Re: [PATCH] mcstrans: fix inverted alias skip check in compute_trans_from_raw()
Stephen Smalley <[email protected]> Tue, 21 Jul 2026 08:51:35 -0400
| Newsgroups | org.kernel.vger.selinux |
|---|---|
| Message-ID | <CAEjxPJ4aj0zgmUVxT-rTMGQDMBzBOuj3SaQ8=OSZBo9J0v=WjQ@mail.gmail.com> |
On Thu, Jul 16, 2026 at 11:14 AM Stephen Smalley <[email protected]> wrote: > > ebitmap_cmp() returns 1 when the bitmaps are equal. The > skip-if-alias-of-last-bc test incorrectly was comparing to 0 as if it > followed strcmp() semantics. Fix the comparison. > > Signed-off-by: Stephen Smalley <[email protected]> Merged. > --- > mcstrans/src/mcstrans.c | 3 +-- > 1 file changed, 1 insertion(+), 2 deletions(-) > > diff --git a/mcstrans/src/mcstrans.c b/mcstrans/src/mcstrans.c > index e79aa3d5..8ed1aa95 100644 > --- a/mcstrans/src/mcstrans.c > +++ b/mcstrans/src/mcstrans.c > @@ -1623,8 +1623,7 @@ static char *compute_trans_from_raw(const char *level, domain_t *domain) > if (l->sens == bc->level->sens) { > /* skip if alias of last bc */ > if (last && last->level->sens == bc->level->sens && > - ebitmap_cmp(&last->level->cat, &bc->level->cat) == > - 0) > + ebitmap_cmp(&last->level->cat, &bc->level->cat)) > continue; > > /* compute bits not consumed by base classification */ > -- > 2.55.0 >