Annotation of pgp/src/mpilib.h, revision 1.1.1.3

1.1.1.2   root        1: /*     C include file for MPI multiprecision integer math routines.
                      2: 
                      3:        Boulder Software Engineering
                      4:        3021 Eleventh Street
                      5:        Boulder, CO 80304
                      6:        (303) 541-0140
                      7: 
                      8:        (c) Copyright 1986-92 by Philip Zimmermann.  All rights reserved.
                      9:        The author assumes no liability for damages resulting from the use
                     10:        of this software, even if the damage results from defects in this
                     11:        software.  No warranty is expressed or implied.
                     12: 
                     13:        These routines implement all of the multiprecision arithmetic necessary
                     14:        for Rivest-Shamir-Adleman (RSA) public key cryptography, as well as
                     15:        other number-theoretic algorithms such as ElGamal, Diffie-Hellman,
                     16:        or Rabin.
                     17: 
                     18:        Although originally developed in Microsoft C for the IBM PC, this code
                     19:        contains few machine dependencies.  It assumes 2's complement
                     20:        arithmetic.  It can be adapted to 8-bit, 16-bit, or 32-bit machines,
                     21:        lowbyte-highbyte order or highbyte-lowbyte order.  This version
                     22:        has been converted to ANSI C.
                     23: 
                     24:        Modified 8 Apr 92 - HAJK - Implement new VAX/VMS primitive support.
                     25:        Modified 29 Nov 92 - Thad Smith
                     26: */
                     27: 
                     28: #include <string.h>
                     29: #include "usuals.h"  /* typedefs for byte, word16, boolean, etc. */
                     30: #include "platform.h" /* customization for different environments */
                     31: 
                     32: /* Platform customization:
                     33:  * A version which runs on almost any computer can be implemented by
                     34:  * defining PORTABLE and MPORTABLE, preferably as a command line
                     35:  * parameter.  Faster versions can be generated by specifying specific
                     36:  * parameters, such as size of unit and MULTUNIT, and by supplying some
                     37:  * of the critical in assembly.  See the file platform.h for more 
                     38:  * details on customization.
                     39:  *
                     40:  * The symbol HIGHFIRST, designating that integers and longs are stored
                     41:  * with the most significant bit in the lowest address, should be defined
                     42:  * on the command line for compiling all files, since it is used by files
                     43:  * other than the mpilib routines.
                     44:  */
                     45:  
                     46: #ifndef ALIGN
                     47: #define ALIGN
                     48: #endif
                     49: 
                     50: #ifndef PEASANT /* if not Russian peasant modulo multiply algorithm */
                     51: #ifndef MERRITT /* if not Merritt's modmult */
                     52: #ifndef UPTON   /* if not Upton's modmult */
                     53: #ifndef SMITH
                     54: #define SMITH   /* default: use Smith's modmult algorithm */
                     55: #endif
                     56: #endif
                     57: #endif
                     58: #endif
                     59: 
                     60: #ifdef SMITH
                     61: #define UPTON_OR_SMITH  /* enable common code */
                     62: #endif
                     63: #ifdef UPTON
                     64: #define UPTON_OR_SMITH  /* enable common code */
                     65: #endif
                     66: 
                     67: #ifndef UNIT32
                     68: #ifndef UNIT8
                     69: #define UNIT16                 /* default--use 16-bit units */
                     70: #endif
                     71: #endif
                     72: 
                     73: /***   CAUTION:  If your machine has an unusual word size that is not a
                     74:        power of 2 (8, 16, 32, or 64) bits wide, then the macros here that
                     75:        use the symbol "LOG_UNITSIZE" must be changed.
                     76: ***/
                     77: 
                     78: #ifdef UNIT8
                     79: typedef unsigned char unit;
                     80: typedef signed char signedunit;
                     81: #define UNITSIZE 8 /* number of bits in a unit */
                     82: #define LOG_UNITSIZE 3
                     83: #define uppermostbit ((unit) 0x80)
                     84: #define BYTES_PER_UNIT 1 /* number of bytes in a unit */
                     85: #define units2bits(n) ((n) << 3) /* fast multiply by UNITSIZE */
                     86: #define units2bytes(n) (n)
                     87: #define bits2units(n) (((n)+7) >> 3)
                     88: #define bytes2units(n) (n)
                     89: #endif
                     90: 
                     91: #ifdef UNIT16
                     92: typedef word16 unit;
                     93: typedef short signedunit;
                     94: #define UNITSIZE 16 /* number of bits in a unit */
                     95: #define LOG_UNITSIZE 4
                     96: #define uppermostbit ((unit) 0x8000)
                     97: #define BYTES_PER_UNIT 2 /* number of bytes in a unit */
                     98: #define units2bits(n) ((n) << 4) /* fast multiply by UNITSIZE */
                     99: #define units2bytes(n) ((n) << 1)
                    100: #define bits2units(n) (((n)+15) >> 4)
                    101: #define bytes2units(n) (((n)+1) >> 1)
                    102: #endif
                    103: 
                    104: #ifdef UNIT32
                    105: typedef word32 unit;
                    106: typedef long signedunit;
                    107: #define UNITSIZE 32 /* number of bits in a unit */
                    108: #define LOG_UNITSIZE 5
                    109: #define uppermostbit ((unit) 0x80000000L)
                    110: #define BYTES_PER_UNIT 4 /* number of bytes in a unit */
                    111: #define units2bits(n) ((n) << 5) /* fast multiply by UNITSIZE */
                    112: #define units2bytes(n) ((n) << 2)
                    113: #define bits2units(n) (((n)+31) >> 5)
                    114: #define bytes2units(n) (((n)+3) >> 2)
                    115: #endif
                    116: 
                    117: #define power_of_2(b) ((unit) 1 << (b)) /* computes power-of-2 bit masks */
                    118: #define bits2bytes(n) (((n)+7) >> 3)
                    119: /*     Some C compilers (like the ADSP2101) will not always collapse constant 
                    120:        expressions at compile time if the expressions contain shift operators. */
                    121: /* #define uppermostbit power_of_2(UNITSIZE-1) */
                    122: /* #define UNITSIZE units2bits(1) */ /* number of bits in a unit */
                    123: /* #define bytes2units(n) bits2units((n)<<3) */
                    124: /* #define BYTES_PER_UNIT (UNITSIZE >> 3) */
                    125: /* LOG_UNITSIZE is the log base 2 of UNITSIZE, ie: 4 for 16-bit units */
                    126: /* #define units2bits(n) ((n) << LOG_UNITSIZE) */ /* fast multiply by UNITSIZE */
                    127: /* #define units2bytes(n) ((n) << (LOG_UNITSIZE-3)) */
                    128: /* #define bits2units(n) (((n)+(UNITSIZE-1)) >> LOG_UNITSIZE) */
                    129: /* #define bytes2units(n) (((n)+(BYTES_PER_UNIT-1)) >> (LOG_UNITSIZE-3)) */
                    130: 
                    131: typedef unit *unitptr;
                    132: 
                    133: 
                    134: /*--------------------- Byte ordering stuff -------------------*/
                    135: #ifdef HIGHFIRST
                    136: 
                    137: /* these definitions assume MSB comes first */
                    138: #define tohigher(n)            (-(n))  /* offset towards higher unit */
                    139: #define pre_higherunit(r)      (--(r))
                    140: #define pre_lowerunit(r)       (++(r))
                    141: #define post_higherunit(r)     ((r)--)
                    142: #define post_lowerunit(r)      ((r)++)
                    143: #define bit_index(n)           (global_precision-bits2units((n)+1))
                    144: #define lsbptr(r,prec)         ((r)+(prec)-1)
                    145: #define make_lsbptr(r,prec)    (r) = lsbptr(r,prec)  
                    146: #define unmake_lsbptr(r,prec) (r) = ((r)-(prec)+1) 
                    147: #define msbptr(r,prec)         (r)
                    148: #define make_msbptr(r,prec)    /* (r) = msbptr(r,prec) */
                    149: 
                    150: /* The macro rescale(r,current_precision,new_precision) rescales
                    151:    a multiprecision integer by adjusting r and its precision to new values.
