From: Conor Dooley <conor@kernel.org>
To: Padmarao.Begari@microchip.com
Cc: rick@andestech.com, Conor.Dooley@microchip.com,
ycliang@andestech.com, lukma@denx.de, seanga2@gmail.com,
u-boot@lists.denx.de
Subject: Re: [PATCH v1 2/6] clk: microchip: mpfs: convert parent rate acquistion to get_get_rate()
Date: Wed, 2 Nov 2022 16:41:59 +0000 [thread overview]
Message-ID: <Y2Kd18ZfJ7bmv+YB@spud> (raw)
In-Reply-To: <d9fd0f23f14533e62255a45d5d3a226805d1b5dd.camel@microchip.com>
On Wed, Nov 02, 2022 at 01:20:33PM +0000, Padmarao.Begari@microchip.com wrote:
> Hi Conor,
>
> > On Tue, 2022-10-25 at 08:58 +0100, Conor Dooley wrote:
> > Currently the clock driver for PolarFire SoC takes a very naive
> > approach
> > to the relationship between clocks. It reads the dt to get an input
> > clock, assumes that that is fixed frequency, reads the "clock-
> s/that that/that
Nope, not a typo ;)
>
> > frequency"
> > property & uses that to set up both the "cfg" and "periph" clocks.
> >
> > Simplifying for the sake of incremental fixes, the "correct"
> > parentage for
> > the clocks currently supported in U-Boot is that the "cfg" clocks
> > should
> > be children of the fixed frequency clock in the dt. The AHB clock is
> > one
> > of these "cfg" clocks and is the parent of the "periph" clocks.
> >
> > Instead of passing the clock rate of the fixed-frequency clock to the
> > "cfg" and "periph" registration functions and the name of the
> > parents,
> > pass their actual parents & use clk_get_rate() to determine their
> > parents
> > rates.
> >
> > The "periph" clocks are purely gate clocks and should not be reading
> > the
> > AHB clocks registers to determine their rates, as they can simply
> > report
> > the output of clk_get_rate() on their parent.
> >
> > Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> > ---
> > drivers/clk/microchip/Makefile | 2 +-
> > drivers/clk/microchip/mpfs_clk.c | 18 ++++++++----------
> > drivers/clk/microchip/mpfs_clk.h | 12 ++++--------
> > drivers/clk/microchip/mpfs_clk_cfg.c | 7 +++----
> > drivers/clk/microchip/mpfs_clk_periph.c | 16 ++++------------
> > 5 files changed, 20 insertions(+), 35 deletions(-)
> >
> > diff --git a/drivers/clk/microchip/Makefile
> > b/drivers/clk/microchip/Makefile
> > index 904b345d75..329b2c0c93 100644
> > --- a/drivers/clk/microchip/Makefile
> > +++ b/drivers/clk/microchip/Makefile
> > @@ -1 +1 @@
> > -obj-y += mpfs_clk.o mpfs_clk_cfg.o mpfs_clk_periph.o
> > +obj-y += mpfs_clk.o mpfs_clk_cfg.o mpfs_clk_periph.o
> > mpfs_clk_msspll.o
> > diff --git a/drivers/clk/microchip/mpfs_clk.c
> > b/drivers/clk/microchip/mpfs_clk.c
> > index 67828c9bf4..7ba1218b56 100644
> > --- a/drivers/clk/microchip/mpfs_clk.c
> > +++ b/drivers/clk/microchip/mpfs_clk.c
> > @@ -11,34 +11,32 @@
> > #include <dm/device.h>
> > #include <dm/devres.h>
> > #include <dm/uclass.h>
> > +#include <dt-bindings/clock/microchip-mpfs-clock.h>
> > #include <linux/err.h>
> >
> > #include "mpfs_clk.h"
> >
> > static int mpfs_clk_probe(struct udevice *dev)
> > {
> > - int ret;
> > + struct clk *parent_clk = dev_get_priv(dev);
> > + struct clk clk_ahb = { .id = CLK_AHB };
> The peripheral clock updated code added in this patch but removed it in
> the patch 4, you can update only related code in this patch instead of
> removing it later.
Right. I did that intentionally so that each patch did the minimum
required to fix individual issues. This patch does some reorganisation
so that the follow up patches were only as minimal fixes as possible.
I'll reorganise things if you think that's the better way to go about it
though.
