staging: iio: tsl2x7x: add common function for writing to the control register

There were four places where the same chunk of code was used to write
to the control register. This patch creates a common function
tsl2x7x_write_control_reg() to reduce duplicate code.

The function tsl2x7x_chip_off() did not correctly return an error
code so this patch also corrects that issue.

Signed-off-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: iio: tsl2x7x: add common function for reading chip status

There were three places where the same chunk of code was used to read
the chip status. This patch creates a common function
tsl2x7x_read_status() to reduce duplicate code. This patch also corrects
tsl2x7x_event_handler() to properly check for an error after reading the
chip status.

Signed-off-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: iio: tsl2x7x: add common function for clearing interrupts

There were three places where the same chunk of code was used to clear
interrupts. This patch creates a common function
tsl2x7x_clear_interrupts() to reduce duplicate code.

Signed-off-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: iio: tsl2x7x: remove power functions from tsl2X7X_platform_data

The tsl2X7X_platform_data structure contains the platform_power,
power_on, and power_off function pointers. These power management
functions should not be in the platform data. These functions were
likely used before the regulator framework was put in place. There are
no users of these functions in the mainline kernel.

Signed-off-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Use sign_extend32 function

Use sign_extend32 function instead of manually coding
it.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Adjust a switch statement

Adjust a switch block to explicitly match channels and
return -EINVAL as default case which makes the code
semantically more clear.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Change some macro names

Make some of the macro names according to the names
given in the datasheet of the adis16209 driver.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Remove and add some comments and group the definitions

Remove some unnecessay comments and group the control
register and register field macros together.
Some of the register names does not make it's puporse
very clear and hence, add some comments for more
information.
Also there are certain unit based comments which are not
providing sufficient information, so expand those comments.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: accel: adis16201: Add a blank space before returns

Adding a blank space before/after some returns improves readability.

Signed-off-by: Himanshu Jha <himanshujha199640@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: accel: adis16201: Prefer alphabetical sequence of header files

Arrange header files in alphabetical sequence to improve readability.

Signed-off-by: Himanshu Jha <himanshujha199640@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: Replace occurrences of magic number 0 by IIO_CHAN_INFO_RAW

Usually, functions responsible for reading raw data typically relies on
values from iio_chan_info_enum to correctly identify the type of data to
be read. There is a set of a device driver that uses the magic number 0
instead of IIO_CHAN_INFO_RAW. This patch improves the readability by
replaces the magic number 0 for the appropriate IIO_CHAN_INFO_RAW in six
devices driver in the IIO subsystem.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging:iio:meter: Aligns open parenthesis

This patch fixes the checkpatch.pl checks:

staging/iio/meter/ade7854-spi.c:19: CHECK: Alignment should match open
parenthesis
staging/iio/meter/ade7854-spi.c:44: CHECK: Alignment should match open
parenthesis
staging/iio/meter/ade7854-spi.c:70: CHECK: Alignment should match open
parenthesis
staging/iio/meter/ade7854-spi.c:97: CHECK: Alignment should match open
parenthesis
staging/iio/meter/ade7854-spi.c:125: CHECK: Alignment should match open
parenthesis

...

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging:iio:meter: Remove unused macro IIO_DEV_ATTR_CH_OFF

This patch removes the macro IIO_DEV_ATTR_CH_OFF. The macro
IIO_DEV_ATTR_CH_OFF is not required, due to the replace of it by the
direct use of IIO_DEVICE_ATTR in files staging/iio/meter/ade7759.c and
staging/iio/meter/ade7753.c.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging:iio:meter: Replaces IIO_DEV_ATTR_CH_OFF by IIO_DEVICE_ATTR

The macro IIO_DEV_ATTR_CH_OFF is a wrapper for IIO_DEVICE_ATTR, with a
tiny change in the name definition. This extra macro does not improve
the readability and also creates some checkpatch errors.

