add support for ASM330LHHX
Subsystems (3)
Lorenzo Bianconi
maintainercover add support for ASM330LHHX
Add support for ASM330LHHX IMU mems sensor
https://www.st.com/resource/en/datasheet/asm330lhhx.pdf
Lorenzo Bianconi (2):
iio: imu: st_lsm6dsx: add support to ASM330LHHX
dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + drivers/iio/imu/st_lsm6dsx/Kconfig | 6 +++--- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h | 2 ++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c | 3 ++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 6 +++++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c | 5 +++++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c | 5 +++++ 7 files changed, 23 insertions(+), 5 deletions(-)
Signature · 1 line
2.35.1
Lorenzo Bianconi
maintainerPATCH 1/2 iio: imu: st_lsm6dsx: add support to ASM330LHHX
Add support to STM ASM330LHHX (acc + gyro) Mems sensor
https://www.st.com/resource/en/datasheet/asm330lhhx.pdf
--- drivers/iio/imu/st_lsm6dsx/Kconfig | 6 +++--- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h | 2 ++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c | 3 ++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 6 +++++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c | 5 +++++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c | 5 +++++ 6 files changed, 22 insertions(+), 5 deletions(-)
drivers/iio/imu/st_lsm6dsx/Kconfig +3 -3
@@ -11,9 +11,9 @@ config IIO_ST_LSM6DSX help Say yes here to build support for STMicroelectronics LSM6DSx imu sensor. Supported devices: lsm6ds3, lsm6ds3h, lsm6dsl, lsm6dsm, - ism330dlc, lsm6dso, lsm6dsox, asm330lhh, lsm6dsr, lsm6ds3tr-c, - ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, the accelerometer/gyroscope - of lsm9ds1 and lsm6dst. + ism330dlc, lsm6dso, lsm6dsox, asm330lhh, asm330lhhx, lsm6dsr, + lsm6ds3tr-c, ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, + the accelerometer/gyroscope of lsm9ds1 and lsm6dst. To compile this driver as a module, choose M here: the module will be called st_lsm6dsx.
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h +2 -0
@@ -31,6 +31,7 @@#define ST_LSM6DSRX_DEV_NAME "lsm6dsrx"#define ST_LSM6DST_DEV_NAME "lsm6dst"#define ST_LSM6DSOP_DEV_NAME "lsm6dsop" +#define ST_ASM330LHHX_DEV_NAME "asm330lhhx"enum st_lsm6dsx_hw_id {ST_LSM6DS3_ID ,
@@ -49,6 +50,7 @@ enum st_lsm6dsx_hw_id {ST_LSM6DSRX_ID ,ST_LSM6DST_ID ,ST_LSM6DSOP_ID , +ST_ASM330LHHX_ID , ST_LSM6DSX_MAX_ID, };
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c +2 -1
@@ -14,7 +14,8 @@* (e.g. Gx, Gy, Gz, Ax, Ay, Az), then data are repeated depending on the * value of the decimation factor and ODR set for each FIFO data set. * -* LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/LSM6DSRX/ISM330DHCX/LSM6DST/LSM6DSOP: +* LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/LSM6DSRX/ISM330DHCX/ +* LSM6DST/LSM6DSOP: * The FIFO buffer can be configured to store data from gyroscope and * accelerometer. Each sample is queued with a tag (1B) indicating data * source (gyroscope, accelerometer, hw timer).
