Re: [PATCH 18/26] ovl: avoid pointless cred reference count bump

[Christian Brauner]

On Sun, Nov 24, 2024 at 03:59:40PM +0100, Amir Goldstein wrote:
> On Sun, Nov 24, 2024 at 2:44 PM Christian Brauner <brauner@xxxxxxxxxx> wrote:
> >
> > No need for the extra reference count bump.
> >
> > Signed-off-by: Christian Brauner <brauner@xxxxxxxxxx>
> > ---
> >  fs/overlayfs/copy_up.c | 6 ++----
> >  1 file changed, 2 insertions(+), 4 deletions(-)
> >
> > diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c
> > index 439bd9a5ceecc4d2f4dc5dfda7cea14c3d9411ba..39f08531abc7e99c32e709a46988939f072a9abe 100644
> > --- a/fs/overlayfs/copy_up.c
> > +++ b/fs/overlayfs/copy_up.c
> > @@ -741,17 +741,15 @@ static int ovl_prep_cu_creds(struct dentry *dentry, struct ovl_cu_creds *cc)
> >                 return err;
> >
> >         if (cc->new)
> > -               cc->old = override_creds(get_new_cred(cc->new));
> > +               cc->old = override_creds(cc->new);
> >
> >         return 0;
> >  }
> >
> >  static void ovl_revert_cu_creds(struct ovl_cu_creds *cc)
> >  {
> > -       if (cc->new) {
> > +       if (cc->new)
> >                 put_cred(revert_creds(cc->old));
> > -               put_cred(cc->new);
> > -       }
> 
> Same comment here, I think this will read more clearly as
>                revert_creds(cc->old));
>                put_cred(cc->new);
> 
> and better reflects the counterpart of ovl_prep_cu_creds().

Ok, done.