Re: [PATCH v3 2/2] pinctrl: renesas: rzg2l: Configure the interrupt type on resume

[claudiu beznea <claudiu.beznea@xxxxxxxxx>]

On 18.04.2024 17:07, Geert Uytterhoeven wrote:
> Hi Claudiu,
> 
> On Wed, Mar 20, 2024 at 11:43 AM Claudiu <claudiu.beznea@xxxxxxxxx> wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@xxxxxxxxxxxxxx>
>>
>> Commit dce0919c83c3 ("irqchip/renesas-rzg2l: Do not set TIEN and TINT
>> source at the same time") removed the setup of TINT from
>> rzg2l_irqc_irq_enable(). To address the spourious interrupt issue the setup
>> of TINT has been moved in rzg2l_tint_set_edge() though
>> rzg2l_disable_tint_and_set_tint_source(). With this, the interrupts are
>> not properly re-configured after a suspend-to-RAM cycle. To address
>> this issue and avoid spurious interrupts while resumming set the
>> interrupt type before enabling it.
>>
>> Fixes: dce0919c83c3 ("irqchip/renesas-rzg2l: Do not set TIEN and TINT source at the same time")
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@xxxxxxxxxxxxxx>
> 
> Thanks for your patch!
> 
>> --- a/drivers/pinctrl/renesas/pinctrl-rzg2l.c
>> +++ b/drivers/pinctrl/renesas/pinctrl-rzg2l.c
>> @@ -2045,7 +2045,9 @@ static void rzg2l_gpio_irq_restore(struct rzg2l_pinctrl *pctrl)
>>
>>         for (unsigned int i = 0; i < RZG2L_TINT_MAX_INTERRUPT; i++) {
>>                 struct irq_data *data;
>> +               unsigned long flags;
>>                 unsigned int virq;
>> +               int ret;
>>
>>                 if (!pctrl->hwirq[i])
>>                         continue;
>> @@ -2063,17 +2065,17 @@ static void rzg2l_gpio_irq_restore(struct rzg2l_pinctrl *pctrl)
>>                         continue;
>>                 }
>>
>> -               if (!irqd_irq_disabled(data)) {
>> -                       unsigned long flags;
>> -
>> -                       /*
>> -                        * This has to be atomically executed to protect against a concurrent
>> -                        * interrupt.
>> -                        */
>> -                       raw_spin_lock_irqsave(&pctrl->lock.rlock, flags);
>> +               /*
>> +                * This has to be atomically executed to protect against a concurrent
>> +                * interrupt.
>> +                */
>> +               raw_spin_lock_irqsave(&pctrl->lock.rlock, flags);
>> +               ret = rzg2l_gpio_irq_set_type(data, irqd_get_trigger_type(data));
>> +               if (ret)
>> +                       dev_crit(pctrl->dev, "Failed to set IRQ type for virq=%u\n", virq);
>> +               else if (!irqd_irq_disabled(data))
>>                         rzg2l_gpio_irq_enable(data);
>> -                       raw_spin_unlock_irqrestore(&pctrl->lock.rlock, flags);
>> -               }
>> +               raw_spin_unlock_irqrestore(&pctrl->lock.rlock, flags);
>>         }
>>  }
> 
> LGTM, but I'd rather move the dev_crit() outside (i.e. after) the
> critical section.

I was in balance about having it as proposed. I'll send an update to move
it outside.

Thank you,
Claudiu Beznea

> 
> Gr{oetje,eeting}s,
> 
>                         Geert
>