Re: [PATCH v2 3/3] mfd: intel-m10-bmc: add Max10 BMC chip support for Intel FPGA PAC

[Xu Yilun]

On Tue, Jul 21, 2020 at 11:47:13AM +0800, Xu Yilun wrote:
> On Fri, Jul 17, 2020 at 07:16:09PM +0100, Mark Brown wrote:
> > On Thu, Jul 16, 2020 at 06:42:54PM +0800, Xu Yilun wrote:
> > 
> > > +static const struct spi_device_id m10bmc_spi_id[] = {
> > > +	{ "m10-n3000", M10_N3000 },
> > > +	{ }
> > > +};
> > > +MODULE_DEVICE_TABLE(spi, m10bmc_spi_id);
> > 
> > > +static struct spi_driver intel_m10bmc_spi_driver = {
> > > +	.driver = {
> > > +		.name = "intel-m10-bmc",
> > > +		.dev_groups = m10bmc_dev_groups,
> > > +	},
> > > +	.probe = intel_m10_bmc_spi_probe,
> > > +	.id_table = m10bmc_spi_id,
> > > +};
> > 
> > > +module_spi_driver(intel_m10bmc_spi_driver);
> > 
> > This device has no ACPI information - how will it be instantiated?
> 
> In our case, The m10-bmc is connected to the intel FPGA (PAC N3000),
> which uses the Device Feature List (DFL) mechanism to enumerate features
> (devices) on FPGA. Each feature in DFL has a feature_id. And for this
> m10-n3000 feature (feature_id = 0xd), it contains a spi-altera & a
> m10-n3000 chip. So the DFL subsystem would help enumerate the info.

And some DFL related infomation:
  Documentation/fpga/dfl.rst
  drivers/fpga/dfl*

Thanks,
Yilun

> 
> Recently I added the platform data for slave information in spi-altera,
> to support this use case.
> 
> Thanks,
> Yilun