                    152:    It can be used to reverse the effects of the normalize
                    153:    routine given above.  See the comments in normalize concerning
                    154:    Intel vs. Motorola LSB/MSB conventions.
                    155:    WARNING:  You can only safely call rescale on registers that
                    156:    you have previously normalized with the above normalize routine,
                    157:    or are known to be big enough for the new precision.  You may
                    158:    specify a new precision that is smaller than the current precision.
                    159:    You must be careful not to specify a new_precision value that is
                    160:    too big, or which adjusts the r pointer out of range.
                    161: */
                    162: #define rescale(r,currentp,newp) r -= ((newp) - (currentp))
                    163: 
                    164: /* The macro normalize(r,precision) "normalizes" a multiprecision integer
                    165:    by adjusting r and precision to new values.  For Motorola-style processors
                    166:    (MSB-first), r is a pointer to the MSB of the register, and must
                    167:    be adjusted to point to the first nonzero unit.  For Intel/VAX-style
                    168:    (LSB-first) processors, r is a  pointer to the LSB of the register,
                    169:    and must be left unchanged.  The precision counter is always adjusted,
                    170:    regardless of processor type.  In the case of precision = 0,
                    171:    r becomes undefined.
                    172: */
                    173: #define normalize(r,prec) \
                    174:   { prec = significance(r); r += (global_precision-(prec)); }
                    175: 
                    176: #else  /* LOWFIRST byte order */
                    177: 
                    178: /* these definitions assume LSB comes first */
                    179: #define tohigher(n)            (n)     /* offset towards higher unit */
                    180: #define pre_higherunit(r)      (++(r))
                    181: #define pre_lowerunit(r)       (--(r))
                    182: #define post_higherunit(r)     ((r)++)
                    183: #define post_lowerunit(r)      ((r)--)
                    184: #define bit_index(n)           (bits2units((n)+1)-1)
                    185: #define lsbptr(r,prec)         (r)
                    186: #define make_lsbptr(r,prec)    /* (r) = lsbptr(r,prec) */  
                    187: #define unmake_lsbptr(r,prec) /* (r) = (r) */  
                    188: #define msbptr(r,prec)         ((r)+(prec)-1)
                    189: #define make_msbptr(r,prec)    (r) = msbptr(r,prec)
                    190: 
                    191: #define rescale(r,currentp,newp) /* nil statement */
                    192: #define normalize(r,prec) prec = significance(r) 
                    193: 
                    194: #endif /* LOWFIRST byte order */
                    195: /*------------------ End byte ordering stuff -------------------*/
                    196: 
                    197: /*     Note that the address calculations require that lsbptr, msbptr, 
                    198:        make_lsbptr, make_msbptr, mp_tstbit, mp_setbit, mp_clrbit, 
                    199:        and bitptr all have unitptr arguments, not byte pointer arguments.      */
                    200: #define bitptr(r,n) &((r)[bit_index(n)])
                    201: #define bitmsk(n) power_of_2((n) & (UNITSIZE-1))
                    202:        /* bitmsk() assumes UNITSIZE is a power of 2 */
                    203: #define mp_tstbit(r,n) (*bitptr(r,n) &   bitmsk(n))
                    204: #define mp_setbit(r,n) (*bitptr(r,n) |=  bitmsk(n))
                    205: #define mp_clrbit(r,n) (*bitptr(r,n) &= ~bitmsk(n))
                    206: #define msunit(r) (*msbptr(r,global_precision))
                    207: #define lsunit(r) (*lsbptr(r,global_precision))
                    208: /* #define mp_tstminus(r) ((msunit(r) & uppermostbit)!=0) */
                    209: #define mp_tstminus(r) ((signedunit) msunit(r) < 0)
                    210: 
                    211: 
                    212:        /* set working precision to specified number of bits. */
                    213: #ifdef mp_setp
                    214: void mp_setp(short nbits);
                    215: #define set_precision(prec)    mp_setp(units2bits(global_precision=(prec)))
                    216: #else
                    217: #define set_precision(prec) (global_precision = (prec))
                    218: #endif
                    219: 
