Merge trunk. Re-generate tclTomMath.h. Use faster exponentiation-function from libtommath 1.0 (in tclExecute.c).

2 files changed, 120 insertions, 149 deletions

diff --git a/generic/tclExecute.c b/generic/tclExecute.c
index 1cfc030..1f78ead 100644
--- a/generic/tclExecute.c
+++ b/generic/tclExecute.c
@@ -8935,7 +8935,7 @@ ExecuteExtendedBinaryMathOp(
 	}
 	Tcl_TakeBignumFromObj(NULL, valuePtr, &big1);
 	mp_init(&bigResult);
-	mp_expt_d(&big1, big2.dp[0], &bigResult);
+	mp_expt_d_ex(&big1, big2.dp[0], &bigResult, 1);
 	mp_clear(&big1);
 	mp_clear(&big2);
 	BIG_RESULT(&bigResult);
diff --git a/generic/tclTomMath.h b/generic/tclTomMath.h
index dd9edaf..6b96d2c 100644
--- a/generic/tclTomMath.h
+++ b/generic/tclTomMath.h
@@ -10,11 +10,12 @@
  * The library is free for all purposes without any express
  * guarantee it works.
  *
- * Tom St Denis, tomstdenis@gmail.com, http://math.libtomcrypt.com
+ * Tom St Denis, tstdenis82@gmail.com, http://math.libtomcrypt.com
  */
 #ifndef BN_H_
 #define BN_H_
 
+#include "tclInt.h"
 #include "tclTomMathDecls.h"
 #ifndef MODULE_SCOPE
 #define MODULE_SCOPE extern
@@ -22,33 +23,15 @@
 
 
 
-#ifndef MIN
-#   define MIN(x,y) ((x)<(y)?(x):(y))
-#endif
-
-#ifndef MAX
-#   define MAX(x,y) ((x)>(y)?(x):(y))
-#endif
-
 #ifdef __cplusplus
 extern "C" {
-
-/* C++ compilers don't like assigning void * to mp_digit * */
-#define  OPT_CAST(x)  (x *)
-
-#else
-
-/* C on the other hand doesn't care */
-#define  OPT_CAST(x)
-
 #endif
 
-
 /* detect 64-bit mode if possible */
-#if defined(NEVER)  /* 128-bit ints fail in too many places */
-#   if !(defined(MP_64BIT) && defined(MP_16BIT) && defined(MP_8BIT))
-#	define MP_64BIT
-#   endif
+#if defined(NEVER) /* 128-bit ints fail in too many places */
+   #if !(defined(MP_32BIT) || defined(MP_16BIT) || defined(MP_8BIT))
+      #define MP_64BIT
+   #endif
 #endif
 
 /* some default configurations.
@@ -61,83 +44,89 @@ extern "C" {
  */
 #ifdef MP_8BIT
 #ifndef MP_DIGIT_DECLARED
-   typedef unsigned char      mp_digit;
+   typedef uint8_t              mp_digit;
 #define MP_DIGIT_DECLARED
 #endif
-   typedef unsigned short     mp_word;
+   typedef uint16_t             mp_word;
+#define MP_SIZEOF_MP_DIGIT      1
+#ifdef DIGIT_BIT
+#error You must not define DIGIT_BIT when using MP_8BIT
+#endif
 #elif defined(MP_16BIT)
 #ifndef MP_DIGIT_DECLARED
-   typedef unsigned short     mp_digit;
+   typedef uint16_t             mp_digit;
 #define MP_DIGIT_DECLARED
 #endif
-   typedef unsigned long      mp_word;
+   typedef uint32_t             mp_word;
+#define MP_SIZEOF_MP_DIGIT      2
+#ifdef DIGIT_BIT
+#error You must not define DIGIT_BIT when using MP_16BIT
+#endif
 #elif defined(MP_64BIT)
    /* for GCC only on supported platforms */
 #ifndef CRYPT
-   typedef unsigned long long ulong64;
-   typedef signed long long   long64;
+   typedef unsigned long long   ulong64;
+   typedef signed long long     long64;
 #endif
 
 #ifndef MP_DIGIT_DECLARED
-   typedef unsigned long      mp_digit;
+   typedef ulong64 mp_digit;
 #define MP_DIGIT_DECLARED
 #endif
-   typedef unsigned long      mp_word __attribute__ ((mode(TI)));
+#if defined(_WIN32)
+   typedef unsigned __int128    mp_word;
+#elif defined(__GNUC__)
+   typedef unsigned long        mp_word __attribute__ ((mode(TI)));
+#else
+   /* it seems you have a problem
+    * but we assume you can somewhere define your own uint128_t */
+   typedef uint128_t            mp_word;
+#endif
 
