aboutsummaryrefslogtreecommitdiffstats
path: root/src/math/mp/mp_karat.cpp
diff options
context:
space:
mode:
authorlloyd <[email protected]>2012-06-19 13:16:58 +0000
committerlloyd <[email protected]>2012-06-19 13:16:58 +0000
commit9d65cb180326ed3009529695da170f3e23c71d0f (patch)
tree33b3c80178101d7701cc8afd1c143d326aa41276 /src/math/mp/mp_karat.cpp
parenta5f0bd2f1a1e80aa4c2f4f1f87c2413f6ac12726 (diff)
In Karatsuba multiplication, we would avoid recursing in cases where
we know one of the sub values was going to be zero. Avoid doing this as it exposes a timing channel. Some bn_asm code was manually inlined into the Karatsuba for doing additions. Just call the normal functions - if these are too slow that should be fixed.
Diffstat (limited to 'src/math/mp/mp_karat.cpp')
-rw-r--r--src/math/mp/mp_karat.cpp58
1 files changed, 10 insertions, 48 deletions
diff --git a/src/math/mp/mp_karat.cpp b/src/math/mp/mp_karat.cpp
index ea0693bf1..9bee7c42d 100644
--- a/src/math/mp/mp_karat.cpp
+++ b/src/math/mp/mp_karat.cpp
@@ -45,7 +45,7 @@ void karatsuba_mul(word z[], const word x[], const word y[], size_t N,
clear_mem(workspace, 2*N);
- if(cmp0 && cmp1)
+ //if(cmp0 && cmp1)
{
if(cmp0 > 0)
bigint_sub3(z0, x0, N2, x1, N2);
@@ -63,30 +63,11 @@ void karatsuba_mul(word z[], const word x[], const word y[], size_t N,
karatsuba_mul(z0, x0, y0, N2, workspace+N);
karatsuba_mul(z1, x1, y1, N2, workspace+N);
- const size_t blocks_of_8 = N - (N % 8);
+ const word ws_carry = bigint_add3_nc(workspace + N, z0, N, z1, N);
+ word z_carry = bigint_add2_nc(z + N2, N, workspace + N, N);
- word ws_carry = 0;
-
- for(size_t j = 0; j != blocks_of_8; j += 8)
- ws_carry = word8_add3(workspace + N + j, z0 + j, z1 + j, ws_carry);
-
- for(size_t j = blocks_of_8; j != N; ++j)
- workspace[N + j] = word_add(z0[j], z1[j], &ws_carry);
-
- word z_carry = 0;
-
- for(size_t j = 0; j != blocks_of_8; j += 8)
- z_carry = word8_add2(z + N2 + j, workspace + N + j, z_carry);
-
- for(size_t j = blocks_of_8; j != N; ++j)
- z[N2 + j] = word_add(z[N2 + j], workspace[N + j], &z_carry);
-
- z[N + N2] = word_add(z[N + N2], ws_carry, &z_carry);
-
- if(z_carry)
- for(size_t j = 1; j != N2; ++j)
- if(++z[N + N2 + j])
- break;
+ z_carry += bigint_add2_nc(z + N + N2, N2, &ws_carry, 1);
+ bigint_add2_nc(z + N + N2, N2, &z_carry, 1);
if((cmp0 == cmp1) || (cmp0 == 0) || (cmp1 == 0))
bigint_add2(z + N2, 2*N-N2, workspace, N);
@@ -122,7 +103,7 @@ void karatsuba_sqr(word z[], const word x[], size_t N, word workspace[])
clear_mem(workspace, 2*N);
- if(cmp)
+ //if(cmp)
{
if(cmp > 0)
bigint_sub3(z0, x0, N2, x1, N2);
@@ -135,30 +116,11 @@ void karatsuba_sqr(word z[], const word x[], size_t N, word workspace[])
karatsuba_sqr(z0, x0, N2, workspace+N);
karatsuba_sqr(z1, x1, N2, workspace+N);
- const size_t blocks_of_8 = N - (N % 8);
-
- word ws_carry = 0;
-
- for(size_t j = 0; j != blocks_of_8; j += 8)
- ws_carry = word8_add3(workspace + N + j, z0 + j, z1 + j, ws_carry);
-
- for(size_t j = blocks_of_8; j != N; ++j)
- workspace[N + j] = word_add(z0[j], z1[j], &ws_carry);
-
- word z_carry = 0;
-
- for(size_t j = 0; j != blocks_of_8; j += 8)
- z_carry = word8_add2(z + N2 + j, workspace + N + j, z_carry);
-
- for(size_t j = blocks_of_8; j != N; ++j)
- z[N2 + j] = word_add(z[N2 + j], workspace[N + j], &z_carry);
-
- z[N + N2] = word_add(z[N + N2], ws_carry, &z_carry);
+ const word ws_carry = bigint_add3_nc(workspace + N, z0, N, z1, N);
+ word z_carry = bigint_add2_nc(z + N2, N, workspace + N, N);
- if(z_carry)
- for(size_t j = 1; j != N2; ++j)
- if(++z[N + N2 + j])
- break;
+ z_carry += bigint_add2_nc(z + N + N2, N2, &ws_carry, 1);
+ bigint_add2_nc(z + N + N2, N2, &z_carry, 1);
/*
* This is only actually required if cmp is != 0, however