Re: [PATCH 2/2] iio: mma8452: add freefall detection for Freescale's accelerometers

From: Martin Kepplinger
Date: Tue Dec 15 2015 - 01:05:45 EST


Am 2015-12-12 um 16:34 schrieb Jonathan Cameron:
> On 08/12/15 16:21, Martin Kepplinger wrote:
>> This adds freefall event detection to the supported devices. It adds
>> the in_accel_x&y&z_mag_falling_en iio event attribute, which activates
>> freefall mode.
>>
>> In freefall mode, the current acceleration magnitude (AND combination
>> of all axis values) is compared to the specified threshold.
>> If it falls under the threshold (in_accel_mag_falling_value),
>> the appropriate IIO event code is generated.
>>
>> The values of rising and falling versions of various sysfs files are
>> shared, which is compliant to the IIO specification.
>>
>> This is what the sysfs "events" directory for these devices looks
>> like after this change:
>>
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_mag_falling_period
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_mag_falling_value
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_mag_rising_period
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_mag_rising_value
>> -r--r--r-- 4096 Oct 23 08:45 in_accel_scale
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_x&y&z_mag_falling_en
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_x_mag_rising_en
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_y_mag_rising_en
>> -rw-r--r-- 4096 Oct 23 08:45 in_accel_z_mag_rising_en
>>
>> Signed-off-by: Martin Kepplinger <martin.kepplinger@xxxxxxxxxxxxxxxxxxxxx>
>> Signed-off-by: Christoph Muellner <christoph.muellner@xxxxxxxxxxxxxxxxxxxxx>
> I think the read back of other events being enabled is broken by this.
> Otherwise a few other minor bits and bobs.
>
> Jonathan
>> ---
>> Other combinations (x&y, x&z or y&z) might be added later. This is the most
>> useful for freefall detection.
>>
>>
>> drivers/iio/accel/mma8452.c | 156 ++++++++++++++++++++++++++++++++++++++------
>> 1 file changed, 137 insertions(+), 19 deletions(-)
>>
>> diff --git a/drivers/iio/accel/mma8452.c b/drivers/iio/accel/mma8452.c
>> index 162bbef..c8ac88c 100644
>> --- a/drivers/iio/accel/mma8452.c
>> +++ b/drivers/iio/accel/mma8452.c
>> @@ -15,7 +15,7 @@
>> *
>> * 7-bit I2C slave address 0x1c/0x1d (pin selectable)
>> *
>> - * TODO: orientation / freefall events, autosleep
>> + * TODO: orientation events, autosleep
>> */
>>
>> #include <linux/module.h>
>> @@ -143,6 +143,14 @@ struct mma_chip_info {
>> u8 ev_count;
>> };
>>
>> +enum {
>> + idx_x,
>> + idx_y,
>> + idx_z,
>> + idx_ts,
>> + idx_xyz,
>> +};
> I would have slightly prefered the change to this enum to have been
> done as a precursor patch as it would have lead to an easily checked
> step 1 followed by the interesting stuff in step 2.

I guess you're right and it would be easier to read.

>> +
>> static int mma8452_drdy(struct mma8452_data *data)
>> {
>> int tries = 150;
>> @@ -409,6 +417,46 @@ fail:
>> return ret;
>> }
>>
>> +static int mma8452_freefall_mode_enabled(struct mma8452_data *data)
>> +{
>> + int val;
>> + const struct mma_chip_info *chip = data->chip_info;
>> +
>> + val = i2c_smbus_read_byte_data(data->client, chip->ev_cfg);
>> + if (val < 0)
>> + return val;
>> +
>> + return !(val & MMA8452_FF_MT_CFG_OAE);
>> +}
>> +
>> +static int mma8452_set_freefall_mode(struct mma8452_data *data, u8 state)
>> +{
>> + int val, ret;
>> + const struct mma_chip_info *chip = data->chip_info;
>> +
>> + val = i2c_smbus_read_byte_data(data->client, chip->ev_cfg);
>> + if (val < 0)
>> + return val;
>> +
>> + if (state && !(mma8452_freefall_mode_enabled(data))) {
>> + val |= BIT(idx_x + chip->ev_cfg_chan_shift);
>> + val |= BIT(idx_y + chip->ev_cfg_chan_shift);
>> + val |= BIT(idx_z + chip->ev_cfg_chan_shift);
>> + val &= ~MMA8452_FF_MT_CFG_OAE;
>> + } else if (!state && mma8452_freefall_mode_enabled(data)) {
>> + val &= ~BIT(idx_x + chip->ev_cfg_chan_shift);
>> + val &= ~BIT(idx_y + chip->ev_cfg_chan_shift);
>> + val &= ~BIT(idx_z + chip->ev_cfg_chan_shift);
>> + val |= MMA8452_FF_MT_CFG_OAE;
>> + }
>> +
>> + ret = mma8452_change_config(data, chip->ev_cfg, val);
>> + if (ret)
>> + return ret;
>> +
>> + return 0;
>> +}
>> +
>> static int mma8452_set_hp_filter_frequency(struct mma8452_data *data,
>> int val, int val2)
>> {
>> @@ -602,6 +650,11 @@ static int mma8452_read_event_config(struct iio_dev *indio_dev,
>> const struct mma_chip_info *chip = data->chip_info;
>> int ret;
>>
>> + switch (chan->channel2) {
>> + case IIO_MOD_X_AND_Y_AND_Z:
>> + return mma8452_freefall_mode_enabled(data);
>> + }
>> +
>> ret = i2c_smbus_read_byte_data(data->client,
>> data->chip_info->ev_cfg);
>> if (ret < 0)
> I may have missed something here, but I think the check for the other events
> being enabled needs updating as well. We are using the same bits in the
> register afterall.

