Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <KL1PR01MB544800D7E51C9209A9BD998BE6669@KL1PR01MB5448.apcprd01.prod.exchangelabs.com>
In-Reply-To: <KL1PR01MB544800D7E51C9209A9BD998BE6669@KL1PR01MB5448.apcprd01.prod.exchangelabs.com>
From:   Linus Walleij <linus.walleij@linaro.org>
Date:   Mon, 24 Apr 2023 08:56:52 +0200
Message-ID: <CACRpkdZ8CVWb6=-sgK3BdW-ssivY1oz4cc_mYoXmyrBMLxsX=A@mail.gmail.com>
Subject: Re: [PATCH v1] gpio: mmio: restroe get multiple gpio mask
To:     Yan Wang <rk.code@outlook.com>
Cc:     brgl@bgdev.pl, Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Yan,

thanks for your patch!

On Sun, Apr 23, 2023 at 11:07=E2=80=AFAM Yan Wang <rk.code@outlook.com> wro=
te:

> Simplify the code,should not modify its logic.

I don't see how it simplifies the code, it seems to me that it is
making it more complex?

> Fixes: 761b5c30c206 ("gpio: mmio: replace open-coded for_each_set_bit()")
> Signed-off-by: Yan Wang <rk.code@outlook.com>
> ---
>  drivers/gpio/gpio-mmio.c | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpio/gpio-mmio.c b/drivers/gpio/gpio-mmio.c
> index d9dff3dc92ae..c2347ef3a4df 100644
> --- a/drivers/gpio/gpio-mmio.c
> +++ b/drivers/gpio/gpio-mmio.c
> @@ -271,10 +271,14 @@ static void bgpio_multiple_get_masks(struct gpio_ch=
ip *gc,
>         *clear_mask =3D 0;
>
>         for_each_set_bit(i, mask, gc->bgpio_bits) {
> -               if (test_bit(i, bits))
> -                       *set_mask |=3D bgpio_line2mask(gc, i);
> -               else
> -                       *clear_mask |=3D bgpio_line2mask(gc, i);
> +               if (*mask =3D=3D 0)
> +                       break;
> +               if (__test_and_clear_bit(i, mask)) {
> +                       if (test_bit(i, bits))
> +                               *set_mask |=3D bgpio_line2mask(gc, i);
> +                       else
> +                               *clear_mask |=3D bgpio_line2mask(gc, i);
> +               }

The intention of the change seems to be to break out of the loop
when all set bits are handled, by successively masking off one
bit at a time from mask. So this is intended as an optimization,
not a simplification.

I think for_each_set_bit() is already skipping over every bit
that is zero, see include/linux/find.h.

So this optimization gains us nothing.

Yours,
Linus Walleij