Re: [PATCH v4 2/3] iio: adc: Support ROHM BD79112 ADC/GPIO

[Jonathan Cameron <jonathan.cameron@huawei.com>]

On Sun, 14 Sep 2025 12:25:06 +0300
Matti Vaittinen <mazziesaccount@gmail.com> wrote:

> On 13/09/2025 15:24, Jonathan Cameron wrote:
> > On Thu, 11 Sep 2025 08:13:03 +0300
> > Matti Vaittinen <mazziesaccount@gmail.com> wrote:
> >   
> >> Morning Jonathan,
> >>
> >> On 10/09/2025 20:46, Jonathan Cameron wrote:  
> >>> On Wed, 10 Sep 2025 14:24:35 +0300
> >>> Matti Vaittinen <mazziesaccount@gmail.com> wrote:
> >>>      
> >>>> The ROHM BD79112 is an ADC/GPIO with 32 channels. The channel inputs can
> >>>> be used as ADC or GPIO. Using the GPIOs as IRQ sources isn't supported.
> >>>>
> >>>> The ADC is 12-bit, supporting input voltages up to 5.7V, and separate I/O
> >>>> voltage supply. Maximum SPI clock rate is 20 MHz (10 MHz with
> >>>> daisy-chain configuration) and maximum sampling rate is 1MSPS.
> >>>>
> >>>> The IC does also support CRC but it is not implemented in the driver.
> >>>>
> >>>> Signed-off-by: Matti Vaittinen <mazziesaccount@gmail.com>  
> >>>
> >>> Hi Matti,
> >>>
> >>> A few trivial things that I'll tidy up if nothing else comes up (I might not
> >>> bother given how trivial they are!)  
> >>
> >> Thanks again!
> >>  
> >>> Also one question. I couldn't immediately follow why any random register
> >>> read is sanity checking if an ADC pin is configured as GPIO.
> >>>      
> >>
> >> Ah. Valid question! I see my comment below is partially wrong.
> >>
> >>  
> >>>> +/*
> >>>> + * Read transaction consists of two 16-bit sequences separated by CSB.
> >>>> + * For register read, 'IOSET' bit must be set. For ADC read, IOSET is cleared
> >>>> + * and ADDR equals the channel number (0 ... 31).
> >>>> + *
> >>>> + * First 16-bit sequence, MOSI as below, MISO data ignored:
> >>>> + * - SCK: | 1 | 2 |   3   |    4   | 5 .. 8 | 9 .. 16 |
> >>>> + * - MOSI:| 0 | 0 | IOSET | RW (1) |  ADDR  |  8'b0   |
> >>>> + *
> >>>> + * CSB released and re-acquired between these sequences
> >>>> + *
> >>>> + * Second 16-bit sequence, MISO as below, MOSI data ignored:
> >>>> + *   For Register read data is 8 bits:
> >>>> + *   - SCK: | 1 .. 8 |   9 .. 16   |
> >>>> + *   - MISO:|  8'b0  | 8-bit data  |
> >>>> + *
> >>>> + *   For ADC read data is 12 bits:
> >>>> + *   - SCK: | 1 .. 4 |   4 .. 16   |
> >>>> + *   - MISO:|  4'b0  | 12-bit data |  
> >>
> >> This is not 100% true. I overlooked the ADC read "status flag" when
> >> adding this comment for the ADC data reading.
> >>
> >> This should be:
> >>
> >>    *   For ADC, read data is 12 bits prepended with a status flag:
> >>    *   - SCK: | 1 |      2      | 3  4 |   4 .. 16   |
> >>    *   - MISO:| 0 | STATUS_FLAG | 2'b0 | 12-bit data |
> >>
> >> The 'STATUS_FLAG' is set if the input pin is configured as a GPIO.  
> > 
> > That's good additional info, but I'm still struggling on why
> > we are effectively providing a 'debug' check in ever register
> > read. My assumption is that it should never fire unless you have
> > a driver bug?  
> 
> Yes, a driver bug or someone accessing the ADC outside the driver.
> 
> I kind of agree the check shouldn't be needed - but I've seen quite a 
> few driver bugs during my career. XD The check is _very_ light weight 
> compared to the SPI access time - but you're right that it is done at 
> every ADC data read - which is 'hot path'. As a result, I am not sure 
> whether to leave or drop it.
Maybe just add a comment along the lines of
/* Lets check this whilst here, but should never happen! */

> 
> Yours,
> 	-- Matti
> 
> > 
> > Jonathan  
> 
>