Add ECRecover support to the Math API #3921
Assignees
Labels
A-cryptography
Area: Cryptography
A-EVM
Area: Native EVM implementation and support
A-transaction-runtime
Area: transaction runtime (transaction and receipts processing, state transition, etc)
C-enhancement
Category: An issue proposing an enhancement or a PR with one.
T-Aurora
Team: issues relevant to the Aurora team
Comments
artob
added
C-enhancement
artob
added
the
T-public-interfaces
artob
added a commit
that referenced
this issue
Feb 17, 2021
artob
added a commit
that referenced
this issue
Feb 19, 2021
artob
added a commit
that referenced
this issue
Feb 19, 2021
joshuajbouw
pushed a commit
that referenced
this issue
Mar 23, 2021
joshuajbouw
pushed a commit
that referenced
this issue
Mar 23, 2021
joshuajbouw
pushed a commit
that referenced
this issue
Mar 26, 2021
joshuajbouw
pushed a commit
that referenced
this issue
Mar 26, 2021
joshuajbouw
pushed a commit
that referenced
this issue
May 20, 2021
joshuajbouw
pushed a commit
that referenced
this issue
May 20, 2021
5 tasks
bowenwang1996
added
T-Aurora
bowenwang1996
added a commit
that referenced
this issue
Jun 30, 2021
* feat: Extend the Math API with EVM precompiles. * feat: Add Math API stubs for ECRecover. (#3921) * test: Add RIPEMD-160 and BLAKE2b to the params estimator. * fix: Improve error output from the params estimator. Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com> * test: Add ECRecover stubs to the params estimator. * fix: Widen ecrecover() parameter type to pass CI checks. * feat(chain): Bump the protocol version. * feat(runtime): Implement ECRecover in the Math API. (#3921) * docs: Document the ecrecover() function. * fix(chain): Guard protocol upgrade behind EVM feature. * add crypto feature gates * fix nightly compile * cargo fmt * fix two ripemd160 missing feature gates * rename crypto_extras feature to protocol_feature equivalents * change repo to our git * put everything under protocol_feature_evm * point to alternative blake2 branch temporarily * Add blake2b f function * add blake2b_f to config * post rebase fixes * wrap all ext_costs in evm feature * add blake_2b_f to runtime estimator * cargo fmt * add new hash algos to feature * add vm-logic test for blake2b_f * fix derive formatting * use same blake2 library * change blake F tests to test 1 round * fix blake f test * cleanup calls_helper * add missing feature gated fields from ExtCostsConfig * add blake2b_f compression docs * remove unused `blake2b_f_byte` as sizes are fixed * add missing blake2b_f to imports * runtime estimator mod docs * Add near-blake2 lib to near-crypto * fix test imports * add more pub methods to blake2 * add blake2 state success tests * cargo fmt * Add initial blake2 logic to vm * Fix ecrecover * Change blake2 `update` to `update_inner` * Update params estimator * Remove duplicate evm feature * Remove pattern * Fix some compilation errors * Fix runtime-params-estimator compile * Fix tests * Reduce blake2s param estimator rounds from 12 to 10 * Add protocol_feature_evm feature to standalone runner * Add estimated gas values * Update estimated gas value note * cargo fmt * Update c2-chacha * Remove nightly features from lib blake2 * Add blake2 pass tests * Fix up ecrecover test * Fix blake2s args * Remove blake2 lib * Update documentation for blake2 * Pin comment for blake2 lib * Fix with_state * Return a bool for ecrecover, not abort * Fix VarBlake2b -> VarBlake2s * Add errors to RPC errors schema * Add bool return to wrapped imports * Take t0 and t1 for blake2b * Charge ripemd160 by blocks, not bytes * Correctly send both t0 and t1 to blake2 * Remove too many rounds error * Remove any near-blake2 errors * Remove hash data overflow error * Rename Blake2InvalidStateLength -> Blake2StateLengthExceeded * Small ecrecover clarifications * Return public key in ecrecover * Correct blake2 message block math * Correct ripemd160 message block math * Read blake2s as u32 not u64 * Remove `state_len` and blake2 len error * Add `saturating_mul` to blake2 math * Don't allocate an extra vec on blake2 * Remove unused import * Fix args in wrapped imports * Fix blake2s test * Remove blake2 and added protocol_feature_evm conditional comps * Remove optional from crypto imports in logic * Revert from near-blake2 to upstream blake2 * Drop libsecp256k1 crate in favour of near-crypto * Improve code quality by using U256 * Pass hash_bytes as ref * Revert "Pass hash_bytes as ref" This reverts commit 434145b. * matklad nit Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com> * Fix ecrecover exports arg sig * Move when malleability check happens * Remove last check in signature values Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com> * Don't use `v` in a check, now unused * Update description of ripemd160 with correct cost * Use u32 for malleability_flag * Introduce protocol_feature_math_extension * Use our own div_ceil implementation * Include protocol_feature_math_extension in runtime-params-estimator/nightly_protocol_features * Nightly protocol version bump * Simplify message_blocks computation * Remove unused div_ceil function * Ignore first byte of result * Remove uneeded return at end of function Co-authored-by: EgorKulikov <egor@egork.net> * Remove check_v and add in option to skip reject upper range * Use `get_vec_from_memory_or_register` over `memory_get_into` * Remove unused * Separate `v`, change args to `u64` * Remove redundant check * 64th byte, not 65 Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com> * Fix test * Update doc cost * Fix estimator * Small fixes * Add missing feature `protocol_feature_math_extension` from neard * Add near-primitives to math extension params estimator * Fix byte length * Fix Cargo.toml * Nit. Typos * Nit. param estimator is missing feature dependency * Add feature support for build.sh of test-contract * Add consts for SECP256K1 malleability values * Remove map err, change to unwrap * Change v check to something more simple * Always do signature check * Check one s bound * Allow for specifying a register * Remove math extension feature from primitives core * Nit. Convering byte to bool * Add measuring support. * Revert "Remove math extension feature from primitives core" This reverts commit ac53dcb. * Const SECP256K1 half + 1 * Revert "Add measuring support." This reverts commit 210ee55. * Update gas costs with correct values * Nit. Method signature typo causing runtime failure * Fix fee. It seems like a copy-paste typo * Undo the wrong fix * Add malleability_flag check as requested in #4380 (comment) * syntax fix Co-authored-by: Arto Bendiken <arto@near.org> Co-authored-by: Aleksey Kladov <aleksey.kladov@gmail.com> Co-authored-by: Michael Birch <michael@near.org> Co-authored-by: Michael Birch <michael.birch@aurora.dev> Co-authored-by: EgorKulikov <egor@egork.net> Co-authored-by: Maksym Zavershynskyi <max@near.org> Co-authored-by: Nikolay Igotti <igotti@gmail.com> Co-authored-by: Bowen Wang <bowenwang1996@users.noreply.github.com> Co-authored-by: Bowen Wang <bowen@near.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
As discussed with @ilblackdragon, @djsatok, and @frankbraun on 2021-02-04, to speed up and lower the cost of the EVM-as-a-contract implementation, we will need to add
ecrecover()as a Math API primitive so that the EVM contract itself will not need to implement this function. More broadly, this function will also be useful to any contracts dealing with Eth1/EVM interoperability.References:
The text was updated successfully, but these errors were encountered: