| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20241018192137.0ff44c23@jic23-huawei>
Date: Fri, 18 Oct 2024 19:21:37 +0100
From: Jonathan Cameron <jic23@...nel.org>
To: Hardevsinh Palaniya <hardevsinh.palaniya@...iconsignals.io>
Cc: jmaneyrol@...ensense.com, lars@...afoo.de, linux-iio@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iio: imu: inv_icm42600: Enable Pedometer Functionality
On Tue, 15 Oct 2024 14:50:03 +0530
Hardevsinh Palaniya <hardevsinh.palaniya@...iconsignals.io> wrote:
> Enables pedometer functionality in the ICM42605 IMU sensor.
>
> The pedometer feature allows for step counting, which is accessible through
> a new sysfs entry. Interrupts are triggered when a step event occurs, enabling
> step event detection.
>
> Signed-off-by: Hardevsinh Palaniya <hardevsinh.palaniya@...iconsignals.io>
Some additional comments from a quick read.
> ---
> drivers/iio/imu/inv_icm42600/inv_icm42600.h | 16 ++
> .../iio/imu/inv_icm42600/inv_icm42600_accel.c | 165 ++++++++++++++++++
> .../iio/imu/inv_icm42600/inv_icm42600_core.c | 36 +++-
> 3 files changed, 211 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/iio/imu/inv_icm42600/inv_icm42600.h b/drivers/iio/imu/inv_icm42600/inv_icm42600.h
> index 3a07e43e4cf1..c3471b73152e 100644
> --- a/drivers/iio/imu/inv_icm42600/inv_icm42600.h
> +++ b/drivers/iio/imu/inv_icm42600/inv_icm42600.h
> @@ -122,6 +122,7 @@ struct inv_icm42600_sensor_conf {
> int filter;
> };
> #define INV_ICM42600_SENSOR_CONF_INIT {-1, -1, -1, -1}
> +#define INV_ICM42600_SENSOR_CONF_APEX { 2, 0, 9, 6}
>
> struct inv_icm42600_conf {
> struct inv_icm42600_sensor_conf gyro;
> @@ -141,6 +142,8 @@ struct inv_icm42600_suspended {
> * @chip: chip identifier.
> * @name: chip name.
> * @map: regmap pointer.
> + * @pedometer_enable status of pedometer function
> + * @pedometer_value status of steps event occurnce
Check the kernel-doc style. Even better run the script over the
files you are touching.
You are missing : here.
> * @vdd_supply: VDD voltage regulator for the chip.
> * @vddio_supply: I/O voltage regulator for the chip.
> * @orientation: sensor chip orientation relative to main hardware.
> @@ -157,6 +160,8 @@ struct inv_icm42600_state {
> enum inv_icm42600_chip chip;
> const char *name;
> struct regmap *map;
> + bool pedometer_enable;
> + bool pedometer_value;
> struct regulator *vdd_supply;
> struct regulator *vddio_supply;
> struct iio_mount_matrix orientation;
> @@ -301,6 +306,15 @@ struct inv_icm42600_sensor_state {
> #define INV_ICM42600_GYRO_ACCEL_CONFIG0_GYRO_FILT(_f) \
> FIELD_PREP(GENMASK(3, 0), (_f))
>
> +/* Pedometer functionality */
> +#define INV_ICM42600_REG_APEX_CONFIG0 0x0056
> +#define INV_ICM42600_DMP_ODR_50Hz BIT(1)
> +#define INV_ICM42600_PED_ENABLE BIT(5)
> +#define INV_ICM42600_REG_INT_STATUS3 0x0038
> +#define INV_ICM42600_STEP_DET_INT BIT(5)
> +#define INV_ICM42600_REG_APEX_DATA 0x0031 // 2 Byte little-endian
/* */ for comments in IIO (and most of the kernel)
Also, put it on the line above rather than making such a long line.
> +
one blank line is enough.
