ring-native 0.0.0

data/vendor/ring/crypto/ec/util-64.c

@@ -0,0 +1,183 @@
+/* Copyright (c) 2015, Google Inc.
+ *
+ * Permission to use, copy, modify, and/or distribute this software for any
+ * purpose with or without fee is hereby granted, provided that the above
+ * copyright notice and this permission notice appear in all copies.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+ * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+ * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
+ * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+ * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
+ * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
+ * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
+
+#include <openssl/base.h>
+
+
+#if defined(OPENSSL_64_BIT) && !defined(OPENSSL_WINDOWS)
+
+#include <openssl/ec.h>
+
+#include "internal.h"
+
+/* Convert an array of points into affine coordinates. (If the point at
+ * infinity is found (Z = 0), it remains unchanged.) This function is
+ * essentially an equivalent to EC_POINTs_make_affine(), but works with the
+ * internal representation of points as used by ecp_nistp###.c rather than
+ * with (BIGNUM-based) EC_POINT data structures. point_array is the
+ * input/output buffer ('num' points in projective form, i.e. three
+ * coordinates each), based on an internal representation of field elements
+ * of size 'felem_size'. tmp_felems needs to point to a temporary array of
+ * 'num'+1 field elements for storage of intermediate values. */
+void ec_GFp_nistp_points_make_affine_internal(
+    size_t num, void *point_array, size_t felem_size, void *tmp_felems,
+    void (*felem_one)(void *out), int (*felem_is_zero)(const void *in),
+    void (*felem_assign)(void *out, const void *in),
+    void (*felem_square)(void *out, const void *in),
+    void (*felem_mul)(void *out, const void *in1, const void *in2),
+    void (*felem_inv)(void *out, const void *in),
+    void (*felem_contract)(void *out, const void *in)) {
+  int i = 0;
+
+#define tmp_felem(I) (&((char *)tmp_felems)[(I)*felem_size])
+#define X(I) (&((char *)point_array)[3 * (I)*felem_size])
+#define Y(I) (&((char *)point_array)[(3 * (I) + 1) * felem_size])
+#define Z(I) (&((char *)point_array)[(3 * (I) + 2) * felem_size])
+
+  if (!felem_is_zero(Z(0))) {
+    felem_assign(tmp_felem(0), Z(0));
+  } else {
+    felem_one(tmp_felem(0));
+  }
+
+  for (i = 1; i < (int)num; i++) {
+    if (!felem_is_zero(Z(i))) {
+      felem_mul(tmp_felem(i), tmp_felem(i - 1), Z(i));
+    } else {
+      felem_assign(tmp_felem(i), tmp_felem(i - 1));
+    }
+  }
+  /* Now each tmp_felem(i) is the product of Z(0) .. Z(i), skipping any
+   * zero-valued factors: if Z(i) = 0, we essentially pretend that Z(i) = 1. */
+
+  felem_inv(tmp_felem(num - 1), tmp_felem(num - 1));
+  for (i = num - 1; i >= 0; i--) {
+    if (i > 0) {
+      /* tmp_felem(i-1) is the product of Z(0) .. Z(i-1), tmp_felem(i)
+       * is the inverse of the product of Z(0) .. Z(i). */
+      /* 1/Z(i) */
+      felem_mul(tmp_felem(num), tmp_felem(i - 1), tmp_felem(i));
+    } else {
+      felem_assign(tmp_felem(num), tmp_felem(0)); /* 1/Z(0) */
+    }
+
+    if (!felem_is_zero(Z(i))) {
+      if (i > 0) {
+        /* For next iteration, replace tmp_felem(i-1) by its inverse. */
+        felem_mul(tmp_felem(i - 1), tmp_felem(i), Z(i));
+      }
+
+      /* Convert point (X, Y, Z) into affine form (X/(Z^2), Y/(Z^3), 1). */
+      felem_square(Z(i), tmp_felem(num));    /* 1/(Z^2) */
+      felem_mul(X(i), X(i), Z(i));           /* X/(Z^2) */
+      felem_mul(Z(i), Z(i), tmp_felem(num)); /* 1/(Z^3) */
+      felem_mul(Y(i), Y(i), Z(i));           /* Y/(Z^3) */
+      felem_contract(X(i), X(i));
+      felem_contract(Y(i), Y(i));
+      felem_one(Z(i));
+    } else {
+      if (i > 0) {
+        /* For next iteration, replace tmp_felem(i-1) by its inverse. */
+        felem_assign(tmp_felem(i - 1), tmp_felem(i));
+      }
+    }
+  }
+}
+
+/* This function looks at 5+1 scalar bits (5 current, 1 adjacent less
+ * significant bit), and recodes them into a signed digit for use in fast point
+ * multiplication: the use of signed rather than unsigned digits means that
+ * fewer points need to be precomputed, given that point inversion is easy (a
+ * precomputed point dP makes -dP available as well).
+ *
+ * BACKGROUND:
+ *
+ * Signed digits for multiplication were introduced by Booth ("A signed binary
+ * multiplication technique", Quart. Journ. Mech. and Applied Math., vol. IV,
+ * pt. 2 (1951), pp. 236-240), in that case for multiplication of integers.
+ * Booth's original encoding did not generally improve the density of nonzero
+ * digits over the binary representation, and was merely meant to simplify the
+ * handling of signed factors given in two's complement; but it has since been
+ * shown to be the basis of various signed-digit representations that do have
+ * further advantages, including the wNAF, using the following general
+ * approach:
+ *
+ * (1) Given a binary representation
+ *
+ *       b_k  ...  b_2  b_1  b_0,
+ *
+ *     of a nonnegative integer (b_k in {0, 1}), rewrite it in digits 0, 1, -1
+ *     by using bit-wise subtraction as follows:
+ *
+ *        b_k b_(k-1)  ...  b_2  b_1  b_0
+ *      -     b_k      ...  b_3  b_2  b_1  b_0
+ *       -------------------------------------
+ *        s_k b_(k-1)  ...  s_3  s_2  s_1  s_0
+ *
+ *     A left-shift followed by subtraction of the original value yields a new
+ *     representation of the same value, using signed bits s_i = b_(i+1) - b_i.
+ *     This representation from Booth's paper has since appeared in the
+ *     literature under a variety of different names including "reversed binary
+ *     form", "alternating greedy expansion", "mutual opposite form", and
+ *     "sign-alternating {+-1}-representation".
+ *
+ *     An interesting property is that among the nonzero bits, values 1 and -1
+ *     strictly alternate.
+ *
+ * (2) Various window schemes can be applied to the Booth representation of
+ *     integers: for example, right-to-left sliding windows yield the wNAF
+ *     (a signed-digit encoding independently discovered by various researchers
+ *     in the 1990s), and left-to-right sliding windows yield a left-to-right
+ *     equivalent of the wNAF (independently discovered by various researchers
+ *     around 2004).
+ *
+ * To prevent leaking information through side channels in point multiplication,
+ * we need to recode the given integer into a regular pattern: sliding windows
+ * as in wNAFs won't do, we need their fixed-window equivalent -- which is a few
+ * decades older: we'll be using the so-called "modified Booth encoding" due to
+ * MacSorley ("High-speed arithmetic in binary computers", Proc. IRE, vol. 49
+ * (1961), pp. 67-91), in a radix-2^5 setting.  That is, we always combine five
+ * signed bits into a signed digit:
+ *
+ *       s_(4j + 4) s_(4j + 3) s_(4j + 2) s_(4j + 1) s_(4j)
+ *
+ * The sign-alternating property implies that the resulting digit values are
+ * integers from -16 to 16.
+ *
+ * Of course, we don't actually need to compute the signed digits s_i as an
+ * intermediate step (that's just a nice way to see how this scheme relates
+ * to the wNAF): a direct computation obtains the recoded digit from the
+ * six bits b_(4j + 4) ... b_(4j - 1).
