Re: [PATCH] vfs: replace ints with enum component_type for LAST_XXX

[Amir Goldstein]

On Wed, Apr 1, 2026 at 12:37 PM Jori Koolstra <jkoolstra@xxxxxxxxx> wrote:
>
>
> > Op 01-04-2026 10:29 CEST schreef Amir Goldstein <amir73il@xxxxxxxxx>:
> >
> > > diff --git a/include/linux/namei.h b/include/linux/namei.h
> > > index 58600cf234bc..e79ff97063e9 100644
> > > --- a/include/linux/namei.h
> > > +++ b/include/linux/namei.h
> > > @@ -16,7 +16,7 @@ enum { MAX_NESTED_LINKS = 8 };
> > >  /*
> > >   * Type of the last component on LOOKUP_PARENT
> > >   */
> > > -enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT};
> > > +enum component_type {NORMAL, ROOT, DOT, DOTDOT};
> > >
> >
> > Nah, I don't think it is a good idea to poluate the global namespace
> > with constant as generic as NORMAL, ROOT
>
> Fair enough. I thought also maybe we can keep this purely internal to namei.c,
> but it turns out there is one user that I missed in fs/smb/server/vfs.c, so that
> one has to be converted still as well.
>
> >
> > At the very least this needs a namespace prefix such as
> > NDT_ (nameidata type)
> >
>
> Maybe I would prefer COMPONENT_XXX.
>

Well that adds zero or negative comprehensibility compared to LAST_ IMO

> > But then I don't think it is worth the churn to rename LAST_ to NDT_
> > so I would drop the constants name change altogether.
> >
>
> I mean if we are going to use a separate type for this might as well clean up
> the semantics a bit. But I would be fine with keeping LAST_XXX as well.

Let's leave it then.

Thanks,
Amir.