Re: [PATCH 06/13] of: unittest: Merge of_unittest_apply{,_revert}_overlay_check()

[Rob Herring <robh@kernel.org>]

On Thu, Jul 27, 2023 at 8:04 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Rob,
>
> On Thu, Jul 20, 2023 at 8:31 PM Rob Herring <robh@kernel.org> wrote:
> > On Wed, Jul 19, 2023 at 05:00:06PM +0200, Geert Uytterhoeven wrote:
> > > of_unittest_apply_overlay_check() and the first part of
> > > of_unittest_apply_revert_overlay_check() are identical.
> > > Reduce code duplication by replacing them by two wrappers around a
> > > common helper.
> > >
> > > Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> > > ---
> > >  drivers/of/unittest.c | 61 ++++++++++++++++---------------------------
> > >  1 file changed, 22 insertions(+), 39 deletions(-)
> >
> > I would do something like this instead:
> >
> > 8<-------------------------------------------------------------------
> > diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
> > index a406a12eb208..a9635935aa26 100644
> > --- a/drivers/of/unittest.c
> > +++ b/drivers/of/unittest.c
> > @@ -2102,7 +2102,7 @@ static int __init of_unittest_apply_overlay(int overlay_nr, int *ovcs_id)
> >  }
> >
> >  /* apply an overlay while checking before and after states */
> > -static int __init of_unittest_apply_overlay_check(int overlay_nr,
> > +static int __init _of_unittest_apply_overlay_check(int overlay_nr,
> >                 int unittest_nr, int before, int after,
> >                 enum overlay_type ovtype)
> >  {
> > @@ -2133,6 +2133,16 @@ static int __init of_unittest_apply_overlay_check(int overlay_nr,
> >                 return -EINVAL;
> >         }
> >
> > +       return ovcs_id;
> > +}
> > +
> > +static int __init of_unittest_apply_overlay_check(int overlay_nr,
> > +               int unittest_nr, int before, int after,
> > +               enum overlay_type ovtype)
> > +{
> > +       int ovcs_id = _of_unittest_apply_overlay_check(overlay_nr, unittest_nr, before, after, ovtype);
> > +       if (ovcs_id < 0)
> > +               return ovcs_id;
> >         return 0;
> >  }
> >
> > @@ -2143,31 +2153,9 @@ static int __init of_unittest_apply_revert_overlay_check(int overlay_nr,
> >  {
> >         int ret, ovcs_id, save_ovcs_id;
> >
> > -       /* unittest device must be in before state */
> > -       if (of_unittest_device_exists(unittest_nr, ovtype) != before) {
> > -               unittest(0, "%s with device @\"%s\" %s\n",
> > -                               overlay_name_from_nr(overlay_nr),
> > -                               unittest_path(unittest_nr, ovtype),
> > -                               !before ? "enabled" : "disabled");
> > -               return -EINVAL;
> > -       }
> > -
> > -       /* apply the overlay */
> > -       ovcs_id = 0;
> > -       ret = of_unittest_apply_overlay(overlay_nr, &ovcs_id);
> > -       if (ret != 0) {
> > -               /* of_unittest_apply_overlay already called unittest() */
> > -               return ret;
> > -       }
> > -
> > -       /* unittest device must be in after state */
> > -       if (of_unittest_device_exists(unittest_nr, ovtype) != after) {
> > -               unittest(0, "%s failed to create @\"%s\" %s\n",
> > -                               overlay_name_from_nr(overlay_nr),
> > -                               unittest_path(unittest_nr, ovtype),
> > -                               !after ? "enabled" : "disabled");
> > -               return -EINVAL;
> > -       }
> > +       ovcs_id = _of_unittest_apply_overlay_check(overlay_nr, unittest_nr, before, after, ovtype);
> > +       if (ovcs_id < 0)
> > +               return ovcs_id;
> >
> >         save_ovcs_id = ovcs_id;
> >         ret = of_overlay_remove(&ovcs_id);
>
> That's what I had done first, before I realized I could reduce it by
> five more lines of code ;-)
>
> mine:  1 file changed, 22 insertions(+), 39 deletions(-)
> yours: 1 file changed, 14 insertions(+), 26 deletions(-)

Less change to review is also worthwhile.

> Anyway, you're the maintainer, so I can update my patch if you insist...

The other thing about this that I noticed is I recall gregkh not
liking the pattern where function parameters change what the function
does (e.g. do_x_or_y(bool do_y)).

So yes, I prefer mine.

Rob