On 07/26/2012 02:35 PM, Mark Brown wrote:

On Thu, Jul 26, 2012 at 01:40:30PM -0600, Stephen Warren wrote:

quoted

+		if (irqd_is_wakeup_set(d)) {
+			/* 1 -- disable, 0 -- enable */
+			switch (irq_data->mask_reg) {

This loop we should just port over into the regmap code.

I assume the best way of doing this is to add new functions
regmap_irq_suspend()/regmap_irq_resume() (which would mask any enabled
interrupts that were not wake enabled); that way, the regmap_irq code
can loop over each register and just write it once. An alternative might
be to implement struct irq_chip's .irq_suspend/.irq_resume ops, but that
might worst-case end up with an I2C write per interrupt.

I see that the MAX8907 IRQ code does this in suspend:

	if (device_may_wakeup(chip->dev))
		enable_irq_wake(i2c->irq);
	else
		disable_irq(i2c->irq);

and this in resume:

	if (device_may_wakeup(chip->dev))
		disable_irq_wake(i2c->irq);
	else
		enable_irq(i2c->irq);

neither of which are done in regmap_irq, since it doesn't explicitly do
anything for suspend/resume at the moment. Are those code blocks
necessary? I see that regmap_irq_sync_unlock() is already calling
irq_set_irq_wake(), which implies that suspend/resume may have already
been completely taken care of?