On Fri, Oct 11, 2024 at 01:34:39PM +0200, Roberto Sassu wrote:
> On Fri, 2024-10-11 at 12:15 +0200, Mickaël Salaün wrote:
> > On Thu, Oct 10, 2024 at 09:20:52PM -0400, Paul Moore wrote:
> > > On Oct 10, 2024 =?UTF-8?q?Micka=C3=ABl=20Sala=C3=BCn?= <[email protected]>
> > > wrote:
> > > >
> > > > Use the new inode_get_ino() helper to log the user space's view of
> > > > inode's numbers instead of the private kernel values.
> > > >
> > > > Cc: Mimi Zohar <[email protected]>
> > > > Cc: Roberto Sassu <[email protected]>
> > > > Cc: Dmitry Kasatkin <[email protected]>
> > > > Cc: Eric Snowberg <[email protected]>
> > > > Signed-off-by: Mickaël Salaün <[email protected]>
> > > > ---
> > > > security/integrity/integrity_audit.c | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > Should we also need to update the inode value used in hmac_add_misc()?
> >
> > I'm not sure what the impact will be wrt backward compatibility. Mimi,
> > Roberto?
>
> Changing the inode number the HMAC was calculated with has the
> potential effect of making the file inaccessible.
>
> In order to use the new inode number, we need to define a new EVM xattr
> type, and update the previous xattr version with the new one. We could
> deprecate the old xattr version after a while (to be discussed with
> Mimi).
That was my though. I don't we should patch hmac_add_misc() because it
is already in the IMA/EVM ABI and not directly reflected to user space.
The issue might be that user space cannot recreate this hmac because
this private inode number is not known to user space, but I don't know
if there is such user space implementation of IMA/EVM.
>
> Roberto
>
> > >
> > > diff --git a/security/integrity/evm/evm_crypto.c
> > > b/security/integrity/evm/evm_crypto.c
> > > index 7c06ffd633d2..68ae454e187f 100644
> > > --- a/security/integrity/evm/evm_crypto.c
> > > +++ b/security/integrity/evm/evm_crypto.c
> > > @@ -155,7 +155,7 @@ static void hmac_add_misc(struct shash_desc *desc,
> > > struct inode *inode,
> > > * signatures
> > > */
> > > if (type != EVM_XATTR_PORTABLE_DIGSIG) {
> > > - hmac_misc.ino = inode->i_ino;
> > > + hmac_misc.ino = inode_get_ino(inode->i_ino);
> > > hmac_misc.generation = inode->i_generation;
> > > }
> > > /* The hmac uid and gid must be encoded in the initial user
> > >
> > > --
> > > paul-moore.com
>
>