+ *
+ * This function takes those five bits as an integer (0 .. 63), writing the
+ * recoded digit to *sign (0 for positive, 1 for negative) and *digit (absolute
+ * value, in the range 0 .. 8).  Note that this integer essentially provides the
+ * input bits "shifted to the left" by one position: for example, the input to
+ * compute the least significant recoded digit, given that there's no bit b_-1,
+ * has to be b_4 b_3 b_2 b_1 b_0 0. */
+void ec_GFp_nistp_recode_scalar_bits(uint8_t *sign, uint8_t *digit,
+                                     uint8_t in) {
+  uint8_t s, d;
+
+  s = ~((in >> 5) - 1); /* sets all bits to MSB(in), 'in' seen as
+                          * 6-bit value */
+  d = (1 << 6) - in - 1;
+  d = (d & s) | (in & ~s);
+  d = (d >> 1) + (d & 1);
+
+  *sign = s & 1;
+  *digit = d;
+}
+
+#endif  /* 64_BIT && !WINDOWS */

data/vendor/ring/crypto/ec/wnaf.c

@@ -0,0 +1,508 @@
+/* Originally written by Bodo Moeller for the OpenSSL project.
+ * ====================================================================
+ * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in
+ *    the documentation and/or other materials provided with the
+ *    distribution.
+ *
+ * 3. All advertising materials mentioning features or use of this
+ *    software must display the following acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
+ * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+/* ====================================================================
+ * Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
+ *
+ * Portions of the attached software ("Contribution") are developed by
+ * SUN MICROSYSTEMS, INC., and are contributed to the OpenSSL project.
+ *
+ * The Contribution is licensed pursuant to the OpenSSL open source
+ * license provided above.
+ *
+ * The elliptic curve binary polynomial software is originally written by
+ * Sheueling Chang Shantz and Douglas Stebila of Sun Microsystems
+ * Laboratories. */
+
+#include <openssl/ec.h>
+
+#include <string.h>
+
+#include <openssl/bn.h>
+#include <openssl/err.h>
+#include <openssl/mem.h>
+#include <openssl/thread.h>
+
+#include "internal.h"
+#include "../internal.h"
+
+
+/* This file implements the wNAF-based interleaving multi-exponentation method
+ * (<URL:http://www.informatik.tu-darmstadt.de/TI/Mitarbeiter/moeller.html#multiexp>);
+ * */
+
+/* Determine the modified width-(w+1) Non-Adjacent Form (wNAF) of 'scalar'.
+ * This is an array  r[]  of values that are either zero or odd with an
+ * absolute value less than  2^w  satisfying
+ *     scalar = \sum_j r[j]*2^j
+ * where at most one of any  w+1  consecutive digits is non-zero
+ * with the exception that the most significant digit may be only
+ * w-1 zeros away from that next non-zero digit.
+ */
+static signed char *compute_wNAF(const BIGNUM *scalar, int w, size_t *ret_len) {
+  int window_val;
+  int ok = 0;
+  signed char *r = NULL;
+  int sign = 1;
+  int bit, next_bit, mask;
+  size_t len = 0, j;
+
+  if (BN_is_zero(scalar)) {
+    r = OPENSSL_malloc(1);
+    if (!r) {
+      OPENSSL_PUT_ERROR(EC, ERR_R_MALLOC_FAILURE);
+      goto err;
+    }
+    r[0] = 0;
+    *ret_len = 1;
+    return r;
+  }
+
+  if (w <= 0 || w > 7) /* 'signed char' can represent integers with absolute
+                          values less than 2^7 */
+  {
+    OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+    goto err;
+  }
+  bit = 1 << w;        /* at most 128 */
+  next_bit = bit << 1; /* at most 256 */
+  mask = next_bit - 1; /* at most 255 */
+
+  if (BN_is_negative(scalar)) {
+    sign = -1;
+  }
+
+  if (scalar->d == NULL || scalar->top == 0) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+    goto err;
+  }
+
+  len = BN_num_bits(scalar);
+  r = OPENSSL_malloc(
+      len +
+      1); /* modified wNAF may be one digit longer than binary representation
