write_event_config callback use an int for state, but it is actually a
boolean. iio_ev_state_store is actually using kstrtobool to check user
input, then gives the converted boolean value to write_event_config.
Fix signature and update all iio drivers to use the new signature.
This patch has been partially written using coccinelle with the
following script:
$ cat iio-bool.cocci
// Options: --all-includes
virtual patch
@c1@
identifier iioinfo;
identifier wecfunc;
@@
static const struct iio_info iioinfo = {
...,
.write_event_config =
(
wecfunc
|
&wecfunc
),
...,
};
@@
identifier c1.wecfunc;
identifier indio_dev, chan, type, dir, state;
@@
int wecfunc(struct iio_dev *indio_dev, const struct iio_chan_spec *chan, enum iio_event_type type, enum iio_event_direction dir,
-int
+bool
state) {
...
}
make coccicheck MODE=patch COCCI=iio-bool.cocci M=drivers/iio
Unfortunately, this script didn't match all files:
* all write_event_config callbacks using iio_device_claim_direct_scoped
were not detected and not patched.
* all files that do not assign and declare the write_event_config
callback in the same file.
iio.h was also manually updated.
The patch was build tested using allmodconfig config.
cc: Julia Lawall <julia.lawall@inria.fr>
Signed-off-by: Julien Stephan <jstephan@baylibre.com>
Link: https://patch.msgid.link/20241031-iio-fix-write-event-config-signature-v2-7-2bcacbb517a2@baylibre.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
state parameter is currently an int, but it is actually a boolean.
iio_ev_state_store is actually using kstrtobool to check user input,
then gives the converted boolean value to write_event_config. The code
in adux1020_write_event_config re-uses state parameter to store an
integer value. To prepare for updating the write_event_config signature
to use a boolean for state, introduce a new local int variable.
Signed-off-by: Julien Stephan <jstephan@baylibre.com>
Link: https://patch.msgid.link/20241031-iio-fix-write-event-config-signature-v2-6-2bcacbb517a2@baylibre.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The raw value conversion to obtain a measurement in lux as
INT_PLUS_MICRO does not calculate the decimal part properly to display
it as micro (in this case microlux). It only calculates the module to
obtain the decimal part from a resolution that is 10000 times the
provided in the datasheet (0.5376 lux/cnt for the veml6030). The
resulting value must still be multiplied by 100 to make it micro.
This bug was introduced with the original implementation of the driver.
Only the illuminance channel is fixed becuase the scale is non sensical
for the intensity channels anyway.
Cc: stable@vger.kernel.org
Fixes: 7b779f573c ("iio: light: add driver for veml6030 ambient light sensor")
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241016-veml6030-fix-processed-micro-v1-1-4a5644796437@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The custom rpr0521_trigger_consumer_store_time() is registered as trigger
handler in the devm_iio_triggered_buffer_setup() function. This function
is called from the calling of the iio_trigger_poll() used in the
sysfs/hrt triggers and it is not used anywhere else in this driver.
The irq handler of the driver is the rpr0521_drdy_irq_handler() which
saves the timestamp and then wakes the irq thread. The irq thread is
the rpr0521_drdy_irq_thread() function which checks if the irq came
from the sensor and wakes up the trigger threaded handler through
iio_trigger_poll_nested() or returns IRQ_NONE in case the irq didn't
come from this sensor.
This means that in the current driver, you can't reach the
rpr0521_trigger_consumer_store_time() when the device's irq is
triggered. This means that the extra check of iio_trigger_using_own()
is redundant since it will always be false so the general
iio_pollfunc_store_time() can be used.
Signed-off-by: Vasileios Amoiridis <vassilisamir@gmail.com>
Link: https://patch.msgid.link/20240922162041.525896-2-vassilisamir@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The veml7700 contains the same sensor as the veml6030 in a different
package with no interrupt line and no pin to select the I2C address.
To handle the lack of the interrupt line and profit from the existing
support for the veml6030, add a specific iio_chan_spec with no
(num_)event_spec(s), and register the device's info from the
veml6030_info_no_irq struct.
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241007-veml7700-v1-3-fb85dd839d63@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The iio_info is identical for veml6030 and veml6035. Moreover,
veml6035_info_no_irq is missing the initialization of the read_avail
member, which is actually a bug if no irq is provided.
Instead of adding the missing initialization, remove the device-specific
iio_info and use the existing one for the veml6030.
