grpc 1.9.1 → 1.10.0.pre1

data/third_party/boringssl/crypto/bn/add.c

@@ -1,377 +0,0 @@
-/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
- * All rights reserved.
- *
- * This package is an SSL implementation written
- * by Eric Young (eay@cryptsoft.com).
- * The implementation was written so as to conform with Netscapes SSL.
- *
- * This library is free for commercial and non-commercial use as long as
- * the following conditions are aheared to.  The following conditions
- * apply to all code found in this distribution, be it the RC4, RSA,
- * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
- * included with this distribution is covered by the same copyright terms
- * except that the holder is Tim Hudson (tjh@cryptsoft.com).
- *
- * Copyright remains Eric Young's, and as such any Copyright notices in
- * the code are not to be removed.
- * If this package is used in a product, Eric Young should be given attribution
- * as the author of the parts of the library used.
- * This can be in the form of a textual message at program startup or
- * in documentation (online or textual) provided with the package.
- *
- * 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 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 acknowledgement:
- *    "This product includes cryptographic software written by
- *     Eric Young (eay@cryptsoft.com)"
- *    The word 'cryptographic' can be left out if the rouines from the library
- *    being used are not cryptographic related :-).
- * 4. If you include any Windows specific code (or a derivative thereof) from
- *    the apps directory (application code) you must include an acknowledgement:
- *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
- *
- * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
- * ANY EXPRESS 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 AUTHOR OR 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.
- *
- * The licence and distribution terms for any publically available version or
- * derivative of this code cannot be changed.  i.e. this code cannot simply be
- * copied and put under another distribution licence
- * [including the GNU Public Licence.] */
-
-#include <openssl/bn.h>
-
-#include <string.h>
-
-#include <openssl/err.h>
-#include <openssl/mem.h>
-
-#include "internal.h"
-
-
-int BN_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b) {
-  const BIGNUM *tmp;
-  int a_neg = a->neg, ret;
-
-  /*  a +  b	a+b
-   *  a + -b	a-b
-   * -a +  b	b-a
-   * -a + -b	-(a+b)
-   */
-  if (a_neg ^ b->neg) {
-    /* only one is negative */
-    if (a_neg) {
-      tmp = a;
-      a = b;
-      b = tmp;
-    }
-
-    /* we are now a - b */
-    if (BN_ucmp(a, b) < 0) {
-      if (!BN_usub(r, b, a)) {
-        return 0;
-      }
-      r->neg = 1;
-    } else {
-      if (!BN_usub(r, a, b)) {
-        return 0;
-      }
-      r->neg = 0;
-    }
-    return 1;
-  }
-
-  ret = BN_uadd(r, a, b);
-  r->neg = a_neg;
-  return ret;
-}
-
-int BN_uadd(BIGNUM *r, const BIGNUM *a, const BIGNUM *b) {
-  int max, min, dif;
-  BN_ULONG *ap, *bp, *rp, carry, t1, t2;
-  const BIGNUM *tmp;
-
-  if (a->top < b->top) {
-    tmp = a;
-    a = b;
-    b = tmp;
-  }
-  max = a->top;
-  min = b->top;
-  dif = max - min;
-
-  if (bn_wexpand(r, max + 1) == NULL) {
-    return 0;
-  }
-
-  r->top = max;
-
-  ap = a->d;
-  bp = b->d;
-  rp = r->d;
-
-  carry = bn_add_words(rp, ap, bp, min);
-  rp += min;
-  ap += min;
-  bp += min;
-
-  if (carry) {
-    while (dif) {
-      dif--;
-      t1 = *(ap++);
-      t2 = (t1 + 1) & BN_MASK2;
-      *(rp++) = t2;
-      if (t2) {
-        carry = 0;
-        break;
-      }
-    }
-    if (carry) {
-      /* carry != 0 => dif == 0 */
-      *rp = 1;
-      r->top++;
-    }
-  }
-
-  if (dif && rp != ap) {
-    while (dif--) {
-      /* copy remaining words if ap != rp */
-      *(rp++) = *(ap++);
-    }
-  }
-
-  r->neg = 0;
-  return 1;
-}
-
-int BN_add_word(BIGNUM *a, BN_ULONG w) {
-  BN_ULONG l;
-  int i;
-
-  w &= BN_MASK2;
-
-  /* degenerate case: w is zero */
-  if (!w) {
-    return 1;
-  }
-
-  /* degenerate case: a is zero */
-  if (BN_is_zero(a)) {
-    return BN_set_word(a, w);
-  }
-
-  /* handle 'a' when negative */
-  if (a->neg) {
-    a->neg = 0;
-    i = BN_sub_word(a, w);
-    if (!BN_is_zero(a)) {
-      a->neg = !(a->neg);
-    }
-    return i;
-  }
-
-  for (i = 0; w != 0 && i < a->top; i++) {
-    a->d[i] = l = (a->d[i] + w) & BN_MASK2;
-    w = (w > l) ? 1 : 0;
-  }
-
-  if (w && i == a->top) {
-    if (bn_wexpand(a, a->top + 1) == NULL) {
-      return 0;
-    }
-    a->top++;
-    a->d[i] = w;
-  }
-
-  return 1;
-}
-
-int BN_sub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b) {
-  int max;
-  int add = 0, neg = 0;
-  const BIGNUM *tmp;
-
-  /*  a -  b	a-b
-   *  a - -b	a+b
-   * -a -  b	-(a+b)
-   * -a - -b	b-a
-   */
-  if (a->neg) {
-    if (b->neg) {
-      tmp = a;
-      a = b;
-      b = tmp;
-    } else {
-      add = 1;
-      neg = 1;
-    }
-  } else {
-    if (b->neg) {
-      add = 1;
-      neg = 0;
-    }
-  }
-
-  if (add) {
-    if (!BN_uadd(r, a, b)) {
-      return 0;
-    }
-
-    r->neg = neg;
-    return 1;
-  }
-
-  /* We are actually doing a - b :-) */
-
-  max = (a->top > b->top) ? a->top : b->top;
-  if (bn_wexpand(r, max) == NULL) {
-    return 0;
-  }
-
-  if (BN_ucmp(a, b) < 0) {
-    if (!BN_usub(r, b, a)) {
-      return 0;
-    }
-    r->neg = 1;
-  } else {
-    if (!BN_usub(r, a, b)) {
-      return 0;
-    }
-    r->neg = 0;
-  }
-
-  return 1;
-}
-
-int BN_usub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b) {
-  int max, min, dif;
-  register BN_ULONG t1, t2, *ap, *bp, *rp;
-  int i, carry;
-
-  max = a->top;
-  min = b->top;
-  dif = max - min;
-
-  if (dif < 0) /* hmm... should not be happening */
-  {
-    OPENSSL_PUT_ERROR(BN, BN_R_ARG2_LT_ARG3);
-    return 0;
-  }
-
-  if (bn_wexpand(r, max) == NULL) {
-    return 0;
-  }
-
-  ap = a->d;
-  bp = b->d;
-  rp = r->d;
-
-  carry = 0;
-  for (i = min; i != 0; i--) {
-    t1 = *(ap++);
-    t2 = *(bp++);
-    if (carry) {
-      carry = (t1 <= t2);
-      t1 = (t1 - t2 - 1) & BN_MASK2;
-    } else {
-      carry = (t1 < t2);
-      t1 = (t1 - t2) & BN_MASK2;
-    }
-    *(rp++) = t1 & BN_MASK2;
-  }
-
-  if (carry) /* subtracted */
-  {
-    if (!dif) {
-      /* error: a < b */
-      return 0;
-    }
-
-    while (dif) {
-      dif--;
-      t1 = *(ap++);
-      t2 = (t1 - 1) & BN_MASK2;
-      *(rp++) = t2;
-      if (t1) {
-        break;
-      }
-    }
-  }
-
-  if (dif > 0 && rp != ap) {
-    OPENSSL_memcpy(rp, ap, sizeof(*rp) * dif);
-  }
-
-  r->top = max;
-  r->neg = 0;
-  bn_correct_top(r);
-
-  return 1;
-}
-
-int BN_sub_word(BIGNUM *a, BN_ULONG w) {
-  int i;
-
-  w &= BN_MASK2;
-
-  /* degenerate case: w is zero */
-  if (!w) {
-    return 1;
-  }
-
-  /* degenerate case: a is zero */
-  if (BN_is_zero(a)) {
-    i = BN_set_word(a, w);
-    if (i != 0) {
-      BN_set_negative(a, 1);
-    }
-    return i;
-  }
-
-  /* handle 'a' when negative */
-  if (a->neg) {
-    a->neg = 0;
-    i = BN_add_word(a, w);
-    a->neg = 1;
-    return i;
-  }
-
-  if ((a->top == 1) && (a->d[0] < w)) {
-    a->d[0] = w - a->d[0];
-    a->neg = 1;
-    return 1;
-  }
-
-  i = 0;
-  for (;;) {
-    if (a->d[i] >= w) {
-      a->d[i] -= w;
-      break;
-    } else {
-      a->d[i] = (a->d[i] - w) & BN_MASK2;
-      i++;
-      w = 1;
-    }
-  }
-
-  if ((a->d[i] == 0) && (i == (a->top - 1))) {
-    a->top--;
-  }
-
-  return 1;
-}

data/third_party/boringssl/crypto/bn/asm/x86_64-gcc.c

@@ -1,532 +0,0 @@
-/* x86_64 BIGNUM accelerator version 0.1, December 2002.
- *
- * Implemented by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
- * project.
- *
- * Rights for redistribution and usage in source and binary forms are
- * granted according to the OpenSSL license. Warranty of any kind is
- * disclaimed.
- *
- * Q. Version 0.1? It doesn't sound like Andy, he used to assign real
- *    versions, like 1.0...
- * A. Well, that's because this code is basically a quick-n-dirty
- *    proof-of-concept hack. As you can see it's implemented with
- *    inline assembler, which means that you're bound to GCC and that
- *    there might be enough room for further improvement.
- *
- * Q. Why inline assembler?
- * A. x86_64 features own ABI which I'm not familiar with. This is
- *    why I decided to let the compiler take care of subroutine
- *    prologue/epilogue as well as register allocation. For reference.
- *    Win64 implements different ABI for AMD64, different from Linux.
- *
- * Q. How much faster does it get?
- * A. 'apps/openssl speed rsa dsa' output with no-asm:
- *
- *	                  sign    verify    sign/s verify/s
- *	rsa  512 bits   0.0006s   0.0001s   1683.8  18456.2
- *	rsa 1024 bits   0.0028s   0.0002s    356.0   6407.0
- *	rsa 2048 bits   0.0172s   0.0005s     58.0   1957.8
- *	rsa 4096 bits   0.1155s   0.0018s      8.7    555.6
- *	                  sign    verify    sign/s verify/s
- *	dsa  512 bits   0.0005s   0.0006s   2100.8   1768.3
- *	dsa 1024 bits   0.0014s   0.0018s    692.3    559.2
- *	dsa 2048 bits   0.0049s   0.0061s    204.7    165.0
- *
- *    'apps/openssl speed rsa dsa' output with this module:
- *
- *	                  sign    verify    sign/s verify/s
- *	rsa  512 bits   0.0004s   0.0000s   2767.1  33297.9
- *	rsa 1024 bits   0.0012s   0.0001s    867.4  14674.7
- *	rsa 2048 bits   0.0061s   0.0002s    164.0   5270.0
- *	rsa 4096 bits   0.0384s   0.0006s     26.1   1650.8
- *	                  sign    verify    sign/s verify/s
- *	dsa  512 bits   0.0002s   0.0003s   4442.2   3786.3
- *	dsa 1024 bits   0.0005s   0.0007s   1835.1   1497.4
- *	dsa 2048 bits   0.0016s   0.0020s    620.4    504.6
- *
- *    For the reference. IA-32 assembler implementation performs
- *    very much like 64-bit code compiled with no-asm on the same
- *    machine.
- */
-
-#include <openssl/bn.h>
-
-/* TODO(davidben): Get this file working on Windows x64. */
-#if !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86_64) && defined(__GNUC__)
-
-#include "../internal.h"
-
-
-#undef mul
-#undef mul_add
-
-#define asm __asm__
-
-/*
- * "m"(a), "+m"(r)	is the way to favor DirectPath µ-code;
- * "g"(0)		let the compiler to decide where does it
- *			want to keep the value of zero;
- */
-#define mul_add(r, a, word, carry)                                     \
-  do {                                                                 \
-    register BN_ULONG high, low;                                       \
-    asm("mulq %3" : "=a"(low), "=d"(high) : "a"(word), "m"(a) : "cc"); \
-    asm("addq %2,%0; adcq %3,%1"                                       \
-        : "+r"(carry), "+d"(high)                                      \
-        : "a"(low), "g"(0)                                             \
-        : "cc");                                                       \
-    asm("addq %2,%0; adcq %3,%1"                                       \
-        : "+m"(r), "+d"(high)                                          \
-        : "r"(carry), "g"(0)                                           \
-        : "cc");                                                       \
-    (carry) = high;                                                    \
-  } while (0)
-
-#define mul(r, a, word, carry)                                         \
-  do {                                                                 \
-    register BN_ULONG high, low;                                       \
-    asm("mulq %3" : "=a"(low), "=d"(high) : "a"(word), "g"(a) : "cc"); \
-    asm("addq %2,%0; adcq %3,%1"                                       \
-        : "+r"(carry), "+d"(high)                                      \
-        : "a"(low), "g"(0)                                             \
-        : "cc");                                                       \
-    (r) = (carry);                                                     \
-    (carry) = high;                                                    \
-  } while (0)
-#undef sqr
-#define sqr(r0, r1, a) asm("mulq %2" : "=a"(r0), "=d"(r1) : "a"(a) : "cc");
-
-BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num,
-                          BN_ULONG w) {
-  BN_ULONG c1 = 0;
-
-  if (num <= 0) {
-    return (c1);
-  }
-
-  while (num & ~3) {
-    mul_add(rp[0], ap[0], w, c1);
-    mul_add(rp[1], ap[1], w, c1);
-    mul_add(rp[2], ap[2], w, c1);
-    mul_add(rp[3], ap[3], w, c1);
-    ap += 4;
-    rp += 4;
-    num -= 4;
-  }
-  if (num) {
-    mul_add(rp[0], ap[0], w, c1);
-    if (--num == 0) {
-      return c1;
-    }
-    mul_add(rp[1], ap[1], w, c1);
-    if (--num == 0) {
-      return c1;
-    }
-    mul_add(rp[2], ap[2], w, c1);
-    return c1;
-  }
-
-  return c1;
-}
-
-BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w) {
-  BN_ULONG c1 = 0;
-
-  if (num <= 0) {
-    return c1;
-  }
-
-  while (num & ~3) {
-    mul(rp[0], ap[0], w, c1);
-    mul(rp[1], ap[1], w, c1);
-    mul(rp[2], ap[2], w, c1);
-    mul(rp[3], ap[3], w, c1);
-    ap += 4;
-    rp += 4;
-    num -= 4;
-  }
-  if (num) {
-    mul(rp[0], ap[0], w, c1);
-    if (--num == 0) {
-      return c1;
-    }
-    mul(rp[1], ap[1], w, c1);
-    if (--num == 0) {
-      return c1;
-    }
-    mul(rp[2], ap[2], w, c1);
-  }
-  return c1;
-}
-
-void bn_sqr_words(BN_ULONG *r, const BN_ULONG *a, int n) {
-  if (n <= 0) {
-    return;
-  }
-
-  while (n & ~3) {
-    sqr(r[0], r[1], a[0]);
-    sqr(r[2], r[3], a[1]);
-    sqr(r[4], r[5], a[2]);
-    sqr(r[6], r[7], a[3]);
-    a += 4;
-    r += 8;
-    n -= 4;
-  }
-  if (n) {
-    sqr(r[0], r[1], a[0]);
-    if (--n == 0) {
-      return;
-    }
-    sqr(r[2], r[3], a[1]);
-    if (--n == 0) {
-      return;
-    }
-    sqr(r[4], r[5], a[2]);
-  }
-}
-
-BN_ULONG bn_add_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
-                      int n) {
-  BN_ULONG ret;
-  size_t i = 0;
-
-  if (n <= 0) {
-    return 0;
-  }
-
-  asm volatile (
-      "	subq	%0,%0		\n" /* clear carry */
-      "	jmp	1f		\n"
-      ".p2align 4			\n"
-      "1:	movq	(%4,%2,8),%0	\n"
-      "	adcq	(%5,%2,8),%0	\n"
-      "	movq	%0,(%3,%2,8)	\n"
-      "	lea	1(%2),%2	\n"
-      "	loop	1b		\n"
-      "	sbbq	%0,%0		\n"
-      : "=&r"(ret), "+c"(n), "+r"(i)
-      : "r"(rp), "r"(ap), "r"(bp)
-      : "cc", "memory");
-
-  return ret & 1;
-}
-
-BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,
-                      int n) {
-  BN_ULONG ret;
-  size_t i = 0;
-
-  if (n <= 0) {
-    return 0;
-  }
-
-  asm volatile (
-      "	subq	%0,%0		\n" /* clear borrow */
-      "	jmp	1f		\n"
-      ".p2align 4			\n"
-      "1:	movq	(%4,%2,8),%0	\n"
-      "	sbbq	(%5,%2,8),%0	\n"
-      "	movq	%0,(%3,%2,8)	\n"
-      "	lea	1(%2),%2	\n"
-      "	loop	1b		\n"
-      "	sbbq	%0,%0		\n"
-      : "=&r"(ret), "+c"(n), "+r"(i)
-      : "r"(rp), "r"(ap), "r"(bp)
-      : "cc", "memory");
-
-  return ret & 1;
-}
-
-/* mul_add_c(a,b,c0,c1,c2)  -- c+=a*b for three word number c=(c2,c1,c0) */
-/* mul_add_c2(a,b,c0,c1,c2) -- c+=2*a*b for three word number c=(c2,c1,c0) */
-/* sqr_add_c(a,i,c0,c1,c2)  -- c+=a[i]^2 for three word number c=(c2,c1,c0) */
-/* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0)
- */
-
-/* Keep in mind that carrying into high part of multiplication result can not
