Re: [PATCH v3] irqchip: Add support for ARMv7-M's NVIC

[Arnd Bergmann <arnd@arndb.de>]

On Friday 19 April 2013, Uwe Kleine-König wrote:
> On Thu, Apr 18, 2013 at 11:38:13AM +0200, Arnd Bergmann wrote:
> > On Thursday 18 April 2013, Uwe Kleine-König wrote:
> > > That is, there are (INTLINESNUM + 1) * 32 irqs for INTLINESNUM < 15. For
> > > INTLINESNUM == 15 there are only 496 and not 16 * 32 == 512. That's the
> > > same on the gic (just with bigger numbers).
> > 
> > Ok, but since you are now using a linear domain, it doesn't actually hurt
> > to register 512 in that special case, right?
> Well, it depends if allocating space for 16 unused unsigned ints hurts
> (maybe not).

As long as SPARSE_IRQ is enabled, the space won't actually be allocated.

> And it makes mapping some irqs successfull while the irq
> doesn't really exist. But probably this doesn't hurt either because the
> problem already exists.
> 
> I don't care much. Is there another advantage beside saving a few source
> lines/instructions?

It just feels strange to read the value from hardware and then override
it anyway. But I agree it's not important either way.

	Arnd