Re: of_match[] warnings

[Bjorn Helgaas <helgaas@kernel.org>]

On Thu, Oct 01, 2020 at 11:43:12AM -0500, Rob Herring wrote:
> On Thu, Oct 1, 2020 at 10:53 AM Bjorn Helgaas <helgaas@kernel.org> wrote:
> >
> > On Thu, Oct 01, 2020 at 07:48:23AM -0500, Rob Herring wrote:
> > > On Wed, Sep 30, 2020 at 5:37 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > > >
> > > > These warnings are sort of annoying.  I guess most of the other
> > > > drivers avoid this by depending on OF as well as COMPILE_TEST.
> > >
> > > Using the of_match_ptr() macro should prevent this.
> >
> > Both drivers *do* use of_match_ptr(), but the of_device_id table is
> > unused when of_match_ptr() throws away the pointer.
> 
> Oh right, it's actually we don't want to use of_match_ptr() in this case.
> 
> > I guess we could add __maybe_unused to squelch the warning.  Ugly, but
> > I do think COMPILE_TEST has some value.
> 
> We do that or an ifdef for drivers that actually work for !OF.
> 
> We also have lot's of 'depends on OF' as proposed too, but that's not
> really my preference given we have empty functions for most
> everything.

So just removing of_match_ptr() as below?  That seems way nicer than
any of the #ifdef, __maybe_unused, or "depends on OF" options.

diff --git a/drivers/pci/controller/dwc/pcie-spear13xx.c b/drivers/pci/controller/dwc/pcie-spear13xx.c
index 62846562da0b..b2ef8ffde79e 100644
--- a/drivers/pci/controller/dwc/pcie-spear13xx.c
+++ b/drivers/pci/controller/dwc/pcie-spear13xx.c
@@ -303,7 +303,7 @@ static struct platform_driver spear13xx_pcie_driver = {
 	.probe		= spear13xx_pcie_probe,
 	.driver = {
 		.name	= "spear-pcie",
-		.of_match_table = of_match_ptr(spear13xx_pcie_of_match),
+		.of_match_table = spear13xx_pcie_of_match,
 		.suppress_bind_attrs = true,
 	},
 };