Re: [PATCH v1 1/1] mfd: intel-lpss: Provide an SSP type to the SPI driver

[Lee Jones]

On Wed, 13 Jul 2022, Andy Shevchenko wrote:

> On Wed, Jul 13, 2022 at 09:10:06AM +0100, Lee Jones wrote:
> > On Sun, 03 Jul 2022, Andy Shevchenko wrote:
> 
> > > +static const struct property_entry spt_spi_properties[] = {
> > > +	PROPERTY_ENTRY_U32("intel,spi-pxa2xx-type", LPSS_SPT_SSP),
> > > +	{ }
> > > +};
> > > +
> > > +static const struct software_node spt_spi_node = {
> > > +	.properties = spt_spi_properties,
> > > +};
> > > +
> > >  static const struct intel_lpss_platform_info spt_info = {
> > >  	.clk_rate = 120000000,
> > > +	.swnode = &spt_spi_node,
> > >  };
> > 
> > IMHO, this is a rubbish interface.
> > 
> > The amount of 10-line changes required to store a 32-bit value is
> > depressing.  Is there not a reduced interface for storing small pieces
> > of data that doesn't require arrays of structs?
> 
> The part of it that makes it longer because of differentiating the hardware,
> but see the difference here:
> 
> In this change:
> 	95 insertions(+), 46 deletions(-)
> 
> In the follow up SPI driver cleaning (not yet submitted):
> 	14 insertions(+), 121 deletions(-)

Not sure we are understanding each other.

The patch is fine.  I'm referring to the software_node interface.

-- 
Lee Jones [李琼斯]
Principal Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog