Re: [i2c] [PATCH 5/7] of/gpio: implement of_dev_gpiochip_{add, remove} calls

2008-10-17 Thread David Brownell 
On Thursday 16 October 2008, Anton Vorontsov wrote:
> +   if (of_gc->chip)
> +   return of_gc->chip;
> +   return &of_gc->gc;

presumably there's a reason not to

of_gc->chip = &of_gc->gc;

when this gets set up, so this can always be a simple

return of_gc->chip;

(and inlined)?  Needlessly complex...



___
i2c mailing list
i2c@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/i2c

Re: [i2c] [PATCH 5/7] of/gpio: implement of_dev_gpiochip_{add, remove} calls

2008-10-17 Thread Anton Vorontsov 
On Fri, Oct 17, 2008 at 01:25:01PM -0700, David Brownell wrote:
> On Thursday 16 October 2008, Anton Vorontsov wrote:
> > +   if (of_gc->chip)
> > +   return of_gc->chip;
> > +   return &of_gc->gc;
> 
> presumably there's a reason not to
> 
>   of_gc->chip = &of_gc->gc;

Yes, there are two reasons:

1. I need some place to insert the huge comment, which you snipped. ;-)

2. I didn't want to break current users. That is, the line you purpose
   ("of_gc->chip = &of_gc->gc;") should be inserted to every of_gpio
   controller. Most of them use of_mm_gpiochip_add(), so these are easy.
   But some (mcu_mpc8349emitx.c) are doing things by themselves.

Thanks,

-- 
Anton Vorontsov
email: [EMAIL PROTECTED]
irc://irc.freenode.net/bd2

___
i2c mailing list
i2c@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/i2c