On Thu, Jul 23, 2020 at 08:42:25AM +0100, David Howells wrote:

Jarkko Sakkinen [off-list ref] wrote:

quoted

Why f1774cb8956a lacked any possible testing? It extends ABI anyway.

I think it is a kind of change that would require more screening before
getting applied.

Yeah.  It went in via a round-about route.  I left off development of it when
the tpm stuff I wrote broke because the tpm2 stuff went in upstream.  I then
handed the patches off to Denis who did the tpm support, but I never got my
stuff finished enough to work out how to do the testsuite (since it would
involve using a tpm).  However, since I did the PKCS#8 testing module as well,
I guess I don't need that to at least test the API.  I'll look at using that
to add some tests.  Any suggestions as to how to do testing via the tpm?

David

The unfortunate thing is that I was not involved with asym_tpm.c review
process in any possible way, which means that at the moment I lack both:

1. Knowledge of crypto/asymmetric_keys.
2. How asym_tpm.c is implemented.

I only became aware of asym_tpm.c's existence last Sep [*].

I'll put to my backlog to try TPM asymmetric keys (earliest when I'm back
from vacation 08/10).

[*] https://lore.kernel.org/linux-integrity/20190926171601.30404-1-jarkko.sakkinen@linux.intel.com/ (local)

/Jarkko