Re: [PATCH v2 1/5] tpm_tis: Allow lowering of IRQ also when locality is not active

[Stefan Berger]

On 6/15/20 11:06 AM, Marc-André Lureau wrote:
> Hi
>
> On Mon, Jun 15, 2020 at 6:23 PM Stefan Berger
> <stefanb@linux.vnet.ibm.com> wrote:
> > From: Stefan Berger <stefanb@linux.ibm.com>
> >
> > This patch fixes a bug that occurs when using interrupts. It
> > allows to lower the IRQ also when a locality is not active.
> Can you quote the specification, or is it purely based on testing
> (Windows & Linux) or checking expected behaviour from Linux code?

There's no spec for this detail. It's purely based on testing.

also: Linux 5.x had interrupt support for a short while but this was 
removed due to some laptops receiving an interrupt storm following 
enablement: https://www.spinics.net/lists/linux-integrity/msg11870.html  
[ reason seems to be misconfiguration of some GPIO pin]

I tested it with this patch: https://lkml.org/lkml/2019/8/20/421

5.4 kernel: 
https://elixir.bootlin.com/linux/v5.4/source/drivers/char/tpm/tpm_tis_core.c#L983


   Stefan




> > Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
> > ---
> >   hw/tpm/tpm_tis_common.c | 4 ----
> >   1 file changed, 4 deletions(-)
> >
> > diff --git a/hw/tpm/tpm_tis_common.c b/hw/tpm/tpm_tis_common.c
> > index 1af4bce139..0f42696f1f 100644
> > --- a/hw/tpm/tpm_tis_common.c
> > +++ b/hw/tpm/tpm_tis_common.c
> > @@ -601,10 +601,6 @@ static void tpm_tis_mmio_write(void *opaque, hwaddr addr,
> >           /* hard wired -- ignore */
> >           break;
> >       case TPM_TIS_REG_INT_STATUS:
> > -        if (s->active_locty != locty) {
> > -            break;
> > -        }
> > -
> >           /* clearing of interrupt flags */
> >           if (((val & TPM_TIS_INTERRUPTS_SUPPORTED)) &&
> >               (s->loc[locty].ints & TPM_TIS_INTERRUPTS_SUPPORTED)) {
> > --
> > 2.24.1