[crypto] Support direct reduction only for Montgomery constant R^2 mod N

The only remaining use case for direct reduction (outside of the unit
tests) is in calculating the constant R^2 mod N used during Montgomery
multiplication.

The current implementation of direct reduction requires a writable
copy of the modulus (to allow for shifting), and both the modulus and
the result buffer must be padded to be large enough to hold (R^2 - N),
which is twice the size of the actual values involved.

For the special case of reducing R^2 mod N (or any power of two mod
N), we can run the same algorithm without needing either a writable
copy of the modulus or a padded result buffer.  The working state
required is only two bits larger than the result buffer, and these
additional bits may be held in local variables instead.

Rewrite bigint_reduce() to handle only this use case, and remove the
no longer necessary uses of double-sized big integers.

Signed-off-by: Michael Brown <mcb30@ipxe.org>
pull/1411/head
Michael Brown 2025-02-13 13:35:45 +00:00
parent 5056e8ad93
commit 8e6b914c53
4 changed files with 247 additions and 259 deletions

View File

@ -27,7 +27,6 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
#include <string.h>
#include <assert.h>
#include <stdio.h>
#include <ipxe/profile.h>
#include <ipxe/bigint.h>
/** @file
@ -35,10 +34,6 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
* Big integer support
*/
/** Modular direct reduction profiler */
static struct profiler bigint_mod_profiler __profiler =
{ .name = "bigint_mod" };
/** Minimum number of least significant bytes included in transcription */
#define BIGINT_NTOA_LSB_MIN 16
@ -180,172 +175,136 @@ void bigint_multiply_raw ( const bigint_element_t *multiplicand0,
}
/**
* Reduce big integer
*
* @v modulus0 Element 0 of big integer modulus
* @v value0 Element 0 of big integer to be reduced
* @v size Number of elements in modulus and value
*/
void bigint_reduce_raw ( bigint_element_t *modulus0, bigint_element_t *value0,
unsigned int size ) {
bigint_t ( size ) __attribute__ (( may_alias ))
*modulus = ( ( void * ) modulus0 );
bigint_t ( size ) __attribute__ (( may_alias ))
*value = ( ( void * ) value0 );
const unsigned int width = ( 8 * sizeof ( bigint_element_t ) );
bigint_element_t *element;
unsigned int modulus_max;
unsigned int value_max;
unsigned int subshift;
int offset;
int shift;
int msb;
int i;
/* Start profiling */
profile_start ( &bigint_mod_profiler );
/* Normalise the modulus
*
* Scale the modulus by shifting left such that both modulus
* "m" and value "x" have the same most significant set bit.
* (If this is not possible, then the value is already less
* than the modulus, and we may therefore skip reduction
* completely.)
*/
value_max = bigint_max_set_bit ( value );
modulus_max = bigint_max_set_bit ( modulus );
shift = ( value_max - modulus_max );
if ( shift < 0 )
goto skip;
subshift = ( shift & ( width - 1 ) );
offset = ( shift / width );
element = modulus->element;
for ( i = ( ( value_max - 1 ) / width ) ; ; i-- ) {
element[i] = ( element[ i - offset ] << subshift );
if ( i <= offset )
break;
if ( subshift ) {
element[i] |= ( element[ i - offset - 1 ]
>> ( width - subshift ) );
}
}
for ( i-- ; i >= 0 ; i-- )
element[i] = 0;
/* Reduce the value "x" by iteratively adding or subtracting
* the scaled modulus "m".
*
* On each loop iteration, we maintain the invariant:
*
* -2m <= x < 2m
*
* If x is positive, we obtain the new value x' by
* subtracting m, otherwise we add m:
*
* 0 <= x < 2m => x' := x - m => -m <= x' < m
* -2m <= x < 0 => x' := x + m => -m <= x' < m
*
* and then halve the modulus (by shifting right):
*
* m' = m/2
*
* We therefore end up with:
*
* -m <= x' < m => -2m' <= x' < 2m'
*
* i.e. we have preseved the invariant while reducing the
* bounds on x' by one power of two.
*
* The issue remains of how to determine on each iteration
* whether or not x is currently positive, given that both
* input values are unsigned big integers that may use all
* available bits (including the MSB).
*
* On the first loop iteration, we may simply assume that x is
* positive, since it is unmodified from the input value and
* so is positive by definition (even if the MSB is set). We
* therefore unconditionally perform a subtraction on the
* first loop iteration.
*
* Let k be the MSB after normalisation. We then have:
*
* 2^k <= m < 2^(k+1)
* 2^k <= x < 2^(k+1)
*
* On the first loop iteration, we therefore have:
*
* x' = (x - m)
* < 2^(k+1) - 2^k
* < 2^k
*
* Any positive value of x' therefore has its MSB set to zero,
* and so we may validly treat the MSB of x' as a sign bit at
* the end of the first loop iteration.
*
* On all subsequent loop iterations, the starting value m is
* guaranteed to have its MSB set to zero (since it has
* already been shifted right at least once). Since we know
* from above that we preserve the loop invariant:
*
* -m <= x' < m
*
* we immediately know that any positive value of x' also has
* its MSB set to zero, and so we may validly treat the MSB of
* x' as a sign bit at the end of all subsequent loop
* iterations.
*
* After the last loop iteration (when m' has been shifted
* back down to the original value of the modulus), we may
* need to add a single multiple of m' to ensure that x' is
* positive, i.e. lies within the range 0 <= x' < m'. To
* allow for reusing the (inlined) expansion of
* bigint_subtract(), we achieve this via a potential
* additional loop iteration that performs the addition and is
* then guaranteed to terminate (since the result will be
* positive).
*/
for ( msb = 0 ; ( msb || ( shift >= 0 ) ) ; shift-- ) {
if ( msb ) {
bigint_add ( modulus, value );
} else {