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +5 -1
@@ -26,7 +26,7 @@* - Gyroscope supported full-scale [dps]: +-125/+-245/+-500/+-1000/+-2000 * - FIFO size: 4KB * -* - LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: +* - LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: * - Accelerometer/Gyroscope supported ODR [Hz]: 12.5, 26, 52, 104, 208, 416, * 833 * - Accelerometer supported full-scale [g]: +-2/+-4/+-8/+-16
@@ -786,6 +786,10 @@ static const struct st_lsm6dsx_settings st_lsm6dsx_sensor_settings[] = { .hw_id = ST_LSM6DST_ID, .name = ST_LSM6DST_DEV_NAME, .wai = 0x6d, + }, { + .hw_id = ST_ASM330LHHX_ID, + .name = ST_ASM330LHHX_DEV_NAME, + .wai = 0x6b, }, }, .channels = {
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c +5 -0
@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_i2c_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_i2c_of_match);
@@ -122,6 +126,7 @@ static const struct i2c_device_id st_lsm6dsx_i2c_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (i2c, st_lsm6dsx_i2c_id_table);
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c +5 -0
@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_spi_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_spi_of_match);
@@ -122,6 +126,7 @@ static const struct spi_device_id st_lsm6dsx_spi_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (spi, st_lsm6dsx_spi_id_table);
Signature · 1 line
2.35.1
Lorenzo Bianconi
maintainerPATCH 2/2 dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
In reply to Lorenzo Bianconi--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)
Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1 -0
@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Signature · 1 line
2.35.1
Jonathan Cameron
maintainerRe: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>
Hi Lorenzo,
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.
Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Jonathan Cameron
maintainerRe: [PATCH 1/2] iio: imu: st_lsm6dsx: add support to ASM330LHHX
In reply to Lorenzo BianconiQuoted 2 lines · On Sat, 2 Apr 2022 12:09:29 +0200 Lorenzo Bianconi <[email protected]> wrote:
Add support to STM ASM330LHHX (acc + gyro) Mems sensor
https://www.st.com/resource/en/datasheet/asm330lhhx.pdf
Use a Datasheet tag as below.
Perhaps mention it's an automotive rated sensor and that it's compatible otherwise with xxx existing part.
Quick glance at the datasheet suggests this part has a sensor hub... Should it be in a the block with parameters for that as you currently have it in the one for no sensorhub I think.
Or is it using an incompatible sensor hub interface?
Quoted 1 line
Datasheet: https://www.st.com/resource/en/datasheet/asm330lhhx.pdf
Quoted 75 lines
Signed-off-by Lorenzo Bianconi <[email protected]>--- drivers/iio/imu/st_lsm6dsx/Kconfig | 6 +++--- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h | 2 ++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c | 3 ++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 6 +++++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c | 5 +++++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c | 5 +++++ 6 files changed, 22 insertions(+), 5 deletions(-)drivers/iio/imu/st_lsm6dsx/Kconfig +3
-3@@ -11,9 +11,9 @@ config IIO_ST_LSM6DSX help Say yes here to build support for STMicroelectronics LSM6DSx imu sensor. Supported devices: lsm6ds3, lsm6ds3h, lsm6dsl, lsm6dsm, - ism330dlc, lsm6dso, lsm6dsox, asm330lhh, lsm6dsr, lsm6ds3tr-c, - ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, the accelerometer/gyroscope - of lsm9ds1 and lsm6dst. + ism330dlc, lsm6dso, lsm6dsox, asm330lhh, asm330lhhx, lsm6dsr, + lsm6ds3tr-c, ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, + the accelerometer/gyroscope of lsm9ds1 and lsm6dst. To compile this driver as a module, choose M here: the module will be called st_lsm6dsx.drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h +2
