@qntx/nostr-native 0.0.0-stage → 0.12.1

package/vendor/secp256k1/src/ecmult_const_impl.h

@@ -0,0 +1,402 @@
+/***********************************************************************
+ * Copyright (c) 2015, 2022 Pieter Wuille, Andrew Poelstra             *
+ * Distributed under the MIT software license, see the accompanying    *
+ * file COPYING or https://www.opensource.org/licenses/mit-license.php.*
+ ***********************************************************************/
+
+#ifndef SECP256K1_ECMULT_CONST_IMPL_H
+#define SECP256K1_ECMULT_CONST_IMPL_H
+
+#include "scalar.h"
+#include "group.h"
+#include "ecmult_const.h"
+#include "ecmult_impl.h"
+
+#if defined(EXHAUSTIVE_TEST_ORDER)
+/* We need 2^ECMULT_CONST_GROUP_SIZE - 1 to be less than EXHAUSTIVE_TEST_ORDER, because
+ * the tables cannot have infinities in them (this breaks the effective-affine technique's
+ * z-ratio tracking) */
+#  if EXHAUSTIVE_TEST_ORDER == 199
+#    define ECMULT_CONST_GROUP_SIZE 4
+#  elif EXHAUSTIVE_TEST_ORDER == 13
+#    define ECMULT_CONST_GROUP_SIZE 3
+#  elif EXHAUSTIVE_TEST_ORDER == 7
+#    define ECMULT_CONST_GROUP_SIZE 2
+#  else
+#    error "Unknown EXHAUSTIVE_TEST_ORDER"
+#  endif
+#else
+/* Group size 4 or 5 appears optimal. */
+#  define ECMULT_CONST_GROUP_SIZE 5
+#endif
+
+#define ECMULT_CONST_TABLE_SIZE (1L << (ECMULT_CONST_GROUP_SIZE - 1))
+#define ECMULT_CONST_GROUPS ((129 + ECMULT_CONST_GROUP_SIZE - 1) / ECMULT_CONST_GROUP_SIZE)
+#define ECMULT_CONST_BITS (ECMULT_CONST_GROUPS * ECMULT_CONST_GROUP_SIZE)
+
+/** Fill a table 'pre' with precomputed odd multiples of a.
+ *
+ *  The resulting point set is brought to a single constant Z denominator, stores the X and Y
+ *  coordinates as ge points in pre, and stores the global Z in globalz.
+ *
+ *  'pre' must be an array of size ECMULT_CONST_TABLE_SIZE.
+ */
+static void nostr_native_secp256k1_ecmult_const_odd_multiples_table_globalz(nostr_native_secp256k1_ge *pre, nostr_native_secp256k1_fe *globalz, const nostr_native_secp256k1_gej *a) {
+    nostr_native_secp256k1_fe zr[ECMULT_CONST_TABLE_SIZE];
+
+    nostr_native_secp256k1_ecmult_odd_multiples_table(ECMULT_CONST_TABLE_SIZE, pre, zr, globalz, a);
+    nostr_native_secp256k1_ge_table_set_globalz(ECMULT_CONST_TABLE_SIZE, pre, zr);
+}
+
+/* Given a table 'pre' with odd multiples of a point, put in r the signed-bit multiplication of n with that point.
+ *
+ * For example, if ECMULT_CONST_GROUP_SIZE is 4, then pre is expected to contain 8 entries:
+ * [1*P, 3*P, 5*P, 7*P, 9*P, 11*P, 13*P, 15*P]. n is then expected to be a 4-bit integer (range 0-15), and its
+ * bits are interpreted as signs of powers of two to look up.
+ *
+ * For example, if n=4, which is 0100 in binary, which is interpreted as [- + - -], so the looked up value is
+ * [ -(2^3) + (2^2) - (2^1) - (2^0) ]*P = -7*P. Every valid n translates to an odd number in range [-15,15],
+ * which means we just need to look up one of the precomputed values, and optionally negate it.
+ */
+#define ECMULT_CONST_TABLE_GET_GE(r,pre,n) do { \
+    unsigned int m = 0; \
+    /* If the top bit of n is 0, we want the negation. */ \
+    volatile unsigned int negative = ((n) >> (ECMULT_CONST_GROUP_SIZE - 1)) ^ 1; \
+    /* Let n[i] be the i-th bit of n, then the index is
+     *     sum(cnot(n[i]) * 2^i, i=0..l-2)
+     * where cnot(b) = b if n[l-1] = 1 and 1 - b otherwise.
+     * For example, if n = 4, in binary 0100, the index is 3, in binary 011.
+     *
+     * Proof:
+     *     Let
+     *         x = sum((2*n[i] - 1)*2^i, i=0..l-1)
+     *           = 2*sum(n[i] * 2^i, i=0..l-1) - 2^l + 1
+     *     be the value represented by n.
+     *     The index is (x - 1)/2 if x > 0 and -(x + 1)/2 otherwise.
