Re: [PATCH 3/4] sh: clock-sh7724: all div6_clks use SH_CLK_DIV6_EXT()

[Paul Mundt <lethal@linux-sh.org>]

On Mon, Dec 05, 2011 at 09:47:17PM -0800, Kuninori Morimoto wrote:
> > On Sun, Dec 04, 2011 at 10:26:30PM -0800, Kuninori Morimoto wrote:
> > > -static struct clk div6_clks[DIV6_NR] = {
> > > -	[DIV6_V] = SH_CLK_DIV6(&div3_clk, VCLKCR, 0),
> > > -	[DIV6_I] = SH_CLK_DIV6(&div3_clk, IRDACLKCR, 0),
> > > -	[DIV6_S] = SH_CLK_DIV6(&div3_clk, SPUCLKCR, CLK_ENABLE_ON_INIT),
> > > +/* Indices are important - they are the actual src selecting values */
> > > +static struct clk *common_parent[] = {
> > > +	[0] = &div3_clk,
> > > +	[1] = NULL,
> > >  };
> > >  
> > > -enum { DIV6_FA, DIV6_FB, DIV6_REPARENT_NR };
> > > +static struct clk *vclkcr_parent[] = {
> > > +	[0] = &div3_clk,
> > > +	[1] = NULL,
> > > +	[2] = &sh7724_dv_clki,
> > > +	[3] = NULL,
> > > +	[4] = &extal_clk,
> > > +	[5] = NULL,
> > > +	[6] = NULL,
> > > +	[7] = NULL,
> > > +};
> > >  
> > You don't really need any of these NULL initializers, they can all be
> > dropped.
> 
> Hmm...
> Then, I can not use ARRAY_SIZE(vclkcr_parent) ?
> 
You can just initialize the array size. There's nothing wrong with having
a vclkcr_parent[8] or some _NR define to that extent as part of the
initializer, even if you're just partially initializing the elements.