Re: [patch 09/13] irqchip: spear_shirq: Kill the clear_reg nonsense

[Thomas Gleixner <tglx@linutronix.de>]

On Fri, 20 Jun 2014, Viresh Kumar wrote:

> On Fri, Jun 20, 2014 at 3:04 AM, Thomas Gleixner <tglx@linutronix.de> wrote:
> > None of the chips has a ACK register.
> 
> I need to recheck on this after looking at datasheets. Arranging for
> them, will revert by tomorrow.
> 
> > The code brainlessly fiddles
> > with the enable register, so it might even reenable a disabled
> > interrupt at least on spear300.
> 
> Ack/Clear register is only configured for SPEAr320, how will it
> make a difference to SPEAr300 ?

Sorry, my bad. misread the code. So this wants a different
changelog.

> And for SPEAr320 as well, the offset mentioned in code for clear
> register is different then ENABLE register.

I still don't see why you'd write something into the status register
on 320, which is RO according to documentation.

> > @@ -150,13 +141,6 @@ static struct spear_shirq spear320_shirq
> >         .nr_irqs        = 7,
> >         .mask           = ((0x1 << 7) - 1) << 0,
> >         .disabled       = 1,
> > -       .regs = {
> > -               .enb_reg = SPEAR320_INT_ENB_MASK_REG,
> > -               .reset_to_enb = 1,
> > -               .status_reg = SPEAR320_INT_STS_MASK_REG,
> > -               .clear_reg = SPEAR320_INT_CLR_MASK_REG,
> > -               .reset_to_clear = 1,
> > -       },
> 
> Was removing .regs completely intentional?
> 
> I don't see these registers getting added again in later patches.

Yes, because that block is NEVER used because disabled = 1

Thanks,

	tglx