Re: [PATCH 4/4] clk: sifive: Refactor __prci_clock array by using macro

From: Zong Li
Date: Thu Nov 05 2020 - 02:16:50 EST

Next message: Tomi Valkeinen: "Re: [PATCH] arm64: dts: ti: k3-am65*/j721e*: Fix unit address format error for dss node"
Previous message: menglong8 . dong: "[PATCH net-next] net: udp: introduce UDP_MIB_MEMERRORS for udp_mem"
In reply to: Stephen Boyd: "Re: [PATCH 4/4] clk: sifive: Refactor __prci_clock array by using macro"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Thu, Nov 5, 2020 at 10:50 AM Stephen Boyd <sboyd@xxxxxxxxxx> wrote:
>
> Quoting Zong Li (2020-10-16 02:18:26)
> > Refactor code by using DEFINE_PRCI_CLOCK to define each clock
> > and reduce duplicate code.
>
> What is duplicate?

Sorry for unclear description, actually, I want to say that we can
remove the repeating code about initializing the data member of
structure.

>
> >
> > Signed-off-by: Zong Li <zong.li@xxxxxxxxxx>
> > ---
> > drivers/clk/sifive/fu540-prci.c | 38 ++++++----------
> > drivers/clk/sifive/fu540-prci.h | 2 +-
> > drivers/clk/sifive/fu740-prci.c | 74 ++++++++++++--------------------
> > drivers/clk/sifive/fu740-prci.h | 2 +-
> > drivers/clk/sifive/sifive-prci.c | 2 +-
> > drivers/clk/sifive/sifive-prci.h | 10 ++++-
> > 6 files changed, 53 insertions(+), 75 deletions(-)
> >
> > diff --git a/drivers/clk/sifive/fu540-prci.c b/drivers/clk/sifive/fu540-prci.c
> > index 840b97bfff85..d43b9a9984f6 100644
> > --- a/drivers/clk/sifive/fu540-prci.c
> > +++ b/drivers/clk/sifive/fu540-prci.c
> > @@ -54,29 +54,19 @@ static const struct clk_ops sifive_fu540_prci_tlclksel_clk_ops = {
> > .recalc_rate = sifive_prci_tlclksel_recalc_rate,
> > };
> >
> > +DEFINE_PRCI_CLOCK(fu540, corepll, hfclk,
> > + &sifive_fu540_prci_wrpll_clk_ops, &__prci_corepll_data);
> > +DEFINE_PRCI_CLOCK(fu540, ddrpll, hfclk,
> > + &sifive_fu540_prci_wrpll_ro_clk_ops, &__prci_ddrpll_data);
> > +DEFINE_PRCI_CLOCK(fu540, gemgxlpll, hfclk,
> > + &sifive_fu540_prci_wrpll_clk_ops, &__prci_gemgxlpll_data);
> > +DEFINE_PRCI_CLOCK(fu540, tlclk, corepll,
> > + &sifive_fu540_prci_tlclksel_clk_ops, NULL);
>
> Readability looks to decrease with this change. Why should all us code
> reviewers suffer because the code author wants to type a few less
> characters? Named initializers are useful so we don't have to hold each
> macro argument in our head and map it to the struct member that is being
> initialized.

Ok, as you mentioned, macro reduce the readability, let me remove this
change in the next version.

