[tpm2] Re: Question: standardized development environment for TPM2 testing, regressions, etc.

[tpm2] Re: Question: standardized development environment for TPM2 testing, regressions, etc.

[matthew]

[-- Attachment #1: Type: text/plain, Size: 338 bytes --]

I guess I should be more specific: I'm writing some test utilities that directly communicate with the TPM2 module, and some other OS-supplied pieces (i.e. sysfs), so this would be relevant for "end-to-end" testing s/w apps that operate similarly. It isn't meant to be any sort of replacement for the existing test suite and/or unit tests.

[tpm2] Re: Question: standardized development environment for TPM2 testing, regressions, etc.

[Roberts, William C]

[-- Attachment #1: Type: text/plain, Size: 1256 bytes --]

We have a bunch of docker images that we use for our testing, but they don't use the
In-kernel RM like a QEMU virtualized TPM can do or any kernel provided interfaces like
sysfs.

Some projects use the SWTPM simulator, others still use the IBM TPM simulator.

It would be nice to add a test to the CI that used the in-kernel RM and sysfs interfaces to add some
more test coverage against externally provided pieces.

> -----Original Message-----
> From: matthew(a)giassa.net <matthew(a)giassa.net>
> Sent: Friday, October 16, 2020 5:30 PM
> To: tpm2(a)lists.01.org
> Subject: [tpm2] Re: Question: standardized development environment for TPM2
> testing, regressions, etc.
> 
> I guess I should be more specific: I'm writing some test utilities that directly
> communicate with the TPM2 module, and some other OS-supplied pieces (i.e.
> sysfs), so this would be relevant for "end-to-end" testing s/w apps that operate
> similarly. It isn't meant to be any sort of replacement for the existing test suite
> and/or unit tests.
> _______________________________________________
> tpm2 mailing list -- tpm2(a)lists.01.org
> To unsubscribe send an email to tpm2-leave(a)lists.01.org
> %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s

[tpm2] Re: Question: standardized development environment for TPM2 testing, regressions, etc.

[Matthew Giassa]

[-- Attachment #1: Type: text/plain, Size: 388 bytes --]

One particular test I was thinking of was injecting bogus messages into the TPM2 event log, and then using tpm2-tools to parse it (i.e. binary ==> ASCII) and then validate (i.e. TPM_GENERATED) against each entry. Also, throwing in an OSCP responder container into the mix and doing some remote attestation tests (including failing cases where we deliberate revoke certs in the responder).

[tpm2] Re: Question: standardized development environment for TPM2 testing, regressions, etc.

[Roberts, William C]

[-- Attachment #1: Type: text/plain, Size: 1008 bytes --]



> -----Original Message-----
> From: Matthew Giassa <matthew(a)giassa.net>
> Sent: Monday, October 19, 2020 11:20 AM
> To: tpm2(a)lists.01.org
> Subject: [tpm2] Re: Question: standardized development environment for TPM2
> testing, regressions, etc.
> 
> One particular test I was thinking of was injecting bogus messages into the TPM2
> event log, and then using tpm2-tools to parse it (i.e. binary ==> ASCII) and then
> validate (i.e. TPM_GENERATED) against each entry. Also, throwing in an OSCP
> responder container into the mix and doing some remote attestation tests
> (including failing cases where we deliberate revoke certs in the responder).

You could actually add a libfuzzer test, we do it pkcs11 and tss for examples.
We add them as unit tests, they run fast.


> _______________________________________________
> tpm2 mailing list -- tpm2(a)lists.01.org
> To unsubscribe send an email to tpm2-leave(a)lists.01.org
> %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s