                    220: 
                    221: #ifdef PEASANT
                    222: 
                    223: /* Define C names for Russian peasant modmult primitives. */
                    224: #define stage_modulus  stage_peasant_modulus
                    225: #define mp_modmult     peasant_modmult
                    226: #define modmult_burn   peasant_burn
                    227: #define SLOP_BITS      PEASANT_SLOP_BITS
                    228: 
                    229: #else  /* not PEASANT */
                    230: #ifdef MERRITT
                    231: /* Define C names for Merritt's modmult primitives. */
                    232: #define stage_modulus  stage_merritt_modulus
                    233: #define mp_modmult     merritt_modmult
                    234: #define modmult_burn   merritt_burn
                    235: #define SLOP_BITS      MERRITT_SLOP_BITS
                    236: 
                    237: #else  /* not PEASANT, MERRITT */
                    238: #ifdef UPTON
                    239: /* Define C names for Upton's modmult primitives. */
                    240: #define stage_modulus  stage_upton_modulus
                    241: #define mp_modmult     upton_modmult
                    242: #define modmult_burn   upton_burn
                    243: #define SLOP_BITS      UPTON_SLOP_BITS
                    244: 
                    245: #else  /* not PEASANT, MERRITT, UPTON */
                    246: #ifdef SMITH
                    247: /* Define C names for Smith's modmult primitives. */
                    248: #define stage_modulus  stage_smith_modulus
                    249: #define mp_modmult     smith_modmult
                    250: #define modmult_burn   smith_burn
                    251: #define SLOP_BITS      SMITH_SLOP_BITS
                    252: 
                    253: #endif /* SMITH */
                    254: #endif /* UPTON */
                    255: #endif /* MERRITT */
                    256: #endif /* PEASANT */
                    257: 
                    258: 
                    259: #define mp_shift_left(r1) mp_rotate_left(r1,(boolean)0)
                    260:        /* multiprecision shift left 1 bit */
                    261: 
                    262: #define mp_add(r1,r2) mp_addc(r1,r2,(boolean)0)
                    263:        /* multiprecision add with no carry */
                    264: 
                    265: #define mp_sub(r1,r2) mp_subb(r1,r2,(boolean)0)
                    266:        /* multiprecision subtract with no borrow */
                    267: 
                    268: #define mp_abs(r) (mp_tstminus(r) ? (mp_neg(r),TRUE) : FALSE)
                    269: 
                    270: #define msub(r,m) if (mp_compare(r,m) >= 0) mp_sub(r,m)
                    271:        /* Prevents r from getting bigger than modulus m */
                    272: 
                    273: #define testeq(r,i)    \
                    274:        ( (lsunit(r)==(i)) && (significance(r)<=1) )
                    275: 
                    276: #define testne(r,i)    \
                    277:        ( (lsunit(r)!=(i)) || (significance(r)>1) )
                    278: 
                    279: #define testge(r,i)    \
                    280:        ( (lsunit(r)>=(i)) || (significance(r)>1) )
                    281: 
                    282: #define testle(r,i)    \
                    283:        ( (lsunit(r)<=(i)) && (significance(r)<=1) )
                    284: 
                    285: #define mp_square(r1,r2) mp_mult(r1,r2,r2)
                    286:        /* Square r2, returning product in r1 */
                    287: 
                    288: #define mp_modsquare(r1,r2) mp_modmult(r1,r2,r2)
                    289:        /* Square r2, returning modulo'ed product in r1 */
                    290: 
                    291: #define countbytes(r) ((countbits(r)+7)>>3)
                    292: 
                    293: /*     SLOP_BITS is how many "carry bits" to allow for intermediate
                    294:        calculation results to exceed the size of the modulus.
                    295:        It is used by modexp to give some overflow elbow room for
                    296:        modmult to use to perform modulo operations with the modulus.
                    297:        The number of slop bits required is determined by the modmult
                    298:        algorithm.  The Russian peasant modmult algorithm only requires
                    299:        1 slop bit, for example.  Note that if we use an external assembly
                    300:        modmult routine, SLOP_BITS may be meaningless or may be defined in a
                    301:        non-constant manner.