This patch fixes the checkpatch.pl errors:

staging/iio/meter/ade7753.c:391: ERROR: Use 4 digit octal (0777) not
decimal permissions
staging/iio/meter/ade7753.c:395: ERROR: Use 4 digit octal (0777) not
decimal permissions
staging/iio/meter/ade7759.c:331: ERROR: Use 4 digit octal (0777) not
decimal permissions
staging/iio/meter/ade7759.c:335: ERROR: Use 4 digit octal (0777) not
decimal permissions

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: iio: meter: Remove reduntant __func__ from debug print

dev_dbg includes the function name & line number by default when dynamic
debugging is enabled. Hence__func__ is reduntant here and removed.

Signed-off-by: HariPrasath Elango <hariprasath.elango@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: iio: Remove unnecessary cast on void pointer

The following Coccinelle script was used to detect this:
@r@
expression x;
void* e;
type T;
identifier f;
@@
(
  *((T *)e)
|
  ((T *)x)[...]
|
  ((T*)x)->f
|

- (T*)
  e
)

Signed-off-by: Arushi Singhal <arushisinghal19971997@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Add _REG postfix for registers

The defined names for registers does not make it very
clear that they are registers and hence, add _REG postfix.
This improves the readability of the code.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Change the definition name

The change in the definition name makes it then obvious
what the units are throughout the driver and there will
be no need of the comment.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

Staging: iio: adis16209: Arrange headers in alphabetical order

Arrange the headers in alphabetical order for cleanup
purpose.

Signed-off-by: Shreeya Patel <shreeya.patel23498@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio:magnetometer: Replace magic number 0 by IIO_CHAN_INFO_RAW

The function magn_3d_read_raw has a switch statement handling multiple
cases per channel. The first case statement uses the magic number 0,
which means IIO_CHAN_INFO_RAW. Additionally, the iio_chan_spec for
magn_3d_channels is configured to be IIO_CHAN_INFO_RAW. Therefore, this
patch replaces the magic number 0 for the appropriate IIO_CHAN_INFO_RAW.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio:dummy: Add extra paragraphs on Kconfig

This patch fixes the checkpatch.pl warning:

drivers/iio/dummy/Kconfig:21: WARNING: please write a paragraph that
describes the config symbol fully
drivers/iio/dummy/Kconfig:27: WARNING: please write a paragraph that
describes the config symbol fully

This patch expands the explanation about IIO_DUMMY_EVGEN by using the
code documentation found in iio/dummy/iio_dummy_evgen.c. In the same
way, the information related to IIO_SIMPLE_DUMMY_BUFFER was extracted
from file iio_simple_dummy_buffer.c.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: dummy: Add correct tabs and spaces to Kconfig

Kconfig from iio/dummy does not follow the coding style recommendations.
According to the coding-style, Lines under a config definition are
indented with one tab, while help text is indented an additional two
spaces. This patch adds the proper tabulation and space.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: light: lv0104cs: Add support for LV0104CS light sensor

This patch adds support for the On Semiconductor LV0104CS ambient
light sensor.

Signed-off-by: Jeff LaBundy <jeff@labundy.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: proximity: sx9500: Add GPIO ACPI mapping table

In order to satisfy GPIO ACPI library requirements convert users of
gpiod_get_index() to correctly behave when there no mapping is provided
by firmware.

Here we add explicit mapping between _CRS GpioIo() resources and
their names used in the driver.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

dt-bindings: iio: adc: stm32-dfsdm: fix types, add missing pinctrl

- Add missing pinctrl description. Support is made optional as dfsdm
  may use internal sources (e.g. via registers)
- Fix typo in IIO STM32 DFSDM filter "MANCH_F" description.
Basically, this should be "falling edge = logic 0", not "1" that applies
to "MANCH_R".
BTW, make the description complete by describing both rising/falling
edges as described in reference manuals.

Fixes: 6c82f947fc97 ("IIO: add DT bindings for stm32 DFSDM filter")
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@st.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: cros_ec: Relax sampling frequency before suspending

If an application set a tight sampling frequency, given the interrupt
use is a wakeup source, suspend will not happen: the kernel will receive
a wake up interrupt and will cancel the suspend process.

Given cros_ec sensors type is non wake up, this patch adds prepare and
complete callbacks to set 1s sampling period just before suspend. This
ensures the sensor hub will not be a source of interrupt during the
suspend process.

Signed-off-by: Gwendal Grignou <gwendal@chromium.org>
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Acked-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging:iio:meter: Add name to function definition arguments

This patch fixes the checkpatch.pl warning:

drivers/staging/iio/meter/ade7854.h:157: WARNING: function definition
argument 'struct device *' should also have an identifier name...

This commit adds arguments names to the signature declared in the
ade7854_state struct. For consistency reason, It also renames all
arguments in function definitions.

Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@gmail.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio temperature/mlx90632: silence a static checker warning

This shouldn't affect runtime at all, but Smatch complains that we
should check if mlx90632_read_ambient_raw() otherwise we
"ambient_new_raw" can be uninitialized.

    drivers/iio/temperature/mlx90632.c:509 mlx90632_calc_ambient_dsp105()
    error: uninitialized symbol 'ambient_new_raw'.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Crt Mori <cmo@melexis.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: accel: bmc150: Remove redundant __func__ in dev_dbg()

Dynamic debug has a run time knob to enable function name printing.
Remove this from dev_dbg() calls.

Furthermore, functional tracing when enabled can show what function is
called, therefore remove empty dev_dbg() calls.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: adc: axp20x_adc: remove !! in favor of ternary condition

!!'s behaviour isn't that obvious and sparse complained about it, so
let's replace it with a ternary condition.

Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

iio: chemical: ccs811: Renamed resistance member in ccs811_reading struct

The resistance member in ccs811_reading struct is an unsigned 16-bit
integer variable used to store RAW_DATA register bytes read from CCS811.
It is kind of misleading to name this struct member as resistance.

About the RAW_DATA register bytes, the CCS811 datasheet states that:
-----
Two byte read only register which contains the latest readings from the
sense resistor.

The most significant 6 bits of the Byte 0 contain the value of the current
through the sensor (0μA to 63μA).

The lower 10 bits contain (as computed from the ADC) the readings of the
voltage across the sensor with the selected current (1023 = 1.65V)"
-----

Hence, the RAW_DATA register byte contains information about electric
current and voltage of the CCS811 sensor. Calling this struct member
'resistance' is kind of misleading, although both electric current and
voltage are needed to calculate the electrical resistance of the sensor
using Ohm's law, V = I x R, in which a new channel type of IIO_RESISTANCE
may be added to the driver in the future.

Signed-off-by: Richard Lai <richard@richardman.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

staging: speakup: remove space after a cast

Remove blank space after a cast to conform to Linux kernel coding style.
Problem found using checkpatch.

Signed-off-by: Santha Meena Ramamoorthy <santhameena13@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: match alignment with open parenthesis

Match alignment with open parenthesis to conform to Linux kernel coding
style. Problem found using checkpatch.

Signed-off-by: Santha Meena Ramamoorthy <santhameena13@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: add spaces around arithmetic operators

Add space around arithmetic operators ('+', '-' and '*') to conform to
Linux kernel coding style. Problem found using checkpatch.

Signed-off-by: Santha Meena Ramamoorthy <santhameena13@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: gdm724x: LTE: Refactor gdm_lte_pdn_table().

Mostly this change just reverses the primary conditional so most of
the code can be pulled back a tab, which fixes some code style
warnings.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: gdm724x: Simplify the struct gdm_endian to a variable.

Since the testing for host endianness and in-driver conversion were
removed in 77e8a50149a2, the gdm_endian struct contains only one member,
and can therefore be simplified to a single u8 variable.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: gdm724x: LTE: Fix trailing open parentheses.

Fix lines with a trailing open parenthesis, which is a coding style issue.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-mc: Move irqchip code out of staging

Now that the fsl-mc bus core infrastructure is out of staging, the
remaining irqchip glue code used (irq-gic-v3-its-fsl-mc-msi.c) goes
to drivers/irqchip.

Create new Kconfig option for irqchip code that depends on
FSL_MC_BUS and ARM_GIC_V3_ITS. This ensures irqchip code only
gets built on ARM64 platforms. We can now remove #ifdef
GENERIC_MSI_DOMAIN_OPS as it was only needed for x86.

