Re: [PATCH v2 0/8] iio: imu: new inv_icm45600 driver

[Sean Nyekjaer]

On Thu, Jul 10, 2025 at 08:57:55AM +0100, Remi Buisson via B4 Relay wrote:
> This series add a new driver for managing InvenSense ICM-456xx 6-axis IMUs.
> This next generation of chips includes new generations of 3-axis gyroscope
> and 3-axis accelerometer, support of I3C in addition to I2C and SPI, and
> intelligent MotionTracking features like pedometer, tilt detection, and
> tap detection.
> 
> This series is delivering a driver supporting gyroscope, accelerometer and
> temperature data, with polling and buffering using hwfifo and watermark,
> on I2C, SPI and I3C busses.
> 
> Gyroscope and accelerometer sensors are completely independent and can have
> different ODRs. Since there is only a single FIFO a specific value is used to
> mark invalid data. For keeping the device standard we are de-multiplexing data
> from the FIFO to 2 IIO devices with 2 buffers, 1 for the accelerometer and 1
> for the gyroscope. This architecture also enables to easily turn each sensor
> on/off without impacting the other. The device interrupt is used to read the
> FIFO and launch parsing of accelerometer and gyroscope data. This driver
> relies on the common Invensense timestamping mechanism to handle correctly
> FIFO watermark and dynamic changes of settings.
> 
> The structure of the driver is quite similar to the inv_icm42600 driver,
> however there are significant reasons for adding a different driver for
> inv_icm45600, such as:
> - A completely different register map.

At one point we asked TDK/Invense for a driver for icm42670. It also
have a completely different register map... Grrr :S

Anyhow, should we combine these drivers in inv_icm42600? Like
st_lsm6dsx?

/Sean

> - Different FIFO management, based on number of samples instead of bytes.
> - Different indirect register access mechanism.
>