> +
> #define INV_ICM42600_REG_TMST_CONFIG 0x0054
> #define INV_ICM42600_TMST_CONFIG_MASK GENMASK(4, 0)
> #define INV_ICM42600_TMST_CONFIG_TMST_TO_REGS_EN BIT(4)
> @@ -373,6 +387,8 @@ struct inv_icm42600_sensor_state {
> #define INV_ICM42600_INTF_CONFIG6_I3C_SDR_EN BIT(0)
>
> /* User bank 4 (MSB 0x40) */
> +#define INV_ICM42600_REG_INT_SOURCE6 0x404D
> +#define INV_ICM42600_STEP_DET_INT1_EN BIT(5)
> #define INV_ICM42600_REG_INT_SOURCE8 0x404F
> #define INV_ICM42600_INT_SOURCE8_FSYNC_IBI_EN BIT(5)
> #define INV_ICM42600_INT_SOURCE8_PLL_RDY_IBI_EN BIT(4)
> diff --git a/drivers/iio/imu/inv_icm42600/inv_icm42600_accel.c b/drivers/iio/imu/inv_icm42600/inv_icm42600_accel.c
> index 56ac19814250..90fe4c9e15ab 100644
> --- a/drivers/iio/imu/inv_icm42600/inv_icm42600_accel.c
> +++ b/drivers/iio/imu/inv_icm42600/inv_icm42600_accel.c
> @@ -160,6 +160,13 @@ static const struct iio_chan_spec_ext_info inv_icm42600_accel_ext_infos[] = {
> {},
> };
> +static int inv_icm42600_steps_read_raw(struct iio_dev *indio_dev,
> + struct iio_chan_spec const *chan,
> + int *val, int *val2, long mask)
> +{
> + struct inv_icm42600_state *st = iio_device_get_drvdata(indio_dev);
> + __le16 steps;
> + int ret;
> +
> + if (mask == IIO_CHAN_INFO_PROCESSED) {
> + ret = iio_device_claim_direct_mode(indio_dev);
> + if (ret)
> + return ret;
> + ret = regmap_bulk_read(st->map, INV_ICM42600_REG_APEX_DATA, &steps, sizeof(steps));
> + if (ret)
> + return ret;
> + iio_device_release_direct_mode(indio_dev);
> + if (ret)
> + return ret;
> + *val = steps;
As the bot pointed out, you need an endian conversion here.
le16_to_cpu()
> + return IIO_VAL_INT;
> + }
> +
> + return -EINVAL;
> +}
> +
> static int inv_icm42600_accel_read_raw(struct iio_dev *indio_dev,
> struct iio_chan_spec const *chan,
> int *val, int *val2, long mask)
> @@ -681,6 +721,8 @@ static int inv_icm42600_accel_read_raw(struct iio_dev *indio_dev,
> break;
> case IIO_TEMP:
> return inv_icm42600_temp_read_raw(indio_dev, chan, val, val2, mask);
> + case IIO_STEPS:
> + return inv_icm42600_steps_read_raw(indio_dev, chan, val, val2, mask);
> default:
> return -EINVAL;
> }
> @@ -824,6 +866,126 @@ static int inv_icm42600_accel_hwfifo_flush(struct iio_dev *indio_dev,
> return ret;
> }
>
> +/*****************Pedometer Functionality**************/
No to structure comments like this. They add little to readability and have
a habit of rapidly becoming wrong.
> +static int inv_icm42600_step_en(struct inv_icm42600_state *st, int state)
> +{
> + struct inv_icm42600_sensor_conf conf = INV_ICM42600_SENSOR_CONF_APEX;
> + int ret, value;
> +
> + if (state) {
> +
> + ret = inv_icm42600_set_accel_conf(st, &conf, NULL);
> + if (ret)
> + return ret;
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_APEX_CONFIG0,
> + INV_ICM42600_DMP_ODR_50Hz);
> + if (ret)
> + return ret;
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_SIGNAL_PATH_RESET,
> + INV_ICM42600_SIGNAL_PATH_RESET_DMP_MEM_RESET);
> + if (ret)
> + return ret;
> + msleep(1);
Document the reason for this value.
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_SIGNAL_PATH_RESET,
> + INV_ICM42600_SIGNAL_PATH_RESET_DMP_INIT_EN);
> + if (ret)
> + return ret;
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_INT_SOURCE6,
> + INV_ICM42600_STEP_DET_INT1_EN);
> + if (ret)
> + return ret;
> +
> + value = INV_ICM42600_DMP_ODR_50Hz | INV_ICM42600_PED_ENABLE;
> + ret = regmap_write(st->map, INV_ICM42600_REG_APEX_CONFIG0, value);
> + if (ret)
> + return ret;
> +
> + st->pedometer_enable = true;
return here.
Then can drop the else.
With two such different paths, even better would be two little functions
to handle the two operations (enable + disable) as will make each individually
easier to read.
> +
> + } else {
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_APEX_CONFIG0, 0);
> + if (ret)
> + return ret;
> +
> + ret = regmap_write(st->map, INV_ICM42600_REG_INT_SOURCE6, 0);
> + if (ret)
> + return ret;
> +
> + st->pedometer_enable = false;
> + }
> +
> + return 0;
> +}
> +
> +static int inv_icm42600_write_event_config(struct iio_dev *indio_dev,
> + const struct iio_chan_spec *chan,
> + enum iio_event_type type,
> + enum iio_event_direction dir, int state)
> +{
> + struct inv_icm42600_state *st = iio_device_get_drvdata(indio_dev);
> + int ret;
> +
> + if(chan->type != IIO_STEPS)
> + return -EINVAL;
> +
> + mutex_lock(&st->lock);
guard() is useful in cases like this.
> +
> + ret = inv_icm42600_step_en(st, state);
> +
> + mutex_unlock(&st->lock);
> + return ret;
> +}
> +
> +static int inv_icm42600_read_event_config(struct iio_dev *indio_dev,
> + const struct iio_chan_spec *chan,
> + enum iio_event_type type,
> + enum iio_event_direction dir)
> +{
> + struct inv_icm42600_state *st = iio_device_get_drvdata(indio_dev);
> + int value;
> +
> + if (chan->type != IIO_STEPS)
> + return -EINVAL;
> +
> + regmap_read(st->map, INV_ICM42600_REG_APEX_CONFIG0, &value);
check return value.