-0@@ -31,6 +31,7 @@#define ST_LSM6DSRX_DEV_NAME "lsm6dsrx"#define ST_LSM6DST_DEV_NAME "lsm6dst"#define ST_LSM6DSOP_DEV_NAME "lsm6dsop" +#define ST_ASM330LHHX_DEV_NAME "asm330lhhx"enum st_lsm6dsx_hw_id {ST_LSM6DS3_ID ,@@ -49,6 +50,7 @@ enum st_lsm6dsx_hw_id {ST_LSM6DSRX_ID ,ST_LSM6DST_ID ,ST_LSM6DSOP_ID , +ST_ASM330LHHX_ID , ST_LSM6DSX_MAX_ID, };drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c +2
-1@@ -14,7 +14,8 @@* (e.g. Gx, Gy, Gz, Ax, Ay, Az), then data are repeated depending on the * value of the decimation factor and ODR set for each FIFO data set. * -* LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/LSM6DSRX/ISM330DHCX/LSM6DST/LSM6DSOP: +* LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/LSM6DSRX/ISM330DHCX/ +* LSM6DST/LSM6DSOP: * The FIFO buffer can be configured to store data from gyroscope and * accelerometer. Each sample is queued with a tag (1B) indicating data * source (gyroscope, accelerometer, hw timer). drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +1
-1@@ -26,7 +26,7 @@* - Gyroscope supported full-scale [dps]: +-125/+-245/+-500/+-1000/+-2000 * - FIFO size: 4KB * -* - LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: +* - LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: * - Accelerometer/Gyroscope supported ODR [Hz]: 12.5, 26, 52, 104, 208, 416, * 833 * - Accelerometer supported full-scale [g]: +-2/+-4/+-8/+-16
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c static const struct st_lsm6dsx_settings st_lsm6dsx_sensor_settings[] = { · line 792
Quoted 7 lines
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +4
-0@@ -786,6 +786,10 @@ static const struct st_lsm6dsx_settings st_lsm6dsx_sensor_settings[] = { .hw_id = ST_LSM6DST_ID, .name = ST_LSM6DST_DEV_NAME, .wai = 0x6d, + }, { + .hw_id = ST_ASM330LHHX_ID, + .name = ST_ASM330LHHX_DEV_NAME, + .wai = 0x6b,
Probably nicer to put it next to he other 0x6b entries.
Quoted 49 lines
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +0
-0}, }, .channels = {drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c +5
-0@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_i2c_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_i2c_of_match);@@ -122,6 +126,7 @@ static const struct i2c_device_id st_lsm6dsx_i2c_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (i2c, st_lsm6dsx_i2c_id_table);drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c +5
-0@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_spi_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_spi_of_match);@@ -122,6 +126,7 @@ static const struct spi_device_id st_lsm6dsx_spi_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (spi, st_lsm6dsx_spi_id_table);
Lorenzo Bianconi
maintainerRe: [PATCH 1/2] iio: imu: st_lsm6dsx: add support to ASM330LHHX
Quoted 8 lines
Quoted 2 lines · On Sat, 2 Apr 2022 12:09:29 +0200 Lorenzo Bianconi <[email protected]> wrote:
Add support to STM ASM330LHHX (acc + gyro) Mems sensor
https://www.st.com/resource/en/datasheet/asm330lhhx.pdfUse a Datasheet tag as below.
Perhaps mention it's an automotive rated sensor and that it's compatible otherwise with xxx existing part.
ack, will do
Quoted 6 lines
Quick glance at the datasheet suggests this part has a sensor hub... Should it be in a the block with parameters for that as you currently have it in the one for no sensorhub I think.
Or is it using an incompatible sensor hub interface?
asm330lhhx supports sensor hub with the same configuration of LSM6DSR or LSM6DSO.
Regards,
Lorenzo
Quoted 138 lines
Quoted 2 lines
Datasheet: https://www.st.com/resource/en/datasheet/asm330lhhx.pdf
Quoted 75 lines
Signed-off-by Lorenzo Bianconi <[email protected]>--- drivers/iio/imu/st_lsm6dsx/Kconfig | 6 +++--- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h | 2 ++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c | 3 ++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 6 +++++- drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c | 5 +++++ drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c | 5 +++++ 6 files changed, 22 insertions(+), 5 deletions(-)drivers/iio/imu/st_lsm6dsx/Kconfig +3