>
> > +
> > /* List of clock controls provided by the PRCI */
> > -struct __prci_clock __prci_init_clocks_fu540[] = {
> > - [PRCI_CLK_COREPLL] = {
> > - .name = "corepll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu540_prci_wrpll_clk_ops,
> > - .pwd = &__prci_corepll_data,
> > - },
> > - [PRCI_CLK_DDRPLL] = {
> > - .name = "ddrpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu540_prci_wrpll_ro_clk_ops,
> > - .pwd = &__prci_ddrpll_data,
> > - },
> > - [PRCI_CLK_GEMGXLPLL] = {
> > - .name = "gemgxlpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu540_prci_wrpll_clk_ops,
> > - .pwd = &__prci_gemgxlpll_data,
> > - },
> > - [PRCI_CLK_TLCLK] = {
> > - .name = "tlclk",
> > - .parent_name = "corepll",
> > - .ops = &sifive_fu540_prci_tlclksel_clk_ops,
> > - },
> > +struct __prci_clock *__prci_init_clocks_fu540[] = {
> > + [PRCI_CLK_COREPLL] = &fu540_corepll,
> > + [PRCI_CLK_DDRPLL] = &fu540_ddrpll,
> > + [PRCI_CLK_GEMGXLPLL] = &fu540_gemgxlpll,
> > + [PRCI_CLK_TLCLK] = &fu540_tlclk,
> > };
> > diff --git a/drivers/clk/sifive/fu540-prci.h b/drivers/clk/sifive/fu540-prci.h
> > index c8271efa7bdc..281200cd8848 100644
> > --- a/drivers/clk/sifive/fu540-prci.h
> > +++ b/drivers/clk/sifive/fu540-prci.h
> > @@ -11,7 +11,7 @@
> >
> > #define NUM_CLOCK_FU540 4
> >
> > -extern struct __prci_clock __prci_init_clocks_fu540[NUM_CLOCK_FU540];
> > +extern struct __prci_clock *__prci_init_clocks_fu540[NUM_CLOCK_FU540];
> >
> > static const struct prci_clk_desc prci_clk_fu540 = {
> > .clks = __prci_init_clocks_fu540,
> > diff --git a/drivers/clk/sifive/fu740-prci.c b/drivers/clk/sifive/fu740-prci.c
> > index 3b87e273c3eb..676cad2c3886 100644
> > --- a/drivers/clk/sifive/fu740-prci.c
> > +++ b/drivers/clk/sifive/fu740-prci.c
> > @@ -71,52 +71,32 @@ static const struct clk_ops sifive_fu740_prci_hfpclkplldiv_clk_ops = {
> > .recalc_rate = sifive_prci_hfpclkplldiv_recalc_rate,
> > };
> >
> > +
> > +DEFINE_PRCI_CLOCK(fu740, corepll, hfclk,
> > + &sifive_fu740_prci_wrpll_clk_ops, &__prci_corepll_data);
> > +DEFINE_PRCI_CLOCK(fu740, ddrpll, hfclk,
> > + &sifive_fu740_prci_wrpll_ro_clk_ops, &__prci_ddrpll_data);
> > +DEFINE_PRCI_CLOCK(fu740, gemgxlpll, hfclk,
> > + &sifive_fu740_prci_wrpll_clk_ops, &__prci_gemgxlpll_data);
> > +DEFINE_PRCI_CLOCK(fu740, dvfscorepll, hfclk,
> > + &sifive_fu740_prci_wrpll_clk_ops, &__prci_dvfscorepll_data);
> > +DEFINE_PRCI_CLOCK(fu740, hfpclkpll, hfclk,
> > + &sifive_fu740_prci_wrpll_clk_ops, &__prci_hfpclkpll_data);
> > +DEFINE_PRCI_CLOCK(fu740, cltxpll, hfclk,
> > + &sifive_fu740_prci_wrpll_clk_ops, &__prci_cltxpll_data);
> > +DEFINE_PRCI_CLOCK(fu740, tlclk, corepll,
> > + &sifive_fu740_prci_tlclksel_clk_ops, NULL);
> > +DEFINE_PRCI_CLOCK(fu740, pclk, hfpclkpll,
> > + &sifive_fu740_prci_hfpclkplldiv_clk_ops, NULL);
> > +
> > /* List of clock controls provided by the PRCI */
> > -struct __prci_clock __prci_init_clocks_fu740[] = {
> > - [PRCI_CLK_COREPLL] = {
> > - .name = "corepll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_clk_ops,
> > - .pwd = &__prci_corepll_data,
> > - },
> > - [PRCI_CLK_DDRPLL] = {
> > - .name = "ddrpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_ro_clk_ops,
> > - .pwd = &__prci_ddrpll_data,
> > - },
> > - [PRCI_CLK_GEMGXLPLL] = {
> > - .name = "gemgxlpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_clk_ops,
> > - .pwd = &__prci_gemgxlpll_data,
> > - },
> > - [PRCI_CLK_DVFSCOREPLL] = {
> > - .name = "dvfscorepll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_clk_ops,
> > - .pwd = &__prci_dvfscorepll_data,
> > - },
> > - [PRCI_CLK_HFPCLKPLL] = {
> > - .name = "hfpclkpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_clk_ops,
