Re: [PATCH v6 5/6] iio: imu: st_lsm6dsx: add motion report function and call from interrupt

[Lorenzo Bianconi <lorenzo@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1533 bytes --]

> > > probably a silly question..are you tracing the interrupt line with an
> > > oscilloscope or a logical analyser? If you dump interrupt counters in
> > > /proc/interrupts will you see an interrupt storm for the selected irq
> > > pin?
> > > 
> > 
> > Not a silly question ;-)
> > 
> > An Oscilloscope :-)
> 
> ack, thx
> 
> Could you please try to carry out the following test?
> 1- set the FIFO watermark to a high level (e.g. 128)
>    $echo 256 > /sys/bus/iio/devices/iio:device{x}/buffer/length
>    $echo 128 > /sys/bus/iio/devices/iio:device{x}/buffer/watermark
> 2- set a low acc odr (e.g 13Hz)
>    $echo 13 > /sys/bus/iio/devices/iio:device{x}/sampling_frequency
> 3- start reading from the accel and generate a wake-upp event
> 
> Is still happen? Are you able to decode bus transaction? (register addresses,
> data read, ..)
> 
> > The interrupt counter is stopping.
> > If I switch to IRQ_TYPE_LEVEL_HIGH, (to test if additional readings of the
> > event and FIFO registers would help. It results in interrupt storm and the
> > line continues to stay high.
> > 

Could you please try to enable the LIR (Latched interrupt - BIT(0) in 0x58)?
Please remember that on ISM330DLC the interrupt will be automatically cleared
reading the wake up src register after a time slice equals to 1/ODR so the it
will be set for longer time if you run the device at low ODR

Regards,
Lorenzo

> 
> @Denis, @Mario, @Armando: any thoughts about this?
> 
> Regards,
> Lorenzo
> 
> > /Sean



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]