> +
> + if (value & INV_ICM42600_PED_ENABLE)
> + return 1;
> + else
> + return 0;
> +}
> +
> +static int inv_icm42600_read_event_value(struct iio_dev *indio_dev,
This isn't to get if the event happened, it's for reading thresholds
etc. Not relevant for a pedometer.
> + const struct iio_chan_spec *chan,
> + enum iio_event_type type,
> + enum iio_event_direction dir,
> + enum iio_event_info info,
> + int *val, int *val2)
> +{
> + struct inv_icm42600_state *st = iio_device_get_drvdata(indio_dev);
> +
> + mutex_lock(&st->lock);
guard()
> +
> + if (type == IIO_EV_TYPE_CHANGE) {
flip logic and can test that before taking the lock
> + if (st->pedometer_value == true) {
> + *val = 1;
> + st->pedometer_value = false;
> + } else {
> + *val = 0;
> + }
> + mutex_unlock(&st->lock);
> + return IIO_VAL_INT;
> + }
> +
> + mutex_unlock(&st->lock);
> + return -EINVAL;
> +}
> +
> static const struct iio_info inv_icm42600_accel_info = {
> .read_raw = inv_icm42600_accel_read_raw,
> .read_avail = inv_icm42600_accel_read_avail,
> @@ -833,6 +995,9 @@ static const struct iio_info inv_icm42600_accel_info = {
> .update_scan_mode = inv_icm42600_accel_update_scan_mode,
> .hwfifo_set_watermark = inv_icm42600_accel_hwfifo_set_watermark,
> .hwfifo_flush_to_buffer = inv_icm42600_accel_hwfifo_flush,
> + .write_event_config = inv_icm42600_write_event_config,
> + .read_event_config = inv_icm42600_read_event_config,
> + .read_event_value = inv_icm42600_read_event_value,
> };
>
> struct iio_dev *inv_icm42600_accel_init(struct inv_icm42600_state *st)
> diff --git a/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c b/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c
> index c3924cc6190e..4d78cb5ca396 100644
> --- a/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c
> +++ b/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c
> @@ -15,7 +15,8 @@
> #include <linux/pm_runtime.h>
> #include <linux/property.h>
> #include <linux/regmap.h>
> -
Keep the blank line and keep the iio headers in their own section.
> +#include <linux/iio/events.h>
> +#include <linux/of_irq.h>
hmm. Can we not use the generic property accessors?
Also you aren't using any interrupt related new stuff in here so I think
this is just spurious.
> #include <linux/iio/iio.h>
>
> #include "inv_icm42600.h"
> @@ -533,6 +534,19 @@ static irqreturn_t inv_icm42600_irq_handler(int irq, void *_data)
>
> mutex_lock(&st->lock);
Probably worth considering use of guard() in here as a precursor patch.
>
> + ret = regmap_read(st->map, INV_ICM42600_REG_INT_STATUS3, &status);
> + if (ret)
> + goto out_unlock;
> +
> + if (status & INV_ICM42600_STEP_DET_INT) {
> + iio_push_event(st->indio_accel, IIO_MOD_EVENT_CODE(IIO_STEPS, 0,
> + IIO_NO_MOD,
> + IIO_EV_TYPE_CHANGE,
> + IIO_EV_DIR_NONE),
> + st->timestamp.accel);
> + st->pedometer_value = true;
> + }
> +
> ret = regmap_read(st->map, INV_ICM42600_REG_INT_STATUS, &status);
> if (ret)
> goto out_unlock;
> @@ -860,12 +876,20 @@ static int inv_icm42600_runtime_suspend(struct device *dev)
> mutex_lock(&st->lock);
>
> /* disable all sensors */
> - ret = inv_icm42600_set_pwr_mgmt0(st, INV_ICM42600_SENSOR_MODE_OFF,
> - INV_ICM42600_SENSOR_MODE_OFF, false,
> - NULL);
> - if (ret)
> - goto error_unlock;
> + if (st->pedometer_enable) {
> + ret = inv_icm42600_set_pwr_mgmt0(st, INV_ICM42600_SENSOR_MODE_OFF,
> + INV_ICM42600_SENSOR_MODE_LOW_POWER,
> + false, NULL);
> + if (ret)
> + goto error_unlock;
> + } else {
>
> + ret = inv_icm42600_set_pwr_mgmt0(st, INV_ICM42600_SENSOR_MODE_OFF,
> + INV_ICM42600_SENSOR_MODE_OFF,
> + false, NULL);
> + if (ret)
> + goto error_unlock;
> + }
> regulator_disable(st->vddio_supply);
>
> error_unlock:
Powered by blists - more mailing lists