Re: [PATCH net-next 04/13] sunhme: Return an ERR_PTR from quattro_pci_find

[Sean Anderson]

On 9/19/22 09:11, Rolf Eike Beer wrote:
> Am Montag, 19. September 2022, 01:26:17 CEST schrieb Sean Anderson:
> > In order to differentiate between a missing bridge and an OOM condition,
> > return ERR_PTRs from quattro_pci_find. This also does some general linting
> > in the area.
> >
> > Signed-off-by: Sean Anderson <seanga2@xxxxxxxxx>
> > ---
> >
> >   drivers/net/ethernet/sun/sunhme.c | 33 +++++++++++++++++++------------
> >   1 file changed, 20 insertions(+), 13 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/sun/sunhme.c
> > b/drivers/net/ethernet/sun/sunhme.c index 1fc16801f520..52247505d08e 100644
> > --- a/drivers/net/ethernet/sun/sunhme.c
> > +++ b/drivers/net/ethernet/sun/sunhme.c
> > @@ -2569,30 +2569,33 @@ static void quattro_sbus_free_irqs(void)
> >   #ifdef CONFIG_PCI
> >   static struct quattro *quattro_pci_find(struct pci_dev *pdev)
> >   {
> > +	int i;
> >   	struct pci_dev *bdev = pdev->bus->self;
> >   	struct quattro *qp;
> >
> > -	if (!bdev) return NULL;
> > +	if (!bdev)
> > +		return ERR_PTR(-ENODEV);
> > +
> >   	for (qp = qfe_pci_list; qp != NULL; qp = qp->next) {
> >   		struct pci_dev *qpdev = qp->quattro_dev;
> >
> >   		if (qpdev == bdev)
> >   			return qp;
> >   	}
> > +
> >   	qp = kmalloc(sizeof(struct quattro), GFP_KERNEL);
> > -	if (qp != NULL) {
> > -		int i;
> > +	if (!qp)
> > +		return ERR_PTR(-ENOMEM);
> >
> > -		for (i = 0; i < 4; i++)
> > -			qp->happy_meals[i] = NULL;
> > +	for (i = 0; i < 4; i++)
> > +		qp->happy_meals[i] = NULL;
>
> I know you are only reindenting it, but I dislike moving the variable up to
> the top of the function. Since the kernel is C99 meanwhile the variable could
> be declared just in the for loop. 

Hm, I thought this style was discouraged.

> And when touching this anyway I think we
> could get rid of the magic "4" by using ARRAY_SIZE(qp->happy_meals). Or just
> replace the whole thing with memset(qp->happy_meals, 0, sizeof(qp-
> > happy_meals)).

Yeah, that avoids the whole problem.

--Sean