Introduce HybridPublicKey combining ECDH and ML-KEM public keys. Use it
for HostKey::V2 and add Keyslot::V2 for hybrid keyslots.
Add tests for the new functionality.
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Timo Keller <tkeller@linux.ibm.com>
Reviewed-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Add derive_aes256_gcm_key_hybrid() function that derives an Aes256GcmKey
and ML-KEM1024 ciphertext from a private ECDH customer key, a public
ECDH target key, and a public ML-KEM target key. This enables hybrid
post-quantum cryptography support.
Assisted-by: IBM Bob:1.0.5
Co-developed-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Timo Keller <tkeller@linux.ibm.com>
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Prepare the API for future host keys that might not have an EC key by
returning Option instead of a direct reference. Additionally, use
&PKeyRef<Public> as the borrowed view on PKey for better API
consistency.
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Timo Keller <tkeller@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
These messages indicate verification failures and are therefore reported
as errors rather than debug output. Unfortunately,
X509StoreContext::init(...) expects the callback to return an OpenSSL
ErrorStack, so it's not possible to propagate these failures through our
own error hierarchy.
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Enforce root CA organization name requirements during certificate chain
validation to prevent accepting certificates from unexpected root CAs.
Add RootCaVerification enum with RootCaOrganizationPinning and
SkipPinning variants. Extend verify_chain() to validate root CA
organization. Default to pinning "DigiCert" for standard trust chains,
skip pinning when custom root CA is provided.
Assisted-by: IBM Bob:1.0.6
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Return an error when the provided request data is smaller than the
expected request length. The previous condition used the wrong
comparison, which could lead to an out-of-bounds slice and panic.
Add a regression test.
Fixes: 34bef977e8 ("rust/pv: User-data signing and verifying")
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Perform certificate verification in two stages. First, verify the
certificate chain without CRL checks. Once the chain has been validated,
download the referenced CRLs and repeat the verification with CRL
checking enabled.
Fixes: c6f621d0dc ("rust: Add library for pv tools")
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
3-second timeout might be too short for slow networks therefore increase
it to 10s. In addition move this constant to a more prominent position
and add documentation.
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add a dedicated HKD verification error for CRL downloads that exceed the
maximum file size of 10 MiB and cover the max_filesize behavior in the
helper tests and add tests for it.
This commit adds a new development/test dependency as it verifies that
the correct messages are being logged.
Fixes: c6f621d0dc ("rust: Add library for pv tools")
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Adapt the script to generate the test certificates to use the prefix
'http://' for the CRL distribution points as otherwise the test would
fail with an upcoming change. Regenerate the certificates and adapt the
test case dist_points' accordingly as the distribution point has
changed.
Note: The files der.crt and der.crl are skipped as that would result in
binary changes which caues trouble with quilt. They are only used
by some basic tests read_crls and read_crts anyway
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Replace the existing mocks with a trait-based test infrastructure. This
allows testing download_first_crl_from_x509() functionally while
avoiding actual network access, resulting in more comprehensive and
realistic test coverage.
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Rename the module to the more common test module name tests and remove a
superfluous comment.
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Use a trait-based abstraction for CRL downloads so the download
implementation can be replaced more easily, for example in tests that
should avoid real network calls.
Assisted-by: IBM Bob:1.0.5
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
The comment mentions certificate but it's a certificate revocation
list (CRL). In addition, use 1200 bytes as vector capacity as the
comment says 1200.
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Although most errors are ignored and the next CRL distribution point is
tried, log the errors as debug messages.
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
The test cases uses the ml_kem functions, therefore disable the
directive for tests.
$ cargo test
...
--> pv/src/openssl_extensions/ml_kem.rs:5:11
|
5 | #![expect(unused)]
| ^^^^^^
|
= note: `#[warn(unfulfilled_lint_expectations)]` on by default
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add a HostKey enum (currently V1(PKey<Public>)) and introduce a
versioned Keyslot enum (V1(KeyslotV1)). Rename the existing Keyslot type
to KeyslotV1 to prepare for future format extensions.
Update pv, pvattest, pvimg, and pvsecret to use the new enums.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Timo Keller <tkeller@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add a Display implementation for SymKey that formats values as
`SymKey(<type>)`, delegating to key_type(). This provides a clearer
and stable human-readable representation for logs and CLI output.
Reviewed-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Timo Keller <tkeller@linux.ibm.com
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Return the Secure Execution header version in addition to tags. This
comes handy as soon as multiple Secure Execution header versions are
supported.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Timo Keller <tkeller@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Split req.rs into multiple files so that additions can be done more
easily.
Assisted-by: IBM Bob:1.0.4
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add interfaces for creating an ML-KEM keypair and encapsulation and
decapsulation.
The public part of it can be used to encapsulate a shared secret in a
ciphertext. The latter can be decapsulated with the private part of the
keypair to recover the shared secret.
In terms of code, a keypair of type `PKey<Private>` can be generated
with `generate_ml_kem` specifying `KeyType::ML_KEM_512`,
`KeyType::ML_KEM_768`, or `KeyType::ML_KEM_1024`.
To encapsulate a shared secret, generate a `PkeyCtx<Public>` with
`PkeyCtx::new`, initialize it with `encapsulate_init`, and encapsulate
with `encapsulate_to_vec`.
To decapsulate the ciphertext to the shared secret, generate a
`PkeyCtx<Private>` with `PkeyCtx::new`, initialize it with
`decapsulate_init`, and decapsulate with `decapsulate_to_vec`.
