platform/kernel/linux-rpi.git
2 years agoMerge tag 'fsi-for-v5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/joel/fsi...
Greg Kroah-Hartman [Fri, 22 Oct 2021 07:45:59 +0000 (09:45 +0200)]
Merge tag 'fsi-for-v5.16' of git://git./linux/kernel/git/joel/fsi into char-misc-next

Joel writes:

FSI changes for v5.16

 - SBEFIFO usersapce interfaces to perform FFDC (First Failure
   Data Capture) and detect timeouts

 - A fix to handle multiple messages in flight

* tag 'fsi-for-v5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/joel/fsi:
  fsi: sbefifo: Use interruptible mutex locking
  fsi: sbefifo: Add sysfs file indicating a timeout error
  docs: ABI: testing: Document the SBEFIFO timeout interface
  hwmon: (occ) Provide the SBEFIFO FFDC in binary sysfs
  docs: ABI: testing: Document the OCC hwmon FFDC binary interface
  fsi: occ: Store the SBEFIFO FFDC in the user response buffer
  fsi: occ: Use a large buffer for responses
  hwmon: (occ) Remove sequence numbering and checksum calculation
  fsi: occ: Force sequence numbering per OCC

2 years agofsi: sbefifo: Use interruptible mutex locking
Eddie James [Tue, 3 Aug 2021 21:30:16 +0000 (16:30 -0500)]
fsi: sbefifo: Use interruptible mutex locking

Some SBE operations have extremely large responses and can require
several minutes to process the response. During this time, the device
lock must be held. If another process attempts an operation, it will
wait for the mutex for longer than the kernel hung task watchdog
allows. Therefore, use the interruptible function to lock the mutex.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20210803213016.44739-1-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agofsi: sbefifo: Add sysfs file indicating a timeout error
Eddie James [Tue, 19 Oct 2021 21:17:49 +0000 (16:17 -0500)]
fsi: sbefifo: Add sysfs file indicating a timeout error

The SBEFIFO timeout error requires special handling in userspace
to do recovery operations. Add a sysfs file to indicate a timeout
error, and notify pollers when a timeout occurs.

This will be used by the openpower-occ-control application.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20211019211749.38059-3-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agodocs: ABI: testing: Document the SBEFIFO timeout interface
Eddie James [Tue, 19 Oct 2021 21:17:48 +0000 (16:17 -0500)]
docs: ABI: testing: Document the SBEFIFO timeout interface

Add documentation for the new sysfs entry that indicates whether or not
the SBE has timed out.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Link: https://lore.kernel.org/r/20211019211749.38059-2-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agohwmon: (occ) Provide the SBEFIFO FFDC in binary sysfs
Eddie James [Tue, 19 Oct 2021 20:53:07 +0000 (15:53 -0500)]
hwmon: (occ) Provide the SBEFIFO FFDC in binary sysfs

Save any FFDC provided by the OCC driver, and provide it to userspace
through a binary sysfs entry. Notify userspace pollers when there is an
error too.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20211019205307.36946-5-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agodocs: ABI: testing: Document the OCC hwmon FFDC binary interface
Eddie James [Tue, 19 Oct 2021 20:53:06 +0000 (15:53 -0500)]
docs: ABI: testing: Document the OCC hwmon FFDC binary interface

Add documentation for the new binary sysfs that will dump the SBEFIFO
FFDC.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Link: https://lore.kernel.org/r/20211019205307.36946-4-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agofsi: occ: Store the SBEFIFO FFDC in the user response buffer
Eddie James [Tue, 19 Oct 2021 20:53:05 +0000 (15:53 -0500)]
fsi: occ: Store the SBEFIFO FFDC in the user response buffer

If the SBEFIFO response indicates an error, store the response in the
user buffer and return an error. Previously, the user had no way of
obtaining the SBEFIFO FFDC.

The user's buffer now contains data in the event of a failure. No change
in the event of a successful transfer.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Link: https://lore.kernel.org/r/20211019205307.36946-3-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agofsi: occ: Use a large buffer for responses
Eddie James [Tue, 19 Oct 2021 20:53:04 +0000 (15:53 -0500)]
fsi: occ: Use a large buffer for responses

Allocate a large buffer for each OCC to handle response data. This
removes memory allocation during an operation, and also allows for
the maximum amount of SBE FFDC.

Previously for the putsram and attn commands, only 32 words would have
been available, and for getsram, only up to the size of the transfer.
SBE FFDC might be up to 8Kb.

The SBE interface expects data to be specified in units of words (4
bytes), defined as OCC_MAX_RESP_WORDS.

This change allows the full FFDC capture to be implemented, where before
it was not available.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Link: https://lore.kernel.org/r/20211019205307.36946-2-eajames@linux.ibm.com
Signed-off-by: Joel Stanley <joel@jms.id.au>
2 years agocounter: Fix use-after-free race condition for events_queue_size write
William Breathitt Gray [Thu, 21 Oct 2021 10:35:40 +0000 (19:35 +0900)]
counter: Fix use-after-free race condition for events_queue_size write

A race condition is possible when writing to events_queue_size where the
events kfifo is freed during the execution of a kfifo_in(), resulting in
a use-after-free. This patch prevents such a scenario by protecting the
events queue in operation with a spinlock and locking before performing
the events queue size adjustment.

The existing events_lock mutex is renamed to events_out_lock to reflect
that it only protects events queue out operations. Because the events
queue in operations can occur in an interrupt context, a new
events_in_lock spinlock is introduced and utilized.

Fixes: feff17a550c7 ("counter: Implement events_queue_size sysfs attribute")
Cc: David Lechner <david@lechnology.com>
Signed-off-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Link: https://lore.kernel.org/r/20211021103540.955639-1-vilhelm.gray@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agocounter: Cleanup lingering atomic.h includes
William Breathitt Gray [Thu, 21 Oct 2021 10:35:14 +0000 (19:35 +0900)]
counter: Cleanup lingering atomic.h includes

With the removal of the chrdev_lock atomic flag, the counter-sysfs.c and
counter-chrdev.c no longer needs to include the atomic.h header file.

Signed-off-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Link: https://lore.kernel.org/r/20211021103514.955622-1-vilhelm.gray@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agoMAINTAINERS: Update the entry for MHI bus
Manivannan Sadhasivam [Tue, 19 Oct 2021 13:39:01 +0000 (19:09 +0530)]
MAINTAINERS: Update the entry for MHI bus

Since Hemant is not carrying out any maintainership duties let's make
him as a dedicated reviewer. Also add the new mailing lists dedicated
for MHI in subspace mailing list server.

Reviewed-by: Hemant Kumar <hemantk@codeaurora.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Link: https://lore.kernel.org/r/20211019133901.173966-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agoMerge tag 'iio-for-5.16a-split-take4' of https://git.kernel.org/pub/scm/linux/kernel...
Greg Kroah-Hartman [Tue, 19 Oct 2021 09:44:28 +0000 (11:44 +0200)]
Merge tag 'iio-for-5.16a-split-take4' of https://git./linux/kernel/git/jic23/iio into char-misc-next

Jonathan writes:

First set of IIO new device and feature support for the 5.16 cycle

Counter subsystem changes now sent separately.

This has been a busy cycle, so lots here and a few more stragglers to
come next week.

Big new feature in this cycle is probably output buffer support.
This has been in the works for a very long time so it's great to see
Mihail pick up the challenge and build upon his predecessors work to finally
bring this feature to mainline.

New device support
------------------

* adi,adxl313
  - New driver and dt bindings for this low power accelerometer.
* adi,adxl355
  - New driver and dt bindings for this accelerometer.
  - Later series adds buffer support.
* asahi-kasei,ak8975
  - Minor additions to driver to support ak09916
* aspeed,aspeed-adc
  - Substantial rework plus feature additions to add support for the
    ast2600 including a new dt bindings doc.
* atmel,at91_sama5d2
  - Rework and support introduced for the sama7g5 parts.
* maxim,max31865
  - New driver and bindings for this RTD temperature sensor chip.
* nxp,imx8qxp
  - New driver and bindings for the ADC found on the i.MX 8QuadXPlus Soc.
* senseair,sunrise
  - New driver and bindings for this family of carbon dioxide gas sensors.
* sensiron,scd4x
  - New driver and bindings for this carbon dioxide gas sensor.

New features
------------

* Output buffer support.  Works in a similar fashion to input buffers, but
  in this case userspace pushes data into the kfifo which is then drained
  to the device when a trigger occurs.  Support added to the ad5766 DAC
  driver.
* Core, devm_iio_map_array_register() to avoid need for
  devm_add_action_or_reset() based cleanup in fully managed allocation
  drivers.
* Core iio_push_to_buffers_with_ts_unaligned() function to safely handle a
  few drivers where it really hard to ensure the correct data alignment in
  an iio_push_to_buffers_with_timestamp() call. Note this uses a bounce
  buffer so should be avoided whenever possible.  Used in the ti,adc108s102,
  invense,mpu3050 and adi,adis16400.  This closes the last   known set
  of drivers with alignment issues at this interface.
* maxim,max1027
  - Substantial rework to this driver main target of which was supporting
    use of other triggers than it's own EOC interrupt.
  - Transfer optimization.
* nxp,fxls8962af
  - Threshold even support including using it as a wakeup source.

Cleanups, minor fixes etc
-------------------------

Chances of a common type to multiple drivers:

* devm_ conversion and drop of .remove() callbacks in:
  - adi,ad5064
  - adi,ad7291
  - adi,ad7303
  - adi,ad7746
  - adi,ad9832
  - adi,adis16080
  - dialog,da9150-gpadc
  - intel,mrfld_adc
  - marvell,berlin2
  - maxim,max1363
  - maxim,max44000
  - nuvoton,nau7802
  - st_sensors (includes a lot of rework!)
  - ti,ads8344
  - ti,lp8788

* devm_platform_ioremap_resource() used to reduce boilerplate
  - cirrus,ep93xx
  - rockchip,saradc
  - stm,stm32-dac

* Use dev_err_probe() in more places to both not print on deferred probe and
  ensure a reason for the deferral is available for debug purposes.
  - adi,ad8801
  - capella,cm36651
  - linear,ltc1660
  - maxim,ds4424
  - maxim,max5821
  - microchip,mcp4922
  - nxp,lpc18xx
  - onnn,noa1305
  - st,lsm9ds0
  - st,st_sensors
  - st,stm32-dac
  - ti,afe4403
  - ti,afe4404
  - ti,dac7311

* Drop error returns in SPI and I2C remove() functions as they are ignored and
  long term plan is to change these all over to returning void. In some cases
  these patches just make it 'obvious' they always return 0 where it was the
  case before but not easy to tell.
  - adi,ad5380
  - adi,ad5446
  - adi,ad5686
  - adi,ad5592r
  - bosch,bma400
  - bosch,bmc150
  - fsl,mma7455
  - honeywell,hmc5843
  - kionix,kxsd9
  - maxim,max5487
  - meas,ms5611
  - ti,afe4403

Driver specific changes

* adi,ad5770r
  - Bring driver inline with documented bindings.
* adi,ad7746
  - Trivial style fix
* adi,ad7949
  - Express some magic values as the underlying parts via new #defines.
  - Make it work with SPI controllers that don't support 14 or 16 bit messages
  - Support selection of voltage reference from dt including expanding the
    dt-bindings to cover this new functionality.
* adi,ad799x
  - Implement selection of external reference voltage on AD7991, AD7995 and
    AD7999.
  - Add missing dt-bindings doc for devices supported by this driver.
* adi,adislib
  - Move interrupt startup to better location in startup flow.
  - Handle devices that cannot mask/unmask the drdy pin and must instead mask
    at the interrupt controller.  Applies to the adis16460 and adis16475 from
    which we then drop equivalent code.
* adi,ltc2983
  - Add support for optional reset pin.
  - Fail to probe if no channels specified in dt binding.
* asahi-kasei,ak8975
  - dt-binding additions of missing vid-supply regulator.
* aspeed,aspeed-adc
  - Typo fix.
* fsl,mma7660
  - Mark acpi_device_id table __maybe_unused to avoid build warning.
* fsl,imx25-gcq
  - Avoid initializing regulators that aren't used.
* invensense,mpu3050
  - Drop a dead protection against a clash with the old input driver.
* invensense,mpu6050
  - Rework code to not use strcpy() and hence avoid possibility of wrong sized
    buffers. Note this wasn't a bug, but the new code is a lot more readable.
  - Mark acpi_device_id table __maybe_unused to avoid build warning.
* kionix,kxcjk1013
  - dt-binding addition to note it supports interrupts.
* marvell,berlin2-adc
  - Enable COMPILE_TEST building.
* maxim,max1027
  - Avoid returning success in an error path.
* nxp,imx8qxp
  - Fix warning when runtime pm not enabled via __maybe_unused.
* ricoh,rn5t618
  - Use the new devm_iio_map_array_register() instead of open coding the same.
* samsung,exynos_adc
  - Improve kconfig help text.
* st,lsm6dsx
  - Move max_fifo_size into the fifo_ops structure where the other configuration
    parameters are found.
* st,st_sensors:
  - Reorder to ensure we turn the power off after removing userspace interfaces.
* senseair,sunrise
  - Add missing I2C dependency.
* ti,twl6030
  - Small code tidy up.

* tag 'iio-for-5.16a-split-take4' of https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio: (148 commits)
  iio: imx8qxp-adc: mark PM functions as __maybe_unused
  iio: pressure: ms5611: Make ms5611_remove() return void
  iio: potentiometer: max5487: Don't return an error in .remove()
  iio: magn: hmc5843: Make hmc5843_common_remove() return void
  iio: health: afe4403: Don't return an error in .remove()
  iio: dac: ad5686: Make ad5686_remove() return void
  iio: dac: ad5592r: Make ad5592r_remove() return void
  iio: dac: ad5446: Make ad5446_remove() return void
  iio: dac: ad5380: Make ad5380_remove() return void
  iio: accel: mma7455: Make mma7455_core_remove() return void
  iio: accel: kxsd9: Make kxsd9_common_remove() return void
  iio: accel: bmi088: Make bmi088_accel_core_remove() return void
  iio: accel: bmc150: Make bmc150_accel_core_remove() return void
  iio: accel: bma400: Make bma400_remove() return void
  drivers:iio:dac:ad5766.c: Add trigger buffer
  iio: triggered-buffer: extend support to configure output buffers
  iio: kfifo-buffer: Add output buffer support
  iio: Add output buffer support
  iio: documentation: Document scd4x calibration use
  drivers: iio: chemical: Add support for Sensirion SCD4x CO2 sensor
  ...

2 years agocounter: drop chrdev_lock
David Lechner [Sun, 17 Oct 2021 18:55:21 +0000 (13:55 -0500)]
counter: drop chrdev_lock

This removes the chrdev_lock from the counter subsystem. This was
intended to prevent opening the chrdev more than once. However, this
doesn't work in practice since userspace can duplicate file descriptors
and pass file descriptors to other processes. Since this protection
can't be relied on, it is best to just remove it.

