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

From
Jiri Pirko <jiri@resnulli.us>
Date
2026-10-23 18:15:25
Also in
kvm, linux-coco, linux-doc, lkml
Reply to
Edgecombe, Rick P, 2026-10-22
Fri, Oct 23, 2026 at 12:07:28AM +0200, rick.p.edgecombe@intel.com wrote:
On Thu, 2026-10-15 at 16:50 +0200, Jiri Pirko wrote:
quoted
Connect the attestation family to the native TDX service. Expose
TDREPORT, Quote, MRTD and RTMR operations while preserving the existing
native interfaces.
Along the lines of some of the other commenters I'm not clear on the benefits of
the unified TD/VM attestation part of this.

Especially when you talk about a TD report, it is pretty specific to TDX. So
calling user code is going to need to know about TDX anyway. In fact the report
True. For this case, attestation subsystem creates just unified
transport: EVIDENCE_GET. Nice thing is that the same op can be used for
other evidence providing instances as well, like SPDM, armcca, tpm, etc.

won't have all the details of the TD either. Going forward it's main function is
to serve as a nonce transport blob and it only has TD details because of legacy
constraints about SGX based attestation. So just a quirky TDX specific part of
the flow on the way to get a quote.

Can you help me connect the dots of why the unification for the TD/VM
attestation helps us get to a wider attestation vision?


One idea... Not sure on it. We are currently working on support for DICE based
attestation for TDX. This is an industry standard and AFAIU other vendors will
(want to?) support it too. Perhaps if we could have a single way to ask for for
DICE based quote, it could make a cross vendor interface more useful? Callers
could just call and get some standard format they can understand. It might take
some work on the TDX side to ensure the DICE format quote.
Yes, that sounds useful and fits the proposed interface.

DEVICE_GET advertises the supported evidence formats, and callers select
one through EVIDENCE_GET. Providers implementing the same standardized
DICE evidence profile could advertise the same format identifier,
regardless of whether the underlying platform is TDX or something else:

EVIDENCE_GET(challenge=nonce,
             object[]={PRIMARY(<agreed-DICE-profile>)})

We would not need a separate DICE operation or provider family. TDX could
advertise this alongside its existing evidence formats.

Makes sense?
Regarding the device attestation part of this, DaveJ (CCed) is exploring this
currently and can share some findings in a bit.
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