we are using the same bits but reading is not broken. Freefall-mode is
defined by one bit. If the user wants to read x&y&z status (which only
exists in FALLING), I check this bit. Only one mode (direction) can be
active which these devices.

Only the new part is added and everything else stays the same.

I can't disable the individual axis (events) in RISING direction while
x&y&z freefall is enabled in FALLING, but that's ok according to the iio
sysfs doc. (RISING basically is irrelevant then. the problem is, it
isn't really here; I only assume the user isn't interested).

So what I could (and probably should) do is, I could implement this: A
change to RISING has no affect, while x&y&z FALLING is enabled. This
way, x&y&z stays x&y&z, and can't become, say, x&y, just because the
user changed z in RISING (which _should_ have no effect).

>
>> @@ -620,6 +673,11 @@ static int mma8452_write_event_config(struct iio_dev *indio_dev,
>> const struct mma_chip_info *chip = data->chip_info;
>> int val;
>>
>> + switch (chan->channel2) {
>> + case IIO_MOD_X_AND_Y_AND_Z:
>> + return mma8452_set_freefall_mode(data, state);
>> + }
>> +
>> val = i2c_smbus_read_byte_data(data->client, chip->ev_cfg);
>> if (val < 0)
>> return val;
>> @@ -630,7 +688,6 @@ static int mma8452_write_event_config(struct iio_dev *indio_dev,
>> val &= ~BIT(chan->scan_index + chip->ev_cfg_chan_shift);
>>
>> val |= chip->ev_cfg_ele;
>> - val |= MMA8452_FF_MT_CFG_OAE;
>>
>> return mma8452_change_config(data, chip->ev_cfg, val);
>> }
>> @@ -639,12 +696,26 @@ static void mma8452_transient_interrupt(struct iio_dev *indio_dev)
>> {
>> struct mma8452_data *data = iio_priv(indio_dev);
>> s64 ts = iio_get_time_ns();
>> - int src;
>> + int src, cfg;
>>
>> src = i2c_smbus_read_byte_data(data->client, data->chip_info->ev_src);
>> if (src < 0)
>> return;
>>
>> + cfg = i2c_smbus_read_byte_data(data->client, data->chip_info->ev_cfg);
>> + if (cfg < 0)
>> + return;
> This strikes me as odd. I'd assume ev_cfg could be cached and avoid the
> read here? It's basically saying if we enable a free fall event all
> the individual ones are disabled...

It is. And nothing is really cached here; we always read any status from
the device, consistently in this driver.

If freefall event is enabled, all others have to be disabled. This is a
hardware limitation. The device operates in a different mode when
freefall mode is enabled.

Please get back to me in case of doubt.

>> +
>> + if (!(cfg & MMA8452_FF_MT_CFG_OAE)) {
>> + iio_push_event(indio_dev,
>> + IIO_MOD_EVENT_CODE(IIO_ACCEL, 0,
>> + IIO_MOD_X_AND_Y_AND_Z,
>> + IIO_EV_TYPE_MAG,
>> + IIO_EV_DIR_FALLING),
>> + ts);
>> + return;
>> + }
>> +
>> if (src & data->chip_info->ev_src_xe)
>> iio_push_event(indio_dev,
>> IIO_MOD_EVENT_CODE(IIO_ACCEL, 0, IIO_MOD_X,
>> @@ -738,6 +809,27 @@ static int mma8452_reg_access_dbg(struct iio_dev *indio_dev,
>> return 0;
>> }
>>
>> +static const struct iio_event_spec mma8452_freefall_event[] = {
>> + {
>> + .type = IIO_EV_TYPE_MAG,
>> + .dir = IIO_EV_DIR_FALLING,
>> + .mask_separate = BIT(IIO_EV_INFO_ENABLE),
>> + .mask_shared_by_type = BIT(IIO_EV_INFO_VALUE) |
>> + BIT(IIO_EV_INFO_PERIOD) |
>> + BIT(IIO_EV_INFO_HIGH_PASS_FILTER_3DB)
>> + },
>> +};
>> +
>> +static const struct iio_event_spec mma8652_freefall_event[] = {
>> + {
>> + .type = IIO_EV_TYPE_MAG,
>> + .dir = IIO_EV_DIR_FALLING,
>> + .mask_separate = BIT(IIO_EV_INFO_ENABLE),
>> + .mask_shared_by_type = BIT(IIO_EV_INFO_VALUE) |
>> + BIT(IIO_EV_INFO_PERIOD)
>> + },
>> +};
>> +
>> static const struct iio_event_spec mma8452_transient_event[] = {
>> {
>> .type = IIO_EV_TYPE_MAG,
>> @@ -774,6 +866,24 @@ static struct attribute_group mma8452_event_attribute_group = {
>> .attrs = mma8452_event_attributes,
>> };
>>
>> +#define MMA8452_FREEFALL_CHANNEL(modifier, idx) { \
>> + .type = IIO_ACCEL, \
>> + .modified = 1, \
>> + .channel2 = modifier, \
>> + .scan_index = idx, \
>> + .event_spec = mma8452_freefall_event, \
>> + .num_event_specs = ARRAY_SIZE(mma8452_freefall_event), \
>> +}
>> +
>> +#define MMA8652_FREEFALL_CHANNEL(modifier, idx) { \
>> + .type = IIO_ACCEL, \
>> + .modified = 1, \
>> + .channel2 = modifier, \
>> + .scan_index = idx, \
> This isn't a 'real' channel with data to push into the buffer,
> so I'd expect to see an scan_index of -1 to stop it appearing
> under there. Same for the other varients.

