Re: [PATCH v6 7/7] tpm: pass an array of tpm_bank_list structures to tpm_pcr_extend()

From: Jarkko Sakkinen
Date: Wed Dec 12 2018 - 13:27:43 EST

On Thu, Dec 06, 2018 at 07:38:30PM +0100, Roberto Sassu wrote:
> On 12/5/2018 1:14 AM, Jarkko Sakkinen wrote:
> > On Tue, Dec 04, 2018 at 09:21:38AM +0100, Roberto Sassu wrote:
> > > The new tpm_bank_list structure has been preferred to the tpm_digest
> > > structure, to let the caller specify the size of the digest (which may be
> > > unknown to the TPM driver).
> >
> > Why is that? Didn't previous commit query these?
> Since the TPM driver pads/truncates the first digest passed by the
> caller to extend PCRs for which no digest was provided, it must know
> which amount of data it can use. It is possible that the algorithm of
> the first digest is unknown for the TPM driver, if the caller of
> tpm_pcr_extend() didn't check chip->allocated_banks.
> By requiring that the caller passes also the digest size, this problem
> does not arise. It seems reasonable to me to pass this information, as
> the caller calculated the digest and it should know the digest size.

OK. I noticed something other things that look to alarming:

1. The function does not fail if alg_id is not found. This will go
2. The function does not fail if there is a mismatch with the digest