Refactoring adding psk method by ElsaLopez133 · Pull Request #415 · lake-rs/lakers

ElsaLopez133 · 2026-03-12T09:40:44Z

Adding PSK method after refactoring.

edhoc.rs is kept as the main EDHOC protocol file and depending on the method we branch to different functions that are stored in edhoc/statstat.rsandedhoc/psk.rs. Therefore, edhoc/statstat.rsandedhoc/psk.rs` contain only the parts that differ between the Static-Static and PSK methods.
These method-specific files return result structs back to edhoc.rs, so the central flow can continue without duplicating the shared protocol logic.
We store method-specific fields in enums such as WaitM3MethodSpecifics, ProcessingM2MethodSpecifics, ProcessedM2MethodSpecifics, and ProcessingM3MethodSpecifics. These enums allow the common state structs to carry only the extra data required by the selected method. The use of enums was chosen to avoid having Option values.

…rors

For each function we match on state and create method-specific fiunctions. For example, we have r_prepare_message_2 and we match based on emthod to either use t_prepare_message_2_statstat or r_prepare_message_2_psk

We avoid repetition of shared-values among methods This affects ProcessingM2 and ProcessingM3

Adding enum in prepare_message_2 to define inputs based on method

- Made static keys I and R method dependednt (not needed ofr PSK) - Added high level identity enums - Made i_verify_message_2 take optional I - Added ProcessedM2MethodSpecifics and ParsedMessage2Detials - Updated i_parse_message_2 and high level initiaotr API - Adapted lakers-python and lakers-c

1. Remove old comments (returns c_r) 2. Made r_verify_message_3 method distinction in the outer level (so r_verify_message_3_statstat does not do repeated matches) 3. Add a FIXMEE on the error type. 4. Modified r: Option<> to be method-enum, matching the current sytle and avoiding repetition (ResponderIdentity is the enum) 5. ParsedMessage2Detials: removed ead_2 6. Changed pub to pub(crate) in statstat.rs

1. Follow the same pattern for InitiatorIdentity (similar to RepsonderIdentity) instead of making i: Option<> 2. Fix in lakers-python/src/repsonder.rs: minimize the use of ? by extracting the value only once 3. Avoid wildcard imports 4. Add FIXME for the error type. Need to find a good one 5. Removing unnecesary method in ProcessedM2

Using struct+union to replicate the enum style in Rust

- Adding a SAFETY comment in lakers-c/src/lib - Adding a FIXME to define a new error type (method_specifics doesnt match with the what was agreed) - Correcting prk_4e3m and prk_3e2m to avoid matching on symmetric credential in statstat (since now it is only the stat) - Remove _ead_2, since it is used

Adding PSK specific functions Adding coap examples Adding edhoc/psk.rs module

Fixing unused variables Fixing lakers-c and lakers-python Adding FIXME for error declaration Fixing coap examples Matching in edhoc.rs instead of edhoc/psk.rs Fix lakers-c: Pm2Psk defined twice

(cherry pick from pr-psa-memory-error)

- simplify compute_th_3 to use Option<&[u8]> for CRED_R - unify encode_plaintext_2 with Option<(&[u8], &BytesMac2)> - unify encode_plaintext_3 with Option<(&[u8], &BytesMac3)> - unify message_3 AAD handling with Option<(&[u8], &[u8], &[u8])>

refactor edhoc method-specific message processing Keep match-based dispatch in edhoc.rs and move shared flow out of statstat/psk for: - r_prepare_message_2 - r_parse_message_3 - r_verify_message_3 - i_parse_message_2 - i_verify_message_2 - i_prepare_message_3

WilliamTakeshi · 2026-03-18T10:24:18Z

+                ParsedMessage2Details::StatStat { id_cred_r } => id_cred_r,
+                ParsedMessage2Details::Psk {} => IdCred::new(),
+            };
            *id_cred_r_out = id_cred_r;


on the PSK case, why write an empty id_cred_r in id_cred_r_out. I thought we would just do nothing.

So I addressed that by adding a flag has_id_cred_r_out. The flag now tells whether ID_CRED_R is present. I still initialize id_cred_r_out to a default empty value in the PSK case just to avoid leaving output memory undefined on the C side, but maybe this is not strictly needed?

Good question, I don't know what is the right thing to do on C, but I guess better than adding a flag has_id_cred_r_out it should be something like method and be a 3 for Statstat and 4 for Psk.

And yes, reading again I think your code is correct, they sohuld still have the empty value initialized in the PSK case, we just ignore it.

WilliamTakeshi

Besides the comments I put here,

can you also check if nRF52 code still runs??? I don't have one with me right now :((
write an full e2e test like test_handshake(). It will show if everything works and give an example for future. Also you can use to test, like the one that I said when message3 is empty you can force the responder to crash

WilliamTakeshi · 2026-03-18T14:36:28Z


    pub fn set_identity(&mut self, i: BytesP256ElemLen, cred_i: Credential) {
-        self.i = Some(i);
+        self.i = Some(InitiatorIdentity::StatStat { i });


do you need a set_identity for PSK? or its the default?

