Re: [PATCH] ovl: avoid possible NULL dereference

[Dan Carpenter]

On Wed, Sep 27, 2023 at 05:02:26PM +0300, Amir Goldstein wrote:
> On Mon, Sep 25, 2023 at 7:52 AM Su Hui <suhui@xxxxxxxxxxxx> wrote:
> >
> > smatch warn:
> > fs/overlayfs/copy_up.c:450 ovl_set_origin() warn:
> > variable dereferenced before check 'fh' (see line 449)
> >
> > If 'fh' is NULL, passing NULL instead of 'fh->buf'.
> >
> > Signed-off-by: Su Hui <suhui@xxxxxxxxxxxx>
> > ---
> >  fs/overlayfs/copy_up.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c
> > index d1761ec5866a..086f9176b4d4 100644
> > --- a/fs/overlayfs/copy_up.c
> > +++ b/fs/overlayfs/copy_up.c
> > @@ -446,7 +446,7 @@ int ovl_set_origin(struct ovl_fs *ofs, struct dentry *lower,
> >         /*
> >          * Do not fail when upper doesn't support xattrs.
> >          */
> > -       err = ovl_check_setxattr(ofs, upper, OVL_XATTR_ORIGIN, fh->buf,
> > +       err = ovl_check_setxattr(ofs, upper, OVL_XATTR_ORIGIN, fh ? fh->buf : NULL,
> >                                  fh ? fh->fb.len : 0, 0);
> >         kfree(fh);
> >
> > --
> > 2.30.2
> 
> After discussing this with Dan Carpenter, this is not a kernel bug,
> it is a smatch bug.

Yeah.  Sorry about that, Su Hui.  The ->buf struct member is not a
pointer, it's an array.  So this isn't really a dereference, it's just
pointer math and foo = fh->buf won't crash even if fh is NULL.

I have written a fix for this in Smatch.  I'll test it a bit before I
push it.

regards,
dan carpenter