path: root/camellia-set-encrypt-key.c

/* camellia-set-encrypt-key.c
 *
 * Key setup for the camellia block cipher.
 */
/*
 * Copyright (C) 2006,2007
 * NTT (Nippon Telegraph and Telephone Corporation).
 *
 * Copyright (C) 2010 Niels Möller
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 */

/*
 * Algorithm Specification 
 *  http://info.isl.ntt.co.jp/crypt/eng/camellia/specifications.html
 */

/* Based on camellia.c ver 1.2.0, see
   http://info.isl.ntt.co.jp/crypt/eng/camellia/dl/camellia-LGPL-1.2.0.tar.gz.
 */
#if HAVE_CONFIG_H
# include "config.h"
#endif

#include <assert.h>
#include <limits.h>

#include "camellia-internal.h"

#include "macros.h"

/* key constants */

#define SIGMA1 0xA09E667F3BCC908BULL
#define SIGMA2 0xB67AE8584CAA73B2ULL
#define SIGMA3 0xC6EF372FE94F82BEULL
#define SIGMA4 0x54FF53A5F1D36F1CULL
#define SIGMA5 0x10E527FADE682D1DULL
#define SIGMA6 0xB05688C2B3E6C1FDULL

#define CAMELLIA_SP1110(INDEX) (_nettle_camellia_table.sp1110[(int)(INDEX)])
#define CAMELLIA_SP0222(INDEX) (_nettle_camellia_table.sp0222[(int)(INDEX)])
#define CAMELLIA_SP3033(INDEX) (_nettle_camellia_table.sp3033[(int)(INDEX)])
#define CAMELLIA_SP4404(INDEX) (_nettle_camellia_table.sp4404[(int)(INDEX)])

#define CAMELLIA_F(x, k, y) do {		\
    uint32_t __yl, __yr;			\
    uint64_t __i = (x) ^ (k);			\
    __yl					\
      = CAMELLIA_SP1110( __i & 0xff)		\
      ^ CAMELLIA_SP0222((__i >> 24) & 0xff)	\
      ^ CAMELLIA_SP3033((__i >> 16) & 0xff)	\
      ^ CAMELLIA_SP4404((__i >> 8) & 0xff);	\
    __yr					\
      = CAMELLIA_SP1110( __i >> 56)		\
      ^ CAMELLIA_SP0222((__i >> 48) & 0xff)	\
      ^ CAMELLIA_SP3033((__i >> 40) & 0xff)	\
      ^ CAMELLIA_SP4404((__i >> 32) & 0xff);	\
    __yl ^= __yr;				\
    __yr = ROL32(24, __yr);			\
    __yr ^= __yl;				\
    (y) = ((uint64_t) __yl << 32) | __yr;	\
  } while (0)

#if ! HAVE_NATIVE_64_BIT
#define CAMELLIA_F_HALF_INV(x) do {            \
    uint32_t __t, __w;                         \
    __t = (x) >> 32;                           \
    __w = __t ^(x);                            \
    __w = ROL32(8, __w);                       \
    (x) = ((uint64_t) __w << 32) | (__t ^ __w);        \
  } while (0)
#endif

void
camellia_set_encrypt_key(struct camellia_ctx *ctx,
			 unsigned length, const uint8_t *key)
{
  uint64_t k0, k1;

  uint64_t subkey[34];
  uint64_t w, kw2, kw4;
  
  uint32_t dw, tl, tr;
  unsigned i;

  k0 = READ_UINT64(key);
  k1 = READ_UINT64(key +  8);
  
  if (length == 16)
    {
      ctx->nkeys = 24;
      /**
       * generate KL dependent subkeys
       */
      subkey[0] = k0; subkey[1] = k1;
      ROL128(15, k0, k1);
      subkey[4] = k0; subkey[5] = k1;
      ROL128(30, k0, k1);
      subkey[10] = k0; subkey[11] = k1;
      ROL128(15, k0, k1);
      subkey[13] = k1;
      ROL128(17, k0, k1);
      subkey[16] = k0; subkey[17] = k1;
      ROL128(17, k0, k1);
      subkey[18] = k0; subkey[19] = k1;
      ROL128(17, k0, k1);
      subkey[22] = k0; subkey[23] = k1;

      /* generate KA. D1 is k0, d2 is k1. */
      /* FIXME: Make notation match the spec better. */
      /* For the 128-bit case, KR = 0, the construction of KA reduces to:

