Re: [PATCH v2 3/3] tpm: parse TPM event logs based on EFI table

[Jason Gunthorpe]

On Mon, Sep 18, 2017 at 02:38:03PM +0200, Thiebaud Weksteen wrote:
> On Tue, Sep 12, 2017 at 10:48 AM, Thiebaud Weksteen <tweek@xxxxxxxxxx> wrote:
> > On Mon, Sep 11, 2017 at 10:47:50AM -0600, Jason Gunthorpe wrote:
> >> On Mon, Sep 11, 2017 at 12:00:22PM +0200, Thiebaud Weksteen wrote:
> >>
> >> >     chip->bin_log_seqops.chip = chip;
> >> > -   if (chip->flags & TPM_CHIP_FLAG_TPM2)
> >> > +
> >> > +   if (log_version == EFI_TCG2_EVENT_LOG_FORMAT_TCG_2 ||
> >> > +       (!log_version && (chip->flags & TPM_CHIP_FLAG_TPM2)))
> >> >             chip->bin_log_seqops.seqops =
> >> >                     &tpm2_binary_b_measurements_seqops;
> >>
> >> Lets have all the read_log_* versions return the postitive log_version
> >> and get rid of the chip->flags check here.
> >>
> >> ie Doesn't ACPI always return the TPM 1 version?
> >
> > That is my understanding. Ashley, Nayna, could you confirm the format
> > version expected by tpm_of? Could it be both?
> >
> 
> I've changed the returned code for ACPI but not for DeviceTree.
> Without confirmation for tpm_of, I am reluctant to modify the current
> behaviour.

Move the TPM_CHIP_FLAG_TPM2 check into tpm_of to keep the current behavior
but still return the code..

Jason