aboutsummaryrefslogtreecommitdiffstats
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to 'src')
-rw-r--r--src/lib/block/cast/cast128.cpp180
1 files changed, 135 insertions, 45 deletions
diff --git a/src/lib/block/cast/cast128.cpp b/src/lib/block/cast/cast128.cpp
index f7910f034..584cd988a 100644
--- a/src/lib/block/cast/cast128.cpp
+++ b/src/lib/block/cast/cast128.cpp
@@ -16,7 +16,7 @@ namespace {
/*
* CAST-128 Round Type 1
*/
-inline uint32_t R1(uint32_t R, uint32_t MK, uint8_t RK)
+inline uint32_t F1(uint32_t R, uint32_t MK, uint8_t RK)
{
const uint32_t T = rotl_var(MK + R, RK);
return (CAST_SBOX1[get_byte(0, T)] ^ CAST_SBOX2[get_byte(1, T)]) -
@@ -26,7 +26,7 @@ inline uint32_t R1(uint32_t R, uint32_t MK, uint8_t RK)
/*
* CAST-128 Round Type 2
*/
-inline uint32_t R2(uint32_t R, uint32_t MK, uint8_t RK)
+inline uint32_t F2(uint32_t R, uint32_t MK, uint8_t RK)
{
const uint32_t T = rotl_var(MK ^ R, RK);
return (CAST_SBOX1[get_byte(0, T)] - CAST_SBOX2[get_byte(1, T)] +
@@ -36,7 +36,7 @@ inline uint32_t R2(uint32_t R, uint32_t MK, uint8_t RK)
/*
* CAST-128 Round Type 3
*/
-inline uint32_t R3(uint32_t R, uint32_t MK, uint8_t RK)
+inline uint32_t F3(uint32_t R, uint32_t MK, uint8_t RK)
{
const uint32_t T = rotl_var(MK - R, RK);
return ((CAST_SBOX1[get_byte(0, T)] + CAST_SBOX2[get_byte(1, T)]) ^
@@ -52,29 +52,74 @@ void CAST_128::encrypt_n(const uint8_t in[], uint8_t out[], size_t blocks) const
{
verify_key_set(m_RK.empty() == false);
- BOTAN_PARALLEL_FOR(size_t i = 0; i < blocks; ++i)
+ while(blocks >= 2)
+ {
+ uint32_t L0, R0, L1, R1;
+ load_be(in, L0, R0, L1, R1);
+
+ L0 ^= F1(R0, m_MK[ 0], m_RK[ 0]);
+ L1 ^= F1(R1, m_MK[ 0], m_RK[ 0]);
+ R0 ^= F2(L0, m_MK[ 1], m_RK[ 1]);
+ R1 ^= F2(L1, m_MK[ 1], m_RK[ 1]);
+ L0 ^= F3(R0, m_MK[ 2], m_RK[ 2]);
+ L1 ^= F3(R1, m_MK[ 2], m_RK[ 2]);
+ R0 ^= F1(L0, m_MK[ 3], m_RK[ 3]);
+ R1 ^= F1(L1, m_MK[ 3], m_RK[ 3]);
+ L0 ^= F2(R0, m_MK[ 4], m_RK[ 4]);
+ L1 ^= F2(R1, m_MK[ 4], m_RK[ 4]);
+ R0 ^= F3(L0, m_MK[ 5], m_RK[ 5]);
+ R1 ^= F3(L1, m_MK[ 5], m_RK[ 5]);
+ L0 ^= F1(R0, m_MK[ 6], m_RK[ 6]);
+ L1 ^= F1(R1, m_MK[ 6], m_RK[ 6]);
+ R0 ^= F2(L0, m_MK[ 7], m_RK[ 7]);
+ R1 ^= F2(L1, m_MK[ 7], m_RK[ 7]);
+ L0 ^= F3(R0, m_MK[ 8], m_RK[ 8]);
+ L1 ^= F3(R1, m_MK[ 8], m_RK[ 8]);
+ R0 ^= F1(L0, m_MK[ 9], m_RK[ 9]);
+ R1 ^= F1(L1, m_MK[ 9], m_RK[ 9]);
