Jonathan writes:

Second set of IIO fixes for the 4.5 cycle.  These ones are mostly
dependent on patches from the recent merge cycle.

* adc, imu and iio staging drivers
  - !HAS_IOMEM dependency fixes
* dht11
  - use boottime clock for time measurement to avoid incorrect measurements
   due to clock updates.
* lidar
  - correct a return value for short i2c transfers.

Jonathan writes:

First set of IIO fixes for the 4.5 cycle.

This set comprises those not dependent on patches in the 4.5 merge cycle.
A second set will follow shortly with ones that are.

* core in kernel interfaces
  - fix a possible NULL dereference that is a theoretical possibility
    via odd usage of iio_channel_release. Pretty much a hardening of
    the interface, but observed in the wild with the twl4030_charger
    driver.
* acpi-als
  - report the data as processed as it is in lux.  This fixes a wrong
    use of the IIO ABI.  However, old _raw version retained to avoid
    breaking any userspace in the wild that is relying on that (none
    known but it doesn't hurt us much to retain it)
* ade7753
  - fix some error handling to avoid use of unitialized data.
* ltr501
  - use a signed return type for ltr501_match_samp_freq so as to allow
    returning of an error code.
* mcp4725
  - set name field of struct iio_dev to ensure the sysfs name attribute
    doesn't give NULL.
* mpl115
  - temperature offset sign is wrong.
* stk8ba50
  - IIO_TRIGGER dependency added
* ti_am335x_adc
  - Label buffer as a software buffer. It's actually a hybrid of a
    true hardware buffer feeding a kfifo, but the meaning of these fields
    has changed a little recently and in this case it should be labeled
    a software buffer ensure it is allowed to use the kfifo.
* vf610_adc
  - HAS_IOMEM dependency

The ade7753_spi_read_reg_16() will either successfully read a value
from SPI, or return a failure code without delivering data. However,
the ade7753_stop_device() and ade7753_reset() functions use the returned
data without checking for an error condition first. Gcc detects this
as a possible bug and warns about it:

drivers/staging/iio/meter/ade7753.c: In function 'ade7753_remove':
drivers/staging/iio/meter/ade7753.c:348:6: error: 'val' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  val |= BIT(4);  /* AD converters can be turned off */
      ^
drivers/staging/iio/meter/ade7753.c:345:6: note: 'val' was declared here
  u16 val;
      ^
drivers/staging/iio/meter/ade7753.c: In function 'ade7753_probe':
drivers/staging/iio/meter/ade7753.c:222:6: error: 'val' may be used uninitialized in this function [-Werror=maybe-uninitialized]

In both cases, we can avoids the warning by checking the return code
before using the data.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Not every arch has io memory.
So, unbreak the build by fixing the dependencies.

Signed-off-by: Richard Weinberger <richard@nod.at>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

Pull kbuild updates from Michal Marek:
 - Make <modname>-m in makefiles work like <modname>-y and fix the
   fallout
 - Minor genksyms fix
 - Fix race with make -j install modules_install
 - Move -Wsign-compare from make W=1 to W=2
 - Other minor fixes

* 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kbuild: Demote 'sign-compare' warning to W=2
  Makefile: revert "Makefile: Document ability to make file.lst and file.S" partially
  kbuild: Do not run modules_install and install in paralel
  genksyms: Handle string literals with spaces in reference files
  fixdep: constify strrcmp arguments
  ath10k: Fix build with CONFIG_THERMAL=m
  Revert "drm: Hack around CONFIG_AGP=m build failures"
  kbuild: Allow to specify composite modules with modname-m
  staging/ad7606: Actually build the interface modules

Jonathan writes:

Third set of new stuff for IIO in the 4.5 cycle.

New driver features
- us5182
  * Add interrupt support and rising / falling threshold events.

Cleanups / fixes to new stuff / minor additions
* Expose the IIO value formatting function for drivers to
  make use of internally.
- ina2xx
   * Fix wrong channel order
   * Fix incorrect reporting of endianness
   * Adding documentation of ABI unique to this device
- mma8452
  * Drop an unused register description
  * Use an enum for the channel index to aid readability
- sca3000
  * Use standard NULL comparison style
- us5182
  * fix an inconsistency in status of enable (a bug with no real effect until
    above patches are applied)
  * refactor the read_raw function to improve maintainability / readability.

Jonathan writes:

Second set of IIO new drivers, functionality and cleanups for the 4.5 cycle.

The big one here is the configfs support which has been a long time in the
works but should allow for cleaner ways to do instantiation of those elements
of IIO that aren't directly connected to specific hardware. Lots of cool new
stuff we can use this for in the works!

