Remote attestation with PCRs and AMD SEV-SNP on GCP using RHCOS #39
Conversation
| FROM quay.io/rkaufman/kbs-tpm-snp:latest as kbc | ||
| FROM quay.io/openshift-release-dev/ocp-v4.0-art-dev:419.96.202505021444-0-coreos | ||
|
|
||
| COPY usr /usr |
There was a problem hiding this comment.
You can copy directory by directory based on the argument/label. We should really maintain a single Containerfile
|
@iroykaufman can you please use the clevis pin instead of the old script running the attestation agent? It should also be easier for you. We should align to the latest way to perform attestation |
iroykaufman
force-pushed
the
trustee-on-GCP
branch
2 times, most recently
from
0c280f3 to
e2e5133
Compare
coreos/justfile
Outdated
| label := if os == "scos" { scos_label } else { fcos_label } | ||
| archive := os + ".ociarchive" | ||
| platform := "qemu" | ||
| kbc_image := "quay.io/afrosi_rh/kbs-client-image:latest" |
There was a problem hiding this comment.
can we use our own built trustee agent?https://quay.io/repository/confidential-clusters/trustee-attester
There was a problem hiding this comment.
This version support TPM as an additional device?
There was a problem hiding this comment.
Yes, it is built from the latest guest component, but you need to register the trusted AK in trustee.
There was a problem hiding this comment.
I checked again in the upstream and it's look like the TPM can only be primary device. This is where I look https://github.com/confidential-containers/guest-components/blob/main/attestation-agent/attester/src/lib.rs .
I created this PR#1093 to add this option
There was a problem hiding this comment.
Sorry, I might have misunderstood the previous comment. Isn't the tpm the attester used on GCP?
There was a problem hiding this comment.
Not only, also SEV-SNP is use for attestation. I add the option for a vm with TPM and another TEE to set the TPM as an additional device for attestation.
There was a problem hiding this comment.
Ah I see. We didn't detect this because the local dev environment isn't a real CVM. Would you mind to fork the guest component in our org and add your patch there? In this way, we know which version is working and build the images out of it. Once, the PR is merged we can delete the repository.
We should try to have common images where we do our testing otherwise, it becomes very messy
/cc @Jakob-Naucke
There was a problem hiding this comment.
I don't have permissions to add a fork.
There was a problem hiding this comment.
I added you, you should have an invitation
iroykaufman
force-pushed
the
trustee-on-GCP
branch
3 times, most recently
from
c84a30a to
ca6e9e4
Compare
Signed-off-by: Roy Kaufman <[email protected]>
Also, use the shared Containerfile for the custom image. Signed-off-by: Roy Kaufman <[email protected]>
using AK registration Signed-off-by: Roy Kaufman <[email protected]>
iroykaufman
force-pushed
the
trustee-on-GCP
branch
from
ca6e9e4 to
986fc71
Compare
Signed-off-by: Roy Kaufman <[email protected]>
iroykaufman
force-pushed
the
trustee-on-GCP
branch
from
986fc71 to
02ee44c
Compare
3 participants
See README.md for details.