bigint_subtract ( modulus, value );
}
msb = bigint_msb_is_set ( value );
if ( shift > 0 )
bigint_shr ( modulus );
}
skip:
/* Sanity check */
assert ( ! bigint_is_geq ( value, modulus ) );
/* Stop profiling */
profile_stop ( &bigint_mod_profiler );
}
/**
* Reduce supremum of big integer representation
* Reduce big integer R^2 modulo N
*
* @v modulus0 Element 0 of big integer modulus
* @v result0 Element 0 of big integer to hold result
* @v size Number of elements in modulus and value
* @v size Number of elements in modulus and result
*
* Reduce the value 2^k (where k is the bit width of the big integer
* representation) modulo the specified modulus.
* Reduce the value R^2 modulo N, where R=2^n and n is the number of
* bits in the representation of the modulus N, including any leading
* zero bits.
*/
void bigint_reduce_supremum_raw ( bigint_element_t *modulus0,
bigint_element_t *result0,
unsigned int size ) {
bigint_t ( size ) __attribute__ (( may_alias ))
*modulus = ( ( void * ) modulus0 );
void bigint_reduce_raw ( const bigint_element_t *modulus0,
bigint_element_t *result0, unsigned int size ) {
const bigint_t ( size ) __attribute__ (( may_alias ))
*modulus = ( ( const void * ) modulus0 );
bigint_t ( size ) __attribute__ (( may_alias ))
*result = ( ( void * ) result0 );
const unsigned int width = ( 8 * sizeof ( bigint_element_t ) );
unsigned int shift;
int max;
int sign;
int msb;
int carry;
/* Calculate (2^k) mod N via direct reduction of (2^k - N) mod N */
/* We have the constants:
*
* N = modulus
*
* n = number of bits in the modulus (including any leading zeros)
*
* R = 2^n
*
* Let r be the extension of the n-bit result register by a
* separate two's complement sign bit, such that -R <= r < R,
* and define:
*
* x = r * 2^k
*
* as the value being reduced modulo N, where k is a
* non-negative integer bit shift.
*
* We want to reduce the initial value R^2=2^(2n), which we
* may trivially represent using r=1 and k=2n.
*
* We then iterate over decrementing k, maintaining the loop
* invariant:
*
* -N <= r < N
*
* On each iteration we must first double r, to compensate for
* having decremented k:
*
* k' = k - 1
*
* r' = 2r
*
* x = r * 2^k = 2r * 2^(k-1) = r' * 2^k'
*
* Note that doubling the n-bit result register will create a
* value of n+1 bits: this extra bit needs to be handled
* separately during the calculation.
*
* We then subtract N (if r is currently non-negative) or add
* N (if r is currently negative) to restore the loop
* invariant:
*
* 0 <= r < N => r" = 2r - N => -N <= r" < N
* -N <= r < 0 => r" = 2r + N => -N <= r" < N
*
* Note that since N may use all n bits, the most significant
* bit of the n-bit result register is not a valid two's
* complement sign bit for r: the extra sign bit therefore
* also needs to be handled separately.
*
* Once we reach k=0, we have x=r and therefore:
*
* -N <= x < N
*
* After this last loop iteration (with k=0), we may need to
* add a single multiple of N to ensure that x is positive,
* i.e. lies within the range 0 <= x < N.
*
* Since neither the modulus nor the value R^2 are secret, we
* may elide approximately half of the total number of
* iterations by constructing the initial representation of
* R^2 as r=2^m and k=2n-m (for some m such that 2^m < N).
*/
/* Initialise x=R^2 */
memset ( result, 0, sizeof ( *result ) );
bigint_subtract ( modulus, result );
bigint_reduce ( modulus, result );
max = ( bigint_max_set_bit ( modulus ) - 2 );
if ( max < 0 ) {
/* Degenerate case of N=0 or N=1: return a zero result */
return;
}
bigint_set_bit ( result, max );
shift = ( ( 2 * size * width ) - max );
sign = 0;
/* Iterate as described above */
while ( shift-- ) {
/* Calculate 2r, storing extra bit separately */
msb = bigint_shl ( result );
/* Add or subtract N according to current sign */
if ( sign ) {
carry = bigint_add ( modulus, result );
} else {
carry = bigint_subtract ( modulus, result );
}
/* Calculate new sign of result
*
* We know the result lies in the range -N <= r < N
* and so the tuple (old sign, msb, carry) cannot ever
* take the values (0, 1, 0) or (1, 0, 0). We can
* therefore treat these as don't-care inputs, which
* allows us to simplify the boolean expression by
* ignoring the old sign completely.
*/
assert ( ( sign == msb ) || carry );
sign = ( msb ^ carry );
}
/* Add N to make result positive if necessary */
if ( sign )
bigint_add ( modulus, result );
/* Sanity check */
assert ( ! bigint_is_geq ( result, modulus ) );
}
/**
@ -805,12 +764,9 @@ void bigint_mod_exp_raw ( const bigint_element_t *base0,
( ( void * ) result0 );
const unsigned int width = ( 8 * sizeof ( bigint_element_t ) );
struct {
union {
bigint_t ( 2 * size ) padded_modulus;
struct {
bigint_t ( size ) modulus;
bigint_t ( size ) stash;
};
struct {
bigint_t ( size ) modulus;
bigint_t ( size ) stash;
};
union {
bigint_t ( 2 * size ) full;
@ -833,7 +789,7 @@ void bigint_mod_exp_raw ( const bigint_element_t *base0,
}
/* Factor modulus as (N * 2^scale) where N is odd */
bigint_grow ( modulus, &temp->padded_modulus );
bigint_copy ( modulus, &temp->modulus );
for ( scale = 0 ; ( ! bigint_bit_is_set ( &temp->modulus, 0 ) ) ;
scale++ ) {
bigint_shr ( &temp->modulus );
@ -844,10 +800,10 @@ void bigint_mod_exp_raw ( const bigint_element_t *base0,
submask = ~submask;
/* Calculate (R^2 mod N) */
bigint_reduce_supremum ( &temp->padded_modulus, &temp->product.full );
bigint_copy ( &temp->product.low, &temp->stash );
bigint_reduce ( &temp->modulus, &temp->stash );
/* Initialise result = Montgomery(1, R^2 mod N) */
bigint_grow ( &temp->stash, &temp->product.full );
bigint_montgomery ( &temp->modulus, &temp->product.full, result );
/* Convert base into Montgomery form */

View File

@ -188,10 +188,6 @@ static void weierstrass_init ( struct weierstrass_curve *curve ) {
( ( void * ) curve->mont[0] );
bigint_t ( size ) __attribute__ (( may_alias )) *temp =
( ( void * ) curve->prime[1] );
bigint_t ( size * 2 ) __attribute__ (( may_alias )) *prime_double =
( ( void * ) prime );
bigint_t ( size * 2 ) __attribute__ (( may_alias )) *square_double =
( ( void * ) square );
bigint_t ( size * 2 ) __attribute__ (( may_alias )) *product =
( ( void * ) temp );
bigint_t ( size ) __attribute__ (( may_alias )) *two =
@ -206,15 +202,8 @@ static void weierstrass_init ( struct weierstrass_curve *curve ) {
DBGC ( curve, "WEIERSTRASS %s N = %s\n",
curve->name, bigint_ntoa ( prime ) );
/* Calculate Montgomery constant R^2 mod N
*
* We rely on the fact that the subsequent big integers in the
* cache (i.e. the first prime multiple, and the constant "1")
* have not yet been written to, and so can be treated as
* being the (zero) upper halves required to hold the
* double-width value R^2.
*/
bigint_reduce_supremum ( prime_double, square_double );
/* Calculate Montgomery constant R^2 mod N */
bigint_reduce ( prime, square );
DBGC ( curve, "WEIERSTRASS %s R^2 = %s mod N\n",
curve->name, bigint_ntoa ( square ) );