New core stuff (basically all configfs support related)
* Configfs support
  - Core support (was waiting for a configfs patch that went in around 4.4rc2)
  - A little fixlet to add a configfs.h to contain a reference to the
    configfs_subsystem structure.
* Some infrastructure to simplify handling of software based triggers
  (i.e. ones with no actual hardware associated with them)
* A high resolution timer based trigger.  This has been around for years
    but until the configfs support was ready we didn't have a sensible way
    of instantiating instances of it (the method used for the sysfs_trigger
    has never been really satisfactory)

New Device Support
* AMS iAQ Volatile Organic Compounds sensor support.
* Freescale imx7d ADC driver
* Maxim MAX30100 oximeter driver (note that for these devices most of the
  smart stuff will be in userspace - effectively they are just light sensors
  with some interesting led synchronization as far as the kernel is concerned).
* Microchip mcp3421 support added to the mcp3422 driver.
* TI adc124s021 support added to the adc128s052 driver.
* TI ina219, inda226 power monitors. Note that there is an existing hwmon driver
  for these parts, the usecase is somewhat different so it is unclear at this
  point if the hwmon driver will eventually be replaced by a bridge from
  this driver.  In the meantime the Kconfig dependencies should prevent both
  from being built.

New driver functionality
* us8152d power management support.

Cleanups, fixups
* Use list_for_each_entry_safe instead of list_for_each_safe with the entry
  bit coded longhand.
* Select IRQ_WORK for IIO_DUMMY_EVGEN.  This is a fix that somehow got lost
  when the driver was moved so lets do it again.
* st-accel - drop an unused define.
* vz89x, lidar - optimize i2c transactions by using a single i2c tranfers
  instead of multiple calls where supported (fall back to smbus calls as
  before if not).
* Use dev_get_platdata() in staging drivers: tsl2x7x, adcs and frequency
  drivers instead of direct access to the structure element.

The variable u8 **rx_p, is a pointer-to-pointer and hence the check
should
be "if (!*rx_p)" and not "if (!rx_p)".
In the earlier version, checkpatch.pl gave the following check, which
was incorrect:
CHECK: Comparison to NULL could be written "!rx_p"
+       if (*rx_p == NULL) {

Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

The change to using BIT(20) for one of the bit shifts resulted
in a constant becoming unsigned. When combined with the existing
signed constants in a couple of places, this caused possible trouble,
hence the warnings:

drivers/staging/iio/adc/mxs-lradc.c: In function ‘mxs_lradc_complete_touch_event’:
drivers/staging/iio/adc/mxs-lradc.c:325:5: warning: large integer implicitly truncated to unsigned type [-Woverflow]
     (((x) << LRADC_DELAY_TRIGGER_LRADCS_OFFSET) & \
     ^
drivers/staging/iio/adc/mxs-lradc.c:734:7: note: in expansion of macro ‘LRADC_DELAY_TRIGGER’
       LRADC_DELAY_TRIGGER(1 << TOUCHSCREEN_VCHANNEL1) |
       ^
  LD [M]  drivers/staging/iio/accel/adis16201.o
drivers/staging/iio/adc/mxs-lradc.c: In function ‘mxs_lradc_buffer_preenable’:
drivers/staging/iio/adc/mxs-lradc.c:322:42: warning: large integer implicitly truncated to unsigned type [-Woverflow]
 #define LRADC_DELAY_TRIGGER_LRADCS_MASK  (0xff << 24)
                                          ^
drivers/staging/iio/adc/mxs-lradc.c:1308:29: note: in expansion of macro ‘LRADC_DELAY_TRIGGER_LRADCS_MASK’
  mxs_lradc_reg_clear(lradc, LRADC_DELAY_TRIGGER_LRADCS_MASK |
                             ^
drivers/staging/iio/adc/mxs-lradc.c: In function ‘mxs_lradc_buffer_postdisable’:
drivers/staging/iio/adc/mxs-lradc.c:322:42: warning: large integer implicitly truncated to unsigned type [-Woverflow]
 #define LRADC_DELAY_TRIGGER_LRADCS_MASK  (0xff << 24)
                                          ^
drivers/staging/iio/adc/mxs-lradc.c:1327:29: note: in expansion of macro ‘LRADC_DELAY_TRIGGER_LRADCS_MASK’
  mxs_lradc_reg_clear(lradc, LRADC_DELAY_TRIGGER_LRADCS_MASK |

The simplest fix is to force LRADC_DELAY_TRIGGER_LRADCS_MASK to be
a shift of an unsigned 0xff rather than a signed one.

This is ugly considering it is the only constant in the driver which
is so forced, but it does deal with the issue.

Fixes: e0c961bdaf27 (iio: adc: mxs-lradc: Prefer using the BIT macro)
Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.

Signed-off-by: Nizam Haider <nijamh@cdac.in>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>