-#  define DIGIT_BIT          60
+   #define DIGIT_BIT            60
 #else
    /* this is the default case, 28-bit digits */
-   
+
    /* this is to make porting into LibTomCrypt easier :-) */
 #ifndef CRYPT
-#  if defined(_MSC_VER) || defined(__BORLANDC__)
-      typedef unsigned __int64   ulong64;
-      typedef signed __int64     long64;
-#  else
-      typedef unsigned long long ulong64;
-      typedef signed long long   long64;
-#  endif
+   typedef unsigned long long   ulong64;
+   typedef signed long long     long64;
 #endif
 
 #ifndef MP_DIGIT_DECLARED
-   typedef unsigned int      mp_digit;
+   typedef uint32_t             mp_digit;
 #define MP_DIGIT_DECLARED
 #endif
-   typedef ulong64            mp_word;
+   typedef ulong64              mp_word;
 
-#ifdef MP_31BIT   
+#ifdef MP_31BIT
    /* this is an extension that uses 31-bit digits */
-#  define DIGIT_BIT          31
+   #define DIGIT_BIT            31
 #else
    /* default case is 28-bit digits, defines MP_28BIT as a handy macro to test */
-#  define DIGIT_BIT          28
-#  define MP_28BIT
-#endif   
-#endif
-
-/* define heap macros */
-#if 0 /* these are macros in tclTomMathDecls.h */
-#ifndef CRYPT
-   /* default to libc stuff */
-#  ifndef XMALLOC
-#     define XMALLOC  malloc
-#     define XFREE    free
-#     define XREALLOC realloc
-#     define XCALLOC  calloc
-#  else
-      /* prototypes for our heap functions */
-      extern void *XMALLOC(size_t n);
-      extern void *XREALLOC(void *p, size_t n);
-      extern void *XCALLOC(size_t n, size_t s);
-      extern void XFREE(void *p);
-#  endif
+   #define DIGIT_BIT            28
+   #define MP_28BIT
 #endif
 #endif
 
-
 /* otherwise the bits per digit is calculated automatically from the size of a mp_digit */
 #ifndef DIGIT_BIT
-#   define DIGIT_BIT     ((int)((CHAR_BIT * sizeof(mp_digit) - 1)))  /* bits per digit */
+   #define DIGIT_BIT     (((CHAR_BIT * MP_SIZEOF_MP_DIGIT) - 1))  /* bits per digit */
+   typedef uint_least32_t mp_min_u32;
+#else
+   typedef mp_digit mp_min_u32;
+#endif
+
+/* platforms that can use a better rand function */
+#if defined(__FreeBSD__) || defined(__OpenBSD__) || defined(__NetBSD__) || defined(__DragonFly__)
+    #define MP_USE_ALT_RAND 1
+#endif
+
+/* use arc4random on platforms that support it */
+#ifdef MP_USE_ALT_RAND
+    #define MP_GEN_RANDOM()    arc4random()
+#else
+    #define MP_GEN_RANDOM()    rand()
 #endif
 
 #define MP_DIGIT_BIT     DIGIT_BIT
@@ -180,15 +169,15 @@ MODULE_SCOPE int KARATSUBA_MUL_CUTOFF,
 
 /* default precision */
 #ifndef MP_PREC
-#  ifndef MP_LOW_MEM
-#     define MP_PREC                 32     /* default digits of precision */
-#  else
-#     define MP_PREC                 8      /* default digits of precision */
-#  endif
+   #ifndef MP_LOW_MEM
+      #define MP_PREC                 32     /* default digits of precision */
+   #else
+      #define MP_PREC                 8      /* default digits of precision */
+   #endif
 #endif
 
 /* size of comba arrays, should be at least 2 * 2**(BITS_PER_WORD - BITS_PER_DIGIT*2) */
-#define MP_WARRAY               (1 << (sizeof(mp_word) * CHAR_BIT - 2 * DIGIT_BIT + 1))
+#define MP_WARRAY               (1 << (((sizeof(mp_word) * CHAR_BIT) - (2 * DIGIT_BIT)) + 1))
 
 /* the infamous mp_int structure */
 #ifndef MP_INT_DECLARED
@@ -209,9 +198,7 @@ typedef int ltm_prime_callback(unsigned char *dst, int len, void *dat);
 #define SIGN(m)    ((m)->sign)
 
 /* error code to char* string */
-/*
-char *mp_error_to_string(int code);
-*/
+const char *mp_error_to_string(int code);
 
 /* ---> init and deinit bignum functions <--- */
 /* init a bignum */
@@ -256,8 +243,9 @@ int mp_init_size(mp_int *a, int size);
 