> > - .pwd = &__prci_hfpclkpll_data,
> > - },
> > - [PRCI_CLK_CLTXPLL] = {
> > - .name = "cltxpll",
> > - .parent_name = "hfclk",
> > - .ops = &sifive_fu740_prci_wrpll_clk_ops,
> > - .pwd = &__prci_cltxpll_data,
> > - },
> > - [PRCI_CLK_TLCLK] = {
> > - .name = "tlclk",
> > - .parent_name = "corepll",
> > - .ops = &sifive_fu740_prci_tlclksel_clk_ops,
> > - },
> > - [PRCI_CLK_PCLK] = {
> > - .name = "pclk",
> > - .parent_name = "hfpclkpll",
> > - .ops = &sifive_fu740_prci_hfpclkplldiv_clk_ops,
> > - },
> > +struct __prci_clock *__prci_init_clocks_fu740[] = {
> > + [PRCI_CLK_COREPLL] = &fu740_corepll,
> > + [PRCI_CLK_DDRPLL] = &fu740_ddrpll,
> > + [PRCI_CLK_GEMGXLPLL] = &fu740_gemgxlpll,
> > + [PRCI_CLK_DVFSCOREPLL] = &fu740_dvfscorepll,
> > + [PRCI_CLK_HFPCLKPLL] = &fu740_hfpclkpll,
> > + [PRCI_CLK_CLTXPLL] = &fu740_cltxpll,
> > + [PRCI_CLK_TLCLK] = &fu740_tlclk,
> > + [PRCI_CLK_PCLK] = &fu740_pclk,
> > };
>
> I suppose this is fine and then non-macro structs above this array of
> pointers.
>
> > diff --git a/drivers/clk/sifive/fu740-prci.h b/drivers/clk/sifive/fu740-prci.h
> > index 13ef971f7764..3f03295f719a 100644
> > --- a/drivers/clk/sifive/fu740-prci.h
> > +++ b/drivers/clk/sifive/fu740-prci.h
> > @@ -11,7 +11,7 @@
> >
> > #define NUM_CLOCK_FU740 8
> >
> > -extern struct __prci_clock __prci_init_clocks_fu740[NUM_CLOCK_FU740];
> > +extern struct __prci_clock *__prci_init_clocks_fu740[NUM_CLOCK_FU740];
> >
> > static const struct prci_clk_desc prci_clk_fu740 = {
> > .clks = __prci_init_clocks_fu740,
> > diff --git a/drivers/clk/sifive/sifive-prci.c b/drivers/clk/sifive/sifive-prci.c
> > index 4098dbc5881a..2ef3f9f91b33 100644
> > --- a/drivers/clk/sifive/sifive-prci.c
> > +++ b/drivers/clk/sifive/sifive-prci.c
> > @@ -431,7 +431,7 @@ static int __prci_register_clocks(struct device *dev, struct __prci_data *pd,
> >
> > /* Register PLLs */
> > for (i = 0; i < desc->num_clks; ++i) {
> > - pic = &(desc->clks[i]);
> > + pic = desc->clks[i];
>
> This is related how?
>
> >
> > init.name = pic->name;
> > init.parent_names = &pic->parent_name;
> > diff --git a/drivers/clk/sifive/sifive-prci.h b/drivers/clk/sifive/sifive-prci.h
> > index bc0646bc9c3e..e6c9f72e20de 100644
> > --- a/drivers/clk/sifive/sifive-prci.h
> > +++ b/drivers/clk/sifive/sifive-prci.h
> > @@ -253,6 +253,14 @@ struct __prci_clock {
> > struct __prci_data *pd;
> > };
> >
> > +#define DEFINE_PRCI_CLOCK(_platform, _name, _parent, _ops, _pwd) \
> > + static struct __prci_clock _platform##_##_name = { \
> > + .name = #_name, \
> > + .parent_name = #_parent, \
> > + .ops = _ops, \
> > + .pwd = _pwd, \
> > + } \
> > +
> > #define clk_hw_to_prci_clock(pwd) container_of(pwd, struct __prci_clock, hw)
> >
> > /*
> > @@ -261,7 +269,7 @@ struct __prci_clock {
> > * @num_clks: the number of element of clks
> > */
> > struct prci_clk_desc {
> > - struct __prci_clock *clks;
> > + struct __prci_clock **clks;
>
> Huh? Nothing in the commit text mentions this.
>

Because I introduce the macro in this patch, so the type of array
__prci_init_clocks_fuXXX are changed to pointer which point to
__prci_clock, the related use need to be changed as well. It would be
recover due to discarding this patch.

> > size_t num_clks;
> > };

Next message: Tomi Valkeinen: "Re: [PATCH] arm64: dts: ti: k3-am65*/j721e*: Fix unit address format error for dss node"
Previous message: menglong8 . dong: "[PATCH net-next] net: udp: introduce UDP_MIB_MEMERRORS for udp_mem"
In reply to: Stephen Boyd: "Re: [PATCH 4/4] clk: sifive: Refactor __prci_clock array by using macro"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]