As found by "gcc -Wmaybe-uninitialized", the latest change to the
driver lacked an initalization for the return code in one of the
added cases:

drivers/staging/iio/cdc/ad7746.c: In function ‘ad7746_read_raw’:
drivers/staging/iio/cdc/ad7746.c:655:2: error: ‘ret’ may be used uninitialized in this function [-Werror=maybe-uninitialized]

This sets it to IIO_VAL_INT, which I think is what we want here.

Fixes: 2296c062 ("staging: iio: cdc: ad7746: implement IIO_CHAN_INFO_SAMP_FREQ")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

This adds a new driver for the Invensense MPU-3050 gyroscope.
This driver is based on information from the rough input driver
in drivers/input/misc/mpu3050.c and the scratch misc driver
posted by Nathan Royer in 2011. Some years have passed but this
is finally a fully-fledged driver for this gyroscope. It was
developed and tested on the Qualcomm APQ8060 Dragonboard.

The driver supports both raw and buffered input. It also
supports the internal trigger mechanism by registering a trigger
that can fire in response to the internal sample engine of the
component. In addition to reading out the gyroscope sensor
values, the driver also supports reading the temperature from
the sensor.

The driver currently only supports I2C but the MPU-3050 can
also be used from SPI, so the I2C portions are split in their
own file and we just use regmap to access all registers, so
it will be trivial to plug in SPI support if/when someone has
a system requiring this.

To conserve power, the driver utilizes the runtime PM
framework and will put the sensor in off mode and disable the
regulators when unused, after a timeout of 10 seconds.

The fullscale can be set for the sensor to 250, 500, 1000 or
2000 deg/s. This corresponds to scale values of rougly 0.000122,
0.000275, 0.000512 or 0.001068. By writing such values (or close
to these) into "in_anglevel_scale", the corresponding fullscale
can be chosen. It will default to 2000 deg/s (~35 rad/s).

The gyro component can have DC offsets on all axes. These can be
compensated using the standard sysfs ABI property
"in_anglevel_[xyz]_calibbias". This is in positive/negative
values of the raw values, so a suitable calibration bias can be
determined by userspace by reading the "in_anglevel_[xyz]_raw"
for a few iterations while holding the sensor still, create an
average integer, and writing the negative inverse of that into
"in_anglevel_[xyz]_calibbias". After this the hardware will
automatically subtract the bias, also when using buffered
readings.

Since the MPU-3050 has an outgoing I2C port it needs to act as
an I2C mux. This means that the device is switching I2C traffic
to devices beyond it. On my system this is the only way to reach
the accelerometer. The "sensor fusion" ability of the MPU-3050
to directly talk to the device on the outgoing I2C port is
currently not used by the driver, but it has code to allow I2C
traffic to pass through so that the Linux kernel can reach the
device on the other side with a kernel driver.

Example usage with the native trigger:

$ generic_buffer -a -c10 -n mpu3050
iio device number being used is 0
iio trigger number being used is 0
No channels are enabled, enabling all channels
Enabling: in_anglvel_z_en
Enabling: in_timestamp_en
Enabling: in_anglvel_y_en
Enabling: in_temp_en
Enabling: in_anglvel_x_en
/sys/bus/iio/devices/iio:device0 mpu3050-dev0
29607.142578 -0.117493 0.074768 0.012817 180788797150
29639.285156 -0.117493 0.076904 0.013885 180888982335
29696.427734 -0.116425 0.076904 0.012817 180989178039
29742.857422 -0.117493 0.076904 0.012817 181089377742
29764.285156 -0.116425 0.077972 0.012817 181189574187
29860.714844 -0.115356 0.076904 0.012817 181289772705
29864.285156 -0.117493 0.076904 0.012817 181389971520
29910.714844 -0.115356 0.076904 0.013885 181490170483
29917.857422 -0.116425 0.076904 0.011749 181590369742
29975.000000 -0.116425 0.076904 0.012817 181690567075
Disabling: in_anglvel_z_en
Disabling: in_timestamp_en
Disabling: in_anglvel_y_en
Disabling: in_temp_en
Disabling: in_anglvel_x_en

The first column is the temperature in millidegrees, then the x,y,z
axes in succession followed by the timestamp. Also tested successfully
using the HRTimer trigger.

