fix(persistent-tee): emit v1 report_data in mock-prove journal#74
Merged
Conversation
The piltover `feat/tee-persistent` branch landed the v1 report_data schema (piltover#16, "tee-config-hash-v1"): `validate_input` for `TeeInput` now recomputes a 10-field Poseidon over [version, mode, transition fields, katana_tee_config_hash] for the first 32 bytes of `report_data`, and decodes the second 32 bytes as `katana_tee_config_hash`. The mock-prove journal builder was still writing the legacy 7-field commitment with a zeroed second half, so every settlement attempt reverted at fee estimation with `tee: config hash half mismatch` and saya never submitted an `update_state` tx. - compute_appchain_commitment: hash the v1 10-field array (adds KATANA_TEE_REPORT_VERSION, KATANA_TEE_APPCHAIN_MODE, and katana_tee_config_hash). - build_raw_report: take config_hash and pack it into bytes [32..64] alongside the commitment in [0..32]. - prover.rs: thread `attestation.katana_tee_config_hash` into both helpers in the mock_prove branch. - Tests: replace the limb2/limb3-zero assertion with a both-halves round-trip; pin the inlined version-tag felt values to their ASCII source strings. Verified end-to-end against a live L2/L3 setup: saya-tee --mock-prove batched and settled blocks 2..=11 in a single update_state, advancing piltover's block_number to 0xb with execution_status SUCCEEDED. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
kariy
changed the title
report_data in mock-prove journal
4 tasks
…nt helper `compute_appchain_commitment` now takes 8 fields (was 7), tripping clippy's default threshold. Bundling into a struct would obscure the parity check against the on-chain Cairo call shape (`piltover` `validate_input` lines 198-207). Apply a localized `#[allow]` instead. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
kariy
added a commit
to dojoengine/katana
that referenced
this pull request
May 4, 2026
`dojoengine/saya#74` updated `saya-tee --mock-prove` to emit the v1 report_data layout this PR enforces. Without that bump the e2e job reverts at fee estimation with `tee: config hash half mismatch`. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
kariy
added a commit
to dojoengine/katana
that referenced
this pull request
May 4, 2026
…#556) `tee_generateQuote` responses now carry a `katana_tee_config_hash` — Pedersen-array hash of [`KatanaTeeConfig1`, chain_id, fee_token] — bound into both halves of the SEV-SNP `report_data`: report_data[0..32] = Poseidon([REPORT_VERSION, MODE, ...transition fields..., katana_tee_config_hash]) report_data[32..64] = katana_tee_config_hash (raw felt252) The hash is precomputed at `TeeApi::new` from the chain spec, with no caller input. The on-chain Piltover verifier recomputes it from its own config and rejects attestations from any other appchain configuration at verify time, instead of letting the misconfiguration surface later as a state-transition mismatch. The hash function mirrors the existing `compute_starknet_os_config_hash` (same Pedersen-array shape, same chain_id and STRK fee_token sources). Wire-format breaks (v1-only — legacy zero-config reports are dropped): - `tee_generateQuote(prev, block, Option<Felt>)` → `tee_generateQuote(prev, block)` - `--katana-tee-config-hash` removed from `katana rpc tee generate-quote` - Legacy 7-field appchain / 8-field sharding Poseidon shapes deleted Downstream coordination: - `dojoengine/saya#74` updates `saya-tee --mock-prove` to emit the matching v1 layout. Without it, every `update_state` reverts with `tee: config hash half mismatch`. - Piltover's matching v1 verification is in `cartridge-gg/piltover` on `feat/tee-persistent` (merged as `tee-config-hash-v1`).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
saya-tee --mock-proveis broken against the current piltoverfeat/tee-persistentbranch: every settlement attempt reverts at fee estimation withtee: config hash half mismatch, so saya never submits anupdate_statetx. This PR rebuilds the mock-prove journal under the v1 report_data schema piltover#16 introduced, restoring end-to-endsettlement.
Why
After piltover#16 (
tee-config-hash-v1, merged into thefeat/tee-persistentbranch),validate_inputforTeeInputchecks two halves of the 64-bytereport_data:Poseidon(['KatanaTeeReport1', 'KatanaTeeAppchain', prev_state_root, state_root, prev_block_hash, block_hash, prev_block_number, block_number, messages_commitment, katana_tee_config_hash])— 10 fields, including the version + mode tags and the config hash. - second 32 bytes must equaltee_input.katana_tee_config_hash.mock_proof.rswas still emitting the legacy 7-field Poseidon and a zero second half (the oldassert!(report_data.limb2 == 0 && report_data.limb3 == 0)shape), so:0in the second half but compares against the (non-zero)tee_input.katana_tee_config_hash— revert withTEE_CONFIG_HASH_HALF_MISMATCH.The fee-estimation revert is silent in saya logs beyond a single
Failed estimationline, which is why the symptom presented as "saya never sends a tx."