Mark Brown wrote at Saturday, December 03, 2011 4:17 AM:
> On Fri, Dec 02, 2011 at 03:08:40PM -0700, Stephen Warren wrote:
>
> > + switch (irqd_get_trigger_type(irq_data)) {
> > + case IRQ_TYPE_NONE:
> > + /*
> > + * We assume the controller imposes no restrictions,
> > + * so we are able to select active-high
> > + */
> > + /* Fall-through */
> > + case IRQ_TYPE_LEVEL_HIGH:
> > + pdata->irq_active_low = false;
> > + break;
> > + case IRQ_TYPE_LEVEL_LOW:
> > + pdata->irq_active_low = true;
> > + break;
> > + default:
> > + dev_err(&i2c->dev,
> > + "Unsupported IRQ_TYPE %x\n",
> > + irqd_get_trigger_type(irq_data));
> > + return -EINVAL;
>
> Actually, it occurs to me that we should treat the default case in the
> same way as IRQ_TYPE_NONE - even if the interrupt controller defaults to
> an edge triggered mode which we can't use it might also support a level
> triggered interrupt we can use. If it doesn't then we'll just fail
> later on when we request the IRQ so the end result should be the same.
Yes, deferring the failure to later seems reasonable; it's probably better
to work somehow if the HW supports it rather than being anal about e.g.
incorrect device tree content.
> Don't worry about respinning for this unless the series needs to get
> resent for some other reason let's just fix it incrementally (I'll do
> that myself when I apply unless you want to).
I'm fine if you change that when applying it.
Thanks.
--
nvpublic
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
