Re: [PATCH 5/5] clk: mediatek: Add USB clock support in MT8173 APMIXEDSYS

[James Liao <jamesjj.liao-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>]

Hi Sascha,

On Tue, 2015-05-26 at 11:41 +0200, Sascha Hauer wrote:
> On Tue, May 26, 2015 at 05:11:15PM +0800, James Liao wrote:
> > On Tue, 2015-05-26 at 10:05 +0200, Sascha Hauer wrote:
> > > On Thu, May 21, 2015 at 03:12:56PM +0800, James Liao wrote:
> > > > +static void __init mtk_clk_register_apmixedsys_special(struct device_node *node,
> > > > +			struct clk_onecell_data *clk_data)
> > > > +{
> > > > +	void __iomem *base;
> > > > +	struct clk *clk;
> > > > +
> > > > +	base = of_iomap(node, 0);
> > > > +	if (!base) {
> > > > +		pr_err("%s(): ioremap failed\n", __func__);
> > > > +		return;
> > > > +	}
> > > > +
> > > > +	clk = mtk_clk_register_ref2usb_tx("ref2usb_tx", base + 0x8);
> > > 
> > > The function seems to be for one special clock only. Why do you pass the
> > > name to it? They will never be called with another name, right?
> > 
> > This function decides clock name and associates clock ID for special
> > clocks. In fact there may be another "special clocks" need to add into
> > apmixedsys. I think it's a better way to group clock names and clock IDs
> > in the same function for maintenance.
> 
> How can a function with ref2usb_tx in its name ever register a clock
> with another name? Then it seems the function name is wrong.

I mean mtk_clk_register_apmixedsys_special() decides the clock names and
clock ID bindings.

mtk_clk_register_ref2usb_tx() is only used by ref2usb_tx clock. Other
special clocks will not share the implementation with ref2usb_tx. But we
can set names and clock IDs for different special clocks in the same
function (mtk_clk_register_apmixedsys_special()) to avoid inconsistence
naming.


Best regards,

James