Re: [PATCH 1/2] mtd: delete unneeded call to platform_get_drvdata

[Brian Norris <computersforpeace@gmail.com>]

+ Mike

On Sat, May 17, 2014 at 03:12:02PM +0800, Julia Lawall wrote:
> On Sat, 17 May 2014, Fabio Estevam wrote:
> > On Sat, May 17, 2014 at 3:32 AM, Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> > > From: Julia Lawall <Julia.Lawall@lip6.fr>
> > >
> > > Platform_get_drvdata is an accessor function, and has no purpose if its
> > > result is not used.
> > >
> > > The semantic patch that fixes this problem is as follows:
> > > (http://coccinelle.lip6.fr/)
> > >
> > > // <smpl>
> > > @@
> > > identifier x;
> > > type T;
> > > @@
> > > - T x = platform_get_drvdata(...);
> > > ... when != x
> > > // </smpl>
> > >
> > > Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
> > >
> > > ---
> > >  drivers/mtd/nand/bf5xx_nand.c |    4 ----
> > >  1 file changed, 4 deletions(-)
> > >
> > > diff --git a/drivers/mtd/nand/bf5xx_nand.c b/drivers/mtd/nand/bf5xx_nand.c
> > > index b7a2494..b5fbd48 100644
> > > --- a/drivers/mtd/nand/bf5xx_nand.c
> > > +++ b/drivers/mtd/nand/bf5xx_nand.c
> > > @@ -840,15 +840,11 @@ out_err_kzalloc:
> > >
> > >  static int bf5xx_nand_suspend(struct platform_device *dev, pm_message_t pm)
> > >  {
> > > -       struct bf5xx_nand_info *info = platform_get_drvdata(dev);
> > > -
> > >         return 0;
> > >  }
> > >
> > >  static int bf5xx_nand_resume(struct platform_device *dev)
> > >  {
> > > -       struct bf5xx_nand_info *info = platform_get_drvdata(dev);
> > > -
> > >         return 0;
> >
> > In this case bf5xx_nand_suspend/resume could be removed?
> 
> I don't know.  It looks like it is intentional to have a definition that
> returns an indication of success?  The complete set of definitions is:

I'm not so sure. I believe suspend/resume works just fine without the
hooks, if those hooks would otherwise be doing nothing. It is notable
that this driver is not using the modern dev_pm_ops form of the
suspend/resume callbacks, so maybe it just hasn't aged gracefully--and
possibly was never supported properly in the first place.

I'll either take this patch as-is, or look to killing off the
suspend/resume callbacks entirely here, depending on Mike's feedback.

> #ifdef CONFIG_PM
> 
> static int bf5xx_nand_suspend(struct platform_device *dev, pm_message_t
> pm)
> {
>         struct bf5xx_nand_info *info = platform_get_drvdata(dev);
> 
> 	return 0;
> }
> 
> static int bf5xx_nand_resume(struct platform_device *dev)
> {
>         struct bf5xx_nand_info *info = platform_get_drvdata(dev);
> 
>         return 0;
> }
> 
> #else
> #define bf5xx_nand_suspend NULL
> #define bf5xx_nand_resume NULL
> #endif

Thanks,
Brian