-3@@ -11,9 +11,9 @@ config IIO_ST_LSM6DSX help Say yes here to build support for STMicroelectronics LSM6DSx imu sensor. Supported devices: lsm6ds3, lsm6ds3h, lsm6dsl, lsm6dsm, - ism330dlc, lsm6dso, lsm6dsox, asm330lhh, lsm6dsr, lsm6ds3tr-c, - ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, the accelerometer/gyroscope - of lsm9ds1 and lsm6dst. + ism330dlc, lsm6dso, lsm6dsox, asm330lhh, asm330lhhx, lsm6dsr, + lsm6ds3tr-c, ism330dhcx, lsm6dsrx, lsm6ds0, lsm6dsop, + the accelerometer/gyroscope of lsm9ds1 and lsm6dst. To compile this driver as a module, choose M here: the module will be called st_lsm6dsx.drivers/iio/imu/st_lsm6dsx/st_lsm6dsx.h +2
-0@@ -31,6 +31,7 @@#define ST_LSM6DSRX_DEV_NAME "lsm6dsrx"#define ST_LSM6DST_DEV_NAME "lsm6dst"#define ST_LSM6DSOP_DEV_NAME "lsm6dsop" +#define ST_ASM330LHHX_DEV_NAME "asm330lhhx"enum st_lsm6dsx_hw_id {ST_LSM6DS3_ID ,@@ -49,6 +50,7 @@ enum st_lsm6dsx_hw_id {ST_LSM6DSRX_ID ,ST_LSM6DST_ID ,ST_LSM6DSOP_ID , +ST_ASM330LHHX_ID , ST_LSM6DSX_MAX_ID, };drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c +2
-1@@ -14,7 +14,8 @@* (e.g. Gx, Gy, Gz, Ax, Ay, Az), then data are repeated depending on the * value of the decimation factor and ODR set for each FIFO data set. * -* LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/LSM6DSRX/ISM330DHCX/LSM6DST/LSM6DSOP: +* LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/LSM6DSRX/ISM330DHCX/ +* LSM6DST/LSM6DSOP: * The FIFO buffer can be configured to store data from gyroscope and * accelerometer. Each sample is queued with a tag (1B) indicating data * source (gyroscope, accelerometer, hw timer). drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +1
-1@@ -26,7 +26,7 @@* - Gyroscope supported full-scale [dps]: +-125/+-245/+-500/+-1000/+-2000 * - FIFO size: 4KB * -* - LSM6DSO/LSM6DSOX/ASM330LHH/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: +* - LSM6DSO/LSM6DSOX/ASM330LHH/ASM330LHHX/LSM6DSR/ISM330DHCX/LSM6DST/LSM6DSOP: * - Accelerometer/Gyroscope supported ODR [Hz]: 12.5, 26, 52, 104, 208, 416, * 833 * - Accelerometer supported full-scale [g]: +-2/+-4/+-8/+-16 drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c static const struct st_lsm6dsx_settings st_lsm6dsx_sensor_settings[] = { · line 792
Quoted 7 lines
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +4
-0@@ -786,6 +786,10 @@ static const struct st_lsm6dsx_settings st_lsm6dsx_sensor_settings[] = { .hw_id = ST_LSM6DST_ID, .name = ST_LSM6DST_DEV_NAME, .wai = 0x6d, + }, { + .hw_id = ST_ASM330LHHX_ID, + .name = ST_ASM330LHHX_DEV_NAME, + .wai = 0x6b,Probably nicer to put it next to he other
0x6bentries.Quoted 49 lines
drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c +0
-0}, }, .channels = {drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_i2c.c +5
-0@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_i2c_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_i2c_of_match);@@ -122,6 +126,7 @@ static const struct i2c_device_id st_lsm6dsx_i2c_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (i2c, st_lsm6dsx_i2c_id_table);drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_spi.c +5
-0@@ -101,6 +101,10 @@ static const struct of_device_id st_lsm6dsx_spi_of_match[] = { .compatible = "st,lsm6dsop", .data = (void *)ST_LSM6DSOP_ID, }, + { + .compatible = "st,asm330lhhx", + .data = (void *)ST_ASM330LHHX_ID, + }, {}, };MODULE_DEVICE_TABLE (of, st_lsm6dsx_spi_of_match);@@ -122,6 +126,7 @@ static const struct spi_device_id st_lsm6dsx_spi_id_table[] = { { ST_LSM6DSRX_DEV_NAME, ST_LSM6DSRX_ID }, { ST_LSM6DST_DEV_NAME, ST_LSM6DST_ID }, { ST_LSM6DSOP_DEV_NAME, ST_LSM6DSOP_ID }, + { ST_ASM330LHHX_DEV_NAME, ST_ASM330LHHX_ID }, {}, };MODULE_DEVICE_TABLE (spi, st_lsm6dsx_spi_id_table);
Lorenzo Bianconi
maintainerRe: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
In reply to Jonathan CameronQuoted 6 lines
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>Hi Lorenzo,
Quoted 7 lines
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).