+     *     Case x > 0:
+     *         n[l-1] = 1
+     *         index = sum(n[i] * 2^i, i=0..l-1) - 2^(l-1)
+     *               = sum(n[i] * 2^i, i=0..l-2)
+     *     Case x <= 0:
+     *         n[l-1] = 0
+     *          index = -(2*sum(n[i] * 2^i, i=0..l-1) - 2^l + 2)/2
+     *                = 2^(l-1) - 1 - sum(n[i] * 2^i, i=0..l-1)
+     *                = sum((1 - n[i]) * 2^i, i=0..l-2)
+     */ \
+    unsigned int index = ((unsigned int)(-negative) ^ n) & ((1U << (ECMULT_CONST_GROUP_SIZE - 1)) - 1U); \
+    nostr_native_secp256k1_fe neg_y; \
+    VERIFY_CHECK((n) < (1U << ECMULT_CONST_GROUP_SIZE)); \
+    VERIFY_CHECK(index < (1U << (ECMULT_CONST_GROUP_SIZE - 1))); \
+    /* Unconditionally set r->x = (pre)[m].x and r->y = (pre)[m].y because it's either the correct one
+     * or will get replaced in the later iterations, this is needed to make sure `r` is initialized. */ \
+    nostr_native_secp256k1_ge_set_xy((r), &(pre)[m].x, &(pre)[m].y); \
+    for (m = 1; m < ECMULT_CONST_TABLE_SIZE; m++) { \
+        /* This loop is used to avoid secret data in array indices. See
+         * the comment in ecmult_gen_impl.h for rationale. */ \
+        nostr_native_secp256k1_fe_cmov(&(r)->x, &(pre)[m].x, m == index); \
+        nostr_native_secp256k1_fe_cmov(&(r)->y, &(pre)[m].y, m == index); \
+    } \
+    nostr_native_secp256k1_fe_negate(&neg_y, &(r)->y, 1); \
+    nostr_native_secp256k1_fe_cmov(&(r)->y, &neg_y, negative); \
+} while(0)
+
+/* For K as defined in the comment of nostr_native_secp256k1_ecmult_const, we have several precomputed
+ * formulas/constants.
+ * - in exhaustive test mode, we give an explicit expression to compute it at compile time: */
+#ifdef EXHAUSTIVE_TEST_ORDER
+static const nostr_native_secp256k1_scalar nostr_native_secp256k1_ecmult_const_K = ((SECP256K1_SCALAR_CONST(0, 0, 0, (1U << (ECMULT_CONST_BITS - 128)) - 2U, 0, 0, 0, 0) + EXHAUSTIVE_TEST_ORDER - 1U) * (1U + EXHAUSTIVE_TEST_LAMBDA)) % EXHAUSTIVE_TEST_ORDER;
+/* - for the real secp256k1 group we have constants for various ECMULT_CONST_BITS values. */
+#elif ECMULT_CONST_BITS == 129
+/* For GROUP_SIZE = 1,3. */
+static const nostr_native_secp256k1_scalar nostr_native_secp256k1_ecmult_const_K = SECP256K1_SCALAR_CONST(0xac9c52b3ul, 0x3fa3cf1ful, 0x5ad9e3fdul, 0x77ed9ba4ul, 0xa880b9fcul, 0x8ec739c2ul, 0xe0cfc810ul, 0xb51283ceul);
+#elif ECMULT_CONST_BITS == 130
+/* For GROUP_SIZE = 2,5. */
+static const nostr_native_secp256k1_scalar nostr_native_secp256k1_ecmult_const_K = SECP256K1_SCALAR_CONST(0xa4e88a7dul, 0xcb13034eul, 0xc2bdd6bful, 0x7c118d6bul, 0x589ae848ul, 0x26ba29e4ul, 0xb5c2c1dcul, 0xde9798d9ul);
+#elif ECMULT_CONST_BITS == 132
+/* For GROUP_SIZE = 4,6 */
+static const nostr_native_secp256k1_scalar nostr_native_secp256k1_ecmult_const_K = SECP256K1_SCALAR_CONST(0x76b1d93dul, 0x0fae3c6bul, 0x3215874bul, 0x94e93813ul, 0x7937fe0dul, 0xb66bcaaful, 0xb3749ca5ul, 0xd7b6171bul);
+#else
+#  error "Unknown ECMULT_CONST_BITS"
+#endif
+
+static void nostr_native_secp256k1_ecmult_const(nostr_native_secp256k1_gej *r, const nostr_native_secp256k1_ge *a, const nostr_native_secp256k1_scalar *q) {
+    /* The approach below combines the signed-digit logic from Mike Hamburg's
+     * "Fast and compact elliptic-curve cryptography" (https://eprint.iacr.org/2012/309)
+     * Section 3.3, with the GLV endomorphism.
+     *
+     * The idea there is to interpret the bits of a scalar as signs (1 = +, 0 = -), and compute a
+     * point multiplication in that fashion. Let v be an n-bit non-negative integer (0 <= v < 2^n),
+     * and v[i] its i'th bit (so v = sum(v[i] * 2^i, i=0..n-1)). Then define:
+     *
+     *   C_l(v, A) = sum((2*v[i] - 1) * 2^i*A, i=0..l-1)
+     *
+     * Then it holds that C_l(v, A) = sum((2*v[i] - 1) * 2^i*A, i=0..l-1)
+     *                              = (2*sum(v[i] * 2^i, i=0..l-1) + 1 - 2^l) * A
+     *                              = (2*v + 1 - 2^l) * A
+     *
+     * Thus, one can compute q*A as C_256((q + 2^256 - 1) / 2, A). This is the basis for the
+     * paper's signed-digit multi-comb algorithm for multiplication using a precomputed table.
