linux-kernel - Re: [RFC PATCH v2 1/4] dt-bindings: clock: renesas: Document RZ/V2H(P) SoC CPG

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAMuHMdXZV0V7St=bXu8Htq40a4qhjMTAV5yP3KYSi6teaQ7ZRQ@mail.gmail.com>
Date: Wed, 26 Jun 2024 13:45:22 +0200
From: Geert Uytterhoeven <geert@...ux-m68k.org>
To: Krzysztof Kozlowski <krzk@...nel.org>
Cc: "Lad, Prabhakar" <prabhakar.csengg@...il.com>, Michael Turquette <mturquette@...libre.com>, 
	Stephen Boyd <sboyd@...nel.org>, Rob Herring <robh@...nel.org>, 
	Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>, 
	Magnus Damm <magnus.damm@...il.com>, linux-renesas-soc@...r.kernel.org, 
	linux-clk@...r.kernel.org, devicetree@...r.kernel.org, 
	linux-kernel@...r.kernel.org, 
	Fabrizio Castro <fabrizio.castro.jz@...esas.com>, 
	Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>
Subject: Re: [RFC PATCH v2 1/4] dt-bindings: clock: renesas: Document
 RZ/V2H(P) SoC CPG

Hi Krzysztof,

On Wed, Jun 26, 2024 at 11:41 AM Krzysztof Kozlowski <krzk@...nel.org> wrote:
> On 26/06/2024 11:35, Geert Uytterhoeven wrote:
> > On Thu, Jun 13, 2024 at 2:57 PM Krzysztof Kozlowski <krzk@...nel.org> wrote:
> >> On 13/06/2024 11:53, Lad, Prabhakar wrote:
> >>> On Tue, Jun 11, 2024 at 8:02 AM Krzysztof Kozlowski <krzk@...nel.org> wrote:
> >>>> On 11/06/2024 01:32, Prabhakar wrote:
> >>>>> From: Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>
> >>>>>
> >>>>> Document the device tree bindings for the Renesas RZ/V2H(P) SoC
> >>>>> Clock Pulse Generator (CPG).
> >>>>>
> >>>>> CPG block handles the below operations:
> >>>>> - Generation and control of clock signals for the IP modules
> >>>>> - Generation and control of resets
> >>>>> - Control over booting
> >>>>> - Low power consumption and power supply domains
> >>>>>
> >>>>> Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@...renesas.com>
> >
> >>>>> +  '#clock-cells':
> >>>>> +    description: |
> >>>>> +      - For CPG core clocks, the two clock specifier cells must be "CPG_CORE"
> >>>>> +        and a core clock reference, as defined in
> >>>>> +        <dt-bindings/clock/r9a09g057-cpg.h>,
> >>>>
> >>>> So second cell is not used?
> >>>>
> >>> It will be used for blocks using core clocks.
> >>>
> >>>>> +      - For module clocks, the two clock specifier cells must be "CPG_MOD" and
> >>>>> +        a module number.  The module number is calculated as the CLKON register
> >>>>> +        offset index multiplied by 16, plus the actual bit in the register
> >>>>> +        used to turn the CLK ON. For example, for CGC_GIC_0_GICCLK, the
> >>>>> +        calculation is (1 * 16 + 3) = 19.
> >>>>
> >>>> You should not have different values. Make it const: 1 and just use IDs.
> >>>>
> >>> Are you suggesting not to differentiate between core/mod clocks. They
> >>> are differentiated because the MOD clocks can turned ON/OFF but where
> >>> as with the core clocks we cannot turn them ON/OF so the driver needs
> >>> to know this, hence two specifiers are used.
> >>
> >> Every driver knows it... I am really, what is the problem here? Are you
> >> saying the drivers create some unknown clocks?
> >
> > The driver knows for sure which clocks are module clocks, and thus can
> > be used for power management.  To simplify the driver, two separate
> > numbers spaces are used:
> >   1. Core clock numbers come from IDs in the DT binding headers,
> >   2. Module clock numbers come straight[1] from the hardware docs.
> > As the latter are fixed, merging them into a single number space in
> > a future-proof way is hard[2], the bindings use 2 clock cells.
>
> IIUC, your module clock numbers are not DT ABI and should not be put
> into the binding headers. I think that's the case currently, right?

Exactly: they are hardware ABI, just like e.g. GIC interrupt numbers.

> If above is correct, considering your explanation I am fine. Thanks for
> the time to make it clear.

Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds