Re: [PATCH v1] gpio: dwapb: mask/unmask IRQ when disable/enable it

[Linus Walleij <linus.walleij@xxxxxxxxxx>]

Sorry for top posting but I need the help of the irqchip maintainer
Marc Z to hash this out.

The mask/unmask/disable/enable semantics is something that
you need to work with every day to understand right.

Yours,
Linus Walleij

On Mon, Nov 30, 2020 at 10:36 AM Luo Jiaxing <luojiaxing@xxxxxxxxxx> wrote:
>
> The mask and unmask registers are not configured in dwapb_irq_enable() and
> dwapb_irq_disable(). In the following situations, the IRQ will be masked by
> default after the IRQ is enabled:
>
> mask IRQ -> disable IRQ -> enable IRQ
>
> In this case, the IRQ status of GPIO controller is inconsistent with it's
> irq_data too. For example, in __irq_enable(), IRQD_IRQ_DISABLED and
> IRQD_IRQ_MASKED are both clear, but GPIO controller do not perform unmask.
>
> Signed-off-by: Luo Jiaxing <luojiaxing@xxxxxxxxxx>
> ---
>  drivers/gpio/gpio-dwapb.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/drivers/gpio/gpio-dwapb.c b/drivers/gpio/gpio-dwapb.c
> index 2a9046c..ca654eb 100644
> --- a/drivers/gpio/gpio-dwapb.c
> +++ b/drivers/gpio/gpio-dwapb.c
> @@ -270,6 +270,8 @@ static void dwapb_irq_enable(struct irq_data *d)
>         u32 val;
>
>         spin_lock_irqsave(&gc->bgpio_lock, flags);
> +       val = dwapb_read(gpio, GPIO_INTMASK) & ~BIT(irqd_to_hwirq(d));
> +       dwapb_write(gpio, GPIO_INTMASK, val);
>         val = dwapb_read(gpio, GPIO_INTEN);
>         val |= BIT(irqd_to_hwirq(d));
>         dwapb_write(gpio, GPIO_INTEN, val);
> @@ -284,6 +286,8 @@ static void dwapb_irq_disable(struct irq_data *d)
>         u32 val;
>
>         spin_lock_irqsave(&gc->bgpio_lock, flags);
> +       val = dwapb_read(gpio, GPIO_INTMASK) | BIT(irqd_to_hwirq(d));
> +       dwapb_write(gpio, GPIO_INTMASK, val);
>         val = dwapb_read(gpio, GPIO_INTEN);
>         val &= ~BIT(irqd_to_hwirq(d));
>         dwapb_write(gpio, GPIO_INTEN, val);
> --
> 2.7.4
>