Hello,
> > > > > > +bool of_pci_is_supply_present(struct device_node *np)
> > > > > > +{
> > > > > > + struct property *prop;
> > > > > > + char *supply;
> > > > > > +
> > > > > > + if (!np)
> > > > > > + return false;
> > > > >
> > > > > Why do we need to test !np here? It should always be non-NULL.
> > > > >
>
> This doesn't appear to be the case, at least on my x86 machine and on x86
> QEMU with CONFIG_OF enabled.
Thank you for testing! Much appreciated.
We've also received other signal about this, too, per:
- https://lore.kernel.org/r/673f39b0.050a0220.363a1b.0126.GAE@xxxxxxxxxx
- https://lore.kernel.org/r/CA+G9fYurbY3B6ahZ+k+Syp5bZ3a+YQdeX8DRb6Twi4BDEFbUsw@xxxxxxxxxxxxxx
> > > > Right, I think this can be dropped. We check for the OF node in the
> > > > function above.
> > > >
> > >
> > > I think it was a leftover that I didn't cleanup. But I do plan to move this API
> > > to drivers/of once 6.13-rc1 is out. So even if it didn't get dropped now, I will
> > > do it later.
> >
> > I removed the NULL check directly on the branch. Thank you!
I added the NULL check back.
Krzysztof