Hi Jonathan,
if we consider only the features implemented in st_lsm6dsx, asm330lhhx will be 1:1 compatible with lsm6dsr or lsm6dso, so we can just use one of bindings in this section to support it (the only side effect is it will be listed as "lsm6dsr" or "lsm6dso", but I guess it is ok). Agree?
The only difference between asm330lhhx and asm330lhh is the former supports sensor-hub while the latter does not declare it (even if the use the same whoami).
AFAIK there is no way to autodetect if the sensor supports sensor-hub and we can just try to discover slave devices connected. This can have some downside as described in the commit:
commit 35619155d044830357f06f1d2c8188c4530b4d7a
Author: Lorenzo Bianconi <[email protected]> Date: Sat Nov 13 16:23:14 2021 +0100
iio: imu: st_lsm6dsx: add dts property to disable sensor-hub
I would like to merge the sections in st_lsm6dsx_settings struct for lsm6dsr, lsm6dso.. and lsm6dsop, asm330lhh since the only difference is sensor-hub support. I guess we can have 2 option here to avoid any sensor-hub corner cases:
- provide the "st,disable-sensor-hub" in dts to disable sensor-hub for asm330lhh, lsm6dsop (need user changes)
- add a bool variable st_lsm6dsx_settings[].id[] in order to specify if the chip supports sensor-hub.
Which one do you prefer?
Regards,
Lorenzo
Quoted 58 lines
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Re: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Quoted 14 lines · On Sun, 3 Apr 2022 16:56:51 +0200 Lorenzo Bianconi <[email protected]> wrote:
Quoted 6 lines
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>Hi Lorenzo,
Quoted 7 lines
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).
Quoted 6 lines
Hi Jonathan,
if we consider only the features implemented in
st_lsm6dsx, asm330lhhx will be 1:1 compatible with lsm6dsr or lsm6dso, so we can just use one of bindings in this section to support it (the only side effect is it will be listed as "lsm6dsr" or "lsm6dso", but I guess it is ok). Agree?
If the part has more features than the base compatible (or a different WAI) then we can definitely have a backup compatible for it (hence making that subset of features work on an old kernel). We still want to introduce the new compatible so that we get the name right etc going forwards and are in a good position to add the extra features if we ever get around to it.
Quoted 7 lines
The only difference between asm330lhhx and asm330lhh is the former supports sensor-hub while the latter does not declare it (even if the use the same whoami).
AFAIK there is no way to autodetect if the sensor supports sensor-hub and we can just try to discover slave devices connected. This can have some downside as described in the commit:
Ah thanks. I'd forgotten this.
Quoted 80 lines
commit
35619155d044830357f06f1d2c8188c4530b4d7aAuthor: Lorenzo Bianconi <[email protected]> Date: Sat Nov 13 16:23:14 2021 +0100iio: imu:
st_lsm6dsx: add dts property to disable sensor-hubI would like to merge the sections in
st_lsm6dsx_settingsstruct for lsm6dsr, lsm6dso.. and lsm6dsop, asm330lhh since the only difference is sensor-hub support. I guess we can have 2 option here to avoid any sensor-hub corner cases:
- provide the "st,disable-sensor-hub" in dts to disable sensor-hub for asm330lhh, lsm6dsop (need user changes)
- add a bool variable st_lsm6dsx_settings[].id[] in order to specify if the chip supports sensor-hub.Which one do you prefer?
Regards,
LorenzoQuoted 58 lines
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Re: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Quoted 24 lines · Jonathan Cameron
Quoted 14 lines · On Sun, 3 Apr 2022 16:56:51 +0200 Lorenzo Bianconi <[email protected]> wrote:
Quoted 6 lines
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>Hi Lorenzo,
Quoted 7 lines
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).Quoted 6 lines
Hi Jonathan,
if we consider only the features implemented in
st_lsm6dsx, asm330lhhx will be 1:1 compatible with lsm6dsr or lsm6dso, so we can just use one of bindings in this section to support it (the only side effect is it will be listed as "lsm6dsr" or "lsm6dso", but I guess it is ok). Agree?
