On Mon, Jul 20, 2026 at 2:06 PM David Windsor <[email protected]> wrote:
> On Thu, Jul 16, 2026 at 5:55 PM Paul Moore <[email protected]> wrote:
> >
...
> > > +int security_lsmxattr_add(struct lsm_xattrs *xattrs, u64 lsm_id,
> > > + const char *name, const void *value,
> > > + size_t value_len)
> > > +{
> > > + struct xattr *xattr;
> > > + void *xattr_value;
> > > + size_t name_len;
> > > +
> > > + if (!xattrs || !xattrs->xattrs || !name || !value)
> > > + return -EINVAL;
> >
> > Sashiko raised a good point about xattrs->xattrs being NULL not
> > necessarily being a good reason for -EINVAL. If xattrs is NULL, yes,
> > something has gone wrong and -EINVAL seems reasonable, but the
> > xattr->xattrs NULL case does seem like it should simply return early
> > with a value of 0 (see SELinux's handling of this case as an example).
>
> The problem is, SELinux has already set the sid by before calling
> lsm_get_xattr_slot, so we do indeed get an in-core label for the inode
> before returning early here. Returning early here from BPF would not
> mean the same thing: there is no label anywhere, in memory or on disk.
It's important to remember that you are not always guaranteed to have
an lsm_xattrs buffer, for some filesystems that is going to be normal
and expected. In these instances the 'xattrs->xattrs' value is going
to be NULL and each LSM that provides a security_inode_init_security()
LSM callback needs to handle that condition gracefully.
As you already know, the SELinux approach to this is to set the
in-memory inode label, but refrain from setting a presistent label
(these filesystems are typically either transient or don't have a
mechanism to allow for persistent label storage). Each BPF program
will need to decide how they want to handle this in their
security_inode_init_security() callback, but in the
security_lsmxattr_add() code where the only purpose is to populate the
lsm_xattrs strcut with the persistent BPF label/xattr, there is
nothing that can be done, it should simply return early without error.
--
paul-moore.com