Signed-off-by: Stuart Yoder <stuyoder@gmail.com>
[rebased, add dpaa2_eth and dpio #include updates]
Signed-off-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
[rebased, split irqchip to separate patch]
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@nxp.com>
[add Kconfig dependency on ARM_GIC_V3_ITS]
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-mc: Move core bus out of staging

Move the source files out of staging into their final locations:
  -mc.h include file in drivers/staging/fsl-mc/include go to include/linux/fsl
  -source files in drivers/staging/fsl-mc/bus go to drivers/bus/fsl-mc
  -overview.rst, providing an overview of DPAA2, goes to
   Documentation/networking/dpaa2/overview.rst

Update or delete other remaining staging files -- Makefile, Kconfig, TODO.
Update dpaa2_eth and dpio staging drivers.
Add integration bits for the documentation build system.

Signed-off-by: Stuart Yoder <stuyoder@gmail.com>
[rebased, add dpaa2_eth and dpio #include updates]
Signed-off-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
[rebased, split irqchip to separate patch]
Signed-off-by: Bogdan Purcareata <bogdan.purcareata@nxp.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename _WPARxGtk_end_case_ label to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename _WPAPtk_end_case_ label to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pstrStationParam to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pstrDelStaParam to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename strDisconnectNotifInfo to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pstrStatistics to avoid camelCase

Fix "Avoid caseCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pstrSetBeaconParam to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename Handle_SetMulticastFilter to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pu32InactiveTime to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename variables using camelCase in handle_rcvd_ntwrk_info()

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename ptstrJoinBssParam to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename pu8HdnNtwrksWidVal to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: selftest: freeing an error pointer

We should just return directly if memdup_user() fails. The current code
tries to free "param" which is an error pointer so it will Oops.

Fixes: 2baddf262e98 ("staging: lustre: use memdup_user to allocate memory and copy from user")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: socklnd: simplify ksnc_rx_iov_space

ksnc_rx_iov_space is currently a union of two arrays,
one of 'struct kvec', the other of 'struct bio_vec'.

The 'struct bio_vec' option is never used.  The
array of kvec is used to read in a packet header, or
to read data that needs to be skipped so as to synchronize
with a packet boundary.
In each case the target memory location is a virtual address,
never a page, so 'struct bio_vec' is never needed.

When we read into a page, different code steps up a separate
array of 'struct bio_vec'.

So remove the bio_vec option, and remove the union ksock_rxiovspace..

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fid: perform sanity checks before commiting

When fid fetches a new range from the server, it commits
to it (*output = *out) *before* performing sanity checks.
This looks backwards.
Don't commit to a value until it has been found to be sane.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fid: fix up debugfs access to ->lcs_space

lcs_space can change while the lock is not held
if an RPC in underway.  This can be detected by
seq->ls_update being set.

In this case, reading or writing the value should return
-EBUSY.

Also, the D_INFO CDEBUG() which reports the lcs_space being
updated never fires, as it tests the wrong value -
ldebugfs_fid_write_common() returns 'count' on success.

Finally, this return value should be returned from
ldebugfs_fid_space_seq_write(), rather than always returning 'count',
so that errors can be detected.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fid: remove seq_fid_alloc_fini() and simplify

seq_fid_alloc_fini() is tiny and only called
from two places in the one function.  We can move
both those calls earlier and merge them so only
one call is needed.  At that point, there is no
value added by having a separate function.

Also instead of using ++ and -- on ->lcs_update to
toggle between 0 and 1, explicitly set to 0 or 1
as appropriate.

Moving the locking earlier means that the code which updates
seq->lcs_fid is now protected, so
ldebugfs_fid_fid_seq_show() now cannot see a torn value.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fid: use wait_event_cmd()

Rather than open-coding a wait event loop twice,
use wait_event_cmd() to wait, dropping the spinlock
over schedule().
This does require duplicating part of the wait
condition, but that is just three tests on values that
are in registers or in cache, so the cost is small
and the increased readability is large.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fid: convert lcs_mutex to a spinlock

There is only one place where this lock is held
while the task might sleep - in
  ldebugfs_fid_space_seq_write()
while ldebugfs_fid_write_common() is called.

This call can easily be taken out of the locked region
by asking it to parse the user data into a local variable,
and then copying that variable into ->lcs_space while
holding the lock.

Note that ldebugfs_gid_write_common returns >0 on
success, so use that to gate updating ->lcs_space.