Quoted 5 lines
If the part has more features than the base compatible (or a different WAI) then we can definitely have a backup compatible for it (hence making that subset of features work on an old kernel). We still want to introduce the new compatible so that we get the name right etc going forwards and are in a good position to add the extra features if we ever get around to it.
ack. I did not completely get what you mean here with "backup compatible". Do you mean:
- use "st,lsm6dsr" for asm330lhhx on older kernels and add "st,asm330lhhx" on new ones. Do you have any pointer on how to document it?
or
- add a "wildcard" compatible string for this kind of devices. Do you have any pointers?
Regards,
Lorenzo
Quoted 93 lines
Quoted 9 lines
The only difference between asm330lhhx and asm330lhh is the former supports sensor-hub while the latter does not declare it (even if the use the same whoami).
AFAIK there is no way to autodetect if the sensor supports sensor-hub and we can just try to discover slave devices connected. This can have some downside as described in the commit:Ah thanks. I'd forgotten this.
Quoted 80 lines
commit
35619155d044830357f06f1d2c8188c4530b4d7aAuthor: Lorenzo Bianconi <[email protected]> Date: Sat Nov 13 16:23:14 2021 +0100iio: imu:
st_lsm6dsx: add dts property to disable sensor-hubI would like to merge the sections in
st_lsm6dsx_settingsstruct for lsm6dsr, lsm6dso.. and lsm6dsop, asm330lhh since the only difference is sensor-hub support. I guess we can have 2 option here to avoid any sensor-hub corner cases:
- provide the "st,disable-sensor-hub" in dts to disable sensor-hub for asm330lhh, lsm6dsop (need user changes)
- add a bool variable st_lsm6dsx_settings[].id[] in order to specify if the chip supports sensor-hub.Which one do you prefer?
Regards,
LorenzoQuoted 58 lines
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Re: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Quoted 31 lines · On Mon, 4 Apr 2022 11:33:17 +0200 Lorenzo Bianconi…
Quoted 24 lines · Jonathan Cameron
Quoted 14 lines · On Sun, 3 Apr 2022 16:56:51 +0200 Lorenzo Bianconi <[email protected]> wrote:
Quoted 6 lines
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>Hi Lorenzo,
Quoted 7 lines
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).Quoted 6 lines
Hi Jonathan,
if we consider only the features implemented in
st_lsm6dsx, asm330lhhx will be 1:1 compatible with lsm6dsr or lsm6dso, so we can just use one of bindings in this section to support it (the only side effect is it will be listed as "lsm6dsr" or "lsm6dso", but I guess it is ok). Agree?Quoted 5 lines
If the part has more features than the base compatible (or a different WAI) then we can definitely have a backup compatible for it (hence making that subset of features work on an old kernel). We still want to introduce the new compatible so that we get the name right etc going forwards and are in a good position to add the extra features if we ever get around to it.
Quoted 4 lines
ack. I did not completely get what you mean here with "backup compatible". Do you mean:
- use "st,lsm6dsr" for asm330lhhx on older kernels and add "st,asm330lhhx" on new ones. Do you have any pointer on how to document it?
Take a look at the mpu6050 patches.
properties:
compatible:
- enum:
- - invensense,iam20680 - - invensense,icm20608 - - invensense,icm20609 - - invensense,icm20689 - - invensense,icm20602 - - invensense,icm20690 - - invensense,mpu6000 - - invensense,mpu6050 - - invensense,mpu6500 - - invensense,mpu6515 - - invensense,mpu6880 - - invensense,mpu9150 - - invensense,mpu9250 - - invensense,mpu9255 + oneOf: + - enum: + - invensense,iam20680 + - invensense,icm20608 + - invensense,icm20609 + - invensense,icm20689 + - invensense,icm20602 + - invensense,icm20690 + - invensense,mpu6000 + - invensense,mpu6050 + - invensense,mpu6500 + - invensense,mpu6515 + - invensense,mpu6880 + - invensense,mpu9150 + - invensense,mpu9250 + - invensense,mpu9255
+ - items:
+ - const: invensense,icm20608d
+ - const: invensense,icm20608
Which ends up expecting
compatible = "invensense,icm20608d", "invensense,icm20608" and will try matching on the first. If that fails it will try with the second value.