Suggested-by: Greg KH <gregkh@linuxfoundation.org>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Signed-off-by: David Lechner <david@lechnology.com>
Link: https://lore.kernel.org/r/20211017185521.3468640-1-david@lechnology.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agocounter/counter-sysfs: use sysfs_emit everywhere
David Lechner [Sun, 17 Oct 2021 19:01:06 +0000 (14:01 -0500)]
counter/counter-sysfs: use sysfs_emit everywhere

In the counter subsystem, we are already using sysfs_emit(), but there
were a few places where we were still using sprintf() in *_show()
functions. For consistency and added protections, use sysfs_emit()
everywhere.

Suggested-by: Greg KH <gregkh@linuxfoundation.org>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Signed-off-by: David Lechner <david@lechnology.com>
Link: https://lore.kernel.org/r/20211017190106.3472645-1-david@lechnology.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agochar: xillybus: fix msg_ep UAF in xillyusb_probe()
Ziyang Xuan [Sat, 16 Oct 2021 05:20:47 +0000 (13:20 +0800)]
char: xillybus: fix msg_ep UAF in xillyusb_probe()

When endpoint_alloc() return failed in xillyusb_setup_base_eps(),
'xdev->msg_ep' will be freed but not set to NULL. That lets program
enter fail handling to cleanup_dev() in xillyusb_probe(). Check for
'xdev->msg_ep' is invalid in cleanup_dev() because 'xdev->msg_ep' did
not set to NULL when was freed. So the UAF problem for 'xdev->msg_ep'
is triggered.

==================================================================
BUG: KASAN: use-after-free in fifo_mem_release+0x1f4/0x210
CPU: 0 PID: 166 Comm: kworker/0:2 Not tainted 5.15.0-rc5+ #19
Call Trace:
 dump_stack_lvl+0xe2/0x152
 print_address_description.constprop.0+0x21/0x140
 ? fifo_mem_release+0x1f4/0x210
 kasan_report.cold+0x7f/0x11b
 ? xillyusb_probe+0x530/0x700
 ? fifo_mem_release+0x1f4/0x210
 fifo_mem_release+0x1f4/0x210
 ? __sanitizer_cov_trace_pc+0x1d/0x50
 endpoint_dealloc+0x35/0x2b0
 cleanup_dev+0x90/0x120
 xillyusb_probe+0x59a/0x700
...

Freed by task 166:
 kasan_save_stack+0x1b/0x40
 kasan_set_track+0x1c/0x30
 kasan_set_free_info+0x20/0x30
 __kasan_slab_free+0x109/0x140
 kfree+0x117/0x4c0
 xillyusb_probe+0x606/0x700

Set 'xdev->msg_ep' to NULL after being freed in xillyusb_setup_base_eps()
to fix the UAF problem.

Fixes: a53d1202aef1 ("char: xillybus: Add driver for XillyUSB (Xillybus variant for USB)")
Cc: stable <stable@vger.kernel.org>
Acked-by: Eli Billauer <eli.billauer@gmail.com>
Signed-off-by: Ziyang Xuan <william.xuanziyang@huawei.com>
Link: https://lore.kernel.org/r/20211016052047.1611983-1-william.xuanziyang@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agobinder: don't detect sender/target during buffer cleanup
Todd Kjos [Fri, 15 Oct 2021 23:38:11 +0000 (16:38 -0700)]
binder: don't detect sender/target during buffer cleanup

When freeing txn buffers, binder_transaction_buffer_release()
attempts to detect whether the current context is the target by
comparing current->group_leader to proc->tsk. This is an unreliable
test. Instead explicitly pass an 'is_failure' boolean.

Detecting the sender was being used as a way to tell if the
transaction failed to be sent.  When cleaning up after
failing to send a transaction, there is no need to close
the fds associated with a BINDER_TYPE_FDA object. Now
'is_failure' can be used to accurately detect this case.

Fixes: 44d8047f1d87 ("binder: use standard functions to allocate fds")
Cc: stable <stable@vger.kernel.org>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Todd Kjos <tkjos@google.com>
Link: https://lore.kernel.org/r/20211015233811.3532235-1-tkjos@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2 years agoMerge tag 'misc-habanalabs-next-2021-10-18' of https://git.kernel.org/pub/scm/linux...
Greg Kroah-Hartman [Tue, 19 Oct 2021 07:37:59 +0000 (09:37 +0200)]
Merge tag 'misc-habanalabs-next-2021-10-18' of https://git./linux/kernel/git/ogabbay/linux into char-misc-next

Oded writes:

This tag contains habanalabs driver changes for v5.16:

- Add a new uAPI (under the memory ioctl) to request from the driver
  to export a DMA-BUF object that represents a memory region on
  the device's DRAM. This is needed to enable peer-to-peer over PCIe
  between habana device and an RDMA adapter (e.g. mlnx5 or efa
  rdma adapter).

- Add debugfs node to dynamically configure CS timeout. Up until now,
  it was only configurable through kernel module parameter.

- Fetch more comprehensive power information from the firmware.

- Always take timestamp when waiting for user interrupt, as the user
  needs that information to optimize the graph runtime compilation.

- Modify user interrupt to look on 64-bit user value as fence, instead
  of 32-bit.

- Bypass reset in case of repeated h/w error event after device reset.
  This is to prevent endless loop of resets to the device.

- Fix several bugs in multi CS completion code.

- Fix race condition in fd close/open.

- Update to latest firmware headers

- Add select CRC32 in kconfig

- Small fixes, cosmetics

* tag 'misc-habanalabs-next-2021-10-18' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux: (25 commits)
  habanalabs: refactor fence handling in hl_cs_poll_fences
  habanalabs: context cleanup cosmetics
  habanalabs: simplify wait for interrupt with timestamp flow
  habanalabs: initialize hpriv fields before adding new node
  habanalabs: Unify frequency set/get functionality
  habanalabs: select CRC32
  habanalabs: add support for dma-buf exporter
  habanalabs: define uAPI to export FD for DMA-BUF
  habanalabs: fix NULL pointer dereference
  habanalabs: fix race condition in multi CS completion
  habanalabs: use only u32
  habanalabs: update firmware files
  habanalabs: bypass reset for continuous h/w error event
  habanalabs: take timestamp on wait for interrupt
  habanalabs: prevent race between fd close/open
  habanalabs: refactor reset log message
  habanalabs: define soft-reset as inference op
  habanalabs: fix debugfs device memory MMU VA translation
  habanalabs: add support for a long interrupt target value
  habanalabs: remove redundant cs validity checks
  ...

2 years agoiio: imx8qxp-adc: mark PM functions as __maybe_unused
Arnd Bergmann [Wed, 13 Oct 2021 14:43:26 +0000 (16:43 +0200)]
iio: imx8qxp-adc: mark PM functions as __maybe_unused

Without CONFIG_PM_SLEEP, the runtime suspend/resume functions
are unused, producing a warning:

drivers/iio/adc/imx8qxp-adc.c:433:12: error: 'imx8qxp_adc_runtime_resume' defined but not used [-Werror=unused-function]
  433 | static int imx8qxp_adc_runtime_resume(struct device *dev)
      |            ^~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/iio/adc/imx8qxp-adc.c:419:12: error: 'imx8qxp_adc_runtime_suspend' defined but not used [-Werror=unused-function]
  419 | static int imx8qxp_adc_runtime_suspend(struct device *dev)
      |            ^~~~~~~~~~~~~~~~~~~~~~~~~~~

Mark them as __maybe_unused to shut up the compiler.

