[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAL_Jsq+8KSGL1963Zk1-z6-=u-wPOBO52cZg29Y3fUjop09fhg@mail.gmail.com>
Date: Mon, 9 Nov 2020 16:55:51 -0600
From: Rob Herring <robh@...nel.org>
To: Michael Walle <michael@...le.cc>
Cc: linux-clk <linux-clk@...r.kernel.org>, devicetree@...r.kernel.org,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
Michael Turquette <mturquette@...libre.com>,
Stephen Boyd <sboyd@...nel.org>,
Shawn Guo <shawnguo@...nel.org>, Li Yang <leoyang.li@....com>,
"Y . b . Lu" <yangbo.lu@....com>,
Xiaowei Bao <xiaowei.bao@....com>,
Ashish Kumar <Ashish.Kumar@....com>,
Vladimir Oltean <vladimir.oltean@....com>
Subject: Re: [PATCH v3 3/9] clk: qoriq: provide constants for the type
On Mon, Nov 9, 2020 at 4:39 PM Michael Walle <michael@...le.cc> wrote:
>
> Am 2020-11-09 23:05, schrieb Rob Herring:
> > On Sun, Nov 08, 2020 at 07:51:07PM +0100, Michael Walle wrote:
> >> To avoid future mistakes in the device tree for the clockgen module,
> >> add
> >> constants for the clockgen subtype as well as a macro for the PLL
> >> divider.
> >>
> >> Signed-off-by: Michael Walle <michael@...le.cc>
> >> ---
> >> Changes since v2:
> >> - new patch
> >>
> >> drivers/clk/clk-qoriq.c | 13 +++++++------
> >> include/dt-bindings/clock/fsl,qoriq-clockgen.h | 15 +++++++++++++++
> >> 2 files changed, 22 insertions(+), 6 deletions(-)
> >> create mode 100644 include/dt-bindings/clock/fsl,qoriq-clockgen.h
> >>
> >> diff --git a/drivers/clk/clk-qoriq.c b/drivers/clk/clk-qoriq.c
> >> index 46101c6a20f2..70aa521e7e7f 100644
> >> --- a/drivers/clk/clk-qoriq.c
> >> +++ b/drivers/clk/clk-qoriq.c
> >> @@ -7,6 +7,7 @@
> >>
> >> #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
> >>
> >> +#include <dt-bindings/clock/fsl,qoriq-clockgen.h>
> >> #include <linux/clk.h>
> >> #include <linux/clk-provider.h>
> >> #include <linux/clkdev.h>
> >> @@ -1368,33 +1369,33 @@ static struct clk *clockgen_clk_get(struct
> >> of_phandle_args *clkspec, void *data)
> >> idx = clkspec->args[1];
> >>
> >> switch (type) {
> >> - case 0:
> >> + case QORIQ_CLK_SYSCLK:
> >> if (idx != 0)
> >> goto bad_args;
> >> clk = cg->sysclk;
> >> break;
> >> - case 1:
> >> + case QORIQ_CLK_CMUX:
> >> if (idx >= ARRAY_SIZE(cg->cmux))
> >> goto bad_args;
> >> clk = cg->cmux[idx];
> >> break;
> >> - case 2:
> >> + case QORIQ_CLK_HWACCEL:
> >> if (idx >= ARRAY_SIZE(cg->hwaccel))
> >> goto bad_args;
> >> clk = cg->hwaccel[idx];
> >> break;
> >> - case 3:
> >> + case QORIQ_CLK_FMAN:
> >> if (idx >= ARRAY_SIZE(cg->fman))
> >> goto bad_args;
> >> clk = cg->fman[idx];
> >> break;
> >> - case 4:
> >> + case QORIQ_CLK_PLATFORM_PLL:
> >> pll = &cg->pll[PLATFORM_PLL];
> >> if (idx >= ARRAY_SIZE(pll->div))
> >> goto bad_args;
> >> clk = pll->div[idx].clk;
> >> break;
> >> - case 5:
> >> + case QORIQ_CLK_CORECLK:
> >> if (idx != 0)
> >> goto bad_args;
> >> clk = cg->coreclk;
> >> diff --git a/include/dt-bindings/clock/fsl,qoriq-clockgen.h
> >> b/include/dt-bindings/clock/fsl,qoriq-clockgen.h
> >> new file mode 100644
> >> index 000000000000..ddec7d0bdc7f
> >> --- /dev/null
> >> +++ b/include/dt-bindings/clock/fsl,qoriq-clockgen.h
> >> @@ -0,0 +1,15 @@
> >> +/* SPDX-License-Identifier: GPL-2.0-only */
> >> +
> >> +#ifndef DT_CLOCK_FSL_QORIQ_CLOCKGEN_H
> >> +#define DT_CLOCK_FSL_QORIQ_CLOCKGEN_H
> >> +
> >> +#define QORIQ_CLK_SYSCLK 0
> >> +#define QORIQ_CLK_CMUX 1
> >> +#define QORIQ_CLK_HWACCEL 2
> >> +#define QORIQ_CLK_FMAN 3
> >> +#define QORIQ_CLK_PLATFORM_PLL 4
> >> +#define QORIQ_CLK_CORECLK 5
> >> +
> >> +#define QORIQ_CLK_PLL_DIV(x) ((x) - 1)
> >
> > This is not used and doesn't seem like part of the ABI (shared with dts
> > files).
>
> TBH I haven't found a nice way to integrate this macro into the clock
> driver. It is used in the device tree for the type PLATFORM_PLL.
> Previously, you had "<&clockgen 4 1>", where 4 is the PLATFORM_PLL and 1
> is actually "div-by-2". Thus I replaced it by <&clockgen
> QORIQ_CLK_PLATFORM_PLL QORIQ_CLK_PLL_DIV(2)>. (I just realized that
> QORIQ_CLK_PLL_DIV_BY(2) might be a better name.)
Ah okay. I guess this is fine.
Acked-by: Rob Herring <robh@...nel.org>
Rob
Powered by blists - more mailing lists