So make that change, and convert lcs_mutex to a spinlock
named lcs_lock.  spinlocks are slightly cheaper than mutexes
and using one makes is clear that the lock is only held for
a short time.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: ptlrpc: list_for_each improvements.

1/ use list_for_each_entry_safe() instead of
   list_for_each_safe() and similar.

2/ use list_first_entry() and list_last_entry() where appropriate.

3/ When removing everything from a list, use
     while ((x = list_first_entry_or_null()) {
   as it makes the intent clear

4/ No need to take a spinlock in a structure that is about
   to be freed - we must have exclusive access at this stage.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fix assorted checkpatch errors

Possibly the most interesting is the for-loop with no body.
Rearranging and initializing end_dirent on each iteration of
the outer while, makes the intent clearer.

Reviewed-by: "Eremin, Dmitry" <dmitry.eremin@intel.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: remove phantom struct cfs_crypto_hash_desc

There is no "struct cfs_crypto_hash_desc" structure.  There
are only pointers to this structure, which are cast back and
forth to struct ahash_request.
So discard cfs_crypto_hash_desc, and just use ahash_request directly.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: discard libcfs_kvzalloc and linux-mem.c

The only interesting difference between libcfs_kvzalloc()
and kvzalloc() is that the former appears to work
with GFP_NOFS, which the latter gives a WARN_ON_ONCE()
when that is attempted.

Each libcfs_kvzalloc() should really be analysed
and either converted to a kzalloc() call if the size is never
more than a page, or to use GFP_KERNEL if no locks are held.

If there is ever a case where locks are held and a large allocation
is needed, then some other technique should be used.

It might be nice to not always blindly zero pages too...

For now, just convert libcfs_kvzalloc() calls to
kvzalloc(), and let the warning remind us that there is work to do.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: improve some libcfs_kvzalloc calls.

Using vmalloc with GFP_NOFS is not supported as vmalloc
performs some internal allocations with GFP_KERNEL.

So in cases where the size passed to libcfs_kvzalloc()
is clearly at most 1 page, convert to kzalloc().
In cases where the call clearly doesn't hold any
filesystem locks, convert to GFP_KERNEL.

Unfortunately there are many more that are not easy to fix.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: discard lu_buf allocation library.

This library code is unnecessarily generic, but also
not generic enough.  Library code that performs
allocations should always take a gfp_flags argument.

So discard the library and in the one file where it is used,
just use kzalloc or krealloc as needed.
In this context, it is clear that vmalloc is never needed.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: discard libcfs_kvzalloc_cpt()

This function is used precisely once, and is sufficiently
trivial that it may as well be open-coded.

Doing so helpfully highlights the similarity
between the new kvzalloc_node() call and the already existing
kzalloc_node() call in the same function.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: make signal-blocking functions inline

cfs_block_sigsinv() and cfs_restore_sigs() are now
simple enough to inline them.
This means we can discard linux-prim.c

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: improve API and implementation of blocking signals.

According to comment for set_current_blocked() in
kernel/signal.c, changing ->blocked directly is wrong.
sigprocmask() should be called instead.

So change cfs_block_sigsinv() and cfs_restore_sigs()
to use sigprocmask().
For consistency, change them to pass the sigset_t by reference
rather than by value.

Also fix cfs_block_sigsinv() so that it correctly blocks
signals above 32 on a 32bit host.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: simplify linux-prim.c

cfs_block_sigs() is never used.
cfs_clear_sigpending() is never used.
cfs_block_allsigs() is no longer used.

So those three functions can go.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: lnet: remove cfs_block_allsigs calls.

Both places that cfs_block_allsigs() is used here,
the goal is to turn an interruptible wait into an
uninterruptible way.
So instead of blocking the signals, change TASK_INTERRUPTIBLE to
TASK_NOLOAD.
In each case, no other functions called while signals are blocked
will sleep - just the one that has been fixed.
In one case, an extra 'interruptible' flag needs to be passed
down so the waiting decision can be made at the right place.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: remove unnecessary cfs_block_allsigs() calls

Threads started by kthread_run() ignore all signals,
as kthreadd() calls ignore_signals(), and this is
inherited by all children.
So there is no need to call cfs_block_allsigs() in functions
that are only run from kthread_run().

