Re: regmap I3C support

[Guenter Roeck]

On 11/15/24 07:06, Mark Brown wrote:
> On Thu, Nov 14, 2024 at 10:58:05PM -0800, Guenter Roeck wrote:
> > On Thu, Nov 14, 2024 at 05:26:19PM +0000, Mark Brown wrote:
>
> > > Right, so the fact that I3C depends on I2C deals with a lot of the
> > > problems that plague the I2C/SPI combination.  Ugh.  I guess the helper
> > > should be OK and there's not much doing for I2C/SPI.
>
> > Is it really that difficult for I2C and SPI ? The patch below seems to work
> > for the LTC2947 driver. It doesn't even need dummies (the compiler drops
> > the unused code), though I am not sure if that can be relied on. I thought
> > that dummy functions are needed, but maybe I am wrong.
> >
> > The Kconfig for the combined ltc2947 driver is
> >
> > config SENSORS_LTC2947
> >          tristate "Analog Devices LTC2947 High Precision Power and Energy Monitor"
> >          depends on I2C || SPI
> >          depends on I2C || I2C=n
> >          select REGMAP_I2C if I2C
> >          select REGMAP_SPI if SPI
>
> This prevents building the driver in if I2C=m which isn't always
> desirable, and IIRC the randconfig people kept turning issues up.  You
> can make things work well enough for normal configurations, it's all
> edge cases that cause issues.

You mean if SPI=y and I2C=m ? Good point. That isn't normally a problem
for sensor drivers, but I understand that it may be undesirable for others.

Thanks,
Guenter