- * overflow, because it cannot be all-ones. */
-#define mul_add_c(a, b, c0, c1, c2)          \
-  do {                                       \
-    BN_ULONG t1, t2;                \
-    asm("mulq %3" : "=a"(t1), "=d"(t2) : "a"(a), "m"(b) : "cc"); \
-    asm("addq %3,%0; adcq %4,%1; adcq %5,%2" \
-        : "+r"(c0), "+r"(c1), "+r"(c2)       \
-        : "r"(t1), "r"(t2), "g"(0)           \
-        : "cc");                             \
-  } while (0)
-
-#define sqr_add_c(a, i, c0, c1, c2)                           \
-  do {                                                        \
-    BN_ULONG t1, t2;                                          \
-    asm("mulq %2" : "=a"(t1), "=d"(t2) : "a"((a)[i]) : "cc"); \
-    asm("addq %3,%0; adcq %4,%1; adcq %5,%2"                  \
-        : "+r"(c0), "+r"(c1), "+r"(c2)                        \
-        : "r"(t1), "r"(t2), "g"(0)                            \
-        : "cc");                                              \
-  } while (0)
-
-#define mul_add_c2(a, b, c0, c1, c2)         \
-  do {                                       \
-    BN_ULONG t1, t2;                                                    \
-    asm("mulq %3" : "=a"(t1), "=d"(t2) : "a"(a), "m"(b) : "cc");        \
-    asm("addq %3,%0; adcq %4,%1; adcq %5,%2" \
-        : "+r"(c0), "+r"(c1), "+r"(c2)       \
-        : "r"(t1), "r"(t2), "g"(0)           \
-        : "cc");                             \
-    asm("addq %3,%0; adcq %4,%1; adcq %5,%2" \
-        : "+r"(c0), "+r"(c1), "+r"(c2)       \
-        : "r"(t1), "r"(t2), "g"(0)           \
-        : "cc");                             \
-  } while (0)
-
-#define sqr_add_c2(a, i, j, c0, c1, c2) mul_add_c2((a)[i], (a)[j], c0, c1, c2)
-
-void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b) {
-  BN_ULONG c1, c2, c3;
-
-  c1 = 0;
-  c2 = 0;
-  c3 = 0;
-  mul_add_c(a[0], b[0], c1, c2, c3);
-  r[0] = c1;
-  c1 = 0;
-  mul_add_c(a[0], b[1], c2, c3, c1);
-  mul_add_c(a[1], b[0], c2, c3, c1);
-  r[1] = c2;
-  c2 = 0;
-  mul_add_c(a[2], b[0], c3, c1, c2);
-  mul_add_c(a[1], b[1], c3, c1, c2);
-  mul_add_c(a[0], b[2], c3, c1, c2);
-  r[2] = c3;
-  c3 = 0;
-  mul_add_c(a[0], b[3], c1, c2, c3);
-  mul_add_c(a[1], b[2], c1, c2, c3);
-  mul_add_c(a[2], b[1], c1, c2, c3);
-  mul_add_c(a[3], b[0], c1, c2, c3);
-  r[3] = c1;
-  c1 = 0;
-  mul_add_c(a[4], b[0], c2, c3, c1);
-  mul_add_c(a[3], b[1], c2, c3, c1);
-  mul_add_c(a[2], b[2], c2, c3, c1);
-  mul_add_c(a[1], b[3], c2, c3, c1);
-  mul_add_c(a[0], b[4], c2, c3, c1);
-  r[4] = c2;
-  c2 = 0;
-  mul_add_c(a[0], b[5], c3, c1, c2);
-  mul_add_c(a[1], b[4], c3, c1, c2);
-  mul_add_c(a[2], b[3], c3, c1, c2);