View File

@ -146,6 +146,28 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
bigint_is_geq_raw ( (value)->element, (reference)->element, \
size ); } )
/**
* Set bit in big integer
*
* @v value Big integer
* @v bit Bit to set
*/
#define bigint_set_bit( value, bit ) do { \
unsigned int size = bigint_size (value); \
bigint_set_bit_raw ( (value)->element, size, bit ); \
} while ( 0 )
/**
* Clear bit in big integer
*
* @v value Big integer
* @v bit Bit to set
*/
#define bigint_clear_bit( value, bit ) do { \
unsigned int size = bigint_size (value); \
bigint_clear_bit_raw ( (value)->element, size, bit ); \
} while ( 0 )
/**
* Test if bit is set in big integer
*
@ -243,29 +265,17 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
} while ( 0 )
/**
* Reduce big integer
* Reduce big integer R^2 modulo N
*
* @v modulus Big integer modulus
* @v value Big integer to be reduced
* @v result Big integer to hold result
*/
#define bigint_reduce( modulus, value ) do { \
#define bigint_reduce( modulus, result ) do { \
unsigned int size = bigint_size (modulus); \
bigint_reduce_raw ( (modulus)->element, (value)->element, \
bigint_reduce_raw ( (modulus)->element, (result)->element, \
size ); \
} while ( 0 )
/**
* Reduce supremum of big integer representation
*
* @v modulus0 Big integer modulus
* @v result0 Big integer to hold result
*/
#define bigint_reduce_supremum( modulus, result ) do { \
unsigned int size = bigint_size (modulus); \
bigint_reduce_supremum_raw ( (modulus)->element, \
(result)->element, size ); \
} while ( 0 )
/**
* Compute inverse of odd big integer modulo any power of two
*
@ -369,6 +379,42 @@ typedef void ( bigint_ladder_op_t ) ( const bigint_element_t *operand0,
unsigned int size, const void *ctx,
void *tmp );
/**
* Set bit in big integer
*
* @v value0 Element 0 of big integer
* @v size Number of elements
* @v bit Bit to set
*/
static inline __attribute__ (( always_inline )) void
bigint_set_bit_raw ( bigint_element_t *value0, unsigned int size,
unsigned int bit ) {
bigint_t ( size ) __attribute__ (( may_alias )) *value =
( ( void * ) value0 );
unsigned int index = ( bit / ( 8 * sizeof ( value->element[0] ) ) );
unsigned int subindex = ( bit % ( 8 * sizeof ( value->element[0] ) ) );
value->element[index] |= ( 1UL << subindex );
}
/**
* Clear bit in big integer
*
* @v value0 Element 0 of big integer
* @v size Number of elements
* @v bit Bit to clear
*/
static inline __attribute__ (( always_inline )) void
bigint_clear_bit_raw ( bigint_element_t *value0, unsigned int size,
unsigned int bit ) {
bigint_t ( size ) __attribute__ (( may_alias )) *value =
( ( void * ) value0 );
unsigned int index = ( bit / ( 8 * sizeof ( value->element[0] ) ) );
unsigned int subindex = ( bit % ( 8 * sizeof ( value->element[0] ) ) );
value->element[index] &= ~( 1UL << subindex );
}
/**
* Test if bit is set in big integer
*
@ -442,11 +488,8 @@ void bigint_multiply_raw ( const bigint_element_t *multiplicand0,
const bigint_element_t *multiplier0,
unsigned int multiplier_size,
bigint_element_t *result0 );
void bigint_reduce_raw ( bigint_element_t *modulus0, bigint_element_t *value0,
unsigned int size );
void bigint_reduce_supremum_raw ( bigint_element_t *modulus0,
bigint_element_t *value0,
unsigned int size );
void bigint_reduce_raw ( const bigint_element_t *modulus0,
bigint_element_t *result0, unsigned int size );
void bigint_mod_invert_raw ( const bigint_element_t *invertend0,
bigint_element_t *inverse0, unsigned int size );
int bigint_montgomery_relaxed_raw ( const bigint_element_t *modulus0,

