Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.34 as permitted sender) client-ip=23.128.96.34;
MIME-Version: 1.0
References: <20231129063730.31184-1-liuhaoran14@163.com>
In-Reply-To: <20231129063730.31184-1-liuhaoran14@163.com>
From:   Geert Uytterhoeven <geert@linux-m68k.org>
Date:   Wed, 29 Nov 2023 09:10:41 +0100
Message-ID: <CAMuHMdVw1oT7L7G5wQUtYK0qAdQ1VayL-DNi=jtGdQ1KER9Ygg@mail.gmail.com>
Subject: Re: [PATCH] [sh/highlander] psw: Add error handling in psw_irq_handler
To:     Haoran Liu <liuhaoran14@163.com>
Cc:     ysato@users.sourceforge.jp, dalias@libc.org,
        glaubitz@physik.fu-berlin.de, linux-sh@vger.kernel.org,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

Hi Haoran,

On Wed, Nov 29, 2023 at 7:38 AM Haoran Liu <liuhaoran14@163.com> wrote:
> This patch adds error handling for the platform_get_drvdata call
> within the psw_irq_handler function in
> arch/sh/boards/mach-highlander/psw.c. Previously, the absence of
> error checking could lead to unexpected behavior if
> platform_get_drvdata returned a null pointer.
>
> Signed-off-by: Haoran Liu <liuhaoran14@163.com>

Thanks for your patch!

> --- a/arch/sh/boards/mach-highlander/psw.c
> +++ b/arch/sh/boards/mach-highlander/psw.c
> @@ -21,6 +21,12 @@ static irqreturn_t psw_irq_handler(int irq, void *arg)
>         unsigned int l, mask;
>         int ret = 0;
>
> +       if (!psw) {
> +               pr_err("psw_irq_handler: No push_switch data associated
> +                       with platform_device\n");
> +               return IRQ_NONE;
> +       }
> +
>         l = __raw_readw(PA_DBSW);
>
>         /* Nothing to do if there's no state change */

This means the button is pressed, and the interrupt fired, in
between the calls to request_irq() and platform_set_drvdata() in
arch/sh/drivers/push-switch.c:switch_drv_probe()?

The same issue can happen with arch/sh/boards/mach-landisk/psw.c.
I think the proper solution is to fix this inside the push switch
driver, by moving the call to request_irq() after the call to
platform_set_drvdata() (and doing the reverse in switch_drv_remove()).

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