Cc: Nick Vaccaro <nvaccaro@google.com>
Cc: Ge Gao <ggao@invensense.com>
Cc: Anna Si <asi@invensense.com>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: Crestez Dan Leonard <leonard.crestez@intel.com>
Cc: Daniel Baluta <daniel.baluta@intel.com>
Cc: Gregor Boirie <gregor.boirie@parrot.com>
Cc: Peter Rosin <peda@axentia.se>
Cc: Peter Meerwald-Stadler <pmeerw@pmeerw.net>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

This adds device tree bindings for the MPU-3050 gyroscope. Since it
is the first set of bindings for a gyroscope, the folder for it
is also created.

Cc: devicetree@vger.kernel.org
Cc: Peter Rosin <peda@axentia.se>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

gb_audio_manager_module_descriptor's intf_id field maintains the
information about the interface on which module is connected hence
having an extra slot field is redundant.

Thus remove the slot field and its associated code.
Signed-off-by: Pankaj Bharadiya <pankaj.bharadiya@gmail.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Vaibhav Agarwal <vaibhav.sr@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes the following checkpath.pl warning
WARNING: Block comments use a trailing */ on a separate line
Signed-off-by: Rahul Krishnan  <mrahul.krishnan@gmail.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix size field of arpc message request by using the header size and not
the pointer size.
Signed-off-by: Rui Miguel Silva <rmfrfs@gmail.com>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix checkpatch warning:

WARNING: Prefer kstrto<type> to single variable sscanf

kstrto* is designed to convert string to numerical value and makes
it easier to understand what the code does.
Signed-off-by: Elise Lennion <elise.lennion@gmail.com>
Reviewed-by: Mark Greer <mgreer@animalcreek.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The removed parentheses are unnecessary and don't add readability.

Found using Coccinelle semantic patch:
@@ expression e, e1, e2; @@
e &=
(
        (e1 == e2)
|
-       (e1)
+       e1
)

@@ expression e, e1, e2; @@
e |=
(
        (e1 == e2)
|
-       (e1)
+       e1
)
Signed-off-by: Elise Lennion <elise.lennion@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The removed parentheses are unnecessary and don't add readability.

Found using Coccinelle semantic patch:
@@ expression e, e1, e2; @@
e +=
(
        (e1 == e2)
|
-       (e1)
+       e1
)

@@ expression e, e1, e2; @@
e =
(
        (e1 == e2)
|
-       (e1)
+       e1
)
Signed-off-by: Elise Lennion <elise.lennion@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch add const qualifier at the declaration of the structure.
The structure become a read-only data, this increase the security.

Found with Coccinelle:
@r disable optional_qualifier@
identifier s,i;
@@
* static struct s i ={...};
Signed-off-by: Mihaela Muraru <mihaela.muraru21@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The removed parentheses are unnecessary and don't add readability.

Found using Coccinelle semantic patch:
@@ expression e, e1, e2; @@
e =
(
        (e1 == e2)
|
-       (e1)
+       e1
)
Signed-off-by: Elise Lennion <elise.lennion@gmail.com>
Acked-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

When compiling vchiq_core.c for 64 bit, the compiler
emits a few warnings that are not actual issues. This
change adds a few casts to remove the extra unnecessary
warnings.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

A VCHIQ_SERVICE_HANDLE_T which is an int is stuffed into a
VCHI_SERVICE_HANDLE_T which is a pointer, passed around, then
converted back to a VCHIQ_SERVICE_HANDLE_T. Since the data is
always actually a VCHIQ_SERVICE_HANDLE_T(int), never actually a
pointer, it is safe to simply cast the two back in forth.

Note that pointers are never stuffed into an int.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The arm processor core and the GPU have a shared data structure.
This structure contains pointers to base linux kernel objects such as
events.  The size of the pointer changes between 32 bit and 64 bit,
so it is necessary to convert these pointers to offsets from the
beginning of the state structure.

