Help with fitting unusual devices into IIO framework

Help with fitting unusual devices into IIO framework

[Sergei Shtylyov]

Hello.

     Recently I have posted a driver for drivers/misc/ driving a pair of 
devices: Renesas R-Car Gyro-ADC and speed-pulse interfaces [1]. Arnd 
Bergmann, one of the maintainers told me that such driver would fit 
better into the IIO infrastructure. I have studied this infrastructure
for some time and only found in it a place for the Gyro-ADC interface.
    Let me describe the devices to be driven in some detail. Gyro-ADC 
i/f is an interface to multichannel serial ADC (of 3 certain models)
which can also calculate and average of each 8 samples and sum up 10 of 
these average values. Speed-pulse interface can count PWM-type input 
signal pulses and period between them. Though  being two separate 
devices, these two have to be driven together because of the shared 
start/stop register (located in Gyro-ADC i/f still) and the clock with 
1.25 ms period supplied from the Gyro-ADC i/f to the speed-pulse i/f;
at the same time speed-pulse i/f generates interrupt with 100 ms period 
to be used by both interfaces.
    At this time, only speed-pulse interface support is required, 
Gyro-ADC can be largely ignored (though it has to be initialized). 
Perhaps we could add a new I/O channel type for speed-pulse i/f? What's 
your thoughts?

[1] marc.info/?l=linux-kernel&m=137367311823027

WBR, Sergei

Re: Help with fitting unusual devices into IIO framework

[Jonathan Cameron]

On 07/15/2013 10:54 PM, Sergei Shtylyov wrote:
> Hello.
> 
One quick point first, if you could persuade your email client to wrap at 80 characters
that  would be great.

> Recently I have posted a driver for drivers/misc/ driving a pair of devices: Renesas R-Car > Gyro-ADC and speed-pulse interfaces [1]. Arnd Bergmann, one of the maintainers told me
> that such driver would fit better into the IIO infrastructure. I have studied this
> infrastructure for some time and only found in it a place for the Gyro-ADC
> interface.

> Let me describe the devices to be driven in some detail. Gyro-ADC i/f is an
> interface to multichannel serial ADC (of 3 certain models) which can also calculate and
> average of each 8 samples and sum up 10 of these average values.
That is indeed fairly standard although some of the ABI for summed values is still under
discussion.

http://thread.gmane.org/gmane.linux.kernel.iio/7978/focus=79864 for example.

> Speed-pulse interface
> can count PWM-type input signal pulses and period between them. Though  being
> two separate devices, these two have to be driven together because of the shared 
> start/stop register (located in Gyro-ADC i/f still) and the clock with 1.25 ms period
> supplied from the Gyro-ADC i/f to the speed-pulse i/f; at the same time speed-pulse i/f
> generates interrupt with 100 ms period to be used by both interfaces. At this time, only
> speed-pulse interface support is required, Gyro-ADC can be largely ignored (though it
> has to be initialized). Perhaps  we could add a new I/O channel type for speed-pulse i/f?
> What's your thoughts?

I certainly have no problem with adding this device type.
Would you like to propose a suitable ABI addition to
/Documentation/ABI/test/sysfs-bus-iio to cover this type of device.

I do vaguely wonder if we would be better treating these entirely generically as pulse
counters rather than specifically as being for speed measurement? (Note I have only
the sketchiest idea of what this device actually does ;)
> 
> [1] marc.info/?l=linux-kernel&m=137367311823027
> 
> WBR, Sergei
> -- 
> To unsubscribe from this list: send the line "unsubscribe linux-iio" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html