Quoted 103 lines
or
- add a "wildcard" compatible string for this kind of devices. Do you have any pointers?
Regards,
LorenzoQuoted 93 lines
Quoted 9 lines
The only difference between asm330lhhx and asm330lhh is the former supports sensor-hub while the latter does not declare it (even if the use the same whoami).
AFAIK there is no way to autodetect if the sensor supports sensor-hub and we can just try to discover slave devices connected. This can have some downside as described in the commit:Ah thanks. I'd forgotten this.
Quoted 80 lines
commit
35619155d044830357f06f1d2c8188c4530b4d7aAuthor: Lorenzo Bianconi <[email protected]> Date: Sat Nov 13 16:23:14 2021 +0100iio: imu:
st_lsm6dsx: add dts property to disable sensor-hubI would like to merge the sections in
st_lsm6dsx_settingsstruct for lsm6dsr, lsm6dso.. and lsm6dsop, asm330lhh since the only difference is sensor-hub support. I guess we can have 2 option here to avoid any sensor-hub corner cases:
- provide the "st,disable-sensor-hub" in dts to disable sensor-hub for asm330lhh, lsm6dsop (need user changes)
- add a bool variable st_lsm6dsx_settings[].id[] in order to specify if the chip supports sensor-hub.Which one do you prefer?
Regards,
LorenzoQuoted 58 lines
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1
Re: [PATCH 2/2] dt-bindings: iio: imu: st_lsm6dsx: add asm330lhhx device bindings
Quoted 78 lines
Quoted 31 lines · On Mon, 4 Apr 2022 11:33:17 +0200 Lorenzo Bianconi…
Quoted 24 lines · Jonathan Cameron
Quoted 14 lines · On Sun, 3 Apr 2022 16:56:51 +0200 Lorenzo Bianconi <[email protected]> wrote:
Quoted 6 lines
Quoted 1 line · On Sat, 2 Apr 2022 12:09:30 +0200 Lorenzo Bianconi <[email protected]> wrote:
Signed-off-by Lorenzo Bianconi <[email protected]>Hi Lorenzo,
Quoted 7 lines
This runs in to the same feedback that was recently had for
https://lore.kernel.org/all/?q=Add+support+for+ICM-20608-D
but in a more extreme sense as this one presents the same whoami value as for other sensors already supported. Things are made more fun by the fact that sensors with the same WAI seem to have different features (presence or not of a sensor hub - is there any documented way to detect that?).Quoted 6 lines
Hi Jonathan,
if we consider only the features implemented in
st_lsm6dsx, asm330lhhx will be 1:1 compatible with lsm6dsr or lsm6dso, so we can just use one of bindings in this section to support it (the only side effect is it will be listed as "lsm6dsr" or "lsm6dso", but I guess it is ok). Agree?Quoted 5 lines
If the part has more features than the base compatible (or a different WAI) then we can definitely have a backup compatible for it (hence making that subset of features work on an old kernel). We still want to introduce the new compatible so that we get the name right etc going forwards and are in a good position to add the extra features if we ever get around to it.