+ L0 ^= F2(R0, m_MK[10], m_RK[10]);
+ L1 ^= F2(R1, m_MK[10], m_RK[10]);
+ R0 ^= F3(L0, m_MK[11], m_RK[11]);
+ R1 ^= F3(L1, m_MK[11], m_RK[11]);
+ L0 ^= F1(R0, m_MK[12], m_RK[12]);
+ L1 ^= F1(R1, m_MK[12], m_RK[12]);
+ R0 ^= F2(L0, m_MK[13], m_RK[13]);
+ R1 ^= F2(L1, m_MK[13], m_RK[13]);
+ L0 ^= F3(R0, m_MK[14], m_RK[14]);
+ L1 ^= F3(R1, m_MK[14], m_RK[14]);
+ R0 ^= F1(L0, m_MK[15], m_RK[15]);
+ R1 ^= F1(L1, m_MK[15], m_RK[15]);
+
+ store_be(out, R0, L0, R1, L1);
+
+ blocks -= 2;
+ out += 2 * BLOCK_SIZE;
+ in += 2 * BLOCK_SIZE;
+ }
+
+ if(blocks)
{
uint32_t L, R;
- load_be(in + BLOCK_SIZE*i, L, R);
-
- L ^= R1(R, m_MK[ 0], m_RK[ 0]);
- R ^= R2(L, m_MK[ 1], m_RK[ 1]);
- L ^= R3(R, m_MK[ 2], m_RK[ 2]);
- R ^= R1(L, m_MK[ 3], m_RK[ 3]);
- L ^= R2(R, m_MK[ 4], m_RK[ 4]);
- R ^= R3(L, m_MK[ 5], m_RK[ 5]);
- L ^= R1(R, m_MK[ 6], m_RK[ 6]);
- R ^= R2(L, m_MK[ 7], m_RK[ 7]);
- L ^= R3(R, m_MK[ 8], m_RK[ 8]);
- R ^= R1(L, m_MK[ 9], m_RK[ 9]);
- L ^= R2(R, m_MK[10], m_RK[10]);
- R ^= R3(L, m_MK[11], m_RK[11]);
- L ^= R1(R, m_MK[12], m_RK[12]);
- R ^= R2(L, m_MK[13], m_RK[13]);
- L ^= R3(R, m_MK[14], m_RK[14]);
- R ^= R1(L, m_MK[15], m_RK[15]);
-
- store_be(out + BLOCK_SIZE*i, R, L);
+ load_be(in, L, R);
+
+ L ^= F1(R, m_MK[ 0], m_RK[ 0]);
+ R ^= F2(L, m_MK[ 1], m_RK[ 1]);
+ L ^= F3(R, m_MK[ 2], m_RK[ 2]);
+ R ^= F1(L, m_MK[ 3], m_RK[ 3]);
+ L ^= F2(R, m_MK[ 4], m_RK[ 4]);
+ R ^= F3(L, m_MK[ 5], m_RK[ 5]);
+ L ^= F1(R, m_MK[ 6], m_RK[ 6]);
+ R ^= F2(L, m_MK[ 7], m_RK[ 7]);
+ L ^= F3(R, m_MK[ 8], m_RK[ 8]);
+ R ^= F1(L, m_MK[ 9], m_RK[ 9]);
+ L ^= F2(R, m_MK[10], m_RK[10]);
+ R ^= F3(L, m_MK[11], m_RK[11]);
+ L ^= F1(R, m_MK[12], m_RK[12]);
+ R ^= F2(L, m_MK[13], m_RK[13]);
+ L ^= F3(R, m_MK[14], m_RK[14]);
+ R ^= F1(L, m_MK[15], m_RK[15]);
+
+ store_be(out, R, L);
}
}
@@ -85,29 +130,74 @@ void CAST_128::decrypt_n(const uint8_t in[], uint8_t out[], size_t blocks) const
{
verify_key_set(m_RK.empty() == false);
- BOTAN_PARALLEL_FOR(size_t i = 0; i < blocks; ++i)
+ while(blocks >= 2)
+ {
+ uint32_t L0, R0, L1, R1;
+ load_be(in, L0, R0, L1, R1);
+
+ R0 ^= F1(L0, m_MK[15], m_RK[15]);
+ R1 ^= F1(L1, m_MK[15], m_RK[15]);
+ L0 ^= F3(R0, m_MK[14], m_RK[14]);
+ L1 ^= F3(R1, m_MK[14], m_RK[14]);
+ R0 ^= F2(L0, m_MK[13], m_RK[13]);
+ R1 ^= F2(L1, m_MK[13], m_RK[13]);
+ L0 ^= F1(R0, m_MK[12], m_RK[12]);
