Re: [PATCH] rtc: add support for maxim dallas rtc ds1343 and ds1344

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 7 May 2014 23:15:29 +0100 Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> On Wed, 7 May 2014 14:42:48 -0700
> Andrew Morton <akpm@linux-foundation.org> wrote:
> 
> > On Wed,  7 May 2014 00:33:56 +0530 Raghavendra Ganiga <ravi23ganiga@gmail.com> wrote:
> > 
> > > This is a patch to add support for
> > > maxim dallas rtc ds1343 and ds1344
> > > 
> > > ...
> > >
> > > +static int ds1343_ioctl(struct device *dev, unsigned int cmd, unsigned long arg)
> > > +{
> > > +	switch (cmd) {
> > > +#ifdef RTC_SET_CHARGE
> > > +	case RTC_SET_CHARGE:
> > > +	{
> > > +		int val;
> > > +
> > > +		if (copy_from_user(&val, (int __user *)arg, sizeof(int)))
> > > +			return -EFAULT;
> > > +
> > > +		return regmap_write(priv->map, DS1343_TRICKLE_REG, val);
> > > +	}
> > > +	break;
> > > +#endif
> > > +	}
> > 
> > Why is RTC_SET_CHARGE defined only for m32r?  It seems pretty generic?
> > 
> > > +	return -ENOIOCTLCMD;
> > > +}
> > 
> > It's possible that this should be -ENOTTY.  I always forget the rule,
> > but Alan never forgets anything ;)
> 
> I forget lots 8)
> 
> -ENOTTY if it will be returned to user space, -ENOIOCTLCMD is used by
> some core code internally as a way for the sub drivers to tell the core
> "I didn't handle this, maybe you have a default method ?"

Ah OK, thanks.

drivers/rtc/rtc-dev.c does

	default:
		/* Finally try the driver's ioctl interface */
		if (ops->ioctl) {
			err = ops->ioctl(rtc->dev.parent, cmd, arg);
			if (err == -ENOIOCTLCMD)
				err = -ENOTTY;
		} else
			err = -ENOTTY;

so ENOIOCTLCMD is appropriate here.

Even though rtc is upside down - it prioritises the core's ioctl
parsing ahead of the drivers'.