Re: [PATCH v4 2/6] PCI: spacemit-k1: Add multiple PHY handles support

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Fri, Jul 10, 2026 at 07:51:05AM -0500, Alex Elder wrote:
> On 7/9/26 2:16 AM, Andy Shevchenko wrote:

...

> > > +		if (IS_ERR(k1->phy[i])) {
> > > +			if (PTR_ERR(k1->phy[i]) == -ENODEV)
> > > +				break;
> > > +
> > > +			return PTR_ERR(k1->phy[i]);
> > > +		}
> > 		if (PTR_ERR(k1->phy[i]) == -ENODEV)
> > 			break;
> > 		if (IS_ERR(k1->phy[i]))
> > 			return PTR_ERR(k1->phy[i]);
> 
> I'm not sure this is a huge improvement.  Checking IS_ERR()
> before using PTR_ERR() is comforting (a little along the
> lines of the other issue you mentioned--assigning a result
> before checking for an error).  Anyway, it's a little bit
> of an ugly construct no matter how you do it.

Yeah...

> Here's another possible way to do it.
> 
> 	k1->phy[i] = devm_of_phy_get_by_index(dev, node, i);
> 	if (!IS_ERR(k1->phy[i]))
> 		continue;
> 
> 	ret = PTR_ERR(k1->phy[i]);
> 	if (ret == -ENODEV)
> 		break;

In this case one can simplify to

	k1->phy[i] = devm_of_phy_get_by_index(dev, node, i);
	ret = PTR_ERR_OR_ZERO(k1->phy[i]);
	if (ret == -ENODEV)
		break;
	else if (ret)
		return ret;

-- 
With Best Regards,
Andy Shevchenko