Fixes: 1e23dcaa1a9f ("iio: imx8qxp-adc: Add driver support for NXP IMX8QXP ADC")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20211013144338.2261316-1-arnd@kernel.org
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: pressure: ms5611: Make ms5611_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:23 +0000 (22:32 +0200)]
iio: pressure: ms5611: Make ms5611_remove() return void

Up to now ms5611_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-16-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: potentiometer: max5487: Don't return an error in .remove()
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:22 +0000 (22:32 +0200)]
iio: potentiometer: max5487: Don't return an error in .remove()

The only effect of returning an error in an spi .remove() callback is
that the spi core issues a generic warning message. Instead emit a more
specific error message and return 0 to not report the same issue twice.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-15-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: magn: hmc5843: Make hmc5843_common_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:21 +0000 (22:32 +0200)]
iio: magn: hmc5843: Make hmc5843_common_remove() return void

Up to now hmc5843_common_remove() returns zero unconditionally. Make it
return void instead which makes it easier to see in the callers that
there is no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-14-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: health: afe4403: Don't return an error in .remove()
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:19 +0000 (22:32 +0200)]
iio: health: afe4403: Don't return an error in .remove()

The only effect of returning an error in an spi .remove() callback is
that the spi core issues another warning message. Don't report the same
problem twice and return 0 unconditionally instead. Also degrade the log
level to warning, as nothing really bad is expected from a failure to
put the device in suspend mode.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-12-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ad5686: Make ad5686_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:18 +0000 (22:32 +0200)]
iio: dac: ad5686: Make ad5686_remove() return void

Up to now ad5686_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-11-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ad5592r: Make ad5592r_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:17 +0000 (22:32 +0200)]
iio: dac: ad5592r: Make ad5592r_remove() return void

Up to now ad5592r_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-10-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ad5446: Make ad5446_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:16 +0000 (22:32 +0200)]
iio: dac: ad5446: Make ad5446_remove() return void

Up to now ad5446_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-9-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ad5380: Make ad5380_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:15 +0000 (22:32 +0200)]
iio: dac: ad5380: Make ad5380_remove() return void

Up to now ad5380_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-8-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: mma7455: Make mma7455_core_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:13 +0000 (22:32 +0200)]
iio: accel: mma7455: Make mma7455_core_remove() return void

Up to now mma7455_core_remove() returns zero unconditionally. Make it
return void instead which makes it easier to see in the callers that
there is no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-6-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: kxsd9: Make kxsd9_common_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:12 +0000 (22:32 +0200)]
iio: accel: kxsd9: Make kxsd9_common_remove() return void

Up to now kxsd9_common_remove() returns zero unconditionally. Make it
return void instead which makes it easier to see in the callers that
there is no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-5-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: bmi088: Make bmi088_accel_core_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:11 +0000 (22:32 +0200)]
iio: accel: bmi088: Make bmi088_accel_core_remove() return void

Up to now bmi088_accel_core_remove() returns zero unconditionally. Make
it return void instead which makes it easier to see in the callers that
there is no error to handle.

Also the return value of spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-4-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: bmc150: Make bmc150_accel_core_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:10 +0000 (22:32 +0200)]
iio: accel: bmc150: Make bmc150_accel_core_remove() return void

Up to now bmc150_accel_core_remove() returns zero unconditionally. Make
it return void instead which makes it easier to see in the callers that
there is no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-3-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: bma400: Make bma400_remove() return void
Uwe Kleine-König [Wed, 13 Oct 2021 20:32:09 +0000 (22:32 +0200)]
iio: accel: bma400: Make bma400_remove() return void

When an i2c or spi driver's remove function returns a non-zero error
code nothing happens apart from emitting a generic error message. Make
this error message more device specific and return zero instead in the
remove callbacks. As the return value of bma400_remove() is unused then,
change the function to not yield a return value.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211013203223.2694577-2-u.kleine-koenig@pengutronix.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agodrivers:iio:dac:ad5766.c: Add trigger buffer
Mihail Chindris [Thu, 7 Oct 2021 08:00:37 +0000 (08:00 +0000)]
drivers:iio:dac:ad5766.c: Add trigger buffer

This chip is able to generate waveform and using an
with the output trigger buffer will be easy to generate one.

Signed-off-by: Mihail Chindris <mihail.chindris@analog.com>
Reviewed-by: Alexandru Ardelean <ardeleanalex@gmail.com>
Link: https://lore.kernel.org/r/20211007080035.2531-7-mihail.chindris@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: triggered-buffer: extend support to configure output buffers
Alexandru Ardelean [Thu, 7 Oct 2021 08:00:32 +0000 (08:00 +0000)]
iio: triggered-buffer: extend support to configure output buffers

Now that output (kfifo) buffers are supported, we need to extend the
{devm_}iio_triggered_buffer_setup_ext() parameter list to take a direction
parameter.

This allows us to attach an output triggered buffer to a DAC device.
Unfortunately it's a bit difficult to add another macro to avoid changing 5
drivers where {devm_}iio_triggered_buffer_setup_ext() is used.
Well, it's doable, but may not be worth the trouble vs just updating all
these 5 drivers.

Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Mihail Chindris <mihail.chindris@analog.com>
Link: https://lore.kernel.org/r/20211007080035.2531-4-mihail.chindris@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: kfifo-buffer: Add output buffer support
Lars-Peter Clausen [Thu, 7 Oct 2021 08:00:31 +0000 (08:00 +0000)]
iio: kfifo-buffer: Add output buffer support

Add output buffer support to the kfifo buffer implementation.

The implementation is straight forward and mostly just wraps the kfifo
API to provide the required operations.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Mihail Chindris <mihail.chindris@analog.com>
Link: https://lore.kernel.org/r/20211007080035.2531-3-mihail.chindris@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: Add output buffer support
Mihail Chindris [Thu, 7 Oct 2021 08:00:30 +0000 (08:00 +0000)]
iio: Add output buffer support

Currently IIO only supports buffer mode for capture devices like ADCs. Add
support for buffered mode for output devices like DACs.

The output buffer implementation is analogous to the input buffer
implementation. Instead of using read() to get data from the buffer write()
is used to copy data into the buffer.

poll() with POLLOUT will wakeup if there is space available.

Drivers can remove data from a buffer using iio_pop_from_buffer(), the
function can e.g. called from a trigger handler to write the data to
hardware.

A buffer can only be either a output buffer or an input, but not both. So,
for a device that has an ADC and DAC path, this will mean 2 IIO buffers
(one for each direction).

The direction of the buffer is decided by the new direction field of the
iio_buffer struct and should be set after allocating and before registering
it.

Co-developed-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Co-developed-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Mihail Chindris <mihail.chindris@analog.com>
Link: https://lore.kernel.org/r/20211007080035.2531-2-mihail.chindris@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: documentation: Document scd4x calibration use
Roan van Dijk [Fri, 8 Oct 2021 10:17:06 +0000 (12:17 +0200)]
iio: documentation: Document scd4x calibration use

Add entries from Documentation/ABI/testing/sysfs-bus-iio-scd30
to Documentation/ABI/testing/sysfs-bus-iio. The attributes of the scd4x
and scd30 are common.

Remove Documentation/ABI/testing/sysfs-bus-iio-scd30.

Signed-off-by: Roan van Dijk <roan@protonic.nl>
Link: https://lore.kernel.org/r/20211008101706.755942-5-roan@protonic.nl
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agodrivers: iio: chemical: Add support for Sensirion SCD4x CO2 sensor
Roan van Dijk [Fri, 8 Oct 2021 10:17:05 +0000 (12:17 +0200)]
drivers: iio: chemical: Add support for Sensirion SCD4x CO2 sensor

This is a driver for the SCD4x CO2 sensor from Sensirion. The sensor is
able to measure CO2 concentration, temperature and relative humdity.
The sensor uses a photoacoustic principle for measuring CO2 concentration.
An I2C interface is supported by this driver in order to communicate with
the sensor.

