Re: [PATCH v1 2/6] clk: starfive: jh71x0: Move reset controller registration to jh71x0 common code

From: Jerome Brunet

Date: Mon Nov 16 2026 - 06:01:18 EST


On lun. 16 nov. 2026 at 08:01, Changhuang Liang <changhuang.liang@xxxxxxxxxxxxxxxx> wrote:

> Hi, Jerome
>
>> > jh7110_reset_controller_register() is only tied to the JH7110 system
>> > clock driver by its location. Its logic is generic: hand the CRG
>> > register base to the reset driver through an auxiliary device.
>> >
>> > Move it to clk-starfive-jh71x0.c as jh71x0_reset_controller_register()
>> > so that the clock drivers of other StarFive SoCs can reuse it, and
>> > make
>> > CLK_STARFIVE_JH71X0 select AUXILIARY_BUS accordingly.
>> >
>> > While at it, replace the open-coded auxiliary device allocation, init,
>> > add and release handling with devm_auxiliary_device_create(), which
>> > already implements all of it. The register base is now passed to the
>> > reset driver as platform data of the auxiliary device, so struct
>> > jh71x0_reset_adev and the embedded auxiliary device go away.
>> >
>> > The auxiliary device ID is dropped: every CRG already uses a distinct
>> > auxiliary device name, so the ID is not needed to keep the device
>> > names unique, and it is always 0 now.
>> >
>> > Since the helper now lives in clk-starfive-jh71x0.c, the auxiliary
>> > device names are prefixed with "clk_starfive_jh71x0" instead of
>> > "clk_starfive_jh7110_sys". Update the auxiliary_device_id table of the
>> > JH7110 reset driver in the same patch to keep the tree bisectable.
>> >
>> > Signed-off-by: Changhuang Liang <changhuang.liang@xxxxxxxxxxxxxxxx>
>> > ---
>> > drivers/clk/starfive/Kconfig | 2 +-
>> > .../clk/starfive/clk-starfive-jh7110-aon.c | 2 +-
>> > .../clk/starfive/clk-starfive-jh7110-isp.c | 2 +-
>> > .../clk/starfive/clk-starfive-jh7110-stg.c | 2 +-
>> > .../clk/starfive/clk-starfive-jh7110-sys.c | 59 +------------------
>> > .../clk/starfive/clk-starfive-jh7110-vout.c | 2 +-
>> > drivers/clk/starfive/clk-starfive-jh7110.h | 4 --
>> > drivers/clk/starfive/clk-starfive-jh71x0.c | 24 ++++++++
>> > drivers/clk/starfive/clk-starfive-jh71x0.h | 2 +
>> > .../reset/starfive/reset-starfive-jh7110.c | 19 +++---
>> > include/soc/starfive/reset-starfive-jh71x0.h | 10 ++--
>> > 11 files changed, 45 insertions(+), 83 deletions(-)
>> >
>> > diff --git a/drivers/clk/starfive/Kconfig
>> > b/drivers/clk/starfive/Kconfig index bd29358ffeec..51716655554f 100644
>> > --- a/drivers/clk/starfive/Kconfig
>> > +++ b/drivers/clk/starfive/Kconfig
>> > @@ -2,6 +2,7 @@
>> >
>> > config CLK_STARFIVE_JH71X0
>> > bool
>> > + select AUXILIARY_BUS
>> >
>> > config CLK_STARFIVE_JH7100
>> > bool "StarFive JH7100 clock support"
>> > @@ -32,7 +33,6 @@ config CLK_STARFIVE_JH7110_PLL config
>> > CLK_STARFIVE_JH7110_SYS
>> > bool "StarFive JH7110 system clock support"
>> > depends on ARCH_STARFIVE || COMPILE_TEST
>> > - select AUXILIARY_BUS
>> > select CLK_STARFIVE_JH71X0
>> > select RESET_STARFIVE_JH7110 if RESET_CONTROLLER
>>
>> You've switched to imply for the new controller you are introducing later in
>> this series.
>> You should probably be consistent and align this one too. Not in this change
>> though, it would be unrelated.
>
> I'm not going to put this suggestion into the current series for now. I plan to send it
> as a separate patch later. Do you think that's okay?

Ok

>
>>
>> This change otherwise looks good to me. It requires Philipp's opinion as it is
>> touching reset too.
>>
>> --
>> Jerome
>
> Best Regards,
> Changhuang
>

--
Jerome