Yes, we still need set_identity for PSK. In the PSK case there is no P-256 i value to provide, but set_identity is still how the initiator records that it is using InitiatorIdentity::Psk and stores cred_i, which is required later when preparing message 3.

Address review feedback around PSK support - Modify lakers-c/initiator flow to hanlde both stat/stat and PSK. - Add initiator-side tests in lakers-c for stat/stat and PSK parse/verify behavior. - Add explicit has_id_cred_r in lakers-c for optional ID_CRED_R. - Update set_identity to take InitiatorIdentity, so PSK can set identity state without a stat/stat i value. - Add a full end-to-end PSK handshake test - Replace silent error drops with propagated EDHOCErrors. - Fix parse_message_3 / PSK parsing to stop assuming one-byte ID_CRED_PSK. - Add proper CBOR byte-string header encoding for lengths beyond 255.

- Add a dedicated PSK native C example and modified the Makefile. - Fix the lakers-c FFI EAD item conversion to respect the populated item count - Add a symmetric credential constructor to the C FFI so the PSK example -

Make the Python EDHOC bindings method-aware for both initiator and responder. The initiator binding can now be constructed for stat-stat or PSK, chooses the correct identity shape in verify_message_2, and parses credentials according to the selected EDHOC method. The responder binding now mirrors the Rust API more closely by making the static responder key optional: passing no `r` selects PSK, while providing `r` selects stat-stat. It also parses credentials according to the negotiated method during verification. Update the Python tests to the new responder constructor API

WilliamTakeshi

There is a binary committed on examples/lakers-c-native/lakers_c_native_psk. Probably committed by mistake.

Its getting really good :))

WilliamTakeshi · 2026-03-26T09:19:58Z

    #[new]
-    fn new(r: Vec<u8>, py: Python<'_>, cred_r: super::AutoCredential) -> PyResult<Self> {
+    #[pyo3(signature = (cred_r, r=None))]
+    fn new(py: Python<'_>, cred_r: super::AutoCredential, r: Option<Vec<u8>>) -> PyResult<Self> {


Here we are breaking the API because we are changing the order, we have to document this somewhere, because it will break people updating the Python API. I don't know if this is fine.

True, I did not think of this. Maybe we could make the arguments keyword only? This way postitional argument order is no longer relevant?

WilliamTakeshi · 2026-03-26T09:23:16Z

    #[new]
-    fn new(r: Vec<u8>, py: Python<'_>, cred_r: super::AutoCredential) -> PyResult<Self> {
+    #[pyo3(signature = (cred_r, r=None))]
+    fn new(py: Python<'_>, cred_r: super::AutoCredential, r: Option<Vec<u8>>) -> PyResult<Self> {


Also another issue is, Option<Vec<u8>> "represents" 3 states:

Not provided - None

Provided wo data - Some([])

Provided with data - Some([1,2])

Its not clear the difference between the user using the API with 1. or 2.

But I am fine keeping like this so we can merge PSK and thinking the APIs later :))

- Fixed encode_ciphertext_3a to allow lengths > 23 - Fixed parse_message_3 to decode CBOR bstr correctly - Added comments on psk.rs - lakers-python: removed Python FFI panics - lakers-c: removed duplicated method field from EdhocInitiator

malishav · 2026-05-20T14:26:45Z

Where are we at with this PR? I still see TODOs in the PR description. @ElsaLopez133 can you give us an update?

ElsaLopez133 · 2026-05-20T15:17:10Z

I believe the PR is ready to review. Here is a concise summary of what was done:

edhoc.rs is kept as the main EDHOC protocol file and depending on the method we branch to different functions that are stored in edhoc/statstat.rsandedhoc/psk.rs. Therefore, edhoc/statstat.rsandedhoc/psk.rs` contain only the parts that differ between the Static-Static and PSK methods.
These method-specific files return result structs back to edhoc.rs, so the central flow can continue without duplicating the shared protocol logic.
We store method-specific fields in enums such as WaitM3MethodSpecifics, ProcessingM2MethodSpecifics, ProcessedM2MethodSpecifics, and ProcessingM3MethodSpecifics. These enums allow the common state structs to carry only the extra data required by the selected method. The use of enums was chosen to avoid having Option values.

What would need to be done is to add the Responder logic in the C bindings, but this would be a separate PR I believe (as it is neither there for the stat-stat protocol flow in the main branch)

malishav · 2026-05-20T17:15:18Z

I believe the PR is ready to review. Here is a concise summary of what was done:

edhoc.rs is kept as the main EDHOC protocol file and depending on the method we branch to different functions that are stored in edhoc/statstat.rsandedhoc/psk.rs. Therefore, edhoc/statstat.rsandedhoc/psk.rs` contain only the parts that differ between the Static-Static and PSK methods.