>
> Other than that:
> Reviewed-by: Padmarao Begari <padmarao.begari@microchip.com>
>
> > void __iomem *base;
> > - u32 clk_rate;
> > - const char *parent_clk_name;
> > - struct clk *clk = dev_get_priv(dev);
> > + int ret;
> >
> > base = dev_read_addr_ptr(dev);
> > if (!base)
> > return -EINVAL;
> >
> > - ret = clk_get_by_index(dev, 0, clk);
> > + ret = clk_get_by_index(dev, 0, parent_clk);
> > if (ret)
> > return ret;
> >
> > - dev_read_u32(clk->dev, "clock-frequency", &clk_rate);
> > - parent_clk_name = clk->dev->name;
> > -
> > - ret = mpfs_clk_register_cfgs(base, clk_rate, parent_clk_name);
> > + ret = mpfs_clk_register_cfgs(base, parent_clk);
> > if (ret)
> > return ret;
> >
> > - ret = mpfs_clk_register_periphs(base, clk_rate, "clk_ahb");
> > + clk_request(dev, &clk_ahb);
> > + ret = mpfs_clk_register_periphs(base, &clk_ahb);
> >
> > return ret;
> > }
> > diff --git a/drivers/clk/microchip/mpfs_clk.h
> > b/drivers/clk/microchip/mpfs_clk.h
> > index 442562a5e7..35cfeac92e 100644
> > --- a/drivers/clk/microchip/mpfs_clk.h
> > +++ b/drivers/clk/microchip/mpfs_clk.h
> > @@ -11,22 +11,18 @@
> > * mpfs_clk_register_cfgs() - register configuration clocks
> > *
> > * @base: base address of the mpfs system register.
> > - * @clk_rate: the mpfs pll clock rate.
> > - * @parent_name: a pointer to parent clock name.
> > + * @parent: a pointer to parent clock.
> > * Return: zero on success, or a negative error code.
> > */
> > -int mpfs_clk_register_cfgs(void __iomem *base, u32 clk_rate,
> > - const char *parent_name);
> > +int mpfs_clk_register_cfgs(void __iomem *base, struct clk *parent);
> > /**
> > * mpfs_clk_register_periphs() - register peripheral clocks
> > *
> > * @base: base address of the mpfs system register.
> > - * @clk_rate: the mpfs pll clock rate.
> > - * @parent_name: a pointer to parent clock name.
> > + * @parent: a pointer to parent clock.
> > * Return: zero on success, or a negative error code.
> > */
> > -int mpfs_clk_register_periphs(void __iomem *base, u32 clk_rate,
> > - const char *parent_name);
> > +int mpfs_clk_register_periphs(void __iomem *base, struct clk
> > *parent);
> > /**
> > * divider_get_val() - get the clock divider value
> > *
> > diff --git a/drivers/clk/microchip/mpfs_clk_cfg.c
> > b/drivers/clk/microchip/mpfs_clk_cfg.c
> > index fefddd1413..5739fd66e8 100644
> > --- a/drivers/clk/microchip/mpfs_clk_cfg.c
> > +++ b/drivers/clk/microchip/mpfs_clk_cfg.c
> > @@ -117,8 +117,7 @@ static struct mpfs_cfg_hw_clock mpfs_cfg_clks[] =
> > {
> > CLK_CFG(CLK_AHB, "clk_ahb", 4, 2, mpfs_div_ahb_table, 0),
> > };
> >
> > -int mpfs_clk_register_cfgs(void __iomem *base, u32 clk_rate,
> > - const char *parent_name)
> > +int mpfs_clk_register_cfgs(void __iomem *base, struct clk *parent)
> > {
> > int ret;
> > int i, id, num_clks;
> > @@ -129,9 +128,9 @@ int mpfs_clk_register_cfgs(void __iomem *base,
> > u32 clk_rate,
> > for (i = 0; i < num_clks; i++) {
> > hw = &mpfs_cfg_clks[i].hw;
> > mpfs_cfg_clks[i].sys_base = base;
> > - mpfs_cfg_clks[i].prate = clk_rate;
> > + mpfs_cfg_clks[i].prate = clk_get_rate(parent);
> > name = mpfs_cfg_clks[i].cfg.name;
> > - ret = clk_register(hw, MPFS_CFG_CLOCK, name,
> > parent_name);
> > + ret = clk_register(hw, MPFS_CFG_CLOCK, name, parent-
> > >dev->name);
> > if (ret)
> > ERR_PTR(ret);
> > id = mpfs_cfg_clks[i].cfg.id;
