unisys: remove unused device_map and guest_devices struct

This patch removes unused struct in the controlvm_channel message
and replaces it with a char reserved inorder to mantain same
message size.

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>

staging: unisys: controlvmchannel.h comment aligment and cleanup

This patch is a comment aligment and cleanup for
controlvmchannel.h.

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>

staging: unisys: rename variable to reserverd since it is unused

This patch renames send_irq_handle to reserved1 since this variable
is unused in the guest side.

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>

staging: unisys: remove unused CONTROL_VM messages from enum

This patch removes CONTROL_VM  disk messages from enum since they
are completely unused.

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>

staging: unisys: remove server crust from visorchipset.

The visorchipset driver originally serviced both servers and
clients. This implementation is client only so remove some
more server side implementation.

Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: remove unused #define MAX_SERIAL_NUM

This patch simply removes #define MAX_SERIAL_NUM from iochannel.h

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>

staging: unisys: remove unused #define in controlvmchannel

This patch remove unused controlvmchannel.h #defines

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>

staging: unisys: remove unused enum from controlvmchannel.h

This patch removes this enum since it is unused

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>

staging: unisys: remove typedef GUEST_PHYSICAL_ADDRESS to u64

This patch removes typedef GUEST_PHYSICAL_ADDRESS to u64

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>

staging: unisys: Eliminate unused visorchannel_dump_section()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_get_memregion() isn't used

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Embed memregion in struct channel

This changes the API for visor_memregion_create() to require a
pre-allocated struct memregion. Embedding this in struct channel
avoids a layer of additional kmalloc()'s and error checks.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: move struct memregion to memregion.h

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: decouple visor_memregion_{read, write}()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_clear(): No need to use vmalloc here

Using a page is sufficient, and avoids the cost of vmalloc/vfree

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel_clear(): Avoid 64KB memory leak

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchipset.c: No need to include memregion.h anymore

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchipset: Use ioremap direction rather than heavy visor_memregion

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchipset: parser_init_guts(): Localize memregion usage

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Eliminate unnecessary 'requested' flag

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Use proper errno for mapit()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: memregion: Nothing uses overlap allocations, so nuke it

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorbus: visorchannel_create_guts() always has parent == NULL

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel: visorchannel_create_overlap() is never used

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: fix sig_read_data and sig_read_data functions

The sig_read_data() and sig_write_data() functions are involved in 2 steps
of calls.  They really don't need to be and this makes for much simpler
code.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel some general function cleanups

Just some cleanups for visorchannel.c, and removal of
safe_sig_queue_validate() which is dead code.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: visorchannel cleanup visorchannel_create_guts()

The error handling in this function was broken and while looking at that
I noticed that the whole function was in need of cleanup.  This patch
fixes the error handling, specifically

                if (!p) {
                        visorchannel_destroy(p);
                        channel = NULL;
                }

and does a lot of cleanup.  I also verified that the called functions
returned correct errors, and that led to a change in
visor_memregion_resize(), visorchannel_destroy() and
visor_memregion_destroy().

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: unisys: cleanup visorbus_private.h

visorbus_private.h contains code that is called from visorbus into the
visorchipset code.  Now that the visorchipset code has been brought into
the visorbus directory, many of the declarations are not necessary and
can be cleaned up.

TODO: PARSER_WHICH_STRING enum only has one member used
(PARSERSTRING_NAME).

TODO: crash_obj_type appears to be unnecessary in the overall scheme of
code.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>