+           * (*ret_len will be set to the actual length, i.e. at most
+           * BN_num_bits(scalar) + 1) */
+  if (r == NULL) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_MALLOC_FAILURE);
+    goto err;
+  }
+  window_val = scalar->d[0] & mask;
+  j = 0;
+  while ((window_val != 0) ||
+         (j + w + 1 < len)) /* if j+w+1 >= len, window_val will not increase */
+  {
+    int digit = 0;
+
+    /* 0 <= window_val <= 2^(w+1) */
+
+    if (window_val & 1) {
+      /* 0 < window_val < 2^(w+1) */
+
+      if (window_val & bit) {
+        digit = window_val - next_bit; /* -2^w < digit < 0 */
+
+#if 1 /* modified wNAF */
+        if (j + w + 1 >= len) {
+          /* special case for generating modified wNAFs:
+           * no new bits will be added into window_val,
+           * so using a positive digit here will decrease
+           * the total length of the representation */
+
+          digit = window_val & (mask >> 1); /* 0 < digit < 2^w */
+        }
+#endif
+      } else {
+        digit = window_val; /* 0 < digit < 2^w */
+      }
+
+      if (digit <= -bit || digit >= bit || !(digit & 1)) {
+        OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+        goto err;
+      }
+
+      window_val -= digit;
+
+      /* now window_val is 0 or 2^(w+1) in standard wNAF generation;
+       * for modified window NAFs, it may also be 2^w
+       */
+      if (window_val != 0 && window_val != next_bit && window_val != bit) {
+        OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+        goto err;
+      }
+    }
+
+    r[j++] = sign * digit;
+
+    window_val >>= 1;
+    window_val += bit * BN_is_bit_set(scalar, j + w);
+
+    if (window_val > next_bit) {
+      OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+      goto err;
+    }
+  }
+
+  if (j > len + 1) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+    goto err;
+  }
+  len = j;
+  ok = 1;
+
+err:
+  if (!ok) {
+    OPENSSL_free(r);
+    r = NULL;
+  }
+  if (ok) {
+    *ret_len = len;
+  }
+  return r;
+}
+
+
+/* TODO: table should be optimised for the wNAF-based implementation,
+ *       sometimes smaller windows will give better performance
+ *       (thus the boundaries should be increased)
+ */
+#define EC_window_bits_for_scalar_size(b)                                      \
+  ((size_t)((b) >= 2000 ? 6 : (b) >= 800 ? 5 : (b) >= 300                      \
+                                                   ? 4                         \
+                                                   : (b) >= 70 ? 3 : (b) >= 20 \
+                                                                         ? 2   \
+                                                                         : 1))
+
+
+int ec_wNAF_mul_private(const EC_GROUP *group, EC_POINT *r,
+                        const BIGNUM *g_scalar, const EC_POINT *p,
+                        const BIGNUM *p_scalar, BN_CTX *ctx) {
+  BN_CTX *new_ctx = NULL;;
+  int ret = 0;
+
+  /* We do not want timing information to leak the length of the scalars so we
+   * use equivalent scalars of fixed bit-length. */
+
+  if (ctx == NULL) {
+    new_ctx = BN_CTX_new();
+    if (new_ctx == NULL) {
+      return 0;
+    }
+    ctx = new_ctx;
+  }
+  BN_CTX_start(ctx);
+
+  BIGNUM *g_scalar_new = NULL;
+  if (g_scalar) {
+    g_scalar_new = BN_CTX_get(ctx);
+    if (g_scalar_new == NULL ||
+        !BN_add(g_scalar_new, g_scalar, &group->order)) {
+      goto err;
+    }
+    if (BN_num_bits(g_scalar_new) <= BN_num_bits(&group->order)) {
+      if (!BN_add(g_scalar_new, g_scalar_new, &group->order)) {
+        goto err;
+      }
+    }
+  }
+
+  BIGNUM *p_scalar_new = NULL;
+  if (p_scalar) {
+    p_scalar_new = BN_CTX_get(ctx);
+    if (!p_scalar_new ||
+        !BN_add(p_scalar_new, p_scalar, &group->order)) {
+      goto err;
+    }
+    if (BN_num_bits(p_scalar_new) <= BN_num_bits(&group->order)) {
+      if (!BN_add(p_scalar_new, p_scalar_new, &group->order)) {
+        goto err;
+      }
+    }
+  }
+
+  ret = ec_wNAF_mul_public(group, r, g_scalar_new, p, p_scalar_new, ctx);
+
+err:
+  BN_CTX_end(ctx);
+  BN_CTX_free(new_ctx);
+  return ret;
+}
+
+int ec_wNAF_mul_public(const EC_GROUP *group, EC_POINT *r,
+                       const BIGNUM *g_scalar, const EC_POINT *p,
+                       const BIGNUM *p_scalar, BN_CTX *ctx) {
+  BN_CTX *new_ctx = NULL;
+  const EC_POINT *generator = NULL;
+  EC_POINT *tmp = NULL;
+  size_t total_num;
+  size_t i, j;
+  int k;
+  int r_is_inverted = 0;
+  int r_is_at_infinity = 1;
+  size_t *wsize = NULL;      /* individual window sizes */
+  signed char **wNAF = NULL; /* individual wNAFs */
+  size_t *wNAF_len = NULL;
+  size_t max_len = 0;
+  size_t num_val;
+  EC_POINT **val = NULL; /* precomputation */
+  EC_POINT **v;
+  EC_POINT ***val_sub = NULL; /* pointers to sub-arrays of 'val' */
+  int ret = 0;
+
+  if (ctx == NULL) {
+    ctx = new_ctx = BN_CTX_new();
+    if (ctx == NULL) {
+      goto err;
+    }
+  }
+
+  /* TODO: This function used to take |points| and |scalars| as arrays of
+   * |num| elements. The code below should be simplified to work in terms of |p|
+   * and |p_scalar|. */
+  size_t num = p != NULL ? 1 : 0;
+  const EC_POINT **points = p != NULL ? &p : NULL;
+  const BIGNUM **scalars = p != NULL ? &p_scalar : NULL;
+
+  total_num = num;
+
+  if (g_scalar != NULL) {
+    generator = EC_GROUP_get0_generator(group);
+    if (generator == NULL) {
+      OPENSSL_PUT_ERROR(EC, EC_R_UNDEFINED_GENERATOR);
+      goto err;
+    }
+
+    ++total_num; /* treat 'g_scalar' like 'num'-th element of 'scalars' */
+  }
+
+
+  wsize = OPENSSL_malloc(total_num * sizeof wsize[0]);
+  wNAF_len = OPENSSL_malloc(total_num * sizeof wNAF_len[0]);
+  wNAF = OPENSSL_malloc((total_num + 1) *
+                        sizeof wNAF[0]); /* includes space for pivot */
+  val_sub = OPENSSL_malloc(total_num * sizeof val_sub[0]);
+
+  /* Ensure wNAF is initialised in case we end up going to err. */
+  if (wNAF) {
+    wNAF[0] = NULL; /* preliminary pivot */
+  }
+
+  if (!wsize || !wNAF_len || !wNAF || !val_sub) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_MALLOC_FAILURE);
+    goto err;
+  }
+
+  /* num_val will be the total number of temporarily precomputed points */
+  num_val = 0;
+
+  for (i = 0; i < total_num; i++) {
+    size_t bits;
+
+    bits = i < num ? BN_num_bits(scalars[i]) : BN_num_bits(g_scalar);
+    wsize[i] = EC_window_bits_for_scalar_size(bits);
+    num_val += (size_t)1 << (wsize[i] - 1);
+    wNAF[i + 1] = NULL; /* make sure we always have a pivot */
+    wNAF[i] =
+        compute_wNAF((i < num ? scalars[i] : g_scalar), wsize[i], &wNAF_len[i]);
+    if (wNAF[i] == NULL) {
+      goto err;
+    }
+    if (wNAF_len[i] > max_len) {
+      max_len = wNAF_len[i];
+    }
+  }
+
+  /* All points we precompute now go into a single array 'val'. 'val_sub[i]' is
+   * a pointer to the subarray for the i-th point. */
+  val = OPENSSL_malloc((num_val + 1) * sizeof val[0]);
+  if (val == NULL) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_MALLOC_FAILURE);
+    goto err;
+  }
+  val[num_val] = NULL; /* pivot element */
+
+  /* allocate points for precomputation */
+  v = val;
+  for (i = 0; i < total_num; i++) {
+    val_sub[i] = v;
+    for (j = 0; j < ((size_t)1 << (wsize[i] - 1)); j++) {
+      *v = EC_POINT_new(group);
+      if (*v == NULL) {
+        goto err;
+      }
+      v++;
+    }
+  }
+  if (!(v == val + num_val)) {
+    OPENSSL_PUT_ERROR(EC, ERR_R_INTERNAL_ERROR);
+    goto err;
+  }
+
+  if (!(tmp = EC_POINT_new(group))) {
+    goto err;
+  }
+
+  /* prepare precomputed values:
+   *    val_sub[i][0] :=     points[i]
+   *    val_sub[i][1] := 3 * points[i]
+   *    val_sub[i][2] := 5 * points[i]
+   *    ...
