Re: [PATCH 1/4] gpio: gpiolib: Allow GPIO IRQs to lazy disable

[Maulik Shah]

Hi,

On 5/23/2020 3:12 PM, Marc Zyngier wrote:
> On 2020-05-22 14:19, Maulik Shah wrote:
> > With 'commit 461c1a7d4733 ("gpiolib: override irq_enable/disable")' 
> > gpiolib
> > overrides irqchip's irq_enable and irq_disable callbacks. If irq_disable
> > callback is implemented then genirq takes unlazy path to disable irq.
> >
> > Underlying irqchip may not want to implement irq_disable callback to 
> > lazy
> > disable irq when client drivers invokes disable_irq(). By overriding
> > irq_disable callback, gpiolib ends up always unlazy disabling IRQ.
> >
> > Allow gpiolib to lazy disable IRQs by overriding irq_disable callback 
> > only
> > if irqchip implemented irq_disable. In cases where irq_disable is not
> > implemented irq_mask is overridden. Similarly override irq_enable 
> > callback
> > only if irqchip implemented irq_enable otherwise irq_unmask is 
> > overridden.
> >
> > Fixes: 461c1a7d47 (gpiolib: override irq_enable/disable)
> > Signed-off-by: Maulik Shah <mkshah@xxxxxxxxxxxxxx>
> > ---
> > Âdrivers/gpio/gpiolib.cÂÂÂÂÂ | 59 
> > +++++++++++++++++++++++++++++----------------
> > Âinclude/linux/gpio/driver.h | 13 ++++++++++
> > Â2 files changed, 51 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> > index eaa0e20..a8fdc74 100644
> > --- a/drivers/gpio/gpiolib.c
> > +++ b/drivers/gpio/gpiolib.c
> > @@ -2465,33 +2465,38 @@ static void gpiochip_irq_relres(struct 
> > irq_data *d)
> > ÂÂÂÂ gpiochip_relres_irq(gc, d->hwirq);
> > Â}
> >
> > +static void gpiochip_irq_mask(struct irq_data *d)
> > +{
> > +ÂÂÂ struct gpio_chip *chip = irq_data_get_irq_chip_data(d);
> > +
> > +ÂÂÂ if (chip->irq.irq_mask)
> > +ÂÂÂÂÂÂÂ chip->irq.irq_mask(d);
> > +ÂÂÂ gpiochip_disable_irq(chip, d->hwirq);
> > +}
> > +
> > +static void gpiochip_irq_unmask(struct irq_data *d)
> > +{
> > +ÂÂÂ struct gpio_chip *chip = irq_data_get_irq_chip_data(d);
> > +
> > +ÂÂÂ gpiochip_enable_irq(chip, d->hwirq);
> > +ÂÂÂ if (chip->irq.irq_unmask)
> > +ÂÂÂÂÂÂÂ chip->irq.irq_unmask(d);
> > +}
> > +
> > Âstatic void gpiochip_irq_enable(struct irq_data *d)
> > Â{
> > ÂÂÂÂ struct gpio_chip *gc = irq_data_get_irq_chip_data(d);
> >
> > -ÂÂÂ gpiochip_enable_irq(gc, d->hwirq);
> > -ÂÂÂ if (gc->irq.irq_enable)
> > -ÂÂÂÂÂÂÂ gc->irq.irq_enable(d);
> > -ÂÂÂ else
> > -ÂÂÂÂÂÂÂ gc->irq.chip->irq_unmask(d);
> > +ÂÂÂ gpiochip_enable_irq(chip, d->hwirq);
>
> You really never compiled this, did you?
>
> I've stopped looking at this. Please send something that you will have 
> actually tested.
>
> ÂÂÂÂÂÂÂ M.

Apologies.

I did tested out on internal devices based on kernel 5.4 as well as 
linux-next with sc7180.

While posting i somehow taken patch from kernel 5.4 and messed up this 
patch during merge conflicts.

I fixed this in v2 and posted out changes that cleanly applies on latest 
linux-next tag (next-20200521).

Thanks,
Maulik

--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation