Re: [PATCH] phy: zynqmp: Fix unused-function compiler warning

[Tobias Klauser <tklauser@distanz.ch>]

On 2020-07-01 at 15:44:43 +0200, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote:
> On Wed, Jul 01, 2020 at 07:06:43PM +0530, Vinod Koul wrote:
> > On 01-07-20, 16:19, Laurent Pinchart wrote:
> > > On Wed, Jul 01, 2020 at 11:04:38AM +0200, Tobias Klauser wrote:
> > > > This fixes the following compiler warning when building with
> > > > CONFIG_PM && !CONFIG_PM_SLEEP:
> > > > 
> > > > drivers/phy/xilinx/phy-zynqmp.c:830:12: warning: ‘xpsgtr_resume’ defined but not used [-Wunused-function]
> > > >   830 | static int xpsgtr_resume(struct device *dev)
> > > >       |            ^~~~~~~~~~~~~
> > > > drivers/phy/xilinx/phy-zynqmp.c:819:12: warning: ‘xpsgtr_suspend’ defined but not used [-Wunused-function]
> > > >   819 | static int xpsgtr_suspend(struct device *dev)
> > > >       |            ^~~~~~~~~~~~~~
> > > 
> > > Oops :-S Sorry about that.
> > > 
> > > Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Thanks for the review Laurent.

> > > Vinod or Kishon, can you pick this patch up, or do I need to send a pull
> > > request ? (It's my first driver in the PHY subsystem so I don't know
> > > what the usual practices are there)
> > 
> > patches are welcome :-)
> > 
> > > 
> > > > Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
> > > > ---
> > > >  drivers/phy/xilinx/phy-zynqmp.c | 4 ++--
> > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > 
> > > > diff --git a/drivers/phy/xilinx/phy-zynqmp.c b/drivers/phy/xilinx/phy-zynqmp.c
> > > > index 8babee2ce9ec..22a0ae635797 100644
> > > > --- a/drivers/phy/xilinx/phy-zynqmp.c
> > > > +++ b/drivers/phy/xilinx/phy-zynqmp.c
> > > > @@ -815,7 +815,7 @@ static struct phy *xpsgtr_xlate(struct device *dev,
> > > >   * Power Management
> > > >   */
> > > >  
> > > > -#ifdef CONFIG_PM
> > > > +#ifdef CONFIG_PM_SLEEP
> > 
> > How about marking it as __maybe_unused instead?
> 
> I don't mind either, I'll let Tobias decide, but his patch seems fine,
> is there a drawback in his approach ? If it's just a matter of personal
> preference, I'd rather not require a v2.

I don't mind either, it was just what seemed the more straight-forward
fix. On the other hand, it seems that marking these functions as
__maybe_unused is the more widely used method in other PHY drivers. In
addition it would have the nice side-effect of the code always being
compile-checked regardless of the value of CONFIG_PM_SLEEP.

Will send a v2 using __maybe_unused and will let you decide which one to
pick :)

Thanks