Signed-off-by: Roan van Dijk <roan@protonic.nl>
Link: https://lore.kernel.org/r/20211008101706.755942-4-roan@protonic.nl
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoMAINTAINERS: Add myself as maintainer of the scd4x driver
Roan van Dijk [Fri, 8 Oct 2021 10:17:04 +0000 (12:17 +0200)]
MAINTAINERS: Add myself as maintainer of the scd4x driver

Signed-off-by: Roan van Dijk <roan@protonic.nl>
Link: https://lore.kernel.org/r/20211008101706.755942-3-roan@protonic.nl
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agodt-bindings: iio: chemical: sensirion,scd4x: Add yaml description
Roan van Dijk [Fri, 8 Oct 2021 10:17:03 +0000 (12:17 +0200)]
dt-bindings: iio: chemical: sensirion,scd4x: Add yaml description

Add documentation for the SCD4x carbon dioxide sensor from Sensirion.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Roan van Dijk <roan@protonic.nl>
Link: https://lore.kernel.org/r/20211008101706.755942-2-roan@protonic.nl
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: light: noa1305: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:41:55 +0000 (09:41 +0800)]
iio: light: noa1305: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014156.1491-4-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: light: cm36651: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:41:53 +0000 (09:41 +0800)]
iio: light: cm36651: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014156.1491-2-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: health: afe4404: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:44:02 +0000 (09:44 +0800)]
iio: health: afe4404: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014403.1563-2-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: health: afe4403: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:44:01 +0000 (09:44 +0800)]
iio: health: afe4403: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014403.1563-1-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: st_lsm9ds0: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:40:54 +0000 (09:40 +0800)]
iio: st_lsm9ds0: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014055.1431-2-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: st_sensors: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:40:53 +0000 (09:40 +0800)]
iio: st_sensors: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928014055.1431-1-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ti-dac7311: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:39:01 +0000 (09:39 +0800)]
iio: dac: ti-dac7311: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-8-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: stm32-dac: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:39:00 +0000 (09:39 +0800)]
iio: dac: stm32-dac: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-7-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: mcp4922: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:59 +0000 (09:38 +0800)]
iio: dac: mcp4922: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-6-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: max5821: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:58 +0000 (09:38 +0800)]
iio: dac: max5821: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-5-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ds4424: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:57 +0000 (09:38 +0800)]
iio: dac: ds4424: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-4-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ltc1660: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:56 +0000 (09:38 +0800)]
iio: dac: ltc1660: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Reviewed-by: Marcus Folkesson <marcus.folkesson@gmail.com>
Link: https://lore.kernel.org/r/20210928013902.1341-3-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: lpc18xx_dac: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:55 +0000 (09:38 +0800)]
iio: dac: lpc18xx_dac: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-2-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: dac: ad8801: Make use of the helper function dev_err_probe()
Cai Huoqing [Tue, 28 Sep 2021 01:38:54 +0000 (09:38 +0800)]
iio: dac: ad8801: Make use of the helper function dev_err_probe()

When possible use dev_err_probe help to properly deal with the
PROBE_DEFER error, the benefit is that DEFER issue will be logged
in the devices_deferred debugfs file.
Using dev_err_probe() can reduce code size, and the error value
gets printed.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210928013902.1341-1-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agodt-bindings: iio: ad779x: Add binding document
Florian Boor [Thu, 30 Sep 2021 10:42:49 +0000 (12:42 +0200)]
dt-bindings: iio: ad779x: Add binding document

New binding documentation for AD799x series of I²C ADC ICs.

Signed-off-by: Florian Boor <florian.boor@kernelconcepts.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210930104249.2924336-2-florian.boor@kernelconcepts.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: ad799x: Implement selecting external reference voltage input on AD7991...
Florian Boor [Thu, 30 Sep 2021 10:42:48 +0000 (12:42 +0200)]
iio: adc: ad799x: Implement selecting external reference voltage input on AD7991, AD7995 and AD7999.

Make use of the AD7991_REF_SEL bit and support using the external
reference voltage if 'vref-supply' is present. Use VCC voltage supply
as reference if no extra reference is supplied.

Signed-off-by: Florian Boor <florian.boor@kernelconcepts.de>
Link: https://lore.kernel.org/r/20210930104249.2924336-1-florian.boor@kernelconcepts.de
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: fix error code in max1027_wait_eoc()
Dan Carpenter [Mon, 4 Oct 2021 13:44:54 +0000 (16:44 +0300)]
iio: adc: max1027: fix error code in max1027_wait_eoc()

Return -ETIMEDOUT on timeout instead of success.

Fixes: 1f7b4048b31b ("iio: adc: max1027: Use the EOC IRQ when populated for single reads")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: imu: adis16400: Fix buffer alignment requirements.
Jonathan Cameron [Sun, 13 Jun 2021 15:10:39 +0000 (16:10 +0100)]
iio: imu: adis16400: Fix buffer alignment requirements.

iio_push_to_buffers_with_timestamp() requires that the buffer
is 8 byte alignment to ensure an inserted timestamp is naturally aligned.

This requirement was not met here when burst mode is in use beause
of a leading u16. Use the new iio_push_to_buffers_with_ts_unaligned()
function that has more relaxed requirements.

It is somewhat complex to access that actual data length, but a
safe bound can be found by using scan_bytes - sizeof(timestamp) so that
is used in this path.

More efficient approaches exist, but this ensure correctness at the
cost of using a bounce buffer.

Fixes: 5075e0720d93 ("iio: imu: adis: generalize burst mode support")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210613151039.569883-5-jic23@kernel.org
2 years agoiio: gyro: mpu3050: Fix alignment and size issues with buffers.
Jonathan Cameron [Sun, 13 Jun 2021 15:10:38 +0000 (16:10 +0100)]
iio: gyro: mpu3050: Fix alignment and size issues with buffers.

Fix a set of closely related issues.
1. When using fifo_values() there was not enough space for the timestamp to
   be inserted by iio_push_to_buffers_with_timestamp()
2. fifo_values() did not meet the alignment requirement of
   iio_push_to_buffers_with_timestamp()
3. hw_values did not meet the alignment requirement either.

1 and 2 fixed by using new iio_push_to_buffers_with_ts_unaligned() which has
no alignment or space padding requirements.
3 fixed by introducing a structure that makes the space and alignment
requirements explicit.

Fixes: 3904b28efb2c ("iio: gyro: Add driver for the MPU-3050 gyroscope")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20210613151039.569883-4-jic23@kernel.org
2 years agoiio: adc: ti-adc108s102: Fix alignment of buffer pushed to iio buffers.
Jonathan Cameron [Sun, 13 Jun 2021 15:10:37 +0000 (16:10 +0100)]
iio: adc: ti-adc108s102: Fix alignment of buffer pushed to iio buffers.

Use the newly introduce iio_push_to_buffers_with_ts_unaligned() function
to ensure a bounce buffer is used to provide the required alignment and
space padding needed by the IIO core which requires the timestamp
is naturally aligned.  There will be a performance cost to this change
but it will ensure the driver works on platforms that do not support
unaligned 8 byte assignments, and with consumer drivers that may
assume natural alignment of the timestamp.