+   */
+  for (i = 0; i < total_num; i++) {
+    if (i < num) {
+      if (!ec_GFp_simple_point_copy(val_sub[i][0], points[i])) {
+        goto err;
+      }
+    } else if (!ec_GFp_simple_point_copy(val_sub[i][0], generator)) {
+      goto err;
+    }
+
+    if (wsize[i] > 1) {
+      if (!ec_GFp_simple_dbl(group, tmp, val_sub[i][0], ctx)) {
+        goto err;
+      }
+      for (j = 1; j < ((size_t)1 << (wsize[i] - 1)); j++) {
+        if (!ec_GFp_simple_add(group, val_sub[i][j], val_sub[i][j - 1], tmp,
+                               ctx)) {
+          goto err;
+        }
+      }
+    }
+  }
+
+#if 1 /* optional; EC_window_bits_for_scalar_size assumes we do this step */
+  if (!ec_GFp_simple_points_make_affine(group, num_val, val, ctx)) {
+    goto err;
+  }
+#endif
+
+  r_is_at_infinity = 1;
+
+  for (k = max_len - 1; k >= 0; k--) {
+    if (!r_is_at_infinity && !ec_GFp_simple_dbl(group, r, r, ctx)) {
+      goto err;
+    }
+
+    for (i = 0; i < total_num; i++) {
+      if (wNAF_len[i] > (size_t)k) {
+        int digit = wNAF[i][k];
+        int is_neg;
+
+        if (digit) {
+          is_neg = digit < 0;
+
+          if (is_neg) {
+            digit = -digit;
+          }
+
+          if (is_neg != r_is_inverted) {
+            if (!r_is_at_infinity && !ec_GFp_simple_invert(group, r, ctx)) {
+              goto err;
+            }
+            r_is_inverted = !r_is_inverted;
+          }
+
+          /* digit > 0 */
+
+          if (r_is_at_infinity) {
+            if (!ec_GFp_simple_point_copy(r, val_sub[i][digit >> 1])) {
+              goto err;
+            }
+            r_is_at_infinity = 0;
+          } else {
+            if (!ec_GFp_simple_add(group, r, r, val_sub[i][digit >> 1], ctx)) {
+              goto err;
+            }
+          }
+        }
+      }
+    }
+  }
+
+  if (r_is_at_infinity) {
+    if (!EC_POINT_set_to_infinity(group, r)) {
+      goto err;
+    }
+  } else if (r_is_inverted && !ec_GFp_simple_invert(group, r, ctx)) {
+    goto err;
+  }
+
+  ret = 1;
+
+err:
+  BN_CTX_free(new_ctx);
+  EC_POINT_free(tmp);
+  OPENSSL_free(wsize);
+  OPENSSL_free(wNAF_len);
+  if (wNAF != NULL) {
+    signed char **w;
+
+    for (w = wNAF; *w != NULL; w++) {
+      OPENSSL_free(*w);
+    }
+
+    OPENSSL_free(wNAF);
+  }
+  if (val != NULL) {
+    for (v = val; *v != NULL; v++) {
+      ec_GFp_simple_point_clear_finish(*v);
+      OPENSSL_cleanse(*v, sizeof **v);
+      OPENSSL_free(*v);
+    }
+
+    OPENSSL_free(val);
+  }
+  OPENSSL_free(val_sub);
+  return ret;
+}