+     *
+     * It is appealing to try to combine this with the GLV optimization: the idea that a scalar
+     * s can be written as s1 + lambda*s2, where lambda is a curve-specific constant such that
+     * lambda*A is easy to compute, and where s1 and s2 are small. In particular we have the
+     * nostr_native_secp256k1_scalar_split_lambda function which performs such a split with the resulting s1
+     * and s2 in range (-2^128, 2^128) mod n. This does work, but is uninteresting:
+     *
+     *   To compute q*A:
+     *   - Let s1, s2 = split_lambda(q)
+     *   - Let R1 = C_256((s1 + 2^256 - 1) / 2, A)
+     *   - Let R2 = C_256((s2 + 2^256 - 1) / 2, lambda*A)
+     *   - Return R1 + R2
+     *
+     * The issue is that while s1 and s2 are small-range numbers, (s1 + 2^256 - 1) / 2 (mod n)
+     * and (s2 + 2^256 - 1) / 2 (mod n) are not, undoing the benefit of the splitting.
+     *
+     * To make it work, we want to modify the input scalar q first, before splitting, and then only
+     * add a 2^128 offset of the split results (so that they end up in the single 129-bit range
+     * [0,2^129]). A slightly smaller offset would work due to the bounds on the split, but we pick
+     * 2^128 for simplicity. Let s be the scalar fed to split_lambda, and f(q) the function to
+     * compute it from q:
+     *
+     *   To compute q*A:
+     *   - Compute s = f(q)
+     *   - Let s1, s2 = split_lambda(s)
+     *   - Let v1 = s1 + 2^128 (mod n)
+     *   - Let v2 = s2 + 2^128 (mod n)
+     *   - Let R1 = C_l(v1, A)
+     *   - Let R2 = C_l(v2, lambda*A)
+     *   - Return R1 + R2
+     *
+     * l will thus need to be at least 129, but we may overshoot by a few bits (see
+     * further), so keep it as a variable.
+     *
+     * To solve for s, we reason:
+     *     q*A  = R1 + R2
+     * <=> q*A  = C_l(s1 + 2^128, A) + C_l(s2 + 2^128, lambda*A)
+     * <=> q*A  = (2*(s1 + 2^128) + 1 - 2^l) * A + (2*(s2 + 2^128) + 1 - 2^l) * lambda*A
+     * <=> q*A  = (2*(s1 + s2*lambda) + (2^129 + 1 - 2^l) * (1 + lambda)) * A
+     * <=> q    = 2*(s1 + s2*lambda) + (2^129 + 1 - 2^l) * (1 + lambda) (mod n)
+     * <=> q    = 2*s + (2^129 + 1 - 2^l) * (1 + lambda) (mod n)
+     * <=> s    = (q + (2^l - 2^129 - 1) * (1 + lambda)) / 2 (mod n)
+     * <=> f(q) = (q + K) / 2 (mod n)
+     *            where K = (2^l - 2^129 - 1)*(1 + lambda) (mod n)
+     *
+     * We will process the computation of C_l(v1, A) and C_l(v2, lambda*A) in groups of
+     * ECMULT_CONST_GROUP_SIZE, so we set l to the smallest multiple of ECMULT_CONST_GROUP_SIZE
+     * that is not less than 129; this equals ECMULT_CONST_BITS.
+     */
+
+    /* The offset to add to s1 and s2 to make them non-negative. Equal to 2^128. */
+    static const nostr_native_secp256k1_scalar S_OFFSET = SECP256K1_SCALAR_CONST(0, 0, 0, 1, 0, 0, 0, 0);
+    nostr_native_secp256k1_scalar s, v1, v2;
+    nostr_native_secp256k1_ge pre_a[ECMULT_CONST_TABLE_SIZE];
+    nostr_native_secp256k1_ge pre_a_lam[ECMULT_CONST_TABLE_SIZE];
+    nostr_native_secp256k1_fe global_z;
+    int group, i;
+
+    /* We're allowed to be non-constant time in the point, and the code below (in particular,
+     * nostr_native_secp256k1_ecmult_const_odd_multiples_table_globalz) cannot deal with infinity in a
+     * constant-time manner anyway. */
+    if (nostr_native_secp256k1_ge_is_infinity(a)) {
+        nostr_native_secp256k1_gej_set_infinity(r);
+        return;
+    }
+
+    /* Compute v1 and v2. */
+    nostr_native_secp256k1_scalar_add(&s, q, &nostr_native_secp256k1_ecmult_const_K);
+    nostr_native_secp256k1_scalar_half(&s, &s);
+    nostr_native_secp256k1_scalar_split_lambda(&v1, &v2, &s);
+    nostr_native_secp256k1_scalar_add(&v1, &v1, &S_OFFSET);
+    nostr_native_secp256k1_scalar_add(&v2, &v2, &S_OFFSET);
+
+#ifdef VERIFY
+    /* Verify that v1 and v2 are in range [0, 2^129-1]. */
+    for (i = 129; i < 256; ++i) {
+        VERIFY_CHECK(nostr_native_secp256k1_scalar_get_bits_limb32(&v1, i, 1) == 0);
+        VERIFY_CHECK(nostr_native_secp256k1_scalar_get_bits_limb32(&v2, i, 1) == 0);
+    }
+#endif
+
+    /* Calculate odd multiples of A and A*lambda.
+     * All multiples are brought to the same Z 'denominator', which is stored
+     * in global_z. Due to secp256k1' isomorphism we can do all operations pretending
+     * that the Z coordinate was 1, use affine addition formulae, and correct
+     * the Z coordinate of the result once at the end.
+     */
+    nostr_native_secp256k1_gej_set_ge(r, a);
+    nostr_native_secp256k1_ecmult_const_odd_multiples_table_globalz(pre_a, &global_z, r);
+    for (i = 0; i < ECMULT_CONST_TABLE_SIZE; i++) {
+        nostr_native_secp256k1_ge_mul_lambda(&pre_a_lam[i], &pre_a[i]);
+    }
+
+    /* Next, we compute r = C_l(v1, A) + C_l(v2, lambda*A).
+     *
+     * We proceed in groups of ECMULT_CONST_GROUP_SIZE bits, operating on that many bits
+     * at a time, from high in v1, v2 to low. Call these bits1 (from v1) and bits2 (from v2).
+     *
+     * Now note that ECMULT_CONST_TABLE_GET_GE(&t, pre_a, bits1) loads into t a point equal
+     * to C_{ECMULT_CONST_GROUP_SIZE}(bits1, A), and analogously for pre_lam_a / bits2.
+     * This means that all we need to do is add these looked up values together, multiplied
+     * by 2^(ECMULT_GROUP_SIZE * group).
+     */
+    for (group = ECMULT_CONST_GROUPS - 1; group >= 0; --group) {
+        /* Using the _var get_bits function is ok here, since it's only variable in offset and count, not in the scalar. */
+        unsigned int bits1 = nostr_native_secp256k1_scalar_get_bits_var(&v1, group * ECMULT_CONST_GROUP_SIZE, ECMULT_CONST_GROUP_SIZE);
+        unsigned int bits2 = nostr_native_secp256k1_scalar_get_bits_var(&v2, group * ECMULT_CONST_GROUP_SIZE, ECMULT_CONST_GROUP_SIZE);
+        nostr_native_secp256k1_ge t;
+        int j;
+
+        ECMULT_CONST_TABLE_GET_GE(&t, pre_a, bits1);
+        if (group == ECMULT_CONST_GROUPS - 1) {
+            /* Directly set r in the first iteration. */
+            nostr_native_secp256k1_gej_set_ge(r, &t);
+        } else {
+            /* Shift the result so far up. */
+            for (j = 0; j < ECMULT_CONST_GROUP_SIZE; ++j) {
+                nostr_native_secp256k1_gej_double(r, r);
+            }
+            nostr_native_secp256k1_gej_add_ge(r, r, &t);
+        }
+        ECMULT_CONST_TABLE_GET_GE(&t, pre_a_lam, bits2);
+        nostr_native_secp256k1_gej_add_ge(r, r, &t);
+    }
+
+    /* Map the result back to the secp256k1 curve from the isomorphic curve. */
+    nostr_native_secp256k1_fe_mul(&r->z, &r->z, &global_z);
+}
+
+static int nostr_native_secp256k1_ecmult_const_xonly(nostr_native_secp256k1_fe* r, const nostr_native_secp256k1_fe *n, const nostr_native_secp256k1_fe *d, const nostr_native_secp256k1_scalar *q, int known_on_curve) {
+
+    /* This algorithm is a generalization of Peter Dettman's technique for
+     * avoiding the square root in a random-basepoint x-only multiplication
+     * on a Weierstrass curve:
+     * https://mailarchive.ietf.org/arch/msg/cfrg/7DyYY6gg32wDgHAhgSb6XxMDlJA/
+     *
+     *
+     * === Background: the effective affine technique ===
+     *
+     * Let phi_u be the isomorphism that maps (x, y) on secp256k1 curve y^2 = x^3 + 7 to
+     * x' = u^2*x, y' = u^3*y on curve y'^2 = x'^3 + u^6*7. This new curve has the same order as
+     * the original (it is isomorphic), but moreover, has the same addition/doubling formulas, as
+     * the curve b=7 coefficient does not appear in those formulas (or at least does not appear in
+     * the formulas implemented in this codebase, both affine and Jacobian). See also Example 9.5.2
+     * in https://www.math.auckland.ac.nz/~sgal018/crypto-book/ch9.pdf.
+     *
+     * This means any linear combination of secp256k1 points can be computed by applying phi_u
+     * (with non-zero u) on all input points (including the generator, if used), computing the
+     * linear combination on the isomorphic curve (using the same group laws), and then applying
+     * phi_u^{-1} to get back to secp256k1.
+     *
+     * Switching to Jacobian coordinates, note that phi_u applied to (X, Y, Z) is simply
+     * (X, Y, Z/u). Thus, if we want to compute (X1, Y1, Z) + (X2, Y2, Z), with identical Z
+     * coordinates, we can use phi_Z to transform it to (X1, Y1, 1) + (X2, Y2, 1) on an isomorphic
+     * curve where the affine addition formula can be used instead.