+ L1 ^= F1(R1, m_MK[12], m_RK[12]);
+ R0 ^= F3(L0, m_MK[11], m_RK[11]);
+ R1 ^= F3(L1, m_MK[11], m_RK[11]);
+ L0 ^= F2(R0, m_MK[10], m_RK[10]);
+ L1 ^= F2(R1, m_MK[10], m_RK[10]);
+ R0 ^= F1(L0, m_MK[ 9], m_RK[ 9]);
+ R1 ^= F1(L1, m_MK[ 9], m_RK[ 9]);
+ L0 ^= F3(R0, m_MK[ 8], m_RK[ 8]);
+ L1 ^= F3(R1, m_MK[ 8], m_RK[ 8]);
+ R0 ^= F2(L0, m_MK[ 7], m_RK[ 7]);
+ R1 ^= F2(L1, m_MK[ 7], m_RK[ 7]);
+ L0 ^= F1(R0, m_MK[ 6], m_RK[ 6]);
+ L1 ^= F1(R1, m_MK[ 6], m_RK[ 6]);
+ R0 ^= F3(L0, m_MK[ 5], m_RK[ 5]);
+ R1 ^= F3(L1, m_MK[ 5], m_RK[ 5]);
+ L0 ^= F2(R0, m_MK[ 4], m_RK[ 4]);
+ L1 ^= F2(R1, m_MK[ 4], m_RK[ 4]);
+ R0 ^= F1(L0, m_MK[ 3], m_RK[ 3]);
+ R1 ^= F1(L1, m_MK[ 3], m_RK[ 3]);
+ L0 ^= F3(R0, m_MK[ 2], m_RK[ 2]);
+ L1 ^= F3(R1, m_MK[ 2], m_RK[ 2]);
+ R0 ^= F2(L0, m_MK[ 1], m_RK[ 1]);
+ R1 ^= F2(L1, m_MK[ 1], m_RK[ 1]);
+ L0 ^= F1(R0, m_MK[ 0], m_RK[ 0]);
+ L1 ^= F1(R1, m_MK[ 0], m_RK[ 0]);
+
+ store_be(out, R0, L0, R1, L1);
+
+ blocks -= 2;
+ out += 2 * BLOCK_SIZE;
+ in += 2 * BLOCK_SIZE;
+ }
+
+ if(blocks)
{
uint32_t L, R;
- load_be(in + BLOCK_SIZE*i, L, R);
-
- L ^= R1(R, m_MK[15], m_RK[15]);
- R ^= R3(L, m_MK[14], m_RK[14]);
- L ^= R2(R, m_MK[13], m_RK[13]);
- R ^= R1(L, m_MK[12], m_RK[12]);
- L ^= R3(R, m_MK[11], m_RK[11]);
- R ^= R2(L, m_MK[10], m_RK[10]);
- L ^= R1(R, m_MK[ 9], m_RK[ 9]);
- R ^= R3(L, m_MK[ 8], m_RK[ 8]);
- L ^= R2(R, m_MK[ 7], m_RK[ 7]);
- R ^= R1(L, m_MK[ 6], m_RK[ 6]);
- L ^= R3(R, m_MK[ 5], m_RK[ 5]);
- R ^= R2(L, m_MK[ 4], m_RK[ 4]);
- L ^= R1(R, m_MK[ 3], m_RK[ 3]);
- R ^= R3(L, m_MK[ 2], m_RK[ 2]);
- L ^= R2(R, m_MK[ 1], m_RK[ 1]);
- R ^= R1(L, m_MK[ 0], m_RK[ 0]);
-
- store_be(out + BLOCK_SIZE*i, R, L);
+ load_be(in, L, R);
+
+ L ^= F1(R, m_MK[15], m_RK[15]);
+ R ^= F3(L, m_MK[14], m_RK[14]);
+ L ^= F2(R, m_MK[13], m_RK[13]);
+ R ^= F1(L, m_MK[12], m_RK[12]);
+ L ^= F3(R, m_MK[11], m_RK[11]);
+ R ^= F2(L, m_MK[10], m_RK[10]);
+ L ^= F1(R, m_MK[ 9], m_RK[ 9]);
+ R ^= F3(L, m_MK[ 8], m_RK[ 8]);
+ L ^= F2(R, m_MK[ 7], m_RK[ 7]);
+ R ^= F1(L, m_MK[ 6], m_RK[ 6]);
+ L ^= F3(R, m_MK[ 5], m_RK[ 5]);
+ R ^= F2(L, m_MK[ 4], m_RK[ 4]);
+ L ^= F1(R, m_MK[ 3], m_RK[ 3]);
+ R ^= F3(L, m_MK[ 2], m_RK[ 2]);
+ L ^= F2(R, m_MK[ 1], m_RK[ 1]);
+ R ^= F1(L, m_MK[ 0], m_RK[ 0]);
+
+ store_be(out, R, L);
}
}
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-29 18:49:04 +0000