[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <161706966003.3012082.13602640109563561014@swboyd.mtv.corp.google.com>
Date: Mon, 29 Mar 2021 19:01:00 -0700
From: Stephen Boyd <sboyd@...nel.org>
To: Alain Volmat <avolmat@...com>,
Michael Turquette <mturquette@...libre.com>,
Patrice Chotard <patrice.chotard@...s.st.com>,
Rob Herring <robh+dt@...nel.org>
Cc: Lee Jones <lee.jones@...aro.org>, linux-clk@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, Alain Volmat <avolmat@...com>
Subject: Re: [PATCH v2 00/16] clk: st: embed clock outputs within drivers
Quoting Alain Volmat (2021-03-25 00:50:02)
> Most of ST clock drivers used by STi platform are updated in
> order to introduce clock outputs informations within each drivers
> and thus allow to avoid having to rely on clock-output-names properties
> within DT clock nodes.
> For that purpose, drivers are updated to allow handling both modes
> (with or without clock-output-names).
> Once all DT will have been updated, the legacy mode could be removed
> from the drivers.
> This will also allow, once all STi DT will be corrected, to remove the
> of_clk_detect_critical API from clk core code since STi clock drivers
> are the only drivers using this API.
>
> This serie also contains modifications within STi DTS in order to use
> the newly introduced compatible and remove clock-output-names
> properties.
Can you split the dts changes from the driver changes and only send the
clk driver changes to clk maintainers? I don't intend to apply dts file
changes to the clk tree. Those can go via arm-soc based on some
immutable branch that lives in clk tree until next merge window closes.
>
> Alain Volmat (16):
> clk: st: clkgen-pll: remove used variable of struct clkgen_pll
> clk: st: flexgen: embed soc clock outputs within compatible data
> dt-bindings: clock: st: flexgen: add new introduced compatible
> clk: st: clkgen-pll: embed soc clock outputs within compatible data
> dt-bindings: clock: st: clkgen-pll: add new introduced compatible
> clk: st: clkgen-fsyn: embed soc clock outputs within compatible data
> dt-bindings: clock: st: clkgen-fsyn: add new introduced compatible
> ARM: dts: sti: update flexgen compatible within stih418-clock
> ARM: dts: sti: update flexgen compatible within stih407-clock
> ARM: dts: sti: update flexgen compatible within stih410-clock
> ARM: dts: sti: update clkgen-pll entries in stih407-clock
> ARM: dts: sti: update clkgen-pll entries in stih410-clock
> ARM: dts: sti: update clkgen-pll entries in stih418-clock
> ARM: dts: sti: update clkgen-fsyn entries in stih407-clock
> ARM: dts: sti: update clkgen-fsyn entries in stih410-clock
> ARM: dts: sti: update clkgen-fsyn entries in stih418-clock
>
> .../bindings/clock/st/st,clkgen-pll.txt | 3 +
> .../bindings/clock/st/st,flexgen.txt | 10 +
> .../bindings/clock/st/st,quadfs.txt | 3 +
> arch/arm/boot/dts/stih407-clock.dtsi | 128 +------
> arch/arm/boot/dts/stih410-clock.dtsi | 138 +------
> arch/arm/boot/dts/stih418-clock.dtsi | 136 +------
> drivers/clk/st/clk-flexgen.c | 344 +++++++++++++++++-
> drivers/clk/st/clkgen-fsyn.c | 113 +++++-
> drivers/clk/st/clkgen-pll.c | 121 +++++-
> 9 files changed, 588 insertions(+), 408 deletions(-)
>
> ---
> v2: fix drivers to update some clocks as CLK_IS_CRITICAL
>
Please document in the code why CLK_IS_CRITICAL is used. For example,
"This clk needs to be on so the CPU can keep fetching instructions from
memory" or something like that.
Powered by blists - more mailing lists