BMI160 driver asking for 0 bytes

[arurke]

We have experienced the BMI160 driver (https://github.com/BoschSensortec/BMI160_driver/) calling read() with a length of 0. What should the read() function we supply return in these cases? BMI160_OK?

[zgg]

Hello

Could you please tell us which API/function was causing this? Thanks. 

[arurke]

We have only seen it from bmi160_get_fifo_data()

Note: We always set the length in "struct bmi160_fifo_frame" inside "struct bmi160_dev" to an appropriate value before passing it as parameter to bmi160_get_fifo_data()

[arurke]

Any news on this? Do you need further information zgg?

[sinya]

Hi,

Generally,  the default fifo read length should be initialized.

one example in README.md file is :

int8_t fifo_gyro_header_time_data(struct bmi160_dev *dev)
{
 int8_t rslt = 0;

 /* Declare memory to store the raw FIFO buffer information */
 uint8_t fifo_buff[300];
 
 /* Modify the FIFO buffer instance and link to the device instance */
 struct bmi160_fifo_frame fifo_frame;
 fifo_frame.data = fifo_buff;
 fifo_frame.length = 300;
 dev->fifo = &fifo_frame;
 uint16_t index = 0;

......

}

Back to function bmi160_get_fifo_data(struct bmi160_dev const *dev)

if dev->fifo->length is 0 , it will cause fifo read error in sentence:

 rslt = dev->read(dev->id, addr, dev->fifo->data, dev->fifo->length);

so error should be returned.

After reviewing the API code, it is found that there is potential risk that the bytes_to_read may be 0 if random read the fifo in polling mode.

so it is recommended to modify the  sentence from:

if (dev->fifo->length > bytes_to_read)

to

if ((dev->fifo->length > bytes_to_read) &&  0 != bytes_to_read)