	 D1 = KL >> 64;
	 W = KL & MASK64;
	 D2 = F(D1, Sigma1);
	 W = D2 ^ W
	 D1 = F(W, Sigma2)
	 D2 = D2 ^ F(D1, Sigma3);
	 D1 = D1 ^ F(D2, Sigma4);
	 KA = (D1 << 64) | D2;
      */
      k0 = subkey[0]; w = subkey[1];
      CAMELLIA_F(k0, SIGMA1, k1);
      w ^= k1;
      CAMELLIA_F(w, SIGMA2, k0);
      CAMELLIA_F(k0, SIGMA3, w);
      k1 ^= w;
      CAMELLIA_F(k1, SIGMA4, w);
      k0 ^= w;

      /* generate KA dependent subkeys */
      subkey[2] = k0; subkey[3] = k1;
      ROL128(15, k0, k1);
      subkey[6] = k0; subkey[7] = k1;
      ROL128(15, k0, k1);
      subkey[8] = k0; subkey[9] = k1;
      ROL128(15, k0, k1);
      subkey[12] = k0;
      ROL128(15, k0, k1);
      subkey[14] = k0; subkey[15] = k1;
      ROL128(34, k0, k1);
      subkey[20] = k0; subkey[21] = k1;
      ROL128(17, k0, k1);
      subkey[24] = k0; subkey[25] = k1;
    }
  else
    {
      uint64_t k2, k3;

      ctx->nkeys = 32;
      k2 = READ_UINT64(key + 16);

      if (length == 24)
	k3 = ~k2;
      else
	{
	  assert (length == 32);
	  k3 = READ_UINT64(key + 24);
	}
      /* generate KL dependent subkeys */
      subkey[0] = k0; subkey[1] = k1;
      ROL128(45, k0, k1);
      subkey[12] = k0; subkey[13] = k1;
      ROL128(15, k0, k1);
      subkey[16] = k0; subkey[17] = k1;
      ROL128(17, k0, k1);
      subkey[22] = k0; subkey[23] = k1;
      ROL128(34, k0, k1);
      subkey[30] = k0; subkey[31] = k1;

      /* generate KR dependent subkeys */
      ROL128(15, k2, k3);
      subkey[4] = k2; subkey[5] = k3;
      ROL128(15, k2, k3);
      subkey[8] = k2; subkey[9] = k3;
      ROL128(30, k2, k3);
      subkey[18] = k2; subkey[19] = k3;
      ROL128(34, k2, k3);
      subkey[26] = k2; subkey[27] = k3;
      ROL128(34, k2, k3);

      /* generate KA */
      /* The construction of KA is done as

	 D1 = (KL ^ KR) >> 64
	 D2 = (KL ^ KR) & MASK64
	 W = F(D1, SIGMA1)
	 D2 = D2 ^ W
	 D1 = F(D2, SIGMA2) ^ (KR >> 64)
	 D2 = F(D1, SIGMA3) ^ W ^ (KR & MASK64)
	 D1 = D1 ^ F(W, SIGMA2)
	 D2 = D2 ^ F(D1, SIGMA3)
	 D1 = D1 ^ F(D2, SIGMA4)
      */

      k0 = subkey[0] ^ k2;
      k1 = subkey[1] ^ k3;

      CAMELLIA_F(k0, SIGMA1, w);
      k1 ^= w;

      CAMELLIA_F(k1, SIGMA2, k0);
      k0 ^= k2;

      CAMELLIA_F(k0, SIGMA3, k1);
      k1 ^= w ^ k3;

      CAMELLIA_F(k1, SIGMA4, w);
      k0 ^= w;

      /* generate KB */
      k2 ^= k0; k3 ^= k1;
      CAMELLIA_F(k2, SIGMA5, w);
      k3 ^= w;
      CAMELLIA_F(k3, SIGMA6, w);
      k2 ^= w;

      /* generate KA dependent subkeys */
      ROL128(15, k0, k1);
      subkey[6] = k0; subkey[7] = k1;
      ROL128(30, k0, k1);
      subkey[14] = k0; subkey[15] = k1;
      ROL128(32, k0, k1);
      subkey[24] = k0; subkey[25] = k1;
      ROL128(17, k0, k1);
      subkey[28] = k0; subkey[29] = k1;

      /* generate KB dependent subkeys */
      subkey[2] = k2; subkey[3] = k3;
      ROL128(30, k2, k3);
      subkey[10] = k2; subkey[11] = k3;
      ROL128(30, k2, k3);
      subkey[20] = k2; subkey[21] = k3;
      ROL128(51, k2, k3);
      subkey[32] = k2; subkey[33] = k3;
    }

  /* At this point, the subkey array contains the subkeys as described
     in the spec, 26 for short keys and 34 for large keys. */

  /* absorb kw2 to other subkeys */
  kw2 = subkey[1];

  subkey[3] ^= kw2;
  subkey[5] ^= kw2;
  subkey[7] ^= kw2;
  for (i = 8; i < ctx->nkeys; i += 8)
    {
      /* FIXME: gcc for x86_32 is smart enough to fetch the 32 low bits
	 and xor the result into the 32 high bits, but it still generates
	 worse code than for explicit 32-bit operations. */
      kw2 ^= (kw2 & ~subkey[i+1]) << 32;
      dw = (kw2 & subkey[i+1]) >> 32; kw2 ^= ROL32(1, dw); 

      subkey[i+3] ^= kw2;
      subkey[i+5] ^= kw2;
      subkey[i+7] ^= kw2;
    }
  subkey[i] ^= kw2;
  
  /* absorb kw4 to other subkeys */  
  kw4 = subkey[ctx->nkeys + 1];

  for (i = ctx->nkeys - 8; i > 0; i -= 8)
    {
      subkey[i+6] ^= kw4;
      subkey[i+4] ^= kw4;
      subkey[i+2] ^= kw4;
      kw4 ^= (kw4 & ~subkey[i]) << 32;
      dw = (kw4 & subkey[i]) >> 32; kw4 ^= ROL32(1, dw);      
    }

  subkey[6] ^= kw4;
  subkey[4] ^= kw4;
  subkey[2] ^= kw4;
  subkey[0] ^= kw4;

  /* key XOR is end of F-function */
  ctx->keys[0] = subkey[0] ^ subkey[2];
  ctx->keys[1] = subkey[3];

  ctx->keys[2] = subkey[2] ^ subkey[4];
  ctx->keys[3] = subkey[3] ^ subkey[5];
  ctx->keys[4] = subkey[4] ^ subkey[6];
  ctx->keys[5] = subkey[5] ^ subkey[7];

  for (i = 8; i < ctx->nkeys; i += 8)
    {
      tl = (subkey[i+2] >> 32) ^ (subkey[i+2] & ~subkey[i]);
      dw = tl & (subkey[i] >> 32);
      tr = subkey[i+2] ^ ROL32(1, dw);
      ctx->keys[i-2] = subkey[i-2] ^ ( ((uint64_t) tl << 32) | tr);

      ctx->keys[i-1] = subkey[i];
      ctx->keys[i] = subkey[i+1];

      tl = (subkey[i-1] >> 32) ^ (subkey[i-1] & ~subkey[i+1]);
      dw = tl & (subkey[i+1] >> 32);
      tr = subkey[i-1] ^ ROL32(1, dw);
      ctx->keys[i+1] = subkey[i+3] ^ ( ((uint64_t) tl << 32) | tr);

      ctx->keys[i+2] = subkey[i+2] ^ subkey[i+4];
      ctx->keys[i+3] = subkey[i+3] ^ subkey[i+5];
      ctx->keys[i+4] = subkey[i+4] ^ subkey[i+6];
      ctx->keys[i+5] = subkey[i+5] ^ subkey[i+7];
    }
  ctx->keys[i-2] = subkey[i-2];
  ctx->keys[i-1] = subkey[i] ^ subkey[i-1];

#if !HAVE_NATIVE_64_BIT
  for (i = 0; i < ctx->nkeys; i += 8)
    {
      /* apply the inverse of the last half of F-function */
      CAMELLIA_F_HALF_INV(ctx->keys[i+1]);
      CAMELLIA_F_HALF_INV(ctx->keys[i+2]);
      CAMELLIA_F_HALF_INV(ctx->keys[i+3]);
      CAMELLIA_F_HALF_INV(ctx->keys[i+4]);
      CAMELLIA_F_HALF_INV(ctx->keys[i+5]);
      CAMELLIA_F_HALF_INV(ctx->keys[i+6]);
    }
#endif
}