These method-specific files return result structs back to edhoc.rs, so the central flow can continue without duplicating the shared protocol logic.

We store method-specific fields in enums such as WaitM3MethodSpecifics, ProcessingM2MethodSpecifics, ProcessedM2MethodSpecifics, and ProcessingM3MethodSpecifics. These enums allow the common state structs to carry only the extra data required by the selected method. The use of enums was chosen to avoid having Option values.

What would need to be done is to add the Responder logic in the C bindings, but this would be a separate PR I believe (as it is neither there for the stat-stat protocol flow in the main branch)

Can you update the PR description with this text? thanks

WilliamTakeshi

LGTM

@chrysn , when you have time, can you take a look? I think we can merge this one and slowly improve some parts

ElsaLopez133 added 14 commits March 12, 2026 14:27

Refactoring to support future methods

34e12dc

Added helper functions where method-states are used because of hax er…

65859e6

…rors

Added helper functions for all functions for clarity

c557ed2

For each function we match on state and create method-specific fiunctions. For example, we have r_prepare_message_2 and we match based on emthod to either use t_prepare_message_2_statstat or r_prepare_message_2_psk

Adding a separate submodule for _statstat functions

ded1c64

Replacing outer/inner states for method_specific entry

63781d0

We avoid repetition of shared-values among methods This affects ProcessingM2 and ProcessingM3

Removing method from state ProcessingM3 and ProcessingM2

086f95f

Removing method from state ProcessingM3 and ProcessingM2

f8da180

Adding enum in prepare_message_2 to define inputs based on method

Correcting verify_message_2 for future PSK

c9a7380

Modifying lakers-c

a1d2693

Using struct+union to replicate the enum style in Rust

Adding PSK method

0c8e163

Adding PSK specific functions Adding coap examples Adding edhoc/psk.rs module

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch 3 times, most recently from 1fd46c3 to aaaa14c Compare March 12, 2026 13:51

ElsaLopez133 added 2 commits March 12, 2026 15:47

Adding test vectors for PSK

02b948e

Fixing unused variables Fixing lakers-c and lakers-python Adding FIXME for error declaration Fixing coap examples Matching in edhoc.rs instead of edhoc/psk.rs Fix lakers-c: Pm2Psk defined twice

TEMP: PSA lot exhaustion fix to make tests work

d0192ec

(cherry pick from pr-psa-memory-error)

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch from 6499b4d to d0192ec Compare March 12, 2026 14:48

chrysn mentioned this pull request Mar 15, 2026

Refactoring to support future methods #406

Open

ElsaLopez133 added 2 commits March 16, 2026 17:38

WilliamTakeshi assigned ElsaLopez133 Mar 17, 2026

ElsaLopez133 marked this pull request as ready for review March 17, 2026 13:34

WilliamTakeshi reviewed Mar 18, 2026

View reviewed changes

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch 3 times, most recently from 3875b4d to e4943ee Compare March 18, 2026 16:58

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch 4 times, most recently from 16b9d13 to 65f53f8 Compare March 18, 2026 17:16

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch from 65f53f8 to 0755dc5 Compare March 18, 2026 17:22

ElsaLopez133 added 4 commits March 19, 2026 10:58

Adding test_handshake_psk in lakers-no_std

bfe7991

Fix lakers-c statstat verification by repsecting EAD item count

9e8bb4c

Add PSK native C example and fix lakers-c FFI state conversion

51ac9ad

- Add a dedicated PSK native C example and modified the Makefile. - Fix the lakers-c FFI EAD item conversion to respect the populated item count - Add a symmetric credential constructor to the C FFI so the PSK example -

ElsaLopez133 force-pushed the refactoring-adding-psk-method branch from eb89d72 to cb918be Compare March 20, 2026 12:44

WilliamTakeshi reviewed Mar 26, 2026

View reviewed changes

ElsaLopez133 added 3 commits April 5, 2026 21:43

Fixing review comments

64013ff

- Fixed encode_ciphertext_3a to allow lengths > 23 - Fixed parse_message_3 to decode CBOR bstr correctly - Added comments on psk.rs - lakers-python: removed Python FFI panics - lakers-c: removed duplicated method field from EdhocInitiator

Fixing wrong key in encrytp/decrypt message_3

86aeaa7

adding -psk coap examples

f7aad96

WilliamTakeshi approved these changes Jun 8, 2026

View reviewed changes

WilliamTakeshi requested a review from chrysn June 8, 2026 12:28

Uh oh!

Conversation

ElsaLopez133 commented Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

WilliamTakeshi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

WilliamTakeshi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

malishav commented May 20, 2026

Uh oh!

ElsaLopez133 commented May 20, 2026

Uh oh!

malishav commented May 20, 2026

Uh oh!

WilliamTakeshi left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ElsaLopez133 commented Mar 12, 2026 •

edited

Loading