Steps to a stable 7.0.0

[ionut-arm]

Since we're currently in an alpha release (yet again) it would be good to know what we'd like to achieve before we can tag a stable release. Our focus right now is to remove all not-abstracted FFI types from the tss-esapi interface, replacing them with types that we have more control over.

FFI types still present in the Context interface (not going to link to where they appear, a simple search through the repo should be enough for that):

• TPMT_SIG_SCHEME ( Added SignatureScheme type. #286 )
• TPM2B_ATTEST ( will be lifted from Add TPM2_Certify #221 ; Adds the attest structures #293 covers this )
• TPMA_LOCALITY ( Removes TPMA_LOCALITY from context methods. #294 )
• TPM2_CC ( Create native type for the TPM2_CC constants. #296 )
• TPM2B_SENSITIVE ( Sensitive buffers #306 )

Another type that currently uses other FFI types in its interface is CapabilityData. The types:

• TPM2_PT (just replacing the use in CapabilityData with the existing abstraction, Adds TaggedTpmProprtyList #301)
• TPM2_PT_PCR (Implement native type for TPML_TAGGED_PCR_PROPERTY in CapabilityData. #305, Rust native TPML_TAGGED_PCR_PROPERTY type #308)
• TPMA_ALGORITHM (Updates CapabilityData to use AlgorithmPropertyList #304)
• TPMA_CC(Creates native type for TPML_CCA #315)
• TPML_CCA(Creates native type for TPML_CCA #315)
• TPM2_ECC_CURVE ( Replace type in CapabilityData::EccCurves #312 )

Other bits of work:

• Identify the cause of ActivateCredential error on non-6.1.1 versions (Esys Finish ErrorCode (0x00000101)) #285 and fix it
• Finish off the request in Manually construct PcrData? #277 ( Improved return values of pcr_read. #281 )
• Fix PCR structures issues (PcrSlot wont compile if TPM2_PCR_SELECT_MAX != 4 #310 To few items in PcrSlot compared with what PcrSelectSize can indicate. #311 Fixes some pcr issues. #317 )

I was also wondering about changing a few other things, please list any other ideas/concerns below:

• Some key-bits types seem redundant, for example Sm4KeyBits can only have one value. I wonder if it wouldn't be better to just remove the type completely and have a "default" implementation that produces the correct FFI values wherever they're needed to replace the current conversions
• For structures that can be built using ...Builder structures, we should either document this or create a method on the type to get a new builder, something like pub fn builder() -> ... ( Add builder methods and move NvPublic #316 )

[ionut-arm]

@rshearman - Do you require any specific features in the upcoming release (more specifically, TPM2_Certify)? We can, of course, release extra functionality in a future minor release, but I'm not sure if that's relevant to you at all (i.e. if you use crates.io for dependency management, or if git dependencies are good enough).

[rshearman]

It would be good to have TPM2_Certify in the upcoming release, but I'm decoupled from the tss-esapi release cycle. In other words, it would also be fine to have this in a future minor release.
There's nothing else I can think of that I would need in the near future.

[ionut-arm]

@Superhepper - have you, by any chance, started working on TPM2B_SENSITIVE? Plan on starting that tomorrow, if you're not already on it.

[Superhepper]

Nah I have not started any work on it. I plan to work on the issue with tagged PCR property list

[ionut-arm]

Cool, I've raised #306 to cover the TPM2B_SENSITIVE stuff. I'll have a look at places where we could change references to full ownership in method interface.

[ionut-arm]

@Superhepper - CapabilityData still has TPM2_ECC_CURVE and TPMA_CC to abstract, are you looking at either of them?

[Superhepper]

Right now I am working with TPMA_CC.

[ionut-arm]

Will do the ECC_CURVE then. I assume that's the last of it? Have you seen anything else?

[Superhepper]

No, I usually add stuff if a find things a long the way.

[ionut-arm]

After #316 is merged are we clear to release a 7.0.0, or is there some other stuff to be done?

[Superhepper]

I maybe want to add the compilation bug for PcrSlot. But if people need a released major version rather soon. I think we can make a release without it.

[ionut-arm]

If we can get that fix in next week that would be nice, but otherwise it's not a breaking change, I think, so we can release it as a bugfix update, not even a minor bump.

[Superhepper]

Yeah, just say the release will come on Friday next week. If it makes it, it makes it otherwise it will be added later.

[ionut-arm]

@puiterwijk @lkatalin @rshearman
Please be advised, we're planning to release a new stable version of the tss-crate soon. If you'd like to test any parts of the crate, or to give any feedback on the interface please do so within the next few days.

[lkatalin]

Thanks for the heads up @ionut-arm ! I just ran a test and it seems fine for us with Keylime.

[ionut-arm]

I've released a 7.0.0-beta.1 version of the tss-crate, the stable release will go out next week if no issues are reported until then!

[rshearman]

I've updated our application to use 7.0.0-beta.1 and in the course of that I ran into #322. The only other thing I noticed is that it would be nice if EccCurveIdentifier derived Hash, but that's something that can be easily done later.

[ionut-arm]

I've just released and tagged tss-esapi-7.0.0-beta.2 with the changes requested above by @rshearman

[ionut-arm]

The 7.0.0 stable version has been released and tagged! 🎉 Please let us know if any issues surface, so we can fix them right away!