Yes, thanks (for all the reviewing)!

>
>> + .event_spec = mma8652_freefall_event, \
>> + .num_event_specs = ARRAY_SIZE(mma8652_freefall_event), \
>> +}
>> +
>> #define MMA8452_CHANNEL(axis, idx, bits) { \
>> .type = IIO_ACCEL, \
>> .modified = 1, \
>> @@ -816,31 +926,35 @@ static struct attribute_group mma8452_event_attribute_group = {
>> }
>>
>> static const struct iio_chan_spec mma8452_channels[] = {
>> - MMA8452_CHANNEL(X, 0, 12),
>> - MMA8452_CHANNEL(Y, 1, 12),
>> - MMA8452_CHANNEL(Z, 2, 12),
>> - IIO_CHAN_SOFT_TIMESTAMP(3),
>> + MMA8452_CHANNEL(X, idx_x, 12),
>> + MMA8452_CHANNEL(Y, idx_y, 12),
>> + MMA8452_CHANNEL(Z, idx_z, 12),
>> + IIO_CHAN_SOFT_TIMESTAMP(idx_ts),
>> + MMA8452_FREEFALL_CHANNEL(IIO_MOD_X_AND_Y_AND_Z, idx_xyz),
>> };
>>
>> static const struct iio_chan_spec mma8453_channels[] = {
>> - MMA8452_CHANNEL(X, 0, 10),
>> - MMA8452_CHANNEL(Y, 1, 10),
>> - MMA8452_CHANNEL(Z, 2, 10),
>> - IIO_CHAN_SOFT_TIMESTAMP(3),
>> + MMA8452_CHANNEL(X, idx_x, 10),
>> + MMA8452_CHANNEL(Y, idx_y, 10),
>> + MMA8452_CHANNEL(Z, idx_z, 10),
>> + IIO_CHAN_SOFT_TIMESTAMP(idx_ts),
>> + MMA8452_FREEFALL_CHANNEL(IIO_MOD_X_AND_Y_AND_Z, idx_xyz),
>> };
>>
>> static const struct iio_chan_spec mma8652_channels[] = {
>> - MMA8652_CHANNEL(X, 0, 12),
>> - MMA8652_CHANNEL(Y, 1, 12),
>> - MMA8652_CHANNEL(Z, 2, 12),
>> - IIO_CHAN_SOFT_TIMESTAMP(3),
>> + MMA8652_CHANNEL(X, idx_x, 12),
>> + MMA8652_CHANNEL(Y, idx_y, 12),
>> + MMA8652_CHANNEL(Z, idx_z, 12),
>> + IIO_CHAN_SOFT_TIMESTAMP(idx_ts),
>> + MMA8652_FREEFALL_CHANNEL(IIO_MOD_X_AND_Y_AND_Z, idx_xyz),
>> };
>>
>> static const struct iio_chan_spec mma8653_channels[] = {
>> - MMA8652_CHANNEL(X, 0, 10),
>> - MMA8652_CHANNEL(Y, 1, 10),
>> - MMA8652_CHANNEL(Z, 2, 10),
>> - IIO_CHAN_SOFT_TIMESTAMP(3),
>> + MMA8652_CHANNEL(X, idx_x, 10),
>> + MMA8652_CHANNEL(Y, idx_y, 10),
>> + MMA8652_CHANNEL(Z, idx_z, 10),
>> + IIO_CHAN_SOFT_TIMESTAMP(idx_ts),
>> + MMA8652_FREEFALL_CHANNEL(IIO_MOD_X_AND_Y_AND_Z, idx_xyz),
>> };
>>
>> enum {
>> @@ -1183,6 +1297,10 @@ static int mma8452_probe(struct i2c_client *client,
>> if (ret < 0)
>> goto buffer_cleanup;
>>
>> + ret = mma8452_set_freefall_mode(data, 0);
>> + if (ret)
>> + return ret;
>> +
>> return 0;
>>
>> buffer_cleanup:
>>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/