Hi Omar,
On Thu, 2020-04-16 at 12:02 -0700, Omar Sandoval wrote:
> On Thu, Apr 16, 2020 at 08:22:10AM +0200, Paul Menzel wrote:
> > Dear Omar,
> >
> >
> > Am 16.04.20 um 02:23 schrieb Omar Sandoval:
> > > From: Omar Sandoval <osandov@xxxxxx>
> >
> > Thank you for the patch.
> >
> > > We've encountered a particular model of STMicroelectronics TPM that
> >
> > Please add models you are encountering this with to the commit message.
> >
> > > transiently returns a bad value in the status register. This causes the
> >
> > Have you contacted STMMicroelectronics?
Also how transient is it? Is this something that only happens early,
for example before selftests finishes? Could you get some statistics
here?
> >
> > > kernel to believe that the TPM is ready to receive a command when it
> > > actually isn't, which in turn causes the send to time out in
> > > get_burstcount(). In testing, reading the status register one extra time
> > > convinces the TPM to return a valid value.
Is this only affecting get_burstcount()?
> > >
> > > Signed-off-by: Omar Sandoval <osandov@xxxxxx>
> > > ---
> > > drivers/char/tpm/tpm_tis_core.c | 19 +++++++++++++++++++
> > > 1 file changed, 19 insertions(+)
> > >
> > > diff --git a/drivers/char/tpm/tpm_tis_core.c b/drivers/char/tpm/tpm_tis_core.c
> > > index 27c6ca031e23..5a2f6acaf768 100644
> > > --- a/drivers/char/tpm/tpm_tis_core.c
> > > +++ b/drivers/char/tpm/tpm_tis_core.c
> > > @@ -238,6 +238,25 @@ static u8 tpm_tis_status(struct tpm_chip *chip)
> > > rc = tpm_tis_read8(priv, TPM_STS(priv->locality), &status);
> > > if (rc < 0)
> > > return 0;
> > > + /*
> > > + * Some STMicroelectronics TPMs have a bug where the status register is
> > > + * sometimes bogus (all 1s) if read immediately after the access
> > > + * register is written to. Bits 0, 1, and 5 are always supposed to read
> > > + * as 0, so this is clearly invalid. Reading the register a second time
> > > + * returns a valid value.
> > > + */
> > > + if (unlikely(status == 0xff)) {
> >
> > I’d like to see a debug message here, saying the TPM is buggy. Maybe the
> > model can be printed to, and that the TPM manufacturer should be contacted.
>
> How can I get the model information? (Sorry, I'm not very familiar with
> TPMs, I'm just the guy on the team that ended up tracking this down.)
Ken's post yesterday suggested using a userspace tool.
In general, Linux does support buggy HW, like the iTPM support. As
Paul said, see if there is a vendor solution first. Whatever fix is
upstreamed should be very specific with a clear explanation of the
problem.
thanks,
Mimi