+     * If (X3, Y3, Z3) = (X1, Y1) + (X2, Y2) on that curve, then our answer on secp256k1 is
+     * (X3, Y3, Z3*Z).
+     *
+     * This is the effective affine technique: if we have a linear combination of group elements
+     * to compute, and all those group elements have the same Z coordinate, we can simply pretend
+     * that all those Z coordinates are 1, perform the computation that way, and then multiply the
+     * original Z coordinate back in.
+     *
+     * The technique works on any a=0 short Weierstrass curve. It is possible to generalize it to
+     * other curves too, but there the isomorphic curves will have different 'a' coefficients,
+     * which typically does affect the group laws.
+     *
+     *
+     * === Avoiding the square root for x-only point multiplication ===
+     *
+     * In this function, we want to compute the X coordinate of q*(n/d, y), for
+     * y = sqrt((n/d)^3 + 7). Its negation would also be a valid Y coordinate, but by convention
+     * we pick whatever sqrt returns (which we assume to be a deterministic function).
+     *
+     * Let g = y^2*d^3 = n^3 + 7*d^3. This also means y = sqrt(g/d^3).
+     * Further let v = sqrt(d*g), which must exist as d*g = y^2*d^4 = (y*d^2)^2.
+     *
+     * The input point (n/d, y) also has Jacobian coordinates:
+     *
+     *     (n/d, y, 1)
+     *   = (n/d * v^2, y * v^3, v)
+     *   = (n/d * d*g, y * sqrt(d^3*g^3), v)
+     *   = (n/d * d*g, sqrt(y^2 * d^3*g^3), v)
+     *   = (n*g, sqrt(g/d^3 * d^3*g^3), v)
+     *   = (n*g, sqrt(g^4), v)
+     *   = (n*g, g^2, v)
+     *
+     * It is easy to verify that both (n*g, g^2, v) and its negation (n*g, -g^2, v) have affine X
+     * coordinate n/d, and this holds even when the square root function doesn't have a
+     * deterministic sign. We choose the (n*g, g^2, v) version.
+     *
+     * Now switch to the effective affine curve using phi_v, where the input point has coordinates
+     * (n*g, g^2). Compute (X, Y, Z) = q * (n*g, g^2) there.
+     *
+     * Back on secp256k1, that means q * (n*g, g^2, v) = (X, Y, v*Z). This last point has affine X
+     * coordinate X / (v^2*Z^2) = X / (d*g*Z^2). Determining the affine Y coordinate would involve
+     * a square root, but as long as we only care about the resulting X coordinate, no square root
+     * is needed anywhere in this computation.
+     */
+
+    nostr_native_secp256k1_fe g, i;
+    nostr_native_secp256k1_ge p;
+    nostr_native_secp256k1_gej rj;
+
+    /* Compute g = (n^3 + B*d^3). */
+    nostr_native_secp256k1_fe_sqr(&g, n);
+    nostr_native_secp256k1_fe_mul(&g, &g, n);
+    if (d) {
+        nostr_native_secp256k1_fe b;
+        VERIFY_CHECK(!nostr_native_secp256k1_fe_normalizes_to_zero(d));
+        nostr_native_secp256k1_fe_sqr(&b, d);
+        VERIFY_CHECK(SECP256K1_B <= 8); /* magnitude of b will be <= 8 after the next call */
+        nostr_native_secp256k1_fe_mul_int(&b, SECP256K1_B);
+        nostr_native_secp256k1_fe_mul(&b, &b, d);
+        nostr_native_secp256k1_fe_add(&g, &b);
+        if (!known_on_curve) {
+            /* We need to determine whether (n/d)^3 + 7 is square.
+             *
+             *     is_square((n/d)^3 + 7)
+             * <=> is_square(((n/d)^3 + 7) * d^4)
+             * <=> is_square((n^3 + 7*d^3) * d)
+             * <=> is_square(g * d)
+             */
+            nostr_native_secp256k1_fe c;
+            nostr_native_secp256k1_fe_mul(&c, &g, d);
+            if (!nostr_native_secp256k1_fe_is_square_var(&c)) return 0;
+        }
+    } else {
+        nostr_native_secp256k1_fe_add_int(&g, SECP256K1_B);
+        if (!known_on_curve) {
+            /* g at this point equals x^3 + 7. Test if it is square. */
+            if (!nostr_native_secp256k1_fe_is_square_var(&g)) return 0;
+        }
+    }
+
+    SECP256K1_FE_VERIFY_MAGNITUDE(&g, 2);
+
+    /* Compute base point P = (n*g, g^2), the effective affine version of
+     * (n*g, g^2, v), which has corresponding affine X coordinate n/d. */
+    {
+        nostr_native_secp256k1_fe x, y;
+        nostr_native_secp256k1_fe_mul(&x, &g, n);
+        nostr_native_secp256k1_fe_sqr(&y, &g);
+        nostr_native_secp256k1_ge_set_xy(&p, &x, &y);
+    }
+
+    /* Perform x-only EC multiplication of P with q. */
+    VERIFY_CHECK(!nostr_native_secp256k1_scalar_is_zero(q));
+    nostr_native_secp256k1_ecmult_const(&rj, &p, q);
+    VERIFY_CHECK(!nostr_native_secp256k1_gej_is_infinity(&rj));
+
+    /* The resulting (X, Y, Z) point on the effective-affine isomorphic curve corresponds to
+     * (X, Y, Z*v) on the secp256k1 curve. The affine version of that has X coordinate
+     * (X / (Z^2*d*g)). */
+    nostr_native_secp256k1_fe_sqr(&i, &rj.z);
+    nostr_native_secp256k1_fe_mul(&i, &i, &g);
+    if (d) nostr_native_secp256k1_fe_mul(&i, &i, d);
+    nostr_native_secp256k1_fe_inv(&i, &i);
+    nostr_native_secp256k1_fe_mul(r, &rj.x, &i);
+
+    return 1;
+}
+
+#endif /* SECP256K1_ECMULT_CONST_IMPL_H */

package/vendor/secp256k1/src/ecmult_gen.h

@@ -0,0 +1,145 @@
+/***********************************************************************
+ * Copyright (c) Pieter Wuille, Peter Dettman                          *
+ * Distributed under the MIT software license, see the accompanying    *
+ * file COPYING or https://www.opensource.org/licenses/mit-license.php.*
+ ***********************************************************************/
+
+#ifndef SECP256K1_ECMULT_GEN_H
+#define SECP256K1_ECMULT_GEN_H
+
+#include "hash.h"
+#include "scalar.h"
+#include "group.h"
+
+
+/* Configuration parameters for the signed-digit multi-comb algorithm:
+ *
+ * - COMB_BLOCKS is the number of blocks the input is split into. Each
+ *   has a corresponding table.