View File

@ -185,14 +185,14 @@ void bigint_multiply_sample ( const bigint_element_t *multiplicand0,
bigint_multiply ( multiplicand, multiplier, result );
}
void bigint_reduce_sample ( bigint_element_t *modulus0,
bigint_element_t *value0, unsigned int size ) {
void bigint_reduce_sample ( const bigint_element_t *modulus0,
bigint_element_t *result0, unsigned int size ) {
const bigint_t ( size ) __attribute__ (( may_alias ))
*modulus = ( ( const void * ) modulus0 );
bigint_t ( size ) __attribute__ (( may_alias ))
*modulus = ( ( void * ) modulus0 );
bigint_t ( size ) __attribute__ (( may_alias ))
*value = ( ( void * ) value0 );
*result = ( ( void * ) result0 );
bigint_reduce ( modulus, value );
bigint_reduce ( modulus, result );
}
void bigint_mod_invert_sample ( const bigint_element_t *invertend0,
@ -553,42 +553,35 @@ void bigint_mod_exp_sample ( const bigint_element_t *base0,
} while ( 0 )
/**
* Report result of big integer modular direct reduction test
* Report result of big integer modular direct reduction of R^2 test
*
* @v modulus Big integer modulus
* @v value Big integer to be reduced
* @v expected Big integer expected result
*/
#define bigint_reduce_ok( modulus, value, expected ) do { \
#define bigint_reduce_ok( modulus, expected ) do { \
static const uint8_t modulus_raw[] = modulus; \
static const uint8_t value_raw[] = value; \
static const uint8_t expected_raw[] = expected; \
uint8_t result_raw[ sizeof ( expected_raw ) ]; \
unsigned int size = \
bigint_required_size ( sizeof ( modulus_raw ) ); \
bigint_t ( size ) modulus_temp; \
bigint_t ( size ) value_temp; \
bigint_t ( size ) result_temp; \
{} /* Fix emacs alignment */ \
\
assert ( bigint_size ( &modulus_temp ) == \
bigint_size ( &value_temp ) ); \
bigint_size ( &result_temp ) ); \
assert ( sizeof ( result_temp ) == sizeof ( result_raw ) ); \
bigint_init ( &modulus_temp, modulus_raw, \
sizeof ( modulus_raw ) ); \
bigint_init ( &value_temp, value_raw, sizeof ( value_raw ) ); \
DBG ( "Modular reduce:\n" ); \
DBG ( "Modular reduce R^2:\n" ); \
DBG_HDA ( 0, &modulus_temp, sizeof ( modulus_temp ) ); \
DBG_HDA ( 0, &value_temp, sizeof ( value_temp ) ); \
bigint_reduce ( &modulus_temp, &value_temp ); \
DBG_HDA ( 0, &value_temp, sizeof ( value_temp ) ); \
bigint_done ( &value_temp, result_raw, sizeof ( result_raw ) ); \
bigint_reduce ( &modulus_temp, &result_temp ); \
DBG_HDA ( 0, &result_temp, sizeof ( result_temp ) ); \
bigint_done ( &result_temp, result_raw, \
sizeof ( result_raw ) ); \
\
ok ( memcmp ( result_raw, expected_raw, \
sizeof ( result_raw ) ) == 0 ); \
\
bigint_init ( &value_temp, modulus_raw, \
sizeof ( modulus_raw ) ); \
ok ( memcmp ( &modulus_temp, &value_temp, \
sizeof ( modulus_temp ) ) == 0 ); \
} while ( 0 )
/**
@ -1801,39 +1794,46 @@ static void bigint_test_exec ( void ) {
0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x01 ) );