-  mul_add_c(a[3], b[2], c3, c1, c2);
-  mul_add_c(a[4], b[1], c3, c1, c2);
-  mul_add_c(a[5], b[0], c3, c1, c2);
-  r[5] = c3;
-  c3 = 0;
-  mul_add_c(a[6], b[0], c1, c2, c3);
-  mul_add_c(a[5], b[1], c1, c2, c3);
-  mul_add_c(a[4], b[2], c1, c2, c3);
-  mul_add_c(a[3], b[3], c1, c2, c3);
-  mul_add_c(a[2], b[4], c1, c2, c3);
-  mul_add_c(a[1], b[5], c1, c2, c3);
-  mul_add_c(a[0], b[6], c1, c2, c3);
-  r[6] = c1;
-  c1 = 0;
-  mul_add_c(a[0], b[7], c2, c3, c1);
-  mul_add_c(a[1], b[6], c2, c3, c1);
-  mul_add_c(a[2], b[5], c2, c3, c1);
-  mul_add_c(a[3], b[4], c2, c3, c1);
-  mul_add_c(a[4], b[3], c2, c3, c1);
-  mul_add_c(a[5], b[2], c2, c3, c1);
-  mul_add_c(a[6], b[1], c2, c3, c1);
-  mul_add_c(a[7], b[0], c2, c3, c1);
-  r[7] = c2;
-  c2 = 0;
-  mul_add_c(a[7], b[1], c3, c1, c2);
-  mul_add_c(a[6], b[2], c3, c1, c2);
-  mul_add_c(a[5], b[3], c3, c1, c2);
-  mul_add_c(a[4], b[4], c3, c1, c2);
-  mul_add_c(a[3], b[5], c3, c1, c2);
-  mul_add_c(a[2], b[6], c3, c1, c2);
-  mul_add_c(a[1], b[7], c3, c1, c2);
-  r[8] = c3;
-  c3 = 0;
-  mul_add_c(a[2], b[7], c1, c2, c3);
-  mul_add_c(a[3], b[6], c1, c2, c3);
-  mul_add_c(a[4], b[5], c1, c2, c3);
-  mul_add_c(a[5], b[4], c1, c2, c3);
-  mul_add_c(a[6], b[3], c1, c2, c3);
-  mul_add_c(a[7], b[2], c1, c2, c3);
-  r[9] = c1;
-  c1 = 0;
-  mul_add_c(a[7], b[3], c2, c3, c1);
-  mul_add_c(a[6], b[4], c2, c3, c1);
-  mul_add_c(a[5], b[5], c2, c3, c1);
-  mul_add_c(a[4], b[6], c2, c3, c1);
-  mul_add_c(a[3], b[7], c2, c3, c1);
-  r[10] = c2;
-  c2 = 0;
-  mul_add_c(a[4], b[7], c3, c1, c2);
-  mul_add_c(a[5], b[6], c3, c1, c2);
-  mul_add_c(a[6], b[5], c3, c1, c2);
-  mul_add_c(a[7], b[4], c3, c1, c2);
-  r[11] = c3;
-  c3 = 0;
-  mul_add_c(a[7], b[5], c1, c2, c3);
-  mul_add_c(a[6], b[6], c1, c2, c3);
-  mul_add_c(a[5], b[7], c1, c2, c3);
-  r[12] = c1;
-  c1 = 0;
-  mul_add_c(a[6], b[7], c2, c3, c1);
-  mul_add_c(a[7], b[6], c2, c3, c1);
-  r[13] = c2;
-  c2 = 0;
-  mul_add_c(a[7], b[7], c3, c1, c2);
-  r[14] = c3;
-  r[15] = c1;
-}
-
-void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b) {
-  BN_ULONG c1, c2, c3;
-
-  c1 = 0;
-  c2 = 0;
-  c3 = 0;
-  mul_add_c(a[0], b[0], c1, c2, c3);
-  r[0] = c1;
-  c1 = 0;
-  mul_add_c(a[0], b[1], c2, c3, c1);
-  mul_add_c(a[1], b[0], c2, c3, c1);
-  r[1] = c2;
-  c2 = 0;
-  mul_add_c(a[2], b[0], c3, c1, c2);
-  mul_add_c(a[1], b[1], c3, c1, c2);
-  mul_add_c(a[0], b[2], c3, c1, c2);
-  r[2] = c3;
-  c3 = 0;
-  mul_add_c(a[0], b[3], c1, c2, c3);
-  mul_add_c(a[1], b[2], c1, c2, c3);
-  mul_add_c(a[2], b[1], c1, c2, c3);
-  mul_add_c(a[3], b[0], c1, c2, c3);
-  r[3] = c1;
-  c1 = 0;
-  mul_add_c(a[3], b[1], c2, c3, c1);
-  mul_add_c(a[2], b[2], c2, c3, c1);
