Re: [PATCH v2 3/5] selinux: convert cond_av_list to array

[Ondrej Mosnacek <omosnace@redhat.com>]

On Fri, Jan 31, 2020 at 5:17 AM Paul Moore <paul@paul-moore.com> wrote:
>
> On Fri, Jan 17, 2020 at 3:58 AM Ondrej Mosnacek <omosnace@redhat.com> wrote:
> > Since it is fixed-size after allocation and we know the size beforehand,
> > using a plain old array is simpler and more efficient.
> >
> > Signed-off-by: Ondrej Mosnacek <omosnace@redhat.com>
> > Reviewed-by: Stephen Smalley <sds@tycho.nsa.gov>
> > ---
> >  security/selinux/ss/conditional.c | 125 ++++++++++++------------------
> >  security/selinux/ss/conditional.h |   8 +-
> >  2 files changed, 53 insertions(+), 80 deletions(-)
>
> ...
>
> > diff --git a/security/selinux/ss/conditional.c b/security/selinux/ss/conditional.c
> > index c8a02c9b23ee..b847fd2a6a51 100644
> > --- a/security/selinux/ss/conditional.c
> > +++ b/security/selinux/ss/conditional.c
> > @@ -255,19 +249,17 @@ err:
> >
> >  struct cond_insertf_data {
> >         struct policydb *p;
> > +       struct avtab_node **dst;
> >         struct cond_av_list *other;
> > -       struct cond_av_list *head;
> > -       struct cond_av_list *tail;
> >  };
> >
> >  static int cond_insertf(struct avtab *a, struct avtab_key *k, struct avtab_datum *d, void *ptr)
> >  {
> >         struct cond_insertf_data *data = ptr;
> >         struct policydb *p = data->p;
> > -       struct cond_av_list *other = data->other, *list, *cur;
> >         struct avtab_node *node_ptr;
> > -       u8 found;
> > -       int rc = -EINVAL;
> > +       u32 i;
> > +       bool found;
> >
> >         /*
> >          * For type rules we have to make certain there aren't any
> > @@ -277,7 +269,7 @@ static int cond_insertf(struct avtab *a, struct avtab_key *k, struct avtab_datum
> >         if (k->specified & AVTAB_TYPE) {
> >                 if (avtab_search(&p->te_avtab, k)) {
> >                         pr_err("SELinux: type rule already exists outside of a conditional.\n");
> > -                       goto err;
> > +                       return -EINVAL;
> >                 }
> >                 /*
> >                  * If we are reading the false list other will be a pointer to
> > @@ -287,29 +279,29 @@ static int cond_insertf(struct avtab *a, struct avtab_key *k, struct avtab_datum
> >                  * If we are reading the true list (other == NULL) there shouldn't
> >                  * be any other entries.
> >                  */
> > -               if (other) {
> > +               if (data->other) {
> >                         node_ptr = avtab_search_node(&p->te_cond_avtab, k);
> >                         if (node_ptr) {
> >                                 if (avtab_search_node_next(node_ptr, k->specified)) {
> >                                         pr_err("SELinux: too many conflicting type rules.\n");
> > -                                       goto err;
> > +                                       return -EINVAL;
> >                                 }
> > -                               found = 0;
> > -                               for (cur = other; cur; cur = cur->next) {
> > -                                       if (cur->node == node_ptr) {
> > -                                               found = 1;
> > +                               found = false;
> > +                               for (i = 0; i < data->other->len; i++) {
> > +                                       if (data->other->nodes[i] == node_ptr) {
> > +                                               found = true;
> >                                                 break;
> >                                         }
> >                                 }
>
> Can you explain why you got rid of the "other" variable in this
> function?  It seems like it would be nice in the loop above.

It's probably a result of incremental edits... I agree it looks a bit
nicer with it and it also makes the diff simpler. I'll put it back in
v3.

-- 
Ondrej Mosnacek <omosnace at redhat dot com>
Software Engineer, Security Technologies
Red Hat, Inc.