+ * - COMB_TEETH is the number of bits simultaneously covered by one table.
+ * - COMB_RANGE is the number of bits in supported scalars. For production
+ *   purposes, only 256 is reasonable, but smaller numbers are supported for
+ *   exhaustive test mode.
+ *
+ * The comb's spacing (COMB_SPACING), or the distance between the teeth,
+ * is defined as ceil(COMB_RANGE / (COMB_BLOCKS * COMB_TEETH)). Each block covers
+ * COMB_SPACING * COMB_TEETH consecutive bits in the input.
+ *
+ * The size of the precomputed table is COMB_BLOCKS * (1 << (COMB_TEETH - 1))
+ * nostr_native_secp256k1_ge_storages.
+ *
+ * The number of point additions equals COMB_BLOCKS * COMB_SPACING. Each point
+ * addition involves a cmov from (1 << (COMB_TEETH - 1)) table entries and a
+ * conditional negation.
+ *
+ * The number of point doublings is COMB_SPACING - 1. */
+
+#if defined(EXHAUSTIVE_TEST_ORDER)
+/* We need to control these values for exhaustive tests because
+ * the table cannot have infinities in them (nostr_native_secp256k1_ge_storage
+ * doesn't support infinities) */
+#  undef COMB_BLOCKS
+#  undef COMB_TEETH
+#  if EXHAUSTIVE_TEST_ORDER == 7
+#    define COMB_RANGE 3
+#    define COMB_BLOCKS 1
+#    define COMB_TEETH 2
+#  elif EXHAUSTIVE_TEST_ORDER == 13
+#    define COMB_RANGE 4
+#    define COMB_BLOCKS 1
+#    define COMB_TEETH 2
+#  elif EXHAUSTIVE_TEST_ORDER == 199
+#    define COMB_RANGE 8
+#    define COMB_BLOCKS 2
+#    define COMB_TEETH 3
+#  else
+#    error "Unknown exhaustive test order"
+#  endif
+#  if (COMB_RANGE >= 32) || ((EXHAUSTIVE_TEST_ORDER >> (COMB_RANGE - 1)) != 1)
+#    error "COMB_RANGE != ceil(log2(EXHAUSTIVE_TEST_ORDER+1))"
+#  endif
+#else /* !defined(EXHAUSTIVE_TEST_ORDER) */
+#  define COMB_RANGE 256
+#endif /* defined(EXHAUSTIVE_TEST_ORDER) */
+
+/* Use (11, 6) as default configuration, which results in a 22 kB table. */
+#ifndef COMB_BLOCKS
+#  define COMB_BLOCKS 11
+#  ifdef DEBUG_CONFIG
+#    pragma message DEBUG_CONFIG_MSG("COMB_BLOCKS undefined, assuming default value")
+#  endif
+#endif
+#ifndef COMB_TEETH
+#  define COMB_TEETH 6
+#  ifdef DEBUG_CONFIG
+#    pragma message DEBUG_CONFIG_MSG("COMB_TEETH undefined, assuming default value")
+#  endif
+#endif
+/* Use ceil(COMB_RANGE / (COMB_BLOCKS * COMB_TEETH)) as COMB_SPACING. */
+#define COMB_SPACING CEIL_DIV(COMB_RANGE, COMB_BLOCKS * COMB_TEETH)
+
+/* Range checks on the parameters. */
+
+/* The remaining COMB_* parameters are derived values, don't modify these. */
+/* - The number of bits covered by all the blocks; must be at least COMB_RANGE. */
+#define COMB_BITS (COMB_BLOCKS * COMB_TEETH * COMB_SPACING)
+/* - The number of entries per table. */
+#define COMB_POINTS (1 << (COMB_TEETH - 1))
+
+/* Sanity checks. */
+#if !(1 <= COMB_BLOCKS && COMB_BLOCKS <= 256)
+#  error "COMB_BLOCKS must be in the range [1, 256]"
+#endif
+#if !(1 <= COMB_TEETH && COMB_TEETH <= 8)
+#  error "COMB_TEETH must be in the range [1, 8]"
+#endif
+#if COMB_BITS < COMB_RANGE
+#  error "COMB_BLOCKS * COMB_TEETH * COMB_SPACING is too low"
+#endif
+
+/* These last 2 checks are not strictly required, but prevent gratuitously inefficient
+ * configurations. Note that they compare with 256 rather than COMB_RANGE, so they do
+ * permit somewhat excessive values for the exhaustive test case, where testing with
+ * suboptimal parameters may be desirable. */
+#if (COMB_BLOCKS - 1) * COMB_TEETH * COMB_SPACING >= 256
+#  error "COMB_BLOCKS can be reduced"
+#endif
+#if COMB_BLOCKS * (COMB_TEETH - 1) * COMB_SPACING >= 256
+#  error "COMB_TEETH can be reduced"
+#endif
+
+#ifdef DEBUG_CONFIG
+#  pragma message DEBUG_CONFIG_DEF(COMB_RANGE)
+#  pragma message DEBUG_CONFIG_DEF(COMB_BLOCKS)
+#  pragma message DEBUG_CONFIG_DEF(COMB_TEETH)
+#  pragma message DEBUG_CONFIG_DEF(COMB_SPACING)
+#endif
+
+typedef struct {
