Add fixed-array Scalar batch inversion (#789)

curve25519-dalek/CHANGELOG.md

@@ -5,6 +5,11 @@ major series.
 
 ## 5.x series
 
+## 5.0.0-pre.1
+
+* Rename `Scalar::batch_invert` -> `Scalar::invert_batch` for consistency. Also make it no-alloc.
+* Add an allocating batch inversion called `Scalar::invert_batch_alloc`.
+
 ## 5.0.0-pre.0
 
 * Update edition to 2024

curve25519-dalek/benches/dalek_benchmarks.rs

@@ -393,7 +393,7 @@ mod scalar_benches {
                         (0..size).map(|_| Scalar::random(&mut rng)).collect();
                     b.iter(|| {
                         let mut s = scalars.clone();
-                        Scalar::batch_invert(&mut s);
+                        Scalar::invert_batch_alloc(&mut s);
                     });
                 },
             );

curve25519-dalek/src/scalar.rs

@@ -770,7 +770,7 @@ impl Scalar {
     ///     Scalar::from(11u64),
     /// ];
     ///
-    /// let allinv = Scalar::batch_invert(&mut scalars);
+    /// let allinv = Scalar::invert_batch(&mut scalars);
     ///
     /// assert_eq!(allinv, Scalar::from(3*5*7*11u64).invert());
     /// assert_eq!(scalars[0], Scalar::from(3u64).invert());
@@ -779,8 +779,39 @@ impl Scalar {
     /// assert_eq!(scalars[3], Scalar::from(11u64).invert());
     /// # }
     /// ```
+    pub fn invert_batch<const N: usize>(inputs: &mut [Scalar; N]) -> Scalar {
+        let one: UnpackedScalar = Scalar::ONE.unpack().as_montgomery();
+
+        let mut scratch = [one; N];
+
+        Self::invert_batch_internal(inputs, &mut scratch)
+    }
+
+    /// Given a slice of nonzero (possibly secret) `Scalar`s, compute their inverses in a batch.
+    /// This the allocating form of [`Self::invert_batch`]. See those docs for examples.
+    ///
+    /// # Return
+    ///
+    /// Each element of `inputs` is replaced by its inverse.
+    ///
+    /// The product of all inverses is returned.
+    ///
+    /// # Warning
+    ///
+    /// All input `Scalars` **MUST** be nonzero.  If you cannot
+    /// *prove* that this is the case, you **SHOULD NOT USE THIS
+    /// FUNCTION**.
     #[cfg(feature = "alloc")]
-    pub fn batch_invert(inputs: &mut [Scalar]) -> Scalar {
+    pub fn invert_batch_alloc(inputs: &mut [Scalar]) -> Scalar {
+        let n = inputs.len();
+        let one: UnpackedScalar = Scalar::ONE.unpack().as_montgomery();
+
+        let mut scratch = vec![one; n];
+
+        Self::invert_batch_internal(inputs, &mut scratch)
+    }
+
+    fn invert_batch_internal(inputs: &mut [Scalar], scratch: &mut [UnpackedScalar]) -> Scalar {
         // This code is essentially identical to the FieldElement
         // implementation, and is documented there.  Unfortunately,
         // it's not easy to write it generically, since here we want
@@ -788,11 +819,6 @@ impl Scalar {
         // externally, but there's no corresponding distinction for
         // field elements.
 
-        let n = inputs.len();
-        let one: UnpackedScalar = Scalar::ONE.unpack().as_montgomery();
-
-        let mut scratch = vec![one; n];
-
         // Keep an accumulator of all of the previous products
         let mut acc = Scalar::ONE.unpack().as_montgomery();
 
@@ -826,7 +852,7 @@ impl Scalar {
         }
 
         #[cfg(feature = "zeroize")]
-        Zeroize::zeroize(&mut scratch);
+        Zeroize::zeroize(&mut scratch.iter_mut());
 
         ret
     }
@@ -1845,25 +1871,44 @@ pub(crate) mod test {
         assert_eq!(X, bincode::deserialize(X.as_bytes()).unwrap(),);
     }
 
-    #[cfg(all(debug_assertions, feature = "alloc"))]
+    #[cfg(debug_assertions)]
     #[test]
     #[should_panic]
-    fn batch_invert_with_a_zero_input_panics() {
-        let mut xs = vec![Scalar::ONE; 16];
+    fn invert_batch_with_a_zero_input_panics() {
+        let mut xs = [Scalar::ONE; 16];
         xs[3] = Scalar::ZERO;
         // This should panic in debug mode.
-        Scalar::batch_invert(&mut xs);
+        Scalar::invert_batch(&mut xs);
+    }
+
+    #[test]
+    fn invert_batch_empty() {
+        assert_eq!(Scalar::ONE, Scalar::invert_batch(&mut []));
+    }
+
+    #[test]
+    fn invert_batch_consistency() {
+        let mut x = Scalar::from(1u64);
+        let mut v1: [Scalar; 16] = core::array::from_fn(|_| {
+            let tmp = x;
+            x = x + x;
+            tmp
+        });
+        let v2 = v1;
+
+        let expected: Scalar = v1.iter().product();
+        let expected = expected.invert();
+        let ret = Scalar::invert_batch(&mut v1);
+        assert_eq!(ret, expected);
+
+        for (a, b) in v1.iter().zip(v2.iter()) {
+            assert_eq!(a * b, Scalar::ONE);
+        }
     }
 
     #[test]
     #[cfg(feature = "alloc")]
-    fn batch_invert_empty() {
-        assert_eq!(Scalar::ONE, Scalar::batch_invert(&mut []));
-    }
-
-    #[test]
-    #[cfg(feature = "alloc")]
-    fn batch_invert_consistency() {
+    fn batch_vec_invert_consistency() {
         let mut x = Scalar::from(1u64);
         let mut v1: Vec<_> = (0..16)
             .map(|_| {
@@ -1876,7 +1921,7 @@ pub(crate) mod test {
 
         let expected: Scalar = v1.iter().product();
         let expected = expected.invert();
-        let ret = Scalar::batch_invert(&mut v1);
+        let ret = Scalar::invert_batch_alloc(&mut v1);
         assert_eq!(ret, expected);
 
         for (a, b) in v1.iter().zip(v2.iter()) {