Re: [v2 06/24] clk: mediatek: Add MT8195 apmixedsys clock support

From: Chen-Yu Tsai
Date: Mon Aug 23 2021 - 05:56:41 EST


On Mon, Aug 23, 2021 at 5:21 PM Chen-Yu Tsai <wenst@xxxxxxxxxxxx> wrote:
>
> On Fri, Aug 20, 2021 at 7:20 PM Chun-Jie Chen
> <chun-jie.chen@xxxxxxxxxxxx> wrote:
> >
> > Add MT8195 apmixedsys clock controller which provides Plls
> > generated from SoC 26m and ssusb clock gate control.
> >
> > Signed-off-by: Chun-Jie Chen <chun-jie.chen@xxxxxxxxxxxx>
> > ---
> > drivers/clk/mediatek/Kconfig | 8 +
> > drivers/clk/mediatek/Makefile | 1 +
> > drivers/clk/mediatek/clk-mt8195-apmixedsys.c | 145 +++++++++++++++++++
> > 3 files changed, 154 insertions(+)
> > create mode 100644 drivers/clk/mediatek/clk-mt8195-apmixedsys.c
> >
> > diff --git a/drivers/clk/mediatek/Kconfig b/drivers/clk/mediatek/Kconfig
> > index 576babd86f98..7ba1f4118e0d 100644
> > --- a/drivers/clk/mediatek/Kconfig
> > +++ b/drivers/clk/mediatek/Kconfig
> > @@ -580,6 +580,14 @@ config COMMON_CLK_MT8192_VENCSYS
> > help
> > This driver supports MediaTek MT8192 vencsys clocks.
> >
> > +config COMMON_CLK_MT8195
> > + bool "Clock driver for MediaTek MT8195"
> > + depends on ARM64 || COMPILE_TEST
> > + select COMMON_CLK_MEDIATEK
> > + default ARM64
> > + help
> > + This driver supports MediaTek MT8195 basic clocks.
>
> Since we will have all clocks under the same Kconfig option, please
> drop the word "basic".
>
> > +
> > config COMMON_CLK_MT8516
> > bool "Clock driver for MediaTek MT8516"
> > depends on ARCH_MEDIATEK || COMPILE_TEST
> > diff --git a/drivers/clk/mediatek/Makefile b/drivers/clk/mediatek/Makefile
> > index 15bc045f0b71..d4157cfca865 100644
> > --- a/drivers/clk/mediatek/Makefile
> > +++ b/drivers/clk/mediatek/Makefile
> > @@ -80,5 +80,6 @@ obj-$(CONFIG_COMMON_CLK_MT8192_MSDC) += clk-mt8192-msdc.o
> > obj-$(CONFIG_COMMON_CLK_MT8192_SCP_ADSP) += clk-mt8192-scp_adsp.o
> > obj-$(CONFIG_COMMON_CLK_MT8192_VDECSYS) += clk-mt8192-vdec.o
> > obj-$(CONFIG_COMMON_CLK_MT8192_VENCSYS) += clk-mt8192-venc.o
> > +obj-$(CONFIG_COMMON_CLK_MT8195) += clk-mt8195-apmixedsys.o
> > obj-$(CONFIG_COMMON_CLK_MT8516) += clk-mt8516.o
> > obj-$(CONFIG_COMMON_CLK_MT8516_AUDSYS) += clk-mt8516-aud.o
> > diff --git a/drivers/clk/mediatek/clk-mt8195-apmixedsys.c b/drivers/clk/mediatek/clk-mt8195-apmixedsys.c
> > new file mode 100644
> > index 000000000000..253eb30b22d4
> > --- /dev/null
> > +++ b/drivers/clk/mediatek/clk-mt8195-apmixedsys.c
> > @@ -0,0 +1,145 @@
> > +// SPDX-License-Identifier: (GPL-2.0 OR MIT)

I actually only requested the files related to DT bindings, which includes
`Documentation/device-tree/bindings/*` and `include/dt-bindings/*`, be dual
licensed.

I'm not sure dual licensing the actual driver code makes sense, or is
even valid, since this calls into existing GPL-2.0 code. But I am not
a lawyer.


ChenYu

> > +//
> > +// Copyright (c) 2021 MediaTek Inc.
> > +// Author: Chun-Jie Chen <chun-jie.chen@xxxxxxxxxxxx>
> > +
> > +#include "clk-gate.h"
> > +#include "clk-mtk.h"
> > +
> > +#include <dt-bindings/clock/mt8195-clk.h>
> > +#include <linux/of_device.h>
> > +#include <linux/platform_device.h>
> > +
> > +static const struct mtk_gate_regs apmixed_cg_regs = {
> > + .set_ofs = 0x8,
> > + .clr_ofs = 0x8,
> > + .sta_ofs = 0x8,
> > +};
> > +
> > +#define GATE_APMIXED(_id, _name, _parent, _shift) \
> > + GATE_MTK(_id, _name, _parent, &apmixed_cg_regs, _shift, &mtk_clk_gate_ops_no_setclr_inv)
> > +
> > +static const struct mtk_gate apmixed_clks[] = {
> > + GATE_APMIXED(CLK_APMIXED_PLL_SSUSB26M, "pll_ssusb26m", "clk26m", 1),
> > +};
> > +
> > +#define MT8195_PLL_FMAX (3800UL * MHZ)
> > +#define MT8195_PLL_FMIN (1500UL * MHZ)
> > +#define MT8195_INTEGER_BITS 8
> > +
> > +#define PLL(_id, _name, _reg, _pwr_reg, _en_mask, _flags, \
> > + _rst_bar_mask, _pcwbits, _pd_reg, _pd_shift, \
> > + _tuner_reg, _tuner_en_reg, _tuner_en_bit, \
> > + _pcw_reg, _pcw_shift, _pcw_chg_reg, \
> > + _en_reg, _pll_en_bit) { \
>
> Nit: Even for macro definitions, you could align the lines to the opening
> parenthesis. That would give you more room. And it would make it slightly
> easier to read.
>
> Otherwise,
>
> Reviewed-by: Chen-Yu Tsai <wenst@xxxxxxxxxxxx>
>
> [...]