Delay for IRQ testing is not correct.

[fredjsq]

Code : https://github.com/raspberrypi/linux/blob/7335f3c839417ab9fc40fb205249d415e8ed5aa3/drivers/char/tpm/tpm_tis_core.c#L600C2-L600C17

When IRQ has not been tested, the driver is waiting for 1ms only instead of the ordinal maximum duration of the command (usually the GetCapability). For a GetCapability command, we should wait the IRQ for 750ms.

As a result if a TPM is too slow but still meets the specification requirements, the IRQ is wrongly seen at not working.

[pelwell]

That's not a file that we have modified - it's unchanged from the mainline/upstream kernel code. I would normally suggest you start by emailing the maintainers:

$ scripts/get_maintainer.pl drivers/char/tpm/tpm_tis_core.c
Peter Huewe <peterhuewe@gmx.de> (maintainer:TPM DEVICE DRIVER)
Jarkko Sakkinen <jarkko@kernel.org> (maintainer:TPM DEVICE DRIVER)
Jason Gunthorpe <jgg@ziepe.ca> (reviewer:TPM DEVICE DRIVER)
linux-integrity@vger.kernel.org (open list:TPM DEVICE DRIVER)
linux-kernel@vger.kernel.org (open list)

But in this case we can (slightly cheekily) notify @PeterHuewe and @jarkkojs like this. They may have feedback, but they would be within their rights to expect us to go the more formal route of the kernel mailing lists.

[pelwell]

See also #7054.