                    302: */
                    303: #define PEASANT_SLOP_BITS      1
                    304: #define MERRITT_SLOP_BITS      UNITSIZE
                    305: #define UPTON_SLOP_BITS        (UNITSIZE/2)
                    306: #ifdef  mp_smul /* old version requires MS word = 0 */
                    307: #define SMITH_SLOP_BITS        UNITSIZE
                    308: #else           /* mp_smula or C version of mp_smul */
                    309: #define SMITH_SLOP_BITS        0
                    310: #endif /* mp_smul */
                    311: 
1.1.1.3 ! root      312: /*     MAX_BIT_PRECISION is upper limit that assembly primitives can handle.
        !           313:        It must be less than 32704 bits, or 4088 bytes.  It should be an
        !           314:        integer multiple of UNITSIZE*2.
        !           315: */
        !           316: #if (SLOP_BITS > 0)
        !           317: #define MAX_BIT_PRECISION (1152+(2*UNITSIZE))
        !           318: #else
        !           319: #define MAX_BIT_PRECISION 1152
        !           320: #endif
        !           321: #define MAX_BYTE_PRECISION (MAX_BIT_PRECISION/8)
        !           322: #define MAX_UNIT_PRECISION (MAX_BIT_PRECISION/UNITSIZE)
        !           323: 
1.1.1.2   root      324: 
                    325: /* global_precision is the unit precision last set by set_precision */
                    326: extern short global_precision;
                    327: 
                    328: 
                    329: /*     The "bit sniffer" macros all begin sniffing at the MSB.
                    330:        They are used internally by all the various multiply, divide, 
                    331:        modulo, exponentiation, and square root functions.
                    332: */
                    333: #define sniff_bit(bptr,bitmask)        (*(bptr) & bitmask)
                    334: 
                    335: #define init_bitsniffer(bptr,bitmask,prec,bits) \
                    336: { normalize(bptr,prec); \
                    337:   if (!prec) \
                    338:     return(0); \
                    339:   bits = units2bits(prec); \
                    340:   make_msbptr(bptr,prec); bitmask = uppermostbit; \
                    341:   while (!sniff_bit(bptr,bitmask)) \
                    342:   { bitmask >>= 1; bits--; \
                    343:   } \
                    344: }
                    345: 
                    346: #define bump_bitsniffer(bptr,bitmask) \
                    347: { if (!(bitmask >>= 1)) \
                    348:   { bitmask = uppermostbit; \
                    349:        post_lowerunit(bptr); \
                    350:   } \
                    351: }
                    352: 
                    353: /* bump_2bitsniffers is used internally by mp_udiv. */
                    354: #define bump_2bitsniffers(bptr,bptr2,bitmask) \
                    355: { if (!(bitmask >>= 1)) \
                    356:   { bitmask = uppermostbit; \
                    357:     post_lowerunit(bptr); \
                    358:     post_lowerunit(bptr2); \
                    359:   } \
                    360: }
                    361: 
                    362: /* stuff_bit is used internally by mp_udiv and mp_sqrt. */
                    363: #define stuff_bit(bptr,bitmask)        *(bptr) |= bitmask
                    364: 
                    365: 
                    366: boolean mp_addc
                    367:        (register unitptr r1,register unitptr r2,register boolean carry);
                    368:        /* multiprecision add with carry r2 to r1, result in r1 */
                    369: 
                    370: boolean mp_subb
                    371:        (register unitptr r1,register unitptr r2,register boolean borrow);
                    372:        /* multiprecision subtract with borrow, r2 from r1, result in r1 */
                    373: 
                    374: boolean mp_rotate_left(register unitptr r1,register boolean carry);
                    375:        /* multiprecision rotate left 1 bit with carry, result in r1. */
                    376: 
                    377: void mp_shift_right_bits(register unitptr r1,register short bits);
                    378:        /* multiprecision shift right bits, result in r1. */
                    379: 
                    380: short mp_compare(register unitptr r1,register unitptr r2);
                    381:        /* Compares registers *r1, *r2, and returns -1, 0, or 1 */
                    382: 
                    383: boolean mp_inc(register unitptr r);
                    384:        /* Increment multiprecision integer r. */
                    385: 
                    386: boolean mp_dec(register unitptr r);
                    387:        /* Decrement multiprecision integer r. */
                    388: 
                    389: void mp_neg(register unitptr r);
                    390:        /* Compute 2's complement, the arithmetic negative, of r */
