Re: [PATCH v2 1/2] gpio: mlxbf2: Introduce IRQ support

[Andrew Lunn]

On Wed, Sep 22, 2021 at 02:16:40PM +0000, Asmaa Mnebhi wrote:
> 
> > +static int
> > +mlxbf2_gpio_irq_set_type(struct irq_data *irqd, unsigned int type) {
> > +
> > +	switch (type & IRQ_TYPE_SENSE_MASK) {
> > +	case IRQ_TYPE_EDGE_BOTH:
> > +		fall = true;
> > +		rise = true;
> > +		break;
> > +	case IRQ_TYPE_EDGE_RISING:
> > +		rise = true;
> > +		break;
> > +	case IRQ_TYPE_EDGE_FALLING:
> > +		fall = true;
> > +		break;
> > +	default:
> > +		return -EINVAL;
> > +	}
> 
> > What PHY are you using? I think every one i've looked at are 
> > level triggered, not edge. Using an edge interrupt might work 99% 
> > of the time, but when the timing is just wrong, you can loose an interrupt.
> > Which might mean phylib thinks the link is down, when it fact it is up. 
> > You will need to unplug and replug to recover from that.
> 
> It is the micrel PHY KSZ9031 so it is an active low level interrupt.
> Here, IRQ_TYPE_EDGE* macros are mainly used to decide whether to write the
> YU_GPIO_CAUSE_FALL_EN register vs the YU_GPIO_CAUSE_RISE_EN register.
> These 2 registers are used in both LEVEL/EDGE interrupts.

I assume you also have an YU_GPIO_CAUSE_LOW_EN and
YU_GPIO_CAUSE_HIGH_EN registers? These registers need to be set for
IRQ_TYPE_LEVEL_LOW and IRQ_TYPE_LEVEL_HIGH.

	Andrew