does twl3040-pwrirq.c "need" to be a separate file?

does twl3040-pwrirq.c "need" to be a separate file?

[David Brownell]

Hi Peter,

I see your patch 68d7477caca19c0b52b5d4e85700cd3e6115577f created
pwrirq.c as a separate file and thread.

I'm wondering if there's any particular reason that "bank" of
interrupts shouldn't be handled directly by twl4030-core, and
even by the same IRQ handling thread.

As it stands now the TWL "core" is not especially core-ish in
this respect, and I'd like to see that be resolved (e.g. by a
patch I'll probably write this afternoon) before this code
goes to mainline ...

- Dave

Re: does twl3040-pwrirq.c "need" to be a separate file?

[Peter 'p2' De Schrijver]

Hi David,

On Mon, Sep 29, 2008 at 02:17:12PM -0700, ext David Brownell wrote:
> Hi Peter,
> 
> I see your patch 68d7477caca19c0b52b5d4e85700cd3e6115577f created
> pwrirq.c as a separate file and thread.
> 

I guess choose this solution because it was similar to the GPIO IRQs.
Originally, this was 1 shared IRQ. But I wanted to change this to avoid
every driver having to read PWR_ISR1 and clear his interrupt. This saves
some i2c transactions.

> I'm wondering if there's any particular reason that "bank" of
> interrupts shouldn't be handled directly by twl4030-core, and
> even by the same IRQ handling thread.
> 

I don't think so.

> As it stands now the TWL "core" is not especially core-ish in
> this respect, and I'd like to see that be resolved (e.g. by a
> patch I'll probably write this afternoon) before this code
> goes to mainline ...

Ok. Good.

Cheers,

Peter.

-- 
goa is a state of mind

Re: does twl3040-pwrirq.c "need" to be a separate file?

[David Brownell]

Hi Peter,

On Tuesday 30 September 2008, Peter 'p2' De Schrijver wrote:

> > I see your patch 68d7477caca19c0b52b5d4e85700cd3e6115577f created
> > pwrirq.c as a separate file and thread.
> 
> I guess choose this solution because it was similar to the GPIO IRQs.
> Originally, this was 1 shared IRQ. But I wanted to change this to avoid
> every driver having to read PWR_ISR1 and clear his interrupt. This saves
> some i2c transactions.

Right; modularization is appropriate.  Although it doesn't
seem to have hit all the TWL "subchips" yet ... :)


> > I'm wondering if there's any particular reason that "bank" of
> > interrupts shouldn't be handled directly by twl4030-core, and
> > even by the same IRQ handling thread.
> > 
> 
> I don't think so.
> 
> > As it stands now the TWL "core" is not especially core-ish in
> > this respect, and I'd like to see that be resolved (e.g. by a
> > patch I'll probably write this afternoon) before this code
> > goes to mainline ...
> 
> Ok. Good.

Thanks for the sanity check.

- Dave