elliptic-curve: simplify BatchInvert trait#2455
Merged
Merged
Conversation
The previous `BatchInvert` trait was based on blanket impls bounded by `Field`, unfortunately `k256` needs to provide its own to handle normalization. The new version operates in-place requiring a temporary buffer, whereas the previous version return a value. This winds up making the generic implementation of something like batch normalization significantly simpler, since it can be in complete charge of the storage. It also means we don't need separate methods for arrays vs slices / `Vec`. I'm sure we discussed and considered that at one point in the past and I probably said no at the time, but this style of API makes the downstream code *significantly* simpler. Also, rather than blanket impls it has a generic provided implementation which can work everywhere except `k256`, which needs to do its own normalization.
tarcieri
force-pushed
the
batchinvert-redux
branch
from
94dee16 to
b3b04e0
Compare
tarcieri
added a commit
to RustCrypto/elliptic-curves
that referenced
this pull request
Jun 21, 2026
Companion PR to RustCrypto/traits#2455, which switches to in-place batch inversion. This massively simplifies the batch normalization implementation which no longer requires a litany of generic parameters to abstract over slices versus arrays because it accepts a temporary storage buffer the caller can allocate however they choose. It requires every curve add an `impl BatchInvert for FieldElement` but also lets curves provide their own implementation, which is needed for `k256` due to its lazy normalization.
tarcieri
added a commit
to RustCrypto/elliptic-curves
that referenced
this pull request
Jun 21, 2026
Companion PR to RustCrypto/traits#2455, which switches to in-place batch inversion. This massively simplifies the batch normalization implementation which no longer requires a litany of generic parameters to abstract over slices versus arrays because it accepts a temporary storage buffer the caller can allocate however they choose. It requires every curve add an `impl BatchInvert for FieldElement` but also lets curves provide their own implementation, which is needed for `k256` due to its lazy normalization.
tarcieri
added a commit
to RustCrypto/elliptic-curves
that referenced
this pull request
Jun 21, 2026
Companion PR to RustCrypto/traits#2455, which switches to in-place batch inversion. This massively simplifies the batch normalization implementation which no longer requires a litany of generic parameters to abstract over slices versus arrays because it accepts a temporary storage buffer the caller can allocate however they choose. It requires every curve add an `impl BatchInvert for FieldElement` but also lets curves provide their own implementation, which is needed for `k256` due to its lazy normalization.
tarcieri
added a commit
that referenced
this pull request
Jun 21, 2026
Instead of supplying our own generic implementation of Montgomery's trick, we can use the one in `ff` supplied by `BatchInverter`, which just so happens to have a signature that's very close to the one we switched to in #2455 (great minds think alike). The main difference between that PR and what `BatchInverter` provides is instead of returning a `Choice` in the event of zero elements, it ignores them and returns the inverse of the product of all non-zero field elements.
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.
The previous
BatchInverttrait was based on blanket impls bounded byField, unfortunatelyk256needs to provide its own to handle normalization.The new version operates in-place requiring a temporary buffer, whereas the previous version return a value. This winds up making the generic implementation of something like batch normalization significantly simpler, since it can be in complete charge of the storage. It also means we don't need separate methods for arrays vs slices /
Vec.I'm sure we discussed and considered that at one point in the past and I probably said no at the time, but this style of API makes the downstream code significantly simpler.
Also, rather than blanket impls it has a generic provided implementation which can work everywhere except
k256, which needs to do its own normalization.Companion PR: RustCrypto/elliptic-curves#1829