> On Nov 10, 2016, at 11:08 AM, Kevin Townsend wrote:
>
> Raw data is still important in my opinion. 90% of the time a shares API for
> all accels or mags is the best choice, but there are instances where you will
> want the raw data for DSP filtering, etc., and I think
Raw data is still important in my opinion. 90% of the time a shares API
for all accels or mags is the best choice, but there are instances where
you will want the raw data for DSP filtering, etc., and I think there
should be a mechanism to get standard SI as well as 'raw' values back
from a
Hi Sterling,
I think this makes a lot of sense and the API is critical here. I
definitely agree with the comments below that starting with the API is
the right place. Being able to interrogate devices would also be very
valuable, as well as some high level system to control sampling rates,
I should also mention, another option that we could do is have drivers
for the individual sensors (just have them expose native APIs as
drivers, e.g. bmp2085_read_val()), but also map into a generic sensor
API, that we locate in hw/sensors, as an example.
I’m not sure to what extent people
Hi Kevin,
I have some thoughts, but I’m not sure they all directly address your
concerns :-)
More details on directory re-org and drivers below, but I wanted to
mention upfront: I really think its worth thinking through the sensor
api you mention in your PS, and then doing the directory org
> On Nov 10, 2016, at 9:01 AM, Kevin Townsend wrote:
>
>
>> This avoids the problem of a sensor that is has an accelerometer AND a
>> magnetometer as they will typically both use the same bus. In the case of
>> sensors that can use, say, I2C or SPI, you'd need 2 drivers
This avoids the problem of a sensor that is has an accelerometer AND a
magnetometer as they will typically both use the same bus. In the case of
sensors that can use, say, I2C or SPI, you'd need 2 drivers anyway, so rather
than having /hw/drivers/sensors/bno055/spi/ and
Many sensors support both i2c and SPI though which might pose a problem or
oblige you to have two drivers, one for each bus?
Le jeudi 10 novembre 2016, David G. Simmons a écrit :
> I would actually go about this slightly differently ...
>
> All of these sensors come in,
I would actually go about this slightly differently ...
All of these sensors come in, typically, one flavor. So there are SPI-based
sensors, and ADC-based sensors, etc. Grouping them by the bus they use makes a
certain amount of sense to me.
Hw/drivers/SPI/...
hw/drivers/ADC/...
...
That
Although having the family in the path would have been nice but it seems as
though the sensors performing multiple functions could be confusing.
so, +1 for either "hw/drivers/sensors//” and
"hw/drivers/sensors/”. Reference drivers are nice to have.
Meta-data and SI units can be part of the
> On Nov 9, 2016, at 5:59 PM, Kevin Townsend wrote:
>
> There are no sensor drivers in the system at present, but now that the HAL
> rework is complete I wanted to port a few drivers over to Mynewt just to
> properly test the HW and HAL out. Existing sensor drivers is one
There are no sensor drivers in the system at present, but now that the
HAL rework is complete I wanted to port a few drivers over to Mynewt
just to properly test the HW and HAL out. Existing sensor drivers is one
of the key factors to draw people into any embedded RTOS/system, so I
think it's
12 matches
Mail list logo