Issue found as part of an audit of all calls to
iio_push_to_buffers_with_timestamp()

Fixes: 7e87d11c9bda ("iio: adc: Add support for TI ADC108S102 and ADC128S102")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Link: https://lore.kernel.org/r/20210613151039.569883-3-jic23@kernel.org
2 years agoiio: core: Introduce iio_push_to_buffers_with_ts_unaligned()
Jonathan Cameron [Sun, 13 Jun 2021 15:10:36 +0000 (16:10 +0100)]
iio: core: Introduce iio_push_to_buffers_with_ts_unaligned()

Whilst it is almost always possible to arrange for scan data to be
read directly into a buffer that is suitable for passing to
iio_push_to_buffers_with_timestamp(), there are a few places where
leading data needs to be skipped over.

For these cases introduce a function that will allocate an appropriate
sized and aligned bounce buffer (if not already allocated) and copy
the unaligned data into that before calling
iio_push_to_buffers_with_timestamp() on the bounce buffer.
We tie the lifespace of this buffer to that of the iio_dev.dev
which should ensure no memory leaks occur.

Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210613151039.569883-2-jic23@kernel.org
2 years agoiio: chemical: SENSEAIR_SUNRISE_CO2 depends on I2C
Randy Dunlap [Sat, 2 Oct 2021 23:28:03 +0000 (16:28 -0700)]
iio: chemical: SENSEAIR_SUNRISE_CO2 depends on I2C

Fix kconfig symbol dependency warning:

WARNING: unmet direct dependencies detected for REGMAP_I2C
  Depends on [n]: I2C [=n]
  Selected by [y]:
  - SENSEAIR_SUNRISE_CO2 [=y] && IIO [=y]

Fixes: c397894e24f1 ("iio: chemical: Add Senseair Sunrise 006-0-007 drive")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Jacopo Mondi <jacopo@jmondi.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: linux-iio@vger.kernel.org
Link: https://lore.kernel.org/r/20211002232803.5108-1-rdunlap@infradead.org
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: aspeed: Fix spelling mistake "battey" -> "battery"
Colin Ian King [Fri, 1 Oct 2021 12:00:18 +0000 (13:00 +0100)]
iio: adc: aspeed: Fix spelling mistake "battey" -> "battery"

There is a spelling mistake in a dev_warn message. Fix it.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20211001120018.17570-1-colin.king@canonical.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: ad7291: convert probe to device-managed only
Alexandru Ardelean [Sun, 26 Sep 2021 19:43:15 +0000 (22:43 +0300)]
iio: adc: ad7291: convert probe to device-managed only

This is a simple conversion for to device-managed with using
devm_request_threaded_irq(), disabling the regulator via a
devm_add_action_or_reset() hook and finally using
devm_iio_device_register().

The i2c_set_clientdata() call is removed as it becomes redundant after this
change.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210926194315.7742-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: Kconfig: add COMPILE_TEST dep for berlin2-adc
Alexandru Ardelean [Sun, 26 Sep 2021 19:26:42 +0000 (22:26 +0300)]
iio: adc: Kconfig: add COMPILE_TEST dep for berlin2-adc

Otherwise most build checks will omit this driver from a compile-test due
to it's dependency only on the BERLIN_ARCH symbol.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210926192642.4051329-2-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: berlin2-adc: convert probe to device-managed only
Alexandru Ardelean [Sun, 26 Sep 2021 19:26:41 +0000 (22:26 +0300)]
iio: adc: berlin2-adc: convert probe to device-managed only

This driver requires only a devm_add_action_or_reset() hook for the
power-down of the device, and then devm_iio_device_register() can be used
directly.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210926192642.4051329-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: rn5t618-adc: use devm_iio_map_array_register() function
Alexandru Ardelean [Sun, 26 Sep 2021 16:28:59 +0000 (19:28 +0300)]
iio: adc: rn5t618-adc: use devm_iio_map_array_register() function

This driver already hooks a similar unwind callback via
devm_add_action_or_reset().

They pretty much do the same thing, so this change converts it to the
devm_iio_map_array_register().

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Reviewed-by: Andreas Kemnade <andreas@kemnade.info>
Link: https://lore.kernel.org/r/20210926162859.3567685-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1363: convert probe to full device-managed
Alexandru Ardelean [Sun, 26 Sep 2021 16:21:10 +0000 (19:21 +0300)]
iio: adc: max1363: convert probe to full device-managed

For this conversion, the 2 regulators (being enabled) require each a
devm_add_action_or_reset() hook registration.

For the other functions, there are device-managed variants.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210926162110.3536436-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: nau7802: convert probe to full device-managed
Alexandru Ardelean [Sun, 26 Sep 2021 15:49:32 +0000 (18:49 +0300)]
iio: adc: nau7802: convert probe to full device-managed

This is a trivial conversion to device-managed functions.
The mutex_destroy() calls are redundant, as the data will be free'd anyway.

And the IRQ and IIO register functions both have device-managed
equivalents.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Reviewed-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20210926154932.3287590-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adis16460: make use of the new unmasked_drdy flag
Nuno Sá [Fri, 3 Sep 2021 14:14:22 +0000 (16:14 +0200)]
iio: adis16460: make use of the new unmasked_drdy flag

The library can now handle enabling/disabling IRQs for devices that
cannot unmask the data ready pin. Hence there's no need to provide an
'enable_irq' callback anymore.

The library will also automatically request the IRQ with 'IRQF_NO_AUTOEN'
so that we can also remove that from the driver.

Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210903141423.517028-5-nuno.sa@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adis16475: make use of the new unmasked_drdy flag
Nuno Sá [Fri, 3 Sep 2021 14:14:21 +0000 (16:14 +0200)]
iio: adis16475: make use of the new unmasked_drdy flag

The library can now handle enabling/disabling IRQs for devices that
cannot unmask the data ready pin. Hence there's no need to provide an
'enable_irq' callback anymore.

The library will also automatically request the IRQ with 'IRQF_NO_AUTOEN'
so that we can also remove that from the driver.

Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210903141423.517028-4-nuno.sa@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adis: handle devices that cannot unmask the drdy pin
Nuno Sá [Fri, 3 Sep 2021 14:14:20 +0000 (16:14 +0200)]
iio: adis: handle devices that cannot unmask the drdy pin

Some devices can't mask/unmask the data ready pin and in those cases
each driver was just calling '{dis}enable_irq()' to control the trigger
state. This change, moves that handling into the library by introducing
a new boolean in the data structure that tells the library that the
device cannot unmask the pin.

On top of controlling the trigger state, we can also use this flag to
automatically request the IRQ with 'IRQF_NO_AUTOEN' in case it is set.
So far, all users of the library want to start operation with IRQs/DRDY
pin disabled so it should be fairly safe to do this inside the library.

Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210903141423.517028-3-nuno.sa@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adis: do not disabe IRQs in 'adis_init()'
Nuno Sá [Fri, 3 Sep 2021 14:14:19 +0000 (16:14 +0200)]
iio: adis: do not disabe IRQs in 'adis_init()'

With commit ecb010d441088 ("iio: imu: adis: Refactor adis_initial_startup")
we are doing a HW or SW reset to the device which means that we'll get
the default state of the data ready pin (which is enabled). Hence there's
no point in disabling the IRQ in the init function. Moreover, this
function is intended to initialize internal data structures and not
really do anything on the device.

As a result of this, some devices were left with the data ready pin enabled
after probe which was not the desired behavior. Thus, we move the call to
'adis_enable_irq()' to the initial startup function where it makes more
sense for it to be.

Note that for devices that cannot mask/unmask the pin, it makes no sense
to call the function at this point since the IRQ should not have been
yet requested. This will be improved in a follow up change.