 /* ---> Basic Manipulations <--- */
 #define mp_iszero(a) (((a)->used == 0) ? MP_YES : MP_NO)
-#define mp_iseven(a) (((a)->used == 0 || (((a)->dp[0] & 1) == 0)) ? MP_YES : MP_NO)
-#define mp_isodd(a)  (((a)->used > 0 && (((a)->dp[0] & 1) == 1)) ? MP_YES : MP_NO)
+#define mp_iseven(a) ((((a)->used > 0) && (((a)->dp[0] & 1u) == 0u)) ? MP_YES : MP_NO)
+#define mp_isodd(a)  ((((a)->used > 0) && (((a)->dp[0] & 1u) == 1u)) ? MP_YES : MP_NO)
+#define mp_isneg(a)  (((a)->sign != MP_ZPOS) ? MP_YES : MP_NO)
 
 /* set to zero */
 /*
@@ -274,9 +262,25 @@ void mp_set(mp_int *a, mp_digit b);
 int mp_set_int(mp_int *a, unsigned long b);
 */
 
+/* set a platform dependent unsigned long value */
+/*
+int mp_set_long(mp_int *a, unsigned long b);
+*/
+
+/* set a platform dependent unsigned long long value */
+/*
+int mp_set_long_long(mp_int *a, unsigned long long b);
+*/
+
 /* get a 32-bit value */
 unsigned long mp_get_int(mp_int * a);
 
+/* get a platform dependent unsigned long value */
+unsigned long mp_get_long(mp_int * a);
+
+/* get a platform dependent unsigned long long value */
+unsigned long long mp_get_long_long(mp_int * a);
+
 /* initialize and set a digit */
 /*
 int mp_init_set (mp_int * a, mp_digit b);
@@ -302,6 +306,16 @@ int mp_init_copy(mp_int *a, mp_int *b);
 void mp_clamp(mp_int *a);
 */
 
+/* import binary data */
+/*
+int mp_import(mp_int* rop, size_t count, int order, size_t size, int endian, size_t nails, const void* op);
+*/
+
+/* export binary data */
+/*
+int mp_export(void* rop, size_t* countp, int order, size_t size, int endian, size_t nails, mp_int* op);
+*/
+
 /* ---> digit manipulation <--- */
 
 /* right shift by "b" digits */
@@ -314,7 +328,7 @@ void mp_rshd(mp_int *a, int b);
 int mp_lshd(mp_int *a, int b);
 */
 
-/* c = a / 2**b */
+/* c = a / 2**b, implemented as c = a >> b */
 /*
 int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d);
 */
@@ -324,7 +338,7 @@ int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d);
 int mp_div_2(mp_int *a, mp_int *b);
 */
 
-/* c = a * 2**b */
+/* c = a * 2**b, implemented as c = a << b */
 /*
 int mp_mul_2d(const mp_int *a, int b, mp_int *c);
 */
@@ -334,7 +348,7 @@ int mp_mul_2d(const mp_int *a, int b, mp_int *c);
 int mp_mul_2(mp_int *a, mp_int *b);
 */
 
-/* c = a mod 2**d */
+/* c = a mod 2**b */
 /*
 int mp_mod_2d(const mp_int *a, int b, mp_int *c);
 */
@@ -460,6 +474,9 @@ int mp_div_3(mp_int *a, mp_int *c, mp_digit *d);
 /*
 int mp_expt_d(mp_int *a, mp_digit b, mp_int *c);
 */
+/*
+int mp_expt_d_ex (mp_int * a, mp_digit b, mp_int * c, int fast);
+*/
 
 /* c = a mod b, 0 <= c < b  */
 /*
@@ -515,12 +532,20 @@ int mp_lcm(mp_int *a, mp_int *b, mp_int *c);
 /*
 int mp_n_root(mp_int *a, mp_digit b, mp_int *c);
 */
+/*
+int mp_n_root_ex (mp_int * a, mp_digit b, mp_int * c, int fast);
+*/
 
 /* special sqrt algo */
 /*
 int mp_sqrt(mp_int *arg, mp_int *ret);
 */
 
+/* special sqrt (mod prime) */
+/*
+int mp_sqrtmod_prime(mp_int *arg, mp_int *prime, mp_int *ret);
+*/
+
 /* is number a square? */
 /*
 int mp_is_square(mp_int *arg, int *ret);
@@ -623,7 +648,7 @@ int mp_exptmod(mp_int *a, mp_int *b, mp_int *c, mp_int *d);
 
 /* table of first PRIME_SIZE primes */
 #if defined(BUILD_tcl) || !defined(_WIN32)
-MODULE_SCOPE const mp_digit ltm_prime_tab[];
+MODULE_SCOPE const mp_digit ltm_prime_tab[PRIME_SIZE];
 #endif
 
