ibmswtpm-users

[Ibmswtpm-users] jTSS support upcoming & bug report

[Ronald T. <ron...@ia...>]

Hello,

It's great to welcome another Trusted Computing project here at 
sourceforge! 
You'll find it interesting that I'm currently preparing a release of 
jTSS [1] compatible with the software TPM.

I'm quite impressed with the implementation quality, although I found 
the following minor issues:

* tpm_server and the utilities allow illegal port ranges (i.e. 300005 > 
2^16) to be used.
* TPM_Unseal will return a TPM_AUTH2FAIL if the sealing key is not 
properly authorized. The specs (incl. rev 117) require a plain TPM_AUTHFAIL.
* TPM_Quote2 sometimes returns an invalid TPM_CAP_VERSION_INFO 
structure. This bug occurs only in the context of our full test suite, 
not when I repeat the individual unit test - I think it may be some 
minor memory corruption.

Best Regards,
Ronald

[1] trustedjava.sourceforge.net

-- 
Dipl.-Ing. Ronald Tögl          phone +43 316/873-5502
Trusted Computing Labs          fax   +43 316/873-5520
IAIK                            ron...@ia...
Graz University of Technology   http://www.iaik.tugraz.at

Re: [Ibmswtpm-users] jTSS support upcoming & bug report

[Kenneth G. <kgo...@us...>]

Ronald Tögl <ron...@ia...> wrote on 05/31/2010 02:49:36 AM:

>
> It's great to welcome another Trusted Computing project here at
> sourceforge!
> You'll find it interesting that I'm currently preparing a release of
> jTSS [1] compatible with the software TPM.
>
> I'm quite impressed with the implementation quality, although I found
> the following minor issues:
>
> * tpm_server and the utilities allow illegal port ranges (i.e. 300005 >
> 2^16) to be used.
> * TPM_Unseal will return a TPM_AUTH2FAIL if the sealing key is not
> properly authorized. The specs (incl. rev 117) require a plain
TPM_AUTHFAIL.

I believe that the spec has a typo, since this is the second authorization
value.

I will being this up on a TPM working group call.

> * TPM_Quote2 sometimes returns an invalid TPM_CAP_VERSION_INFO
> structure. This bug occurs only in the context of our full test suite,
> not when I repeat the individual unit test - I think it may be some
> minor memory corruption.

Let me know what you find.  Could you send a TPM trace of that ordinal
when it fails?

If you have trouble synchronizing the TPM and the test suite, a trick I
often use is to trace session handles and nonces.  Since they're random,
it's easy to line up two separate traces.

Also, note that the structure is only returned if addVersion is TRUE.
Otherwise a length of 0 is returned.