Fixes: ccc26bd7d7 ("iio: light: veml6030: add support for veml6035")
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241007-veml7700-v1-1-fb85dd839d63@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Jonathan writes:
IIO: 1st set of fixes for the 6.12 cycle.
Most of this pull request is the result of Javier Carrasco doing a
careful audit for missing Kconfig dependencies that luck has meant
the random builds have never hit. The rest is the usual mix of old
bugs that have surfaced and some fallout from the recent merge window.
adi,ad5686
- Fix binding duplication of compatible strings.
bosch,bma400
- Fix an uninitialized variable in the event tap handling.
bosch,bmi323
- Fix several issues in the register saving and restore on suspend/resume
sensiron,spd500
- Fix missing CRC8 dependency
ti,op3001
- Fix a missing full-scale range value (values above this point were
all reported wrongly)
vishay,veml6030
- Fix a segmentation fault due to some type confusion.
- Fix wrong ambient light sensor resolution.
* tag 'iio-fixes-for-6.12a' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/jic23/iio: (34 commits)
iio: frequency: admv4420: fix missing select REMAP_SPI in Kconfig
iio: frequency: {admv4420,adrf6780}: format Kconfig entries
iio: adc: ad4695: Add missing Kconfig select
iio: adc: ti-ads8688: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: hid-sensors: Fix an error handling path in _hid_sensor_set_report_latency()
iioc: dac: ltc2664: Fix span variable usage in ltc2664_channel_config()
iio: dac: stm32-dac-core: add missing select REGMAP_MMIO in Kconfig
iio: dac: ltc1660: add missing select REGMAP_SPI in Kconfig
iio: dac: ad5770r: add missing select REGMAP_SPI in Kconfig
iio: amplifiers: ada4250: add missing select REGMAP_SPI in Kconfig
iio: frequency: adf4377: add missing select REMAP_SPI in Kconfig
iio: resolver: ad2s1210: add missing select (TRIGGERED_)BUFFER in Kconfig
iio: resolver: ad2s1210 add missing select REGMAP in Kconfig
iio: proximity: mb1232: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: pressure: bm1390: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: magnetometer: af8133j: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: light: bu27008: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: chemical: ens160: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: dac: ad5766: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
iio: dac: ad3552r: add missing select IIO_(TRIGGERED_)BUFFER in Kconfig
...
After commit 0edb555a65 ("platform: Make platform_driver::remove()
return void") .remove() is (again) the right callback to implement for
platform drivers.
Convert all platform drivers below drivers/iio/ to use .remove(), with
the eventual goal to drop struct platform_driver::remove_new(). As
.remove() and .remove_new() have the same prototypes, conversion is done
by just changing the structure member name in the driver initializer.
While touching these files, make indention of the struct initializer
consistent in several files.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@baylibre.com>
Link: https://patch.msgid.link/20241009060056.502059-2-u.kleine-koenig@baylibre.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Move devm_add_action_or_reset() with a device shut down action to the
hardware initialization function to ensure that any error path after
powering on the device leads to a power off. Add struct device *dev
to the argument list to clarify the device the action is registered
against, and use it wherever &client->dev was used.
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241001-veml6035-v3-7-d789f6ff147c@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The resolution of the WHITE channel is not provided by the manufacturer,
neither in the datasheet nor in the application note (even their
proprietary application only processes the ALS channel, giving raw
values for WHITE).
The current implementation assumes that both resolutions are identical,
which is extremely unlikely, especially for photodiodes with different
spectral responses.
Drop the processed information as it is meaningless.
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241001-veml6035-v3-6-d789f6ff147c@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Drop custom attributes by using the standard read_avail() callback to
read scale and integration time. When at it, add the integration time
and scale attributes fro the WHITE channel, as they modify its value as
well. To avoid breaking the current ABI, these attributes must be kept
as separate for both channels even though they are shared under the
hood.
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241001-veml6035-v3-5-d789f6ff147c@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
The veml6030 requires a delay of 4 ms after activating the sensor. That
is done correctly during the hw initialization, but it's missing after
resuming.
Move the delay to the power on function to make sure that it is always
observerd. When at it, use fsleep() instead of usleep_range() as such a
narrow range is not required.
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://patch.msgid.link/20241001-veml6035-v3-1-d789f6ff147c@gmail.com
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