Note that when https://github.com/rust-openssl/rust-openssl/pull/2532
is getting merged into `rust-openssl`, these changes will become
unnecessary.
Assisted-by: IBM Bob:1.0.4
Reviewed-by: Marc Hartmayer <marc@linux.ibm.com>
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Timo Keller <tkeller@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
+ Sort and group the imports
+ Normalize and format comments (100 characters width)
Command used:
$ cargo +nightly fmt --
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
The error texts printed a hardcoded .0 instead of the actual value.
Fix this by using the proper string format arguments.
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Fixes: 61c5d7d431 ("rust/pv: Attestation generation and verification support")
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add EBC (Early Boot Customization) utility functions to pv_core library
for parsing and verifying Add-Secret-Request structures.
Introduce the core library functionality needed for EBC:
- Add ebc_utils module to pv_core with ASR parsing and verification
- Export ebc_utils in pv_core lib.rs
- Re-export ebc_utils in pv lib.rs for downstream consumers
- Update pvsecret Cargo.toml dependencies
The library provides the foundation for tools that work with
integrity-protected ASR structures used in SEL guest customization.
Assisted-by: IBM Bob:1.0.1
Acked-by: Holger Dengler <dengler@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Finn Callies <fcallies@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
warning: unresolved link to `GuestSecret::retrievable`
--> pv/src/uvsecret/guest_secret.rs:56:37
|
56 | /// Create Retrievables using [`GuestSecret::retrievable`]
| ^^^^^^^^^^^^^^^^^^^^^^^^ the enum `GuestSecret` has no variant or associated item named `retrievable`
|
= note: `#[warn(rustdoc::broken_intra_doc_links)]` on by default
warning: unresolved link to `GuestSecret::RetrievableKey::name`
--> pv/src/uvsecret/guest_secret.rs:63:30
|
63 | /// SHA256 hash of [`GuestSecret::RetrievableKey::name`]
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the enum `GuestSecret` has no variant or associated item named `RetrievableKey`
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <marc@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Fix two edge cases for generating and interpreting plaintext secrets:
1. The maximum payload size was two bytes to long. The space for the
length header was forgotten to take into account.
2. One of the checks for if the plaintext secret has a size was too
strict.
Fixes: fd024387d7 ("rust/pv: Retrievable secrets support")
Reviewed-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
The ultravisor supports a new secret type in `add-secret` to update
the customer communication key (CCK). Support this new secret
type (0x16).
[seiden@linux.ibm.com: Constify CCK Header struct usage]
Acked-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jakob Naucke <naucke@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
With the recent upgrade to zerocopy 0.8 it is now possible to create
those structs at compile time. This removes the use of a magic array
constant.
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
This enables some const constructors, Dataful Enums,
Dynamically Sized Types and much more.
v0.8 introduces breaking changes including, but not limited to:
- Rename AsBytes to IntoBytes
- Fine-grain (derive) Traits that need to be implemented on top.
- Rename FromZeroes to FromZeros
for which this patch takes care of as well.
Also a direct FromZeros derive is no longer necessary. As it is touched
anyways, remove it where appropriate.
See: https://github.com/google/zerocopy/discussions/1680
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Warn a user that the secret-id to be added is already in the secret
store, but add it anyways.
This helps users to notice issues before they happen, as retrieve may
not retrieve the expected secret due to duplicated IDs.
Reviewed-by: Finn Callies <fcallies@linux.ibm.com>
Tested-by: Finn Callies <fcallies@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Fix the padding of keys that are smaller than the architected slot for
the retrievable key. Previously the keys where appended with zeros.
However, processing software expects a left-padding.
Affected key types:
* Ed448
* SecP521
While at it, fix documentation in the retrievable key struct.
Fixes: fd024387d7 ("rust/pv: Retrievable secrets support")
Acked-by: Ingo Franzki <ifranzki@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Since there have been breaking changes, let's fix them. Fortunately,
there was only one problem to fix.
Command line used:
$ cargo upgrade --incompatible -p 'thiserror'
Acked-by: Jan Höppner <hoeppner@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Secret IDs identify a secret in the store. Tooling (pvsecret) calculates
them by hashing a user-defined string. With this patch it is now
possible to skip the hash step and directly use the input string as the
ID. Up to the first 31 bytes of the input ASCII-string are used. The last byte
is the NUL char. During list pvsecret tries to interpret the secret
as ASCII string and if possible displays the ASCII characters alongside
the hex number.
Also, use the Upper/Lower Hex formatters for the hexstring formatting of
SecretId. Display will, additionally show the ASCII representation if
applicable.
While at it, use Self wherever possible.
Acked-by: Marc Hartmayer <marc@linux.ibm.com>
Reviewed-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Add metadata about the image to the Secure Execution image. This helps
to identify where the Secure Execution header is located in the image
and therefore it's less prone to errors to locate the header.
This patch adds the support for it to 'pvimg' as well as to the
'pvsecret' and 'pvattest' tools.
Reviewed-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
The new Rust port of genprotimg will have the name 'pvimg' as it has
more functionalities than the original genprotimg tool. As preparation
add a Secure Execution header library and an example how to use it. The
example can be used by the KVM-Unit-Tests for creating the Secure
Execution headers needed by the tests [1].
[1] See dc4f7106f3
more information how to use.
Acked-by: Steffen Eiden <seiden@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
