Re: [PATCH 5/5] LSM: Define workqueue for measuring security module state

[Mimi Zohar <zohar@linux.ibm.com>]

On Mon, 2020-06-15 at 11:47 -0400, Stephen Smalley wrote:
> On Mon, Jun 15, 2020 at 10:59 AM Mimi Zohar <zohar@linux.ibm.com> wrote:
> >
> > On Mon, 2020-06-15 at 09:33 -0400, Stephen Smalley wrote:
> > > On Fri, Jun 12, 2020 at 10:42 PM Lakshmi Ramasubramanian
> > > <nramas@linux.microsoft.com> wrote:
> > > >
> > > > The data maintained by the security modules could be tampered with by
> > > > malware. The LSM needs to periodically query the state of
> > > > the security modules and measure the data when the state is changed.
> > > >
> > > > Define a workqueue for handling this periodic query and measurement.
> > >
> > > Won't this make it difficult/impossible to predict the IMA PCR value?
> > > Unless I missed it, you are going to end up measuring every N minutes
> > > even if there was no change and therefore constantly be extending the
> > > PCR.  That will break attestation or sealing against the IMA PCR.
> >
> > Even if it attempts to add the same measurement to the list multiple
> > times, unless something changed, there should only be one measurement
> > in the list.
> 
> Is the PCR only extended once?

Yes, otherwise you wouldn't be able to verify a quote.
 ima_lookup_digest_entry() first verifies the hash isn't in the cache,
before adding it to the measurement list and then extending the TPM.

Mimi