Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

GetConfComputeGpuAttestationReport does not accept a nonce #128

Open
derpsteb opened this issue Aug 23, 2024 · 2 comments
Open

GetConfComputeGpuAttestationReport does not accept a nonce #128

derpsteb opened this issue Aug 23, 2024 · 2 comments

Comments

@derpsteb
Copy link

Hey,

the function GetConfComputeGpuAttestationReport does currently not accept a nonce for the report generation. This is insecure for CC use cases because an attacker might record a valid report once and replay it to a victim multiple times/on different machines, potentially with a modified GPU attached.

We have implemented this in our fork here. This works fine so it seems to be supported by the runtime.

Could you update that function?
Thanks,
Otto

@klueska
Copy link
Contributor

klueska commented Aug 28, 2024

I checked with the NVML team and they confirmed that if you pass a nonce in the calling struct to the underlying C API, then it will be used by the driver (rather than generating one on the fly).

I'll need to think about the right API for this because we don't like to stray too far from the underlying C API (with the exception of transforming inputs meant as return values to actual return values). In this case though, the argument to the C API is both a valid input AND a place to store the return value...

@derpsteb
Copy link
Author

derpsteb commented Oct 7, 2024

Hey. Did you have a chance to consider this? Would be cool if we could drop our fork.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants