On Thu, Jan 26, 2017 at 9:09 AM, Dave Gerlach <d-gerlach@xxxxxx> wrote: > On 01/25/2017 04:32 PM, Rob Herring wrote: >> >> On Wed, Jan 25, 2017 at 10:59 AM, Dave Gerlach <d-gerlach@xxxxxx> wrote: [...] >>> because genpd_xlate_simple only checks that the phandle is zero so that >>> it >>> can fail if it is not, but there's no functional reason it needs to do >>> this. >>> The genpd framework works as it did before no matter what the cells are >>> set >>> to if using of_genpd_add_provider_simple. Then in the attach_dev callback >>> inside the ti_sci_pm_domains driver instead of doing >>> >>> ret = of_property_read_u32(np, "ti,sci-id", &idx); >>> >>> to read the ti,sc-id for a device into idx we can now do: >>> >>> ret = of_parse_phandle_with_args(np, "power-domains", >>> "#power-domain-cells", 0, &pd_args); >>> idx = pd_args.args[0]; >>> >>> or even simpler from within our driver >>> >>> ret = of_property_read_u32_index(np, "power-domains", 1, &idx); >> >> >> This you should not be doing. The client driver shouldn't care how >> many cells or what their values are. > > > Client drivers in other places use xlate functions, like in the > drivers/reset framework, to interpret the cells. Is doing it this way really > that different? Oh, I was thinking the client driver, not the power domain controller driver. NM. Rob -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
