Re: [PATCH RFC 20/20] attestation: Add the TDX family implementation

From
Peter Fang <hidden>
Date
2026-11-06 21:48:02
Also in
kvm, linux-coco, linux-doc, lkml
Reply to
Jason Gunthorpe, 2026-11-03
On Mon, Nov 02, 2026 at 08:58:51PM -0400, Jason Gunthorpe wrote:
quoted
quoted
If you also want to have a way to query specific types, like AMD's
guid thing, then that's fine too. But I think integrating into systemd
report might be a bit tricky so I'm not sure it is a great idea.

In all cases the format of the report is labled, passed to userspace
and given to the verifier. I see Intel already defined DICE format as
'larger than 8k'.
I'll have to check (or maybe Peter can chime in) if there is some easy label to
look at.
There are many ways we can label the quotes. For example:

- tdx-quote-sgx-1.0 (legacy SGX quote)
- tdx-quote-eat-cwt-1.0 (ECDSA signed DICE quote)
- tdx-quote-eat-cwt-2.0 (ECDSA/MLDSA signed DICE quote)
- ...

Or, at a finer grain level:

- tdx-quote-eat-cwt-1.0-ecdsa
- tdx-quote-eat-cwt-1.0-mldsa
- ...

But regarding this labeling, I have some questions below...
quoted
quoted
We'd probably want to give that a dedicated label in
CTLV land.
Not following what you mean exactly. A CTLV format that contains the report? Or
a user ABI way to request a certain format?
The CTVL output will always be something like {label, data blob}, for
everything so it is self described.

Since the DICE and non-DICE TDX quotes sound non-interoperable they
should have different labels.. I guess? I really don't know the
details of TDX enough to say.
Yep, they're non-interoperable and won't share the same label.
quoted
quoted
One of the things I'm quite adament about on the data labeling is that
if something sees Blob type A then it should at least be able to
verify the signature. Further format variation should be self
describing data within the signed payload. If it is so different that
you can't verify the signature then you have to give it a different
label too.
I'll make sure this made it back to our attestation people. (probably someone
already picked it up)
With TDX DICE quotes, this is already the case (the signature is
verifiable). This may not be possible with legacy quotes, but they're a
thing of the past.
Yeah, I'm quite adament about this. I don't want this to turn into a
funnel for undefined binary blobs that nobody can understand. At a
minimum the signature has to parse, and at the maximum I'd love to
insist on a userspace implementation to dump the contents. We will
see.
Hi Jason,

I think this in general is in line with the TDX attestation design. In
an earlier thread you had said:

  "I think in the way this stuff ended up being used the userspace isn't
  actually aware. [...] The verifier needs to be aware of the format,
  but the collection agent should not need special knowledge."

... which totally makes sense to me.

What pops out at me is later when you talked about data labeling:

  "Userspace would collect all the blobs and the labels then ship them
  to the verifier which has to understand how the data described by the
  label is organized."

For TDX, the direction so far has been to keep the guest as simple as
possible in the attestation process. So the "quote type" knowledge is
pushed into the TDX module and the remote verifier. The host kernel is
mostly tasked with transporting the quote blob from the TDX module back
to the guest. In other words, TDX doesn't have any quote type knowledge
outside of the module itself. And as far as I know, parsing these blobs
is sort of a pain.

The TDX DICE blobs are actually self-describing, meaning the verifier
can follow a schema, parse the blob straight away, and know what's
inside, including how to verify the signature. They're also
distinguishable from the legacy quotes. A launched guest is guaranteed
to have a stable format for its quotes. So this makes me wonder how the
labeling should be done. Do you think it makes sense to just have a
"primary raw TDX quote" label and let the TDX quote blob describe
itself? Also, I think currently there are no plans to add
secondary/supplementary blobs alongside the quote itself.

Thanks,
Peter
This will naturally force people to use the right labels since they
won't be able to make a dumper if they don't.

Jason
Thread (212 messages) 212 messages, 15 authors, last reply 6d ago
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help