Quoted 4 lines
ack. I did not completely get what you mean here with "backup compatible". Do you mean:
- use "st,lsm6dsr" for asm330lhhx on older kernels and add "st,asm330lhhx" on new ones. Do you have any pointer on how to document it?Take a look at the mpu6050 patches.
properties:
compatible:(quoted diff) +19
-15- enum: - - invensense,iam20680 - - invensense,icm20608 - - invensense,icm20609 - - invensense,icm20689 - - invensense,icm20602 - - invensense,icm20690 - - invensense,mpu6000 - - invensense,mpu6050 - - invensense,mpu6500 - - invensense,mpu6515 - - invensense,mpu6880 - - invensense,mpu9150 - - invensense,mpu9250 - - invensense,mpu9255 + oneOf: + - enum: + - invensense,iam20680 + - invensense,icm20608 + - invensense,icm20609 + - invensense,icm20689 + - invensense,icm20602 + - invensense,icm20690 + - invensense,mpu6000 + - invensense,mpu6050 + - invensense,mpu6500 + - invensense,mpu6515 + - invensense,mpu6880 + - invensense,mpu9150 + - invensense,mpu9250 + - invensense,mpu9255 + - items: + - const: invensense,icm20608d + - const: invensense,icm20608
Quoted 5 lines
Which ends up expecting
compatible = "invensense,icm20608d", "invensense,icm20608" and will try matching on the first. If that fails it will try with the second value.
ack, thx for pointing this out. I will fix it in v2.
Regards,
Lorenzo
Quoted 105 lines
Quoted 104 lines
or
- add a "wildcard" compatible string for this kind of devices. Do you have any pointers?
Regards,
LorenzoQuoted 93 lines
Quoted 9 lines
The only difference between asm330lhhx and asm330lhh is the former supports sensor-hub while the latter does not declare it (even if the use the same whoami).
AFAIK there is no way to autodetect if the sensor supports sensor-hub and we can just try to discover slave devices connected. This can have some downside as described in the commit:Ah thanks. I'd forgotten this.
Quoted 80 lines
commit
35619155d044830357f06f1d2c8188c4530b4d7aAuthor: Lorenzo Bianconi <[email protected]> Date: Sat Nov 13 16:23:14 2021 +0100iio: imu:
st_lsm6dsx: add dts property to disable sensor-hubI would like to merge the sections in
st_lsm6dsx_settingsstruct for lsm6dsr, lsm6dso.. and lsm6dsop, asm330lhh since the only difference is sensor-hub support. I guess we can have 2 option here to avoid any sensor-hub corner cases:
- provide the "st,disable-sensor-hub" in dts to disable sensor-hub for asm330lhh, lsm6dsop (need user changes)
- add a bool variable st_lsm6dsx_settings[].id[] in order to specify if the chip supports sensor-hub.Which one do you prefer?
Regards,
LorenzoQuoted 58 lines
As such, we should really be listing this as compatible with one of the parts that is already supported such as the LSM6DSR.
For that we'll need a slightly more complex binding and it would have the side effect that if the match was on that compatible we would list the name as whatever that part is.
I'm not sure that really matters a great deal, but it could in theory create a userspace ABI change if we later needed to add explicit support for the part due to some real differences not indicated by the WAI value.
An extension is whether we should relax the need to match on WAI if the part is considered compatible. I guess that depends on just how compatible we think they are.
So I see several steps to this process.
1) Add fallback compatibles for existing entries to first one with same WAI and same feature set.
2) Add fallback compatibles beyond that to first part introduced with particular characteristics. For this we'd also want to have the driver relax its handling to just warn if the WAI isn't listed for any of the parts that share a particular set of characteristic (so you'll have to loop over the local array again to check):
https://elixir.bootlin.com/linux/latest/source/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c#L1197
Same argument applies as for the mpu6050 that, whilst we should modify that code to cope, it's not a prerequisit for adding the compatible fallback to the binding. Personally I'd like it to be the first patch in the series that modifies the binding though. Note it'll be easy to add the fallbacks for this new part as no mainline trees presumably use it. To 'fix' the rest we'll have to find and update any DTs in mainline.Note this won't stop us needing to add compatibles to newer kernels (at very least to the dt-binding, but probably also the driver), but it should help a newer DT 'work' with an old kernel.
Jonathan
Quoted 16 lines
--- Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml | 1 + 1 file changed, 1 insertion(+)Documentation/devicetree/bindings/iio/imu/st,lsm6dsx.yaml +1
-0@@ -31,6 +31,7 @@ properties: - st,lsm6dsrx - st,lsm6dst - st,lsm6dsop + - st,asm330lhhx reg: maxItems: 1