On Tue, Apr 14, 2020 at 10:44:33AM -0700, Jerry Snitselaar wrote:
> On Fri Apr 10 20, Jason Gunthorpe wrote:
> > On Thu, Apr 09, 2020 at 11:10:44PM +0200, Hans de Goede wrote:
> > > Since commit dda8b2af395b ("tpm: Revert "tpm_tis_core: Set
> > > TPM_CHIP_FLAG_IRQ before probing for interrupts"") we no longer set
> > > the TPM_CHIP_FLAG_IRQ ever.
> >
> > This all used to work..
>
> IIRC this goes all the way back to 570a36097f30 ("tpm: drop 'irq' from struct tpm_vendor_specific")
> when the flag was added. There was never anything initially setting it in the tpm_tis code.
> There were checks added, but the only place it got set was where it did the interrupt test in
> tpm_tis_send, and it can only get down to that code if the flag is set. Stefan's patch was enabling
> the flag, but with the flag enabled some systems were seeing interrupt storms. I believe it has
> been seen with both the t490 and an internal system that Dan Williams was working on at Intel.
> Without access to hw seeing the problem the decision was to revert Stefan's patches while
> we try to figure out the issues.
Thanks Jerry.
Yup, kinda hard to debug irq issues without local access
to the whole platform. They can be nasty issues to debug
in the 1st place.
/Jarkko
