Thomas,

On 09/20/2016 04:02 PM, Thomas Gleixner wrote:

On Tue, 20 Sep 2016, Alexandre Torgue wrote:

quoted

On 09/20/2016 02:44 PM, Thomas Gleixner wrote:

quoted

Free will be called when a interrupt in the child domain is torn down,
i.e. when irq_domain_free_irqs() is called. And it will be called for both
domains like the alloc callback is invoked on both domains via
irq_domain_alloc_irqs().

Thanks Thomas for this clarification (I'm sure now that we need .free
callbacks).
irq_domain_free_irqs() is called in 2 scenario:
1- when issue occurs in irq_create_fwspec_mapping()
2- when irq_dispose_mapping() is called

Case 2 is the one I tested some times ago. In this case, I need to mask
interrupts in .free callback of EXTI (parent) domain to avoid spurious
interrupts.

And why would irq_dispose_mapping() be called on an unmasked, i.e. active,
interrupt? The masking is just papering over that.

Ok. So my test was wrong and irq_dispose_mapping() has to be called when 
irq is masked (for example just  after free_irq()). For sure in this 
case the mask inside exti free callback has no sens (catch :))

I will change .free callback by:

static void stm32_exti_free(struct irq_domain *d, unsigned int virq,
			    unsigned int nr_irqs)
{
	struct irq_data *data = irq_domain_get_irq_data(d, virq);
	irq_domain_reset_irq_data(data);
}


so if you agree I will resend only patches concerning stm32 exti driver 
[1],[2],[3],[4]

Thanks for your time.

alex

Thanks,

	tglx