bigint_reduce_ok ( BIGINT ( 0xaf ),
BIGINT ( 0x00 ),
BIGINT ( 0x00 ) );
bigint_reduce_ok ( BIGINT ( 0xab ),
BIGINT ( 0xab ),
BIGINT ( 0x00 ) );
bigint_reduce_ok ( BIGINT ( 0xcc, 0x9d, 0xa0, 0x79, 0x96, 0x6a, 0x46,
0xd5, 0xb4, 0x30, 0xd2, 0x2b, 0xbf ),
BIGINT ( 0x1d, 0x97, 0x63, 0xc9, 0x97, 0xcd, 0x43,
0xcb, 0x8e, 0x71, 0xac, 0x41, 0xdd ),
BIGINT ( 0x1d, 0x97, 0x63, 0xc9, 0x97, 0xcd, 0x43,
0xcb, 0x8e, 0x71, 0xac, 0x41, 0xdd ) );
bigint_reduce_ok ( BIGINT ( 0x21, 0xfa, 0x4f, 0xce, 0x0f, 0x0f, 0x4d,
0x43, 0xaa, 0xad, 0x21, 0x30, 0xe5 ),
BIGINT ( 0x21, 0xfa, 0x4f, 0xce, 0x0f, 0x0f, 0x4d,
0x43, 0xaa, 0xad, 0x21, 0x30, 0xe5 ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0x00, 0x00, 0x00 ) );
bigint_reduce_ok ( BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0xf3, 0x65, 0x35, 0x41,
0x66, 0x65 ),
BIGINT ( 0xf9, 0x78, 0x96, 0x39, 0xee, 0x98, 0x42,
0x6a, 0xb8, 0x74, 0x0b, 0xe8, 0x5c, 0x76,
0x34, 0xaf ),
0x00 ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0xb3, 0x07, 0xe8, 0xb7,
0x01, 0xf6 ) );
bigint_reduce_ok ( BIGINT ( 0x47, 0xaa, 0x88, 0x00, 0xd0, 0x30, 0x62,
0xfb, 0x5d, 0x55 ),
BIGINT ( 0xfe, 0x30, 0xe1, 0xc6, 0x65, 0x97, 0x48,
0x2e, 0x94, 0xd4 ),
BIGINT ( 0x27, 0x31, 0x49, 0xc3, 0xf5, 0x06, 0x1f,
0x3c, 0x7c, 0xd5 ) );
0x00 ) );
bigint_reduce_ok ( BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x01 ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00 ) );
bigint_reduce_ok ( BIGINT ( 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00,
0x00 ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00 ) );
bigint_reduce_ok ( BIGINT ( 0x80, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00 ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00 ) );
bigint_reduce_ok ( BIGINT ( 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff ),
BIGINT ( 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x01 ) );
bigint_reduce_ok ( BIGINT ( 0x39, 0x18, 0x47, 0xc9, 0xa2, 0x1d, 0x4b,
0xa6 ),
BIGINT ( 0x30, 0x9d, 0xcc, 0xac, 0xd6, 0xf9, 0x2f,
0xa0 ) );
bigint_reduce_ok ( BIGINT ( 0x81, 0x96, 0xdb, 0x36, 0xa6, 0xb7, 0x41,
0x45, 0x92, 0x37, 0x7d, 0x48, 0x1b, 0x2f,
0x3c, 0xa6 ),
BIGINT ( 0x4a, 0x68, 0x25, 0xf7, 0x2b, 0x72, 0x91,
0x6e, 0x09, 0x83, 0xca, 0xf1, 0x45, 0x79,
0x84, 0x18 ) );
bigint_reduce_ok ( BIGINT ( 0x84, 0x2d, 0xe4, 0x1c, 0xc3, 0x11, 0x4f,
0xa0, 0x90, 0x4b, 0xa9, 0xa1, 0xdf, 0xed,
0x4b, 0xe0, 0xb7, 0xfc, 0x5e, 0xd1, 0x91,
0x59, 0x4d, 0xc2, 0xae, 0x2f, 0x46, 0x9e,
0x32, 0x6e, 0xf4, 0x67 ),
BIGINT ( 0x46, 0xdd, 0x36, 0x6c, 0x0b, 0xac, 0x3a,
0x8f, 0x9a, 0x25, 0x90, 0xb2, 0x39, 0xe9,
0xa4, 0x65, 0xc1, 0xd4, 0xc1, 0x99, 0x61,
0x95, 0x47, 0xab, 0x4f, 0xd7, 0xad, 0xd4,
0x3e, 0xe9, 0x9c, 0xfc ) );
bigint_mod_invert_ok ( BIGINT ( 0x01 ), BIGINT ( 0x01 ) );
bigint_mod_invert_ok ( BIGINT ( 0xff, 0xff, 0xff, 0xff, 0xff, 0xff,
0xff, 0xff ),