iio: gyro: adxrs290: Add triggered buffer support #1131

layman-n-ish · 2020-08-19T17:15:40Z

Provide a way for continuous data capture by setting up buffer support. The
data ready signal exposed at the SYNC pin of the ADXRS290 is exploited as
a hardware interrupt which triggers to fill the buffer.

Triggered buffer setup was tested with both hardware trigger (DATA_RDY) and
software triggers (sysfs-trig & hrtimer).

Also, extend support to read/write byte data from/to the device using
debugfs interface.

Signed-off-by: Nishant Malpani nish.malpani25@gmail.com

drivers/iio/gyro/adxrs290.c

commodo · 2020-08-20T09:38:42Z

rest looks good

Insert a missing mutex_init() call during the probe that initializes the driver's local lock to unlocked state. Fixes: ac98be7 ("iio: gyro: Update ADXRS290 driver for channels support") Signed-off-by: Nishant Malpani <nish.malpani25@gmail.com>

Provide a way for continuous data capture by setting up buffer support. The data ready signal exposed at the SYNC pin of the ADXRS290 is exploited as a hardware interrupt which triggers to fill the buffer. Triggered buffer setup was tested with both hardware trigger (DATA_RDY) and software triggers (sysfs-trig & hrtimer). Signed-off-by: Nishant Malpani <nish.malpani25@gmail.com>

Configure ADXRS290 to use a GPIO interrupt line as a DATA_RDY trigger signal. Signed-off-by: Nishant Malpani <nish.malpani25@gmail.com>

Append 'interrupts' as a required property and provide a suitable example for using a GPIO interrupt line. Signed-off-by: Nishant Malpani <nish.malpani25@gmail.com>

Extend support to read/write byte data from/to the device using debugfs iio interface. Signed-off-by: Nishant Malpani <nish.malpani25@gmail.com>

layman-n-ish · 2020-08-21T08:18:07Z

Changelog v2:

rewrite adxrs290_set_mode() with switch-case statements
separate out commit for debufs reg access
rewrite return path in adxrs290_data_rdy_trigger_set_state()
include devm_add_action_or_reset() hook to switch the device into STANDBY mode which gets executed as part of standard resource unwinding; this also removes the need of a explicit CHIP_remove() hook
rework adxrs290_probe_trigger() - early exit if no irq line
separate out commit (as a bug-fix) for the missing mutex_init() call during the probe

commodo · 2020-08-25T05:24:54Z

drivers/iio/gyro/adxrs290.c

@@ -385,6 +385,8 @@ static int adxrs290_probe(struct spi_device *spi)
 	indio_dev->num_channels = ARRAY_SIZE(adxrs290_channels);
 	indio_dev->info = &adxrs290_info;

+	mutex_init(&st->lock);


when sending this upstream, the commit has for the Fixes tag should be the one in the upstream tree;
otherwise this is perfect as-is

layman-n-ish force-pushed the rpi-4.19.y_adxrs290 branch 2 times, most recently from 5651375 to c1864e2 Compare August 19, 2020 18:41