Re: [PATCH] security: fix the key_permission LSM hook function type

[Kees Cook]

On Mon, Jun 15, 2020 at 11:23:44AM -0700, Nick Desaulniers wrote:
> On Mon, Jun 15, 2020 at 11:21 AM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> >
> > On Mon, Jun 15, 2020 at 11:12:32AM -0700, Sami Tolvanen wrote:
> > > Commit 8c0637e950d6 ("keys: Make the KEY_NEED_* perms an enum rather than
> > > a mask") changed the type of the key_permission callback functions, but
> > > didn't change the type of the hook, which trips indirect call checking with
> > > Control-Flow Integrity (CFI). This change fixes the issue by changing the
> > > hook type to match the functions.
> > >
> > > Fixes: 8c0637e950d6 ("keys: Make the KEY_NEED_* perms an enum rather than a mask")
> > > Signed-off-by: Sami Tolvanen <samitolvanen@xxxxxxxxxx>
> >
> > Thanks for fixing this!
> >
> > Acked-by: Kees Cook <keescook@xxxxxxxxxxxx>
> >
> > I wonder if there is any compiler option we can turn on to catch the
> > implicit enum/int casting that would help for these kinds of things
> > without absolutely exploding the build warnings.
> >
> > I see -Wenum-conversion, but that seems to be between enums, not between
> > int.
> >
> > I see this, but it's external:
> > https://noamlewis.wordpress.com/2017/10/05/type-safe-enums-in-c-using-a-clang-plugin/
> 
> Unfortunately, using a looser integral type is the typical workaround
> for -Wenum-conversion warnings (which is on by default, GCC recently
> added that flag, too).  For warning on enum to int, the kernel would
> probably blow up with warnings. I don't know of any such existing
> warning flag.

Yeah, I couldn't find anything either. The closest I could find to avoid
one of the common classes of enum -> int conversion is the bitwise
operations, which is covered by __attribute__((flag_enum)):
https://clang.llvm.org/docs/AttributeReference.html#flag-enum

Ah well. We'll let CFI find them. ;)

-- 
Kees Cook