 /* result=1 if a is divisible by one of the first PRIME_SIZE primes */
@@ -646,7 +671,7 @@ int mp_prime_miller_rabin(mp_int *a, mp_int *b, int *result);
 */
 
 /* This gives [for a given bit size] the number of trials required
- * such that Miller-Rabin gives a prob of failure lower than 2^-96 
+ * such that Miller-Rabin gives a prob of failure lower than 2^-96
  */
 /*
 int mp_prime_rabin_miller_trials(int size);
@@ -673,7 +698,7 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style);
 */
 
 /* makes a truly random prime of a given size (bytes),
- * call with bbs = 1 if you want it to be congruent to 3 mod 4 
+ * call with bbs = 1 if you want it to be congruent to 3 mod 4
  *
  * You have to supply a callback which fills in a buffer with random bytes.  "dat" is a parameter you can
  * have passed to the callback (e.g. a state or something).  This function doesn't use "dat" itself
@@ -686,10 +711,9 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style);
 /* makes a truly random prime of a given size (bits),
  *
  * Flags are as follows:
- * 
+ *
  *   LTM_PRIME_BBS      - make prime congruent to 3 mod 4
  *   LTM_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies LTM_PRIME_BBS)
- *   LTM_PRIME_2MSB_OFF - make the 2nd highest bit zero
  *   LTM_PRIME_2MSB_ON  - make the 2nd highest bit one
  *
  * You have to supply a callback which fills in a buffer with random bytes.  "dat" is a parameter you can
@@ -745,12 +769,14 @@ int mp_toradix_n(mp_int * a, char *str, int radix, int maxlen);
 int mp_radix_size(mp_int *a, int radix, int *size);
 */
 
+#ifndef LTM_NO_FILE
 /*
 int mp_fread(mp_int *a, int radix, FILE *stream);
 */
 /*
 int mp_fwrite(mp_int *a, int radix, FILE *stream);
 */
+#endif
 
 #define mp_read_raw(mp, str, len) mp_read_signed_bin((mp), (str), (len))
 #define mp_raw_size(mp)           mp_signed_bin_size(mp)
@@ -764,69 +790,14 @@ int mp_fwrite(mp_int *a, int radix, FILE *stream);
 #define mp_todecimal(M, S) mp_toradix((M), (S), 10)
 #define mp_tohex(M, S)     mp_toradix((M), (S), 16)
 
-/* lowlevel functions, do not call! */
-/*
-int s_mp_add(mp_int *a, mp_int *b, mp_int *c);
-*/
-/*
-int s_mp_sub(mp_int *a, mp_int *b, mp_int *c);
-*/
-#define s_mp_mul(a, b, c) s_mp_mul_digs(a, b, c, (a)->used + (b)->used + 1)
-/*
-int fast_s_mp_mul_digs(mp_int *a, mp_int *b, mp_int *c, int digs);
-*/
-/*
-int s_mp_mul_digs(mp_int *a, mp_int *b, mp_int *c, int digs);
-*/
-/*
-int fast_s_mp_mul_high_digs(mp_int *a, mp_int *b, mp_int *c, int digs);
-*/
-/*
-int s_mp_mul_high_digs(mp_int *a, mp_int *b, mp_int *c, int digs);
-*/
-/*
-int fast_s_mp_sqr(mp_int *a, mp_int *b);
-*/
-/*
-int s_mp_sqr(mp_int *a, mp_int *b);
-*/
-/*
-int mp_karatsuba_mul(mp_int *a, mp_int *b, mp_int *c);
-*/
-/*
-int mp_toom_mul(mp_int *a, mp_int *b, mp_int *c);
-*/
-/*
-int mp_karatsuba_sqr(mp_int *a, mp_int *b);
-*/
-/*
-int mp_toom_sqr(mp_int *a, mp_int *b);
-*/
-/*
-int fast_mp_invmod(mp_int *a, mp_int *b, mp_int *c);
-*/
-/*
-int mp_invmod_slow (mp_int * a, mp_int * b, mp_int * c);
-*/
-/*
-int fast_mp_montgomery_reduce(mp_int *a, mp_int *m, mp_digit mp);
-*/
-/*
-int mp_exptmod_fast(mp_int *G, mp_int *X, mp_int *P, mp_int *Y, int mode);
-*/
-/*
-int s_mp_exptmod (mp_int * G, mp_int * X, mp_int * P, mp_int * Y, int mode);
-*/
-/*
-void bn_reverse(unsigned char *s, int len);
-*/
-
-#if defined(BUILD_tcl) || !defined(_WIN32)
-MODULE_SCOPE const char *mp_s_rmap;
-#endif
-
 #ifdef __cplusplus
-}
+   }
 #endif
 
 #endif
+
+
+/* $Source$ */
+/* $Revision$ */
+/* $Date$ */
+