+    /* Whether the context has been built. */
+    int built;
+
+    /* Values chosen such that
+     *
+     *   n*G == comb(n + scalar_offset, G/2) + ge_offset.
+     *
+     * This expression lets us use scalar blinding and optimize the comb precomputation. See
+     * ecmult_gen_impl.h for more details. */
+    nostr_native_secp256k1_scalar scalar_offset;
+    nostr_native_secp256k1_ge ge_offset;
+
+    /* Factor used for projective blinding. This value is used to rescale the Z
+     * coordinate of the first table lookup. */
+    nostr_native_secp256k1_fe proj_blind;
+} nostr_native_secp256k1_ecmult_gen_context;
+
+static void nostr_native_secp256k1_ecmult_gen_context_build(nostr_native_secp256k1_ecmult_gen_context* ctx, const nostr_native_secp256k1_hash_ctx *hash_ctx);
+static void nostr_native_secp256k1_ecmult_gen_context_clear(nostr_native_secp256k1_ecmult_gen_context* ctx);
+
+/** Multiply with the generator: R = a*G */
+static void nostr_native_secp256k1_ecmult_gen_gej(const nostr_native_secp256k1_ecmult_gen_context* ctx, nostr_native_secp256k1_gej *r, const nostr_native_secp256k1_scalar *a);
+static void nostr_native_secp256k1_ecmult_gen_ge(const nostr_native_secp256k1_ecmult_gen_context* ctx, nostr_native_secp256k1_ge *r, const nostr_native_secp256k1_scalar *a);
+
+static void nostr_native_secp256k1_ecmult_gen_blind(nostr_native_secp256k1_ecmult_gen_context *ctx, const nostr_native_secp256k1_hash_ctx *hash_ctx, const unsigned char *seed32);
+
+#endif /* SECP256K1_ECMULT_GEN_H */

package/vendor/secp256k1/src/ecmult_gen_compute_table.h

@@ -0,0 +1,14 @@
+/***********************************************************************
+ * Copyright (c) Pieter Wuille, Gregory Maxwell                        *
+ * Distributed under the MIT software license, see the accompanying    *
+ * file COPYING or https://www.opensource.org/licenses/mit-license.php.*
+ ***********************************************************************/
+
+#ifndef SECP256K1_ECMULT_GEN_COMPUTE_TABLE_H
+#define SECP256K1_ECMULT_GEN_COMPUTE_TABLE_H
+
+#include "ecmult_gen.h"
+
+static void nostr_native_secp256k1_ecmult_gen_compute_table(nostr_native_secp256k1_ge_storage* table, const nostr_native_secp256k1_ge* gen, int blocks, int teeth, int spacing);
+
+#endif /* SECP256K1_ECMULT_GEN_COMPUTE_TABLE_H */

package/vendor/secp256k1/src/ecmult_gen_compute_table_impl.h

@@ -0,0 +1,108 @@
+/***********************************************************************
+ * Copyright (c) Pieter Wuille, Gregory Maxwell, Peter Dettman         *
+ * Distributed under the MIT software license, see the accompanying    *
+ * file COPYING or https://www.opensource.org/licenses/mit-license.php.*
+ ***********************************************************************/
+
+#ifndef SECP256K1_ECMULT_GEN_COMPUTE_TABLE_IMPL_H
+#define SECP256K1_ECMULT_GEN_COMPUTE_TABLE_IMPL_H
+
+#include "ecmult_gen_compute_table.h"
+#include "group_impl.h"
+#include "field_impl.h"
+#include "scalar_impl.h"
+#include "ecmult_gen.h"
+#include "util.h"
+
+static void nostr_native_secp256k1_ecmult_gen_compute_table(nostr_native_secp256k1_ge_storage* table, const nostr_native_secp256k1_ge* gen, int blocks, int teeth, int spacing) {
+    size_t points = ((size_t)1) << (teeth - 1);
+    size_t points_total = points * blocks;
+    nostr_native_secp256k1_ge* prec = checked_malloc(&default_error_callback, points_total * sizeof(*prec));
+    nostr_native_secp256k1_gej* ds = checked_malloc(&default_error_callback, teeth * sizeof(*ds));
+    nostr_native_secp256k1_gej* vs = checked_malloc(&default_error_callback, points_total * sizeof(*vs));
+    nostr_native_secp256k1_gej u;
+    size_t vs_pos = 0;
+    nostr_native_secp256k1_scalar half;
+    int block, i;
+
+    VERIFY_CHECK(points_total > 0);
+