Luckly, the GPU does not interpret these pointers/offsets,
but this change is necessary to keep the structure the same since
the GPU code is outside the scope of the linux kernel
and can't be easily changed.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The arm processor core and the GPU have shared data structures.
One of these structures is a list of pages of data for messages.
This structure can not change since it is dependent on the GPU
firmware which is external to the kernel.  Convert the fields
of this structure to fixed length fields.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The dsb() macro for 32 arm compiles to dsb(sy) in the binary file.
This macro is no longer supported on arm64, so instead use dsb(sy)
which is completely binary compatible.
Signed-off-by: Michael Zoran <mzoran@crowfest.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'iio-for-4.10a' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-next

Jonathan writes:

First round of IIO new device support, features and cleanups for the 4.10 cycle.

Fair number of outreachy related patches in here. Some of these may well
have already been picked up by Greg but git will sort that out for us.

Also some good staging cleanup work from other sources. Thanks Brian and Lars
in particular for this.

New device support
* ACCES 104-quad-8
  - New driver for this 8 channel encoder input board. Lots of new ABI with
    this one.
* AD7766
  - New driver supporting AD7766, AD7766-1, AD7766-2, AD7767, AD7767-1 and
    AD7767-2 24 bit ADCs.
* dmard 10
  - New driver for this 3 axis accelerometer.
* Honeywell ABP pressure sensors.
  - New driver covering 56 parts in this series (too many to list here!)
* HTS221
  - New driver to support this relative humidiy and temperature device.
* LMP91000
  - New driver for this potentiostat (form of chemical sensor). Nice example
    of use of the buffered consumer interfaces and the use of a consumer
    provided trigger.
* MiraMEMS DA311
  - New driver for this 3 axis accelerometer.
* MiraMEMS DA280
  - New driver for this 3 axis accelerometer. Follow up caught up with
  vendor prefixes for these.

Staging graduations
* isl29018 light sensor
  - Fixes and cleanups listed below (thanks for your hard work on this Brian!)
* sca3000
  - Fixes and cleanups listed below.  This was one of the small set of drivers
  that went into staging when IIO was first added.  Turns out it had a few
  bugs and needed to be brought into the modern era!  Not clear if I am
  the only person who actually has one of these still wired to a board.

New features (Core)
 - Add an iio_trigger_validate_own_device helper which relies on the device
   and trigger having the same parent.  Convenient to have this for some
   of the more complex trigger / device interactions. Was hand rolled in
   a few drivers already so good to bring it into the core.
 - Add an iio_read_channel_offset in kernel access helper (similar to
   the existing one for scale).
 - IIO_ATTR_{RO, WO, RW} and IIO_DEVICE_ATTR_{RO, WO, RW} macros.  These
   lead some rather contrived function naming, but there is no denying they
   do reduced boilerplate.  I'm going to resist their introduction in
   drivers 'unless' they form part of a larger set of cleanups.
 - Counter channel type and index type.

New features (Drivers)
* hdc100x
  - Triggered buffer support.
* mcp4725
  - Device tree bindings and support.
  - Voltage reference selection.
* ti-adc0832
  - Triggered buffer support.
* ti-adc161s626
  - Add regulator support allowing _scale and _offset values to be established
  and exported.

New features (Tools)
* iio_generic_buffer
  - -A option to force enable all channels rather than faulting if some are
  already enabled (like -a does).  Followup patches tidied this support up.

Cleanups (Core)
 - Use kmalloc_array in iio_scan_mask_set.
 - Take event_attrs field of iio_info structure constant
 - Staging todo list updates. Most of it was long done.
 - MAINTAINERS had a wrong directory listing.

Cleanups (Drivers)
* Missing i2c trivial devices entries.
* ad5592r
  - Fix an endian type related sparse warnings.
* ad7150
  - Constify the event attribute_group structures.
* ad7152
  - Add some blank lines to improve readability.
  - Sampling frequency control via chan-info element rather than hand rolled
  attributes.
  - add a new lock to avoid use of mlock for non state change related locking.
* ad7280
  - Constify atrribute_group structure (second patch covers the event ones)
* ad7606 (Lars is driving most of the cleanup on this with some additions from
  Eva)
  - Fix improper setting of oversampling pins.  This has been broken a very
  long time in this staging driver, so not going to push this back to stable.
  - Implement oversampling configuration via the chan_info mask element.
  - Remove an unused int_vref_mv field.
  - Remove a reundant name field from ad7606_chip_info.
  - Remove default device configuration from platform_data in favour of
  whatever the power on defaults are.
  - Remove out of band error reporting in the kernel log as not providing
  much information.
  - Fix oversampling ratio by having 1 be the value for no oversampling.
  - Avoid allocating buffer for each data capture.
  - Factor out common code between periodic and one-shot capture.
  - Move set_drvdat into common code.
  - Let the common probe function return int rather than jumping through
  an ERR_PTR.
  - Pass struct device * into common remove to simplify code.
  - Always run trigger handler only once per event (no one can remember why
  it was being possibly done twice).
  - Move over to the GPIO descriptor API to shorten and clarify code.
  - Move the buffer code into the main file as it's not optional and is
  now rather short in this driver.
  - Fix the naming of the supply regulator.
  - Rework regulator handling to handle errors including deferred probing.
  - Tidy up a ptr_err or 0 return.
* ad7746
  - Sampling frequency control via info_mask element rather than hand rolled
* ad7758
  - Sampling frequency control via info_mask element rather than hand rolled
  attributes.
* ad7816
  - Constify the event attribute_group structure.
* adt7316
  - Constify the event attribute group structures.
* ak8974
  - Cleanup some sparse warnings about endian types.
* ak8975
  - Cleanup some sparse warnings about endian types.
* bmi160
  - Spare endian warning cleanups.
* isl29018 (towards staging graduation)
  - Remove unusedvariables and defines.
  - Improve consistency of error handling.
  - Signed / unsigned comparison fixes.
  - Use the IIO_DEVICE_ATTR_{RO, RW} macros
  - Fix a race in in_illuminance_scale_available_show.
  - Cleanup exit points of _read_raw
  - Sanity check if in suspended state during a write_raw call as was already
    done for read_raw.
  - Document device tree bidnings.
  - Document infrared supression controls.
  - Add some newlines to improve readability and drop one that shouldn't be
    there.
  - Fix a poorly named functions name.
  - Fix multiline coment syntax.
  - Tidy up a pair or return statements by unifying them.
  - Rename description in Kconfig for consistency with similar drivers.
* lidar
  - cleanup power management by dropping unnecessary call.
* ltr501
  - Use the claim_direct_mode helpers. Fix a race condition along the way.
* max1027
  - Fix a dubious x | !y sparse warning.
  - Use the new iio_trigger_validate_own_device helper.
* max440000
  - Clean up some sparse warnings about endian types.
* mcp4725
  - Use the regulator framework to establish the reference voltage rather than
    getting it from platform data.
  - Tidy up a comment typo.
  - Fix a wrong PTR_ERR query (wrong regulator).
* mma7660
  - Take a mma7660_nscale static.
* mma8452
  - Use the new iio_trigger_validate_own_device helper.
  - Use claim_direct_mode helpers - fix a race condition along the way.
* mpl3115
  - Use claim_direct_mode helpers - fix a race condition along the way.
* ms65611
  - Tidy up regulator error handling and clean out a static warning in the mix.
* sca3000
  - Avoid a potential unitialized variable if a hardware read returns a value
  that isn't actually supported (mostly warning supression).
  - Fix a use before setting of the indio_dev->buffer pointer. Broken for
  a very long time so not going to rush this into stable.
  - Merge buffer file with core file.  We used to always split these.
  Sometimes it's just not worth the hassle. In this case the device's main
  feature is it's hardware fifos so unlikely anyone would want to run it
  without.
  - Drop the sca3000_register_ring_funcs function as it's a pointless wrapper
  once we have only one file.
  - Fix cleaning of flag + setting of size of scan. Without this you can't
  start the buffer twice and expect sensible (or any) results.  Again,
  broken for a long time so not heading for stable.
  - Drop the custom watershed setting ABI - for now we'll just support one
  value.
  - Move to a hybrid hard / soft buffer design (how we've been doing it
  for similar devices for a while now!)
  - Cleanup some unusued variables.
  - Use a fake channel to support core handling of freefall event registration.
  - Cleanup the register defines.
  - Fix an off by one error in axis due to IIO_NO_MOD taking up the 0 value.
  Been broken since first admission of IIO to the staging tree.
  - Add readback of the 3db low pass filter frequency and later writing
    allowing droppign of custom measurement mode attributes as they can
    be represented by the filter choices that is their main characteristic.
  - Drop non standard revision attr and replace with dev_info on probe.
  - Avoid a race in probe.
  - Various formatting fixes.
  - Kernel-docify docs that were very nearly in the write format.
* tsl2583
  - Constify attribute_group structure.
* zpa2326
  - Drop a redundant DEBUG ifdef.