Fixes: ecb010d441088 ("iio: imu: adis: Refactor adis_initial_startup")
Signed-off-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210903141423.517028-2-nuno.sa@analog.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: da9150-gpadc: convert probe to full-device managed
Alexandru Ardelean [Fri, 3 Sep 2021 07:29:17 +0000 (10:29 +0300)]
iio: adc: da9150-gpadc: convert probe to full-device managed

This change converts the probe of this driver to use device-managed
functions only, which means that the remove hook can be removed.
The remove hook has only 2 calls to iio_device_unregister() and
iio_map_array_unregister(). Both these can now be done via devm register
functions, now that there's also a devm_iio_map_array_register() function.

The platform_set_drvdata() can also be removed now.

This change also removes the error print for when the iio_device_register()
call fails. This isn't required now.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210903072917.45769-6-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: lp8788_adc: convert probe to full-device managed
Alexandru Ardelean [Fri, 3 Sep 2021 07:29:16 +0000 (10:29 +0300)]
iio: adc: lp8788_adc: convert probe to full-device managed

This change converts the probe of this driver to use device-managed
functions only, which means that the remove hook can be removed.
The remove hook has only 2 calls to iio_device_unregister() and
iio_map_array_unregister(). Both these can now be done via devm register
functions, now that there's also a devm_iio_map_array_register() function.

The platform_set_drvdata() can also be removed now.

This change also removes the error print for when the iio_device_register()
call fails. This isn't required now.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210903072917.45769-5-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: axp288_adc: convert probe to full device-managed
Alexandru Ardelean [Fri, 3 Sep 2021 07:29:15 +0000 (10:29 +0300)]
iio: adc: axp288_adc: convert probe to full device-managed

This change converts the probe of this driver to use device-managed
functions only, which means that the remove hook can be removed.
The remove hook has only 2 calls to iio_device_unregister() and
iio_map_array_unregister(). Both these can now be done via devm register
functions, now that there's also a devm_iio_map_array_register() function.

The platform_set_drvdata() can also be removed now.

This change also removes the error print for when the iio_device_register()
call fails. This isn't required now.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20210903072917.45769-4-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: intel_mrfld_adc: convert probe to full device-managed
Alexandru Ardelean [Fri, 3 Sep 2021 07:29:14 +0000 (10:29 +0300)]
iio: adc: intel_mrfld_adc: convert probe to full device-managed

The only call in the remove hook is the iio_map_array_unregister() call.
Since we have a devm_iio_map_array_register() function now, we can use that
and remove the remove hook entirely.
The IIO device was registered with the devm_iio_device_register() prior to
this change.

Also, the platform_set_drvdata() can be removed now, since it was used only
in the remove hook.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20210903072917.45769-3-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: inkern: introduce devm_iio_map_array_register() short-hand function
Alexandru Ardelean [Fri, 3 Sep 2021 07:29:13 +0000 (10:29 +0300)]
iio: inkern: introduce devm_iio_map_array_register() short-hand function

This change introduces a device-managed variant to the
iio_map_array_register() function. It's a simple implementation of calling
iio_map_array_register() and registering a callback to
iio_map_array_unregister() with the devm_add_action_or_reset().

The function uses an explicit 'dev' parameter to bind the unwinding to. It
could have been implemented to implicitly use the parent of the IIO device,
however it shouldn't be too expensive to callers to just specify to which
device object to bind this unwind call.
It would make the API a bit more flexible.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210903072917.45769-2-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: fsl-imx25-gcq: initialize regulators as needed
Alexandru Ardelean [Fri, 25 Jun 2021 07:43:25 +0000 (10:43 +0300)]
iio: adc: fsl-imx25-gcq: initialize regulators as needed

The driver tries to initialize all possible regulators from the DT, then
match the external regulators with each channel and then release all unused
regulators.

We can change the logic a bit to initialize regulators only when at least
one channel needs them.

This change creates a mx25_gcq_ext_regulator_setup() function that is
called only for the external regulators. If there's already a reference to
an external regulator, the function will just exit early with no error.

This way, the driver doesn't need to keep any track of these regulators
during init.

Signed-off-by: Alexandru Ardelean <aardelean@deviqon.com>
Link: https://lore.kernel.org/r/20210625074325.9237-1-aardelean@deviqon.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: imu: st_lsm6dsx: move max_fifo_size in st_lsm6dsx_fifo_ops
Lorenzo Bianconi [Sun, 26 Sep 2021 14:02:30 +0000 (16:02 +0200)]
iio: imu: st_lsm6dsx: move max_fifo_size in st_lsm6dsx_fifo_ops

Move max_fifo_size in st_lsm6dsx_fifo_ops in order to have all
FIFO configuration parameters in st_lsm6dsx_fifo_ops structure.
This patch does not introduce any logic change, just small code
rearrangement.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/3262ad9d9d1497e19ea1bab208c495c2b9a98994.1632664866.git.lorenzo@kernel.org
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Don't reject external triggers when there is no IRQ
Miquel Raynal [Tue, 21 Sep 2021 11:54:08 +0000 (13:54 +0200)]
iio: adc: max1027: Don't reject external triggers when there is no IRQ

External triggers do not necessarily need the EOC interrupt to be
populated to work properly. The end of conversion status may either come
from an interrupt or from a sufficient enough extra delay. IRQs are not
mandatory so move the triggered buffer setup out of the IRQ condition
and add the logic to wait enough time for all the requested conversions
to be in the device's FIFO.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-17-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Allow all kind of triggers to be used
Miquel Raynal [Tue, 21 Sep 2021 11:54:07 +0000 (13:54 +0200)]
iio: adc: max1027: Allow all kind of triggers to be used

There is no reason to limit this driver to its internal trigger. The
only difference being, when using an external trigger, the sample
conversion must be manually started.

Drop the ->validate_trigger() hook in order to allow other triggers to
be bound.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-16-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Use the EOC IRQ when populated for single reads
Miquel Raynal [Tue, 21 Sep 2021 11:54:06 +0000 (13:54 +0200)]
iio: adc: max1027: Use the EOC IRQ when populated for single reads

So far the End-Of-Conversion interrupt was only used in conjunction with
the internal trigger to process the data. Let's extend the use of this
interrupt handler to support regular single-shot conversions as well.

Doing so requires writing our own hard IRQ handler. This handler has to
check if buffers are enabled or not:

*** Buffers disabled condition ***

  This means the user requested a single conversion and the sample is
  ready to be retrieved.

    -> This implies adding the relevant completion boilerplate.

*** Buffers enabled condition ***

  Triggers are used. So far there is only support for the internal
  trigger but this trigger might soon be attached to another device as
  well so it is the core duty to decide which handler to call in order
  to process the data. The core will decide to either:

  * Call the internal trigger handler which will extract the data that
    is already present in the ADC FIFOs

  or

  * Call the trigger handler of another driver when using this trigger
    with another device, even though this call will be slightly delayed
    by the fact that the max1027 IRQ is a data-ready interrupt rather
    than a real trigger:

  -> The new handler will manually inform the core about the trigger
     having transitioned by directly calling iio_trigger_poll() (which
     iio_trigger_generic_data_rdy_poll() initially did).

In order for the handler to be "source" agnostic, we also need to change
the private pointer and provide the IIO device instead of the trigger
object.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-15-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Stop requesting a threaded IRQ
Miquel Raynal [Tue, 21 Sep 2021 11:54:05 +0000 (13:54 +0200)]
iio: adc: max1027: Stop requesting a threaded IRQ

The threaded handler is not populated, this means there is nothing
running in process context so let's switch to the regular
devm_request_irq() call instead.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-14-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Introduce an end of conversion helper
Miquel Raynal [Tue, 21 Sep 2021 11:54:04 +0000 (13:54 +0200)]
iio: adc: max1027: Introduce an end of conversion helper

For now this helper only waits for the maximum duration of a single
conversion.