> > diff --git a/drivers/clk/microchip/mpfs_clk_periph.c
> > b/drivers/clk/microchip/mpfs_clk_periph.c
> > index 61d90eb4a8..1488ef503e 100644
> > --- a/drivers/clk/microchip/mpfs_clk_periph.c
> > +++ b/drivers/clk/microchip/mpfs_clk_periph.c
> > @@ -99,16 +99,9 @@ static int mpfs_periph_clk_disable(struct clk *hw)
> > static ulong mpfs_periph_clk_recalc_rate(struct clk *hw)
> > {
> > struct mpfs_periph_hw_clock *periph_hw =
> > to_mpfs_periph_clk(hw);
> > - void __iomem *base_addr = periph_hw->sys_base;
> > - unsigned long rate;
> > - u32 val;
> >
> > - val = readl(base_addr + REG_CLOCK_CONFIG_CR) >> CFG_AHB_SHIFT;
> > - val &= clk_div_mask(CFG_WIDTH);
> > - rate = periph_hw->prate / (1u << val);
> > - hw->rate = rate;
> > + return periph_hw->prate;
> >
> > - return rate;
> > }
> >
> > #define CLK_PERIPH(_id, _name, _shift, _flags) { \
> > @@ -150,8 +143,7 @@ static struct mpfs_periph_hw_clock
> > mpfs_periph_clks[] = {
> > CLK_PERIPH(CLK_CFM, "clk_periph_cfm", 29, 0),
> > };
> >
> > -int mpfs_clk_register_periphs(void __iomem *base, u32 clk_rate,
> > - const char *parent_name)
> > +int mpfs_clk_register_periphs(void __iomem *base, struct clk
> > *parent)
> > {
> > int ret;
> > int i, id, num_clks;
> > @@ -162,9 +154,9 @@ int mpfs_clk_register_periphs(void __iomem *base,
> > u32 clk_rate,
> > for (i = 0; i < num_clks; i++) {
> > hw = &mpfs_periph_clks[i].hw;
> > mpfs_periph_clks[i].sys_base = base;
> > - mpfs_periph_clks[i].prate = clk_rate;
> > + mpfs_periph_clks[i].prate = clk_get_rate(parent);
> > name = mpfs_periph_clks[i].periph.name;
> > - ret = clk_register(hw, MPFS_PERIPH_CLOCK, name,
> > parent_name);
> > + ret = clk_register(hw, MPFS_PERIPH_CLOCK, name, parent-
> > >dev->name);
> > if (ret)
> > ERR_PTR(ret);
> > id = mpfs_periph_clks[i].periph.id;
next prev parent reply other threads:[~2022-11-02 16:42 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-25 7:58 [PATCH v1 0/6] clk: microchip: mpfs: incremental fixes Conor Dooley
2022-10-25 7:58 ` [PATCH v1 1/6] dt-bindings: clk: add missing clk ids for microchip mpfs Conor Dooley
2022-11-02 11:19 ` Leo Liang
2022-11-02 13:20 ` Padmarao.Begari
2022-10-25 7:58 ` [PATCH v1 2/6] clk: microchip: mpfs: convert parent rate acquistion to get_get_rate() Conor Dooley
2022-11-02 11:20 ` Leo Liang
2022-11-02 13:20 ` Padmarao.Begari
2022-11-02 16:41 ` Conor Dooley [this message]
2022-10-25 7:58 ` [PATCH v1 3/6] clk: microchip: mpfs: fix reference clock handling Conor Dooley
2022-11-02 11:21 ` Leo Liang
2022-11-02 13:20 ` Padmarao.Begari
2022-10-25 7:58 ` [PATCH v1 4/6] clk: microchip: mpfs: fix periph clk parentage Conor Dooley
2022-11-02 11:21 ` Leo Liang
2022-11-02 13:20 ` Padmarao.Begari
2022-10-25 7:58 ` [PATCH v1 5/6] clk: microchip: mpfs: fix criticality of peripheral clocks Conor Dooley
2022-11-02 11:22 ` Leo Liang
2022-11-02 13:20 ` Padmarao.Begari
2022-10-25 7:58 ` [PATCH v1 6/6] riscv: dts: fix the mpfs's reference clock frequency Conor Dooley
2022-11-02 11:22 ` Leo Liang
2022-11-02 13:21 ` Padmarao.Begari
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Y2Kd18ZfJ7bmv+YB@spud \
--to=conor@kernel.org \
--cc=Conor.Dooley@microchip.com \
--cc=Padmarao.Begari@microchip.com \
--cc=lukma@denx.de \
--cc=rick@andestech.com \
--cc=seanga2@gmail.com \
--cc=u-boot@lists.denx.de \
--cc=ycliang@andestech.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox