HI Ilias,
On Wed, 2 Apr 2025 at 19:20, Ilias Apalodimas
<ilias.apalodi...@linaro.org> wrote:
>
> Hi Simon,
>
> On Wed, 2 Apr 2025 at 00:28, Simon Glass <s...@chromium.org> wrote:
> >
> > The tpm_auto_start() function is used in tests and assumes that it can
> > open the TPM even if it is already open and a locality claimed. The cr50
> > driver does not use the common TPM2 TIS code so lacks a check for the
> > is_open field of struct tpm_chip and in fact it doesn't use that struct.
> >
> > Add an equivalent check to cr50_i2c_open().
> >
> > This fixes tpm_auto_start() on coral.
>
> Mind if I change this on the commit. it doesn't fix tpm_auto_start()
> it fixes *all* init on that TPM because tpm init && tpminit hangs as
> well
>
> Other than that
>
> Reviewed-by: Ilias Apalodimas <ilias.apalodi...@linaro.org>
OK thanks, yes please go ahead.
>
> >
> > Signed-off-by: Simon Glass <s...@chromium.org>
> > ---
> >
> > drivers/tpm/cr50_i2c.c | 4 ++++
> > 1 file changed, 4 insertions(+)
> >
> > diff --git a/drivers/tpm/cr50_i2c.c b/drivers/tpm/cr50_i2c.c
> > index 08ec179346e..5b2d5ccb146 100644
> > --- a/drivers/tpm/cr50_i2c.c
> > +++ b/drivers/tpm/cr50_i2c.c
> > @@ -737,9 +737,13 @@ static int cr50_i2c_report_state(struct udevice *dev,
> > char *str, int str_max)
> >
> > static int cr50_i2c_open(struct udevice *dev)
> > {
> > + struct cr50_priv *priv = dev_get_priv(dev);
> > char buf[80];
> > int ret;
> >
> > + if (priv->locality != -1)
> > + return -EBUSY;
> > +
> > ret = process_reset(dev);
> > if (ret)
> > return log_msg_ret("reset", ret);
> > --
> > 2.43.0
> >
> > base-commit: e0695b08a71e167bf5ce83ae25acfdebca554b8d
> > branch: tpm
Regards,
Simon
