diff options
Diffstat (limited to 'plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c')
-rw-r--r-- | plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c | 86 |
1 files changed, 43 insertions, 43 deletions
diff --git a/plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c b/plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c index e1f6f58eb5..d412b693fd 100644 --- a/plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c +++ b/plugins/MirOTR/libgcrypt-1.4.6/mpi/mpih-mul.c @@ -35,7 +35,7 @@ #define MPN_MUL_N_RECURSE(prodp, up, vp, size, tspace) \ do { \ - if( (size) < KARATSUBA_THRESHOLD ) \ + if ( (size) < KARATSUBA_THRESHOLD ) \ mul_n_basecase (prodp, up, vp, size); \ else \ mul_n (prodp, up, vp, size, tspace); \ @@ -80,8 +80,8 @@ mul_n_basecase( mpi_ptr_t prodp, mpi_ptr_t up, /* Multiply by the first limb in V separately, as the result can be * stored (not added) to PROD. We also avoid a loop for zeroing. */ v_limb = vp[0]; - if( v_limb <= 1 ) { - if( v_limb == 1 ) + if ( v_limb <= 1 ) { + if ( v_limb == 1 ) MPN_COPY( prodp, up, size ); else MPN_ZERO( prodp, size ); @@ -95,11 +95,11 @@ mul_n_basecase( mpi_ptr_t prodp, mpi_ptr_t up, /* For each iteration in the outer loop, multiply one limb from * U with one limb from V, and add it to PROD. */ - for( i = 1; i < size; i++ ) { + for ( i = 1; i < size; i++ ) { v_limb = vp[i]; - if( v_limb <= 1 ) { + if ( v_limb <= 1 ) { cy = 0; - if( v_limb == 1 ) + if ( v_limb == 1 ) cy = _gcry_mpih_add_n(prodp, prodp, up, size); } else @@ -117,7 +117,7 @@ static void mul_n( mpi_ptr_t prodp, mpi_ptr_t up, mpi_ptr_t vp, mpi_size_t size, mpi_ptr_t tspace ) { - if( size & 1 ) { + if ( size & 1 ) { /* The size is odd, and the code below doesn't handle that. * Multiply the least significant (size - 1) limbs with a recursive * call, and handle the most significant limb of S1 and S2 @@ -167,7 +167,7 @@ mul_n( mpi_ptr_t prodp, mpi_ptr_t up, mpi_ptr_t vp, /* Product M. ________________ * |_(U1-U0)(V0-V1)_| */ - if( _gcry_mpih_cmp(up + hsize, up, hsize) >= 0 ) { + if ( _gcry_mpih_cmp(up + hsize, up, hsize) >= 0 ) { _gcry_mpih_sub_n(prodp, up + hsize, up, hsize); negflg = 0; } @@ -175,7 +175,7 @@ mul_n( mpi_ptr_t prodp, mpi_ptr_t up, mpi_ptr_t vp, _gcry_mpih_sub_n(prodp, up, up + hsize, hsize); negflg = 1; } - if( _gcry_mpih_cmp(vp + hsize, vp, hsize) >= 0 ) { + if ( _gcry_mpih_cmp(vp + hsize, vp, hsize) >= 0 ) { _gcry_mpih_sub_n(prodp + hsize, vp + hsize, vp, hsize); negflg ^= 1; } @@ -211,12 +211,12 @@ mul_n( mpi_ptr_t prodp, mpi_ptr_t up, mpi_ptr_t vp, /* Add/copy Product L (twice) */ cy += _gcry_mpih_add_n(prodp + hsize, prodp + hsize, tspace, size); - if( cy ) + if ( cy ) _gcry_mpih_add_1(prodp + hsize + size, prodp + hsize + size, hsize, cy); MPN_COPY(prodp, tspace, hsize); cy = _gcry_mpih_add_n(prodp + hsize, prodp + hsize, tspace + hsize, hsize); - if( cy ) + if ( cy ) _gcry_mpih_add_1(prodp + size, prodp + size, size, 1); } } @@ -232,8 +232,8 @@ _gcry_mpih_sqr_n_basecase( mpi_ptr_t prodp, mpi_ptr_t up, mpi_size_t size ) /* Multiply by the first limb in V separately, as the result can be * stored (not added) to PROD. We also avoid a loop for zeroing. */ v_limb = up[0]; - if( v_limb <= 1 ) { - if( v_limb == 1 ) + if ( v_limb <= 1 ) { + if ( v_limb == 1 ) MPN_COPY( prodp, up, size ); else MPN_ZERO(prodp, size); @@ -247,11 +247,11 @@ _gcry_mpih_sqr_n_basecase( mpi_ptr_t prodp, mpi_ptr_t up, mpi_size_t size ) /* For each iteration in the outer loop, multiply one limb from * U with one limb from V, and add it to PROD. */ - for( i=1; i < size; i++) { + for ( i=1; i < size; i++) { v_limb = up[i]; - if( v_limb <= 1 ) { + if ( v_limb <= 1 ) { cy_limb = 0; - if( v_limb == 1 ) + if ( v_limb == 1 ) cy_limb = _gcry_mpih_add_n(prodp, prodp, up, size); } else @@ -267,7 +267,7 @@ void _gcry_mpih_sqr_n( mpi_ptr_t prodp, mpi_ptr_t up, mpi_size_t size, mpi_ptr_t tspace) { - if( size & 1 ) { + if ( size & 1 ) { /* The size is odd, and the code below doesn't handle that. * Multiply the least significant (size - 1) limbs with a recursive * call, and handle the most significant limb of S1 and S2 @@ -302,7 +302,7 @@ _gcry_mpih_sqr_n( mpi_ptr_t prodp, /* Product M. ________________ * |_(U1-U0)(U0-U1)_| */ - if( _gcry_mpih_cmp( up + hsize, up, hsize) >= 0 ) + if ( _gcry_mpih_cmp( up + hsize, up, hsize) >= 0 ) _gcry_mpih_sub_n( prodp, up + hsize, up, hsize); else _gcry_mpih_sub_n (prodp, up, up + hsize, hsize); @@ -329,13 +329,13 @@ _gcry_mpih_sqr_n( mpi_ptr_t prodp, /* Add/copy Product L (twice). */ cy += _gcry_mpih_add_n (prodp + hsize, prodp + hsize, tspace, size); - if( cy ) + if ( cy ) _gcry_mpih_add_1(prodp + hsize + size, prodp + hsize + size, hsize, cy); MPN_COPY(prodp, tspace, hsize); cy = _gcry_mpih_add_n (prodp + hsize, prodp + hsize, tspace + hsize, hsize); - if( cy ) + if ( cy ) _gcry_mpih_add_1 (prodp + size, prodp + size, size, 1); } } @@ -348,8 +348,8 @@ _gcry_mpih_mul_n( mpi_ptr_t prodp, { int secure; - if( up == vp ) { - if( size < KARATSUBA_THRESHOLD ) + if ( up == vp ) { + if ( size < KARATSUBA_THRESHOLD ) _gcry_mpih_sqr_n_basecase( prodp, up, size ); else { mpi_ptr_t tspace; @@ -360,7 +360,7 @@ _gcry_mpih_mul_n( mpi_ptr_t prodp, } } else { - if( size < KARATSUBA_THRESHOLD ) + if ( size < KARATSUBA_THRESHOLD ) mul_n_basecase( prodp, up, vp, size ); else { mpi_ptr_t tspace; @@ -382,8 +382,8 @@ _gcry_mpih_mul_karatsuba_case( mpi_ptr_t prodp, { mpi_limb_t cy; - if( !ctx->tspace || ctx->tspace_size < vsize ) { - if( ctx->tspace ) + if ( !ctx->tspace || ctx->tspace_size < vsize ) { + if ( ctx->tspace ) _gcry_mpi_free_limb_space( ctx->tspace, ctx->tspace_nlimbs ); ctx->tspace_nlimbs = 2 * vsize; ctx->tspace = mpi_alloc_limb_space( 2 * vsize, @@ -397,9 +397,9 @@ _gcry_mpih_mul_karatsuba_case( mpi_ptr_t prodp, prodp += vsize; up += vsize; usize -= vsize; - if( usize >= vsize ) { - if( !ctx->tp || ctx->tp_size < vsize ) { - if( ctx->tp ) + if ( usize >= vsize ) { + if ( !ctx->tp || ctx->tp_size < vsize ) { + if ( ctx->tp ) _gcry_mpi_free_limb_space( ctx->tp, ctx->tp_nlimbs ); ctx->tp_nlimbs = 2 * vsize; ctx->tp = mpi_alloc_limb_space( 2 * vsize, gcry_is_secure( up ) @@ -417,12 +417,12 @@ _gcry_mpih_mul_karatsuba_case( mpi_ptr_t prodp, } while( usize >= vsize ); } - if( usize ) { - if( usize < KARATSUBA_THRESHOLD ) { + if ( usize ) { + if ( usize < KARATSUBA_THRESHOLD ) { _gcry_mpih_mul( ctx->tspace, vp, vsize, up, usize ); } else { - if( !ctx->next ) { + if ( !ctx->next ) { ctx->next = gcry_xcalloc( 1, sizeof *ctx ); } _gcry_mpih_mul_karatsuba_case( ctx->tspace, @@ -442,15 +442,15 @@ _gcry_mpih_release_karatsuba_ctx( struct karatsuba_ctx *ctx ) { struct karatsuba_ctx *ctx2; - if( ctx->tp ) + if ( ctx->tp ) _gcry_mpi_free_limb_space( ctx->tp, ctx->tp_nlimbs ); - if( ctx->tspace ) + if ( ctx->tspace ) _gcry_mpi_free_limb_space( ctx->tspace, ctx->tspace_nlimbs ); - for( ctx=ctx->next; ctx; ctx = ctx2 ) { + for ( ctx=ctx->next; ctx; ctx = ctx2 ) { ctx2 = ctx->next; - if( ctx->tp ) + if ( ctx->tp ) _gcry_mpi_free_limb_space( ctx->tp, ctx->tp_nlimbs ); - if( ctx->tspace ) + if ( ctx->tspace ) _gcry_mpi_free_limb_space( ctx->tspace, ctx->tspace_nlimbs ); gcry_free( ctx ); } @@ -479,18 +479,18 @@ _gcry_mpih_mul( mpi_ptr_t prodp, mpi_ptr_t up, mpi_size_t usize, mpi_limb_t cy; struct karatsuba_ctx ctx; - if( vsize < KARATSUBA_THRESHOLD ) { + if ( vsize < KARATSUBA_THRESHOLD ) { mpi_size_t i; mpi_limb_t v_limb; - if( !vsize ) + if ( !vsize ) return 0; /* Multiply by the first limb in V separately, as the result can be * stored (not added) to PROD. We also avoid a loop for zeroing. */ v_limb = vp[0]; - if( v_limb <= 1 ) { - if( v_limb == 1 ) + if ( v_limb <= 1 ) { + if ( v_limb == 1 ) MPN_COPY( prodp, up, usize ); else MPN_ZERO( prodp, usize ); @@ -504,11 +504,11 @@ _gcry_mpih_mul( mpi_ptr_t prodp, mpi_ptr_t up, mpi_size_t usize, /* For each iteration in the outer loop, multiply one limb from * U with one limb from V, and add it to PROD. */ - for( i = 1; i < vsize; i++ ) { + for ( i = 1; i < vsize; i++ ) { v_limb = vp[i]; - if( v_limb <= 1 ) { + if ( v_limb <= 1 ) { cy = 0; - if( v_limb == 1 ) + if ( v_limb == 1 ) cy = _gcry_mpih_add_n(prodp, prodp, up, usize); } else |