                    391: 
                    392: #ifndef mp_move
                    393: #define mp_move(d,s)    memcpy((void*)(d), (void*)(s), \
                    394:                units2bytes(global_precision))
                    395: #endif  
                    396: #ifndef unitfill0
                    397: #define unitfill0(r,ct) memset((void*)(r), 0, units2bytes(ct))
                    398: #endif
                    399: 
                    400: #ifndef mp_burn
                    401: #define mp_burn(r) mp_init(r,0) /* for burning the evidence */
                    402: #define mp_init0(r) mp_init(r,0)
                    403: #endif
                    404: 
                    405: #define empty_array(r)  unitfill0(r, sizeof(r)/sizeof(r[0])/sizeof(unit))
                    406: 
                    407: void mp_init(register unitptr r, word16 value);
                    408:        /* Init multiprecision register r with short value. */
                    409: 
                    410: short significance(register unitptr r);
                    411:        /* Returns number of significant units in r */
                    412: 
                    413: int mp_udiv(register unitptr remainder,register unitptr quotient,
                    414:        register unitptr dividend,register unitptr divisor);
                    415:        /* Unsigned divide, treats both operands as positive. */
                    416: 
                    417: int mp_recip(register unitptr quotient,register unitptr divisor);
                    418:        /* Compute reciprocal as 1/divisor.  Used by faster modmult. */
                    419: 
                    420: int mp_div(register unitptr remainder,register unitptr quotient,
                    421:        register unitptr dividend,register unitptr divisor);
                    422:        /* Signed divide, either or both operands may be negative. */
                    423: 
                    424: word16 mp_shortdiv(register unitptr quotient,
                    425:        register unitptr dividend,register word16 divisor);
                    426:        /* Returns short remainder of unsigned divide. */
                    427: 
                    428: int mp_mod(register unitptr remainder,
                    429:        register unitptr dividend,register unitptr divisor);
                    430:        /* Unsigned divide, treats both operands as positive. */
                    431: 
                    432: word16 mp_shortmod(register unitptr dividend,register word16 divisor);
                    433:        /* Just returns short remainder of unsigned divide. */
                    434: 
                    435: int mp_mult(register unitptr prod,
                    436:        register unitptr multiplicand,register unitptr multiplier);
                    437:        /* Computes multiprecision prod = multiplicand * multiplier */
                    438: 
                    439: int countbits(unitptr r);
                    440:        /* Returns number of significant bits in r. */
                    441: 
                    442: int stage_peasant_modulus(unitptr n);
                    443: int stage_merritt_modulus(unitptr n);
                    444: int stage_upton_modulus(unitptr n);
                    445: int stage_smith_modulus(unitptr n);
                    446:        /* Must pass modulus to stage_modulus before calling modmult. */
                    447: 
                    448: int peasant_modmult(register unitptr prod,
                    449:        unitptr multiplicand,register unitptr multiplier);
                    450: int merritt_modmult(register unitptr prod,
                    451:        unitptr multiplicand,register unitptr multiplier);
                    452: int upton_modmult(register unitptr prod,
                    453:        unitptr multiplicand,register unitptr multiplier);
                    454: int smith_modmult(register unitptr prod,
                    455:        unitptr multiplicand,register unitptr multiplier);
                    456:        /* Performs combined multiply/modulo operation, with global modulus */
                    457:  
                    458:  
                    459: 
                    460: int mp_modexp(register unitptr expout,register unitptr expin,
                    461:        register unitptr exponent,register unitptr modulus);
                    462:        /* Combined exponentiation/modulo algorithm. */
                    463: 
                    464: int rsa_decrypt(unitptr M, unitptr C,
                    465:        unitptr d, unitptr p, unitptr q, unitptr u);
                    466:        /* Uses Chinese Remainder Theorem shortcut for RSA decryption. */
                    467: 
                    468: /****************** end of MPI library ****************************/

unix.superglobalmegacorp.com

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.