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: remove linux-curproc.c

The only functionality remaining here is
cfs_curproc_cap_pack(),
and it can be trivially implemented as an inline
in curproc.h.
So do that and remove the file.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: opencode cfs_cap_{raise, lower, raised}

Each of these functions is used precisely once, so having
a separate exported function seems like overkill.

cfs_cap_raised() is trivial - one line.
cfs_cap_raise() and cfs_cap_lower() are used as a pair
which is more effectively implemented with
override_cred() / revert_creds().

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: replace all CFS_CAP_* macros with CAP_*

Lustre defines a few CFS_CAP_* macros which are exactly the
same as the corresponding CAP_* macro, with one exception.

CFS_CAP_SYS_BOOT is 23
CAP_SYS_BOOT is 22.

CFS_CAP_SYS_BOOT is only used through CFS_CAP_FS_MASK and
causes capability 23 (CAP_SYS_NICE) to be dropped in certain
circumstances.
It is probable that the intention was to drop CAP_SYS_BOOT,
and this is what is now done.

CFS_CAP_CHOWN_MASK and CFS_CAP_SYS_RESOURCE_MASK are never
used, so they have been removed.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: Remove unneeded cast

Fix Coccinelle alert:

drivers/staging//rtl8188eu/os_dep/usb_intf.c:336:13-27: WARNING: casting value returned by memory allocation function to (struct adapter *) is useless.

This issue was detected by using the Coccinelle software.

Signed-off-by: Christopher Diaz Riveros <chrisadr@gentoo.org>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: emxx_udc: Remove unneeded cast

Fix Coccinelle alert:

drivers/staging//emxx_udc/emxx_udc.c:2689:19-21: WARNING: casting value returned by memory allocation function to (u8 *) is useless.

This issue was detected by using the Coccinelle software.

Signed-off-by: Christopher Diaz Riveros <chrisadr@gentoo.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723bs: Remove unneeded cast

Fix Coccinelle alert:

drivers/staging//rtl8723bs/os_dep/sdio_intf.c:340:13-27: WARNING: casting value returned by memory allocation function to (struct adapter *) is useless.

This issue was detected by using the Coccinelle software.

Signed-off-by: Christopher Diaz Riveros <chrisadr@gentoo.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: net: netlogic: Remove unneeded cast

Fix Coccinelle alert:

drivers/staging//netlogic/xlr_net.c:996:12-30: WARNING: casting value returned by memory allocation function to (struct xlr_adapter *) is useless.

This issue was detected by using the Coccinelle software.

Signed-off-by: Christopher Diaz Riveros <chrisadr@gentoo.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: bcm2048: Fix function argument alignment in radio-bcm2048.c.

Fix a coding style problem.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: ks7010: hostif: Convert the ps_confirm_wait_inc() macro to a real function.

Convert the unsafe macro into an inline function.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: ks7010: hostif: Convert SME queue macros to real functions.

Convert the unsafe macros into inline functions.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: ks7010: sdio: Convert RX/TX queue macros into real functions.

Convert the unsafe macros into inline functions.

Signed-off-by: Quytelda Kahja <quytelda@tamalin.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: remove 'if' on field address

Remove 'if' statements testing struct's field address.
Since such statements always return true, they are redundant.

Signed-off-by: Dafna Hirschfeld <dafna3@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: pi433: fix CamelCase for addressFiltering

Fixes checkpatch warning:

  CHECK: Avoid CamelCase: <addressFiltering>

Signed-off-by: Valentin Vidic <Valentin.Vidic@CARNet.hr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: pi433: fix CamelCase for thresholdDecrement

Fixes checkpatch warning:

  CHECK: Avoid CamelCase: <thresholdDecrement>

Signed-off-by: Valentin Vidic <Valentin.Vidic@CARNet.hr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: pi433: fix CamelCase for txStartCondition

Fixes checkpatch warning:

  CHECK: Avoid CamelCase: <txStartCondition>

Signed-off-by: Valentin Vidic <Valentin.Vidic@CARNet.hr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: Use '"%s:", __func__' instead of function name

Replace all occurrences of functions' names in strings by a reference
to __func__, to improve robustness. Problem found with checkpatch.

