Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
MIME-Version: 1.0
References: <20231026063941.1882023-1-peng.fan@oss.nxp.com>
In-Reply-To: <20231026063941.1882023-1-peng.fan@oss.nxp.com>
From:   Geert Uytterhoeven <geert@linux-m68k.org>
Date:   Thu, 26 Oct 2023 08:53:03 +0200
Message-ID: <CAMuHMdUA_OOvQmkZXn21pxC=nTw6wHQZp1P8mNjTca65+KjxAQ@mail.gmail.com>
Subject: Re: [PATCH] clk: clk-conf: bypass setting rate/parent if already same
To:     "Peng Fan (OSS)" <peng.fan@oss.nxp.com>
Cc:     sboyd@kernel.org, abelvesa@kernel.org, mturquette@baylibre.com,
        clement.leger@bootlin.com, sudeep.holla@arm.com,
        cristian.marussi@arm.com, kernel@pengutronix.de,
        pure.logic@nexus-software.ie, a.fatoum@pengutronix.de,
        aford173@gmail.com, absahu@codeaurora.org,
        linux-clk@vger.kernel.org, linux-kernel@vger.kernel.org,
        Peng Fan <peng.fan@nxp.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

Hi Peng,

On Thu, Oct 26, 2023 at 8:35 AM Peng Fan (OSS) <peng.fan@oss.nxp.com> wrote:
> From: Peng Fan <peng.fan@nxp.com>
>
> If the original rate and parent is already the same as what users
> wanna to configure through assigned clock rate and parent, there
> is no need to configure them again which may cause more cpu cycles
> or more SCMI RPC calls.
>
> So check the rate and parent first, and bypass when the original
> rate and parent are same as requested by device tree node.
>
> Signed-off-by: Peng Fan <peng.fan@nxp.com>

Thanks for your patch!

> --- a/drivers/clk/clk-conf.c
> +++ b/drivers/clk/clk-conf.c
> @@ -65,7 +65,11 @@ static int __set_clk_parents(struct device_node *node, bool clk_supplier)
>                         goto err;
>                 }
>
> -               rc = clk_set_parent(clk, pclk);
> +               if (__clk_get_hw(pclk) != __clk_get_hw(clk_get_parent(clk)))
> +                       rc = clk_set_parent(clk, pclk);
> +               else
> +                       rc = 0;

The else branch is not needed, as rc already indicates a non-error
condition.

> +
>                 if (rc < 0)
>                         pr_err("clk: failed to reparent %s to %s: %d\n",
>                                __clk_get_name(clk), __clk_get_name(pclk), rc);
> @@ -112,7 +116,10 @@ static int __set_clk_rates(struct device_node *node, bool clk_supplier)
>                                 return PTR_ERR(clk);
>                         }
>
> -                       rc = clk_set_rate(clk, rate);
> +                       if (clk_get_rate(clk) != rate)
> +                               rc = clk_set_rate(clk, rate);
> +                       else
> +                               rc = 0;

Likewise.

>                         if (rc < 0)
>                                 pr_err("clk: couldn't set %s clk rate to %u (%d), current rate: %lu\n",
>                                        __clk_get_name(clk), rate, rc,

However, one can wonder whether this check should be done in
clk_set_parent() resp. clk_set_rate() instead, so it would benefit
all callers?

Also:

    /**
     * clk_get_rate - obtain the current clock rate (in Hz) for a clock source.
     *                This is only valid once the clock source has been enabled.
     * @clk: clock source
     */

So at least the second part is not guaranteed to work?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-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