[PATCH -next v2] cred: conditionally declare groups-related functions

[Paul Moore]

On Tue, Jun 26, 2018 at 4:13 AM Ondrej Mosnacek <omosnace at redhat.com> wrote:
> po 25. 6. 2018 o 23:04 Paul Moore <paul at paul-moore.com> napísal(a):
> > On Mon, Jun 25, 2018 at 2:56 AM Ondrej Mosnacek <omosnace at redhat.com> wrote:
> > > The groups-related functions declared in include/linux/cred.h are
> > > defined in kernel/groups.c, which is compiled only when
> > > CONFIG_MULTIUSER=y. Move all these function declarations under #ifdef
> > > CONFIG_MULTIUSER to help avoid accidental usage in contexts where
> > > CONFIG_MULTIUSER might be disabled.
> > >
> > > This patch also adds a fallback for groups_search(). Currently this
> > > function is only called from kernel/groups.c itself and
> > > keys/permissions.c, which depends on CONFIG_MULTIUSER. However, the
> > > audit subsystem (which does not depend on CONFIG_MULTIUSER) calls this
> > > function in -next, so the fallback will be needed to avoid compilation
> > > errors or ugly workarounds.
> > >
> > > See also:
> > > https://lkml.org/lkml/2018/6/20/670
> > > https://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/audit.git/commit/?h=next&id=af85d1772e31fed34165a1b3decef340cf4080c0
> > >
> > > Reported-by: Randy Dunlap <rdunlap at infradead.org>
> > > Signed-off-by: Ondrej Mosnacek <omosnace at redhat.com>
> > > ---
> > >  include/linux/cred.h | 15 ++++++++++-----
> > >  1 file changed, 10 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/include/linux/cred.h b/include/linux/cred.h
> > > index 631286535d0f..7eed6101c791 100644
> > > --- a/include/linux/cred.h
> > > +++ b/include/linux/cred.h
> > > @@ -65,6 +65,12 @@ extern void groups_free(struct group_info *);
> > >
> > >  extern int in_group_p(kgid_t);
> > >  extern int in_egroup_p(kgid_t);
> > > +extern int groups_search(const struct group_info *, kgid_t);
> > > +
> > > +extern int set_current_groups(struct group_info *);
> > > +extern void set_groups(struct cred *, struct group_info *);
> > > +extern bool may_setgroups(void);
> > > +extern void groups_sort(struct group_info *);
> > >  #else
> > >  static inline void groups_free(struct group_info *group_info)
> > >  {
> > > @@ -78,12 +84,11 @@ static inline int in_egroup_p(kgid_t grp)
> > >  {
> > >          return 1;
> > >  }
> > > +static inline int groups_search(const struct group_info *group_info, kgid_t grp)
> > > +{
> > > +       return 1;
> > > +}
> > >  #endif
> > > -extern int set_current_groups(struct group_info *);
> > > -extern void set_groups(struct cred *, struct group_info *);
> > > -extern int groups_search(const struct group_info *, kgid_t);
> > > -extern bool may_setgroups(void);
> > > -extern void groups_sort(struct group_info *);
> >
> > I was going through the list of functions to make sure it was safe to
> > move all of these under CONFIG_MULTIUSER and I believe there may be an
> > issue with key_task_permission() and groups_search(), specifically one
> > can disable CONFIG_MULTIUSER yet keep CONFIG_KEYS enabled.
>
> Oh, you're right, CONFIG_KEYS does not depend on CONFIG_MULTIUSER, for
> some reason I was looking at CONFIG_SECURITY instead, which does :/ I
> will need to update the commit message...
>
> > I'm going to guess that the fix is going to be to have CONFIG_KEYS
> > depend on CONFIG_MULTIUSER, but I would talk with the keys folks (I've
> > CC'd David Howells) to see what they would prefer.
>
> I was wondering why the security/keys/permissions.c did not give the
> link error and it turns out that it is due to the definition of
> __kgid_val(), which is hard-coded to 0 if CINFIG_MULTIUSER=n, so the
> compiler just optimizes out most of the code in key_task_permission()
> (including the groups_search() call). Thus, for permissions.c it does
> not matter which fallback we choose (and since this patch always
> defines groups_search(), it will not lead to compile errors in
> permissions.c either).
>
> TL;DR: There is no need to fix anything in the keys code (beyond this
> patch). I also checked the other functions and they are only used in
> places that either depend on MULTIUSER or use also groups_alloc(),
> which is already declared only when CONFIG_MULTIUSER=y.

... and you've already defined a dummy implementation of
groups_search(); that's the important part.  I clearly was in a rush
while reviewing this patch.

/me shakes his head

-- 
paul moore
www.paul-moore.com