Signed-off-by: Eisha Chen-yen-su <chenyensu0@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: Remove a "tracing" call

Remove a "tracing" call as it is not needed anymore
because there is an in-kernel function for that.

Signed-off-by: Eisha Chen-yen-su <chenyensu0@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: ion: Remove check of idev->debug_root

We will go on initial idev if failed to create debug_root, and it does
not matter to check the return value of this debugfs call, just remove it.

Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: ion: Remove check of debug_file

There's no need to check the return value of debug_file for it is just a
debugfs and we will go on the following process if we failed to create
debug_file. So just remove it.

Signed-off-by: Yisheng Xie <xieyisheng1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'iio-for-4.17a' of git://git./linux/kernel/git/jic23/iio into staging-next

Jonathan writes:

First round of new devices, features and cleanups for IIO in the 4.17 cycle.

Outside of IIO
* Strongly typed 64bit int_sqrt function needed by the mlx90632

New device support
* adc081s
  - New driver supporting adc081s, adc101s and adc121s TI ADCs.
* ad5272
  - New driver supproting the ad5272 and ad5274 ADI digital potentiometers
    with DT bindings.
* axp20x_adc
  - support the AXP813 ADC - includes rework patches to prepare for this.
* mlx90632
  - New driver with dt bindings for this IR temperature sensor.

Features
* axp20x_adc
  - Add DT bindings and probing.
* dht11
  - The sensor has a wider range than advertised in the datasheet - support it.
* st_lsm6dsx
  - Add hardware timestamp su9pport.

Cleanups
* ABI docs
  - Update email contact for Matt Ranostay
* SPDX changes
  - Matt Ranostay has moved his drivers over to SPDX.  Currently we are making
    this an author choice in IIO.
* ad7192
  - Disable burnout current on misconfiguration.  No actually effect as
    they simply won't work otherwise.
* ad7476
  - Drop a license definition that was replicating information in SPDX tag.
* ade7758
  - Expand buf_lock to cover both buffer and state protection allowing
    unintented uses of mlock in the core to be removed.
* ade7759
  - Align parameters to opening parenthesis.
* at91_adc
  - Depend on sysfs instead of selecting it - for try wide consistency.
* ccs811
  - trivial naming type for a define.
* ep93xx
  - Drop a redundant return as a result checking platform_get_resource.
* hts221
  - Regmap conversion which simplifies the driver somewhat.
  - Clean up some restricted endian cast warnings.
  - Drop a trailing whitespace from a comment
  - Drop an unnecessary get_unaligned by changing to the right 16bit data type.
* ms5611
  - Fix coding style in the probe function (whitespace)
* st_accel
  - Use strlcpy instead of strncpy to avoid potentially truncating a string.

staging: android: ion: Initialize dma_address of new sg list

Fix the dup_sg_table function to initialize the dma_address of the new
sg list entries instead of the source dma_address entries.

Since ION duplicates the sg_list this issue does not appear to result in
an actual bug.

Signed-off-by: Liam Mark <lmark@codeaurora.org>
Acked-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: use memdup_user to allocate memory and copy from user

Replace a call to kmalloc and a call to copy_from_user with a
call to memdup_user to simplify the code.
Issue found with coccicheck.

Signed-off-by: Dafna Hirschfeld <dafna3@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: pi433: Split subtraction across 2 lines

Split a subtraction across 2 lines in order to make these lines
no longer than 80 columns. Problem found with checkpatch.

Signed-off-by: Eisha Chen-yen-su <chenyensu0@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: pi433: Split some function calls

Split some function calls on several lines in order to make these
lines no longer than 80 columns. Problem found with checkpatch.

Signed-off-by: Eisha Chen-yen-su <chenyensu0@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: remove Unnecessary parentheses around 'hif_set_mc->cnt'

Fix "Unnecessary parentheses around 'hif_set_mc->cnt'" issue found by
checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename wilc_remove_key() parameters to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename GetPeriodicRSSI to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename ListenTimerCB to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename functions starting with TimerCB_ to avoid camelCase

Fix "Avoid camelCase" issues found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: rename functions starting with Handle_ to avoid camelCase

Fix "Avoid camelCase" issue found by checkpatch.pl script.

Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>