In practice, a "temperature measurement" will take twice this
time because it will also carry another analog conversion but as here we
will only care about the temperature conversion which happens first, we
can still only wait for a single sample and get the right data.

This helper will soon be improved to properly handle the end of
conversion interrupt as well as a higher number of samples.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-13-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Separate the IRQ handler from the read logic
Miquel Raynal [Tue, 21 Sep 2021 11:54:03 +0000 (13:54 +0200)]
iio: adc: max1027: Separate the IRQ handler from the read logic

Create a max1027_read_scan() helper which will make clearer the future IRQ
handler updates (no functional change).

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-12-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Prevent single channel accesses during buffer reads
Miquel Raynal [Tue, 21 Sep 2021 11:54:02 +0000 (13:54 +0200)]
iio: adc: max1027: Prevent single channel accesses during buffer reads

When hardware buffers are enabled (the cnvst pin being the trigger), one
should not mess with the device state by requesting a single channel
read.

There is already a iio_buffer_enabled() check in *_read_single_value()
to merely prevent this situation but the check is inconsistent since
buffers can be enabled after the if clause anyway. Instead, use the core
mutex by calling iio_device_claim/release_direct_mode().

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-11-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Create a helper to configure the channels to scan
Miquel Raynal [Tue, 21 Sep 2021 11:54:01 +0000 (13:54 +0200)]
iio: adc: max1027: Create a helper to configure the channels to scan

These bits are meant to be reused for triggered buffers setup.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-10-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Ensure a default cnvst trigger configuration
Miquel Raynal [Tue, 21 Sep 2021 11:54:00 +0000 (13:54 +0200)]
iio: adc: max1027: Ensure a default cnvst trigger configuration

We don't expect the (hardware) cnvst trigger to be enabled at boot time,
this is a user choice made in sysfs and there is a dedicated callback to
enable/disable this trigger. Hence, we can just ensure it is disabled in
the probe at initialization time and then assume that whenever a
->read_raw() call happens, the trigger has been disabled and conversions
will start on register write.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-9-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Simplify the _set_trigger_state() helper
Miquel Raynal [Tue, 21 Sep 2021 11:53:59 +0000 (13:53 +0200)]
iio: adc: max1027: Simplify the _set_trigger_state() helper

The call to max1027_enable_trigger() is the same in both cases thanks to
the 'state' variable, so factorize a little bit to simplify the code and
explain why we call this helper.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-8-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Create a helper to enable/disable the cnvst trigger
Miquel Raynal [Tue, 21 Sep 2021 11:53:58 +0000 (13:53 +0200)]
iio: adc: max1027: Create a helper to enable/disable the cnvst trigger

There are two ways to physically trigger a conversion:
- A falling edge on the cnvst pin
- A write operation on the conversion register

Let's create a helper for this.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210921115408.66711-7-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Rename a helper
Miquel Raynal [Tue, 21 Sep 2021 11:53:57 +0000 (13:53 +0200)]
iio: adc: max1027: Rename a helper

Make it clear that the *_set_trigger_state() hook is responsible for
cnvst based conversions by renaming the helper. This may avoid
confusions with software trigger support that is going to be
introduced.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210921115408.66711-6-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Minimize the number of converted channels
Miquel Raynal [Tue, 21 Sep 2021 11:53:56 +0000 (13:53 +0200)]
iio: adc: max1027: Minimize the number of converted channels

Provide a list of ->available_scan_masks which match the device's
capabilities. Basically, these devices are able to scan from 0 to N, N
being the highest voltage channel requested by the user. The temperature
can be included or not, but cannot be retrieved alone.

The consequence is, instead of reading and pushing to the IIO buffers
all channels each time, the "minimum" number of channels will be scanned
and pushed based on the ->active_scan_mask.

For example, if the user wants channels 1, 4 and 5, all channels from
0 to 5 will be scanned and pushed to the IIO buffers. The core will then
filter out the unneeded samples based on the ->active_scan_mask that has
been selected and only channels 1, 4 and 5 will be available to the user
in the shared buffer.

Provide a comment in the code explaining this logic.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-5-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Drop useless debug messages
Miquel Raynal [Tue, 21 Sep 2021 11:53:55 +0000 (13:53 +0200)]
iio: adc: max1027: Drop useless debug messages

These two debug messages bring absolutely no value, let's drop them.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-4-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Drop extra warning message
Miquel Raynal [Tue, 21 Sep 2021 11:53:54 +0000 (13:53 +0200)]
iio: adc: max1027: Drop extra warning message

Memory allocation errors automatically trigger the right logs, no need
to have our own.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Link: https://lore.kernel.org/r/20210921115408.66711-3-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: adc: max1027: Fix style
Miquel Raynal [Tue, 21 Sep 2021 11:53:53 +0000 (13:53 +0200)]
iio: adc: max1027: Fix style

Follow checkpatch.pl's main advices before hacking into the driver, mainly:
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
WARNING: Prefer 'unsigned int *' to bare use of 'unsigned *'
CHECK: Comparison to NULL could be written "!foo"
CHECK: Alignment should match open parenthesis

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20210921115408.66711-2-miquel.raynal@bootlin.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoMAINTAINERS: Add the driver info of the NXP IMX8QXP
Cai Huoqing [Sat, 25 Sep 2021 02:05:47 +0000 (10:05 +0800)]
MAINTAINERS: Add the driver info of the NXP IMX8QXP

The NXP i.MX 8QuadXPlus SOC has a new ADC IP. After adding
the driver support for it, I add the driver info of the
NXP IMX8QXP ADC to MAINTAINERS file.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210925020555.129-4-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agodt-bindings: iio: adc: Add binding documentation for NXP IMX8QXP ADC
Cai Huoqing [Sat, 25 Sep 2021 02:05:46 +0000 (10:05 +0800)]
dt-bindings: iio: adc: Add binding documentation for NXP IMX8QXP ADC

The NXP i.MX 8QuadXPlus SOC a new ADC IP, so add
binding documentation for NXP IMX8QXP ADC.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210925020555.129-3-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: imx8qxp-adc: Add driver support for NXP IMX8QXP ADC
Cai Huoqing [Sat, 25 Sep 2021 02:05:45 +0000 (10:05 +0800)]
iio: imx8qxp-adc: Add driver support for NXP IMX8QXP ADC

The NXP i.MX 8QuadXPlus SOC has a new ADC IP, so add
driver support for this ADC.

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Link: https://lore.kernel.org/r/20210925020555.129-2-caihuoqing@baidu.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: fxls8962af: add wake on event
Sean Nyekjaer [Mon, 20 Sep 2021 11:42:21 +0000 (13:42 +0200)]
iio: accel: fxls8962af: add wake on event

This adds ways for the SoC to wake from accelerometer wake events.

In the suspend function we skip disabling the sensor if wakeup-source
and events are activated.

If buffered reads are enabled they will be deactivated before suspend.
As the onboard buffer is only holding up to 32 12-bit X/Y/Z data
triplets.

Signed-off-by: Sean Nyekjaer <sean@geanix.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Link: https://lore.kernel.org/r/20210920114221.1595543-2-sean@geanix.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
2 years agoiio: accel: fxls8962af: add threshold event handling
Sean Nyekjaer [Mon, 20 Sep 2021 11:42:20 +0000 (13:42 +0200)]
iio: accel: fxls8962af: add threshold event handling

Add event channels that control the creation of motion events.

Signed-off-by: Sean Nyekjaer <sean@geanix.com>
Link: https://lore.kernel.org/r/20210920114221.1595543-1-sean@geanix.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>