-  mul_add_c(a[1], b[3], c2, c3, c1);
-  r[4] = c2;
-  c2 = 0;
-  mul_add_c(a[2], b[3], c3, c1, c2);
-  mul_add_c(a[3], b[2], c3, c1, c2);
-  r[5] = c3;
-  c3 = 0;
-  mul_add_c(a[3], b[3], c1, c2, c3);
-  r[6] = c1;
-  r[7] = c2;
-}
-
-void bn_sqr_comba8(BN_ULONG *r, const BN_ULONG *a) {
-  BN_ULONG c1, c2, c3;
-
-  c1 = 0;
-  c2 = 0;
-  c3 = 0;
-  sqr_add_c(a, 0, c1, c2, c3);
-  r[0] = c1;
-  c1 = 0;
-  sqr_add_c2(a, 1, 0, c2, c3, c1);
-  r[1] = c2;
-  c2 = 0;
-  sqr_add_c(a, 1, c3, c1, c2);
-  sqr_add_c2(a, 2, 0, c3, c1, c2);
-  r[2] = c3;
-  c3 = 0;
-  sqr_add_c2(a, 3, 0, c1, c2, c3);
-  sqr_add_c2(a, 2, 1, c1, c2, c3);
-  r[3] = c1;
-  c1 = 0;
-  sqr_add_c(a, 2, c2, c3, c1);
-  sqr_add_c2(a, 3, 1, c2, c3, c1);
-  sqr_add_c2(a, 4, 0, c2, c3, c1);
-  r[4] = c2;
-  c2 = 0;
-  sqr_add_c2(a, 5, 0, c3, c1, c2);
-  sqr_add_c2(a, 4, 1, c3, c1, c2);
-  sqr_add_c2(a, 3, 2, c3, c1, c2);
-  r[5] = c3;
-  c3 = 0;
-  sqr_add_c(a, 3, c1, c2, c3);
-  sqr_add_c2(a, 4, 2, c1, c2, c3);
-  sqr_add_c2(a, 5, 1, c1, c2, c3);
-  sqr_add_c2(a, 6, 0, c1, c2, c3);
-  r[6] = c1;
-  c1 = 0;
-  sqr_add_c2(a, 7, 0, c2, c3, c1);
-  sqr_add_c2(a, 6, 1, c2, c3, c1);
-  sqr_add_c2(a, 5, 2, c2, c3, c1);
-  sqr_add_c2(a, 4, 3, c2, c3, c1);
-  r[7] = c2;
-  c2 = 0;
-  sqr_add_c(a, 4, c3, c1, c2);
-  sqr_add_c2(a, 5, 3, c3, c1, c2);
-  sqr_add_c2(a, 6, 2, c3, c1, c2);
-  sqr_add_c2(a, 7, 1, c3, c1, c2);
-  r[8] = c3;
-  c3 = 0;
-  sqr_add_c2(a, 7, 2, c1, c2, c3);
-  sqr_add_c2(a, 6, 3, c1, c2, c3);
-  sqr_add_c2(a, 5, 4, c1, c2, c3);
-  r[9] = c1;
-  c1 = 0;
-  sqr_add_c(a, 5, c2, c3, c1);
-  sqr_add_c2(a, 6, 4, c2, c3, c1);
-  sqr_add_c2(a, 7, 3, c2, c3, c1);
-  r[10] = c2;
-  c2 = 0;
-  sqr_add_c2(a, 7, 4, c3, c1, c2);
-  sqr_add_c2(a, 6, 5, c3, c1, c2);
-  r[11] = c3;
-  c3 = 0;
-  sqr_add_c(a, 6, c1, c2, c3);
-  sqr_add_c2(a, 7, 5, c1, c2, c3);
-  r[12] = c1;
-  c1 = 0;
-  sqr_add_c2(a, 7, 6, c2, c3, c1);
-  r[13] = c2;
-  c2 = 0;
-  sqr_add_c(a, 7, c3, c1, c2);
-  r[14] = c3;
-  r[15] = c1;
-}
-
-void bn_sqr_comba4(BN_ULONG *r, const BN_ULONG *a) {
-  BN_ULONG c1, c2, c3;
-
-  c1 = 0;
-  c2 = 0;
-  c3 = 0;
-  sqr_add_c(a, 0, c1, c2, c3);
-  r[0] = c1;
-  c1 = 0;
-  sqr_add_c2(a, 1, 0, c2, c3, c1);
-  r[1] = c2;
-  c2 = 0;
-  sqr_add_c(a, 1, c3, c1, c2);
-  sqr_add_c2(a, 2, 0, c3, c1, c2);
-  r[2] = c3;
-  c3 = 0;
-  sqr_add_c2(a, 3, 0, c1, c2, c3);
-  sqr_add_c2(a, 2, 1, c1, c2, c3);
-  r[3] = c1;
-  c1 = 0;
-  sqr_add_c(a, 2, c2, c3, c1);
-  sqr_add_c2(a, 3, 1, c2, c3, c1);
-  r[4] = c2;
-  c2 = 0;
-  sqr_add_c2(a, 3, 2, c3, c1, c2);
-  r[5] = c3;
-  c3 = 0;
-  sqr_add_c(a, 3, c1, c2, c3);
-  r[6] = c1;
-  r[7] = c2;
-}
-
-#endif  /* !NO_ASM && X86_64 && __GNUC__ */