+    /* u is the running power of two times gen we're working with, initially gen/2. */
+    nostr_native_secp256k1_scalar_half(&half, &nostr_native_secp256k1_scalar_one);
+    nostr_native_secp256k1_gej_set_infinity(&u);
+    for (i = 255; i >= 0; --i) {
+        /* Use a very simple multiplication ladder to avoid dependency on ecmult. */
+        nostr_native_secp256k1_gej_double_var(&u, &u, NULL);
+        if (nostr_native_secp256k1_scalar_get_bits_limb32(&half, i, 1)) {
+            nostr_native_secp256k1_gej_add_ge_var(&u, &u, gen, NULL);
+        }
+    }
+#ifdef VERIFY
+    {
+        /* Verify that u*2 = gen. */
+        nostr_native_secp256k1_gej double_u;
+        nostr_native_secp256k1_gej_double_var(&double_u, &u, NULL);
+        VERIFY_CHECK(nostr_native_secp256k1_gej_eq_ge_var(&double_u, gen));
+    }
+#endif
+
+    for (block = 0; block < blocks; ++block) {
+        int tooth;
+        /* Here u = 2^(block*teeth*spacing) * gen/2. */
+        nostr_native_secp256k1_gej sum;
+        nostr_native_secp256k1_gej_set_infinity(&sum);
+        for (tooth = 0; tooth < teeth; ++tooth) {
+            /* Here u = 2^((block*teeth + tooth)*spacing) * gen/2. */
+            /* Make sum = sum(2^((block*teeth + t)*spacing), t=0..tooth) * gen/2. */
+            nostr_native_secp256k1_gej_add_var(&sum, &sum, &u, NULL);
+            /* Make u = 2^((block*teeth + tooth)*spacing + 1) * gen/2. */
+            nostr_native_secp256k1_gej_double_var(&u, &u, NULL);
+            /* Make ds[tooth] = u = 2^((block*teeth + tooth)*spacing + 1) * gen/2. */
+            ds[tooth] = u;
+            /* Make u = 2^((block*teeth + tooth + 1)*spacing) * gen/2, unless at the end. */
+            if (block + tooth != blocks + teeth - 2) {
+                int bit_off;
+                for (bit_off = 1; bit_off < spacing; ++bit_off) {
+                    nostr_native_secp256k1_gej_double_var(&u, &u, NULL);
+                }
+            }
+        }
+        /* Now u = 2^((block*teeth + teeth)*spacing) * gen/2
+         *       = 2^((block+1)*teeth*spacing) * gen/2       */
+
+        /* Next, compute the table entries for block number block in Jacobian coordinates.
+         * The entries will occupy vs[block*points + i] for i=0..points-1.
+         * We start by computing the first (i=0) value corresponding to all summed
+         * powers of two times G being negative. */
+        nostr_native_secp256k1_gej_neg(&vs[vs_pos++], &sum);
+        /* And then teeth-1 times "double" the range of i values for which the table
+         * is computed: in each iteration, double the table by taking an existing
+         * table entry and adding ds[tooth]. */
+        for (tooth = 0; tooth < teeth - 1; ++tooth) {
+            size_t stride = ((size_t)1) << tooth;
+            size_t index;
+            for (index = 0; index < stride; ++index, ++vs_pos) {
+                nostr_native_secp256k1_gej_add_var(&vs[vs_pos], &vs[vs_pos - stride], &ds[tooth], NULL);
+            }
+        }
+    }
+    VERIFY_CHECK(vs_pos == points_total);
+
+    /* Convert all points simultaneously from nostr_native_secp256k1_gej to nostr_native_secp256k1_ge. */
+    nostr_native_secp256k1_ge_set_all_gej_var(prec, vs, points_total);
+    /* Convert all points from nostr_native_secp256k1_ge to nostr_native_secp256k1_ge_storage output. */
+    for (block = 0; block < blocks; ++block) {
+        size_t index;
+        for (index = 0; index < points; ++index) {
+            VERIFY_CHECK(!nostr_native_secp256k1_ge_is_infinity(&prec[block * points + index]));
+            nostr_native_secp256k1_ge_to_storage(&table[block * points + index], &prec[block * points + index]);
+        }
+    }
+
+    /* Free memory. */
+    free(vs);
+    free(ds);
+    free(prec);
+}
+
+#endif /* SECP256K1_ECMULT_GEN_COMPUTE_TABLE_IMPL_H */