Cleanups (Tools)
* iio_generic_buffer
  - Fix the ? arguement. Previously it sort of worked as you got the help
    message as a result of it not recognising the arguement.

PTR_ERR should access the value just tested by IS_ERR

Generated by: scripts/coccinelle/tests/odd_ptr_err.cocci

CC: Tomas Novotny <tomas@novotny.cz>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

drivers/staging/iio/adc/ad7606.c:357:1-3: WARNING: PTR_ERR_OR_ZERO can be used

 Use PTR_ERR_OR_ZERO rather than if(IS_ERR(...)) + PTR_ERR

Generated by: scripts/coccinelle/api/ptr_ret.cocci

CC: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Remove pm_runtime_mark_last_busy() call that wasn't need in the probe
since the device should be put to sleep.

Clarification from Matt:
Basically it going to be suspended once pm_runtime_idle() is called,
and setting the last busy is useless and not needed.
Clearly this doesn't affect the device running but just makes the code
more consistent with other uses.
Signed-off-by: Matt Ranostay <matt@ranostay.consulting>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

This adds triggered buffer support for the ti-adc0832 driver.  Tested with
ADC0831 and ADC0832 by using SYSFS trigger.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Cc: Jonathan Cameron <jic23@kernel.org>
Cc: Hartmut Knaack <knaack.h@gmx.de>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Peter Meerwald-Stadler <pmeerw@pmeerw.net>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Triggered buffer support uses the HDC100X's dual acquisition mode
to read both humidity and temperature in one shot.

This patch depends on
	447136effbf4 ("iio: humidity: hdc100x: fix sensor data reads of
		       temp and humidity")
Signed-off-by: Alison Schofield <amsfield22@gmail.com>
Cc: Daniel Baluta <daniel.baluta@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Signed-off-by: Tomas Novotny <tomas@novotny.cz>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

MCP47x6 chip supports selection of a voltage reference (VDD, VREF buffered
or unbuffered). MCP4725 doesn't have this feature thus the eventual setting
is ignored and user is warned.

The setting is stored only in the volatile memory of the chip. You need to
manually store it to the EEPROM of the chip via 'store_eeprom' sysfs entry.
Signed-off-by: Tomas Novotny <tomas@novotny.cz>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Signed-off-by: Tomas Novotny <tomas@novotny.cz>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Number 2 is referencing to the settings with the largest available
resistor.

No functional change.
Signed-off-by: Tomas Novotny <tomas@novotny.cz>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently, this driver ignores all errors from regulator_get(). The way
it is now, it also breaks probe deferral (EPROBE_DEFER). The correct
behavior is to propagate the error to the upper layers so they can
handle it accordingly.

Rework the regulator handling so that it matches the standard behavior.
If the specific design uses a static always-on regulator and does not
explicitly specify it, regulator_get() will return the dummy regulator.
Suggested-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Eva Rachel Retuya <eraretuya@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

The name passed to devm_regulator_get() should match the name of the
supply as specified in the device datasheet. The supply on this device
is called 'AVcc' while currently, the driver uses just 'vcc'.

Use 'avcc' to specify the supply voltage since it is custom to use the
lower-caps version of the datasheet name.
Suggested-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Eva Rachel Retuya <eraretuya@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the ad7606 buffer handling code resides in its own source file.
But this file contains only 4 small functions of which half are just
wrappers around other functions. Buffer support is also always enabled for
this driver, so move them over to the main source file. This reduces the
amount of boilerplate code.

Also rename the main function from ad7606_core.c to ad7606.c since there is
only a single file now.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Convert the ad7606 driver away from the deprecated legacy GPIO API and use
the new GPIO descriptor API.

This also means that the platform data struct is now empty and can be
removed.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the ad7606 driver installs the same function for the hard-irq and
threaded trigger handlers. This was introduced in commit 1caf7cb4
("staging:iio:adc:ad7606 Convert to new channel registration method Update
Add missing call to iio_trigger_notify_done() Set pollfunc top and bottom
half handler"). Unfortunately the commit message does not mention why this
was done and Michael does not remember either.

Since the trigger handler function is idempotent (set a GPIO to 1) running
it twice does not do any harm, but is simply not necessary either. So set
the threaded trigger handler for the driver to NULL.

While we are at it also remove the function description comment that does
no say anything that can't be derived from the function name itself.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the common remove function takes a struct iio_dev *. This
parameter is retrieved by the individual driver remove functions by calling
get_drvdata() on their device. To simplify the code let the common remove
function directly take a struct dev * and do the IIO device in retrieval
the common remove function.

This also aligns the interface with the common probe function.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

The common probe function for the ad7606 currently returns a struct iio_dev
pointer. The returned value is not used by the individual driver probe
functions other than for error checking.

Let the common probe function return a int instead to report the error
value directly (or 0 on success). This allows to simplify the code.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Both the platform_device and SPI driver call set_drvdata() at the end of
their probe function. Move this into the common probe() function to reduce
duplicated code.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Both the periodic buffer based and one-shot sysfs based capture methods
share a large portion of their code. Factor this out into a common helper
function.

Also provide a comment that better explains in more detail what is going on
in the capture function.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the ad7606 driver dynamically allocates and frees a transfer
buffer each time a sample capture is performed in buffered mode, which
introduces unnecessary overhead. The driver state struct already contains a
buffer that is used for transfers in one-shot mode. This buffer is large
enough to hold all samples, but not the timestamp that might be present in
buffered mode. Extend the buffer size to be able to contain the timestamp
and update the buffered capture function to use this buffer.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the ad7606 driver uses a value of 0 for the oversampling ratio to
express that no oversampling is done. Strictly speaking this means though
that no data capture is done at all. Instead change the driver to use a
value of 1, this is in accordance with what other drivers do and what the
IIO spec suggests.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Currently the ad7606 driver prints a error message to the kernel log when
an application writes an invalid value to a sysfs attribute. While for
initial driver development and testing this might be useful it is quite
disadvantageous in a production environment. The write() call to the sysfs
attribute will already return an error if the value was invalid so the
application is aware that the operation failed. And generally speaking it
is impossible for an application to reliably match a log message in the
kernel log to a specific operation it performed, so the message becomes
just noise and might distract from more critical messages.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

While for some very selected setups it might be useful to be able to
provide default configuration data via the platform data, generally this
becomes very impractical as the number of configuration options increases.
So the general policy is to use the power-on default values of the device
and let the application using the device configure it according to its
needs.

Implement this scheme for the ad7606 driver by removing support for
specifying a default configuration via the platform data.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

The name field in the ad7606_chip_info struct is set to the same value as
the as the name field in the corresponding {platform,spi}_device_id table
entry. Remove it from the ad7606_chip_info struct and pass the name from
the ID to the probe function. This slightly reduces the size of the
chip_info table and adding new entries requires less boilerplate.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>