Split up asm modules into appropriate (topic-specific) modules, eg

hash/sha1_amd64 and cipher/serpent_ia32.

Remaining code in asm/ dir is for BigInt, so rename to bigint/ in prep for
all (or most) of BigInt being modularized.

8 files changed, 989 insertions, 0 deletions

diff --git a/src/hash/md4_ia32/md4.cpp b/src/hash/md4_ia32/md4.cpp
new file mode 100644
index 000000000..e3dc79012
--- /dev/null
+++ b/src/hash/md4_ia32/md4.cpp
@@ -0,0 +1,43 @@
+/*************************************************
+* MD4 Source File                                *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/md4.h>
+#include <botan/loadstor.h>
+
+namespace Botan {
+
+extern "C" void md4_core(u32bit[4], const byte[64], u32bit[16]);
+
+/*************************************************
+* MD4 Compression Function                       *
+*************************************************/
+void MD4::hash(const byte input[])
+   {
+   md4_core(digest, input, M);
+   }
+
+/*************************************************
+* Copy out the digest                            *
+*************************************************/
+void MD4::copy_out(byte output[])
+   {
+   for(u32bit j = 0; j != OUTPUT_LENGTH; ++j)
+      output[j] = get_byte(3 - (j % 4), digest[j/4]);
+   }
+
+/*************************************************
+* Clear memory of sensitive data                 *
+*************************************************/
+void MD4::clear() throw()
+   {
+   MDx_HashFunction::clear();
+   M.clear();
+   digest[0] = 0x67452301;
+   digest[1] = 0xEFCDAB89;
+   digest[2] = 0x98BADCFE;
+   digest[3] = 0x10325476;
+   }
+
+}
diff --git a/src/hash/md4_ia32/md4core.S b/src/hash/md4_ia32/md4core.S
new file mode 100644
index 000000000..662e9924a
--- /dev/null
+++ b/src/hash/md4_ia32/md4core.S
@@ -0,0 +1,135 @@
+/*************************************************
+* MD4 Source File                                *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/asm_macr.h>
+
+START_LISTING(md4core.S)
+
+START_FUNCTION(md4_core)
+   SPILL_REGS()
+
+#define PUSHED 4
+
+   ASSIGN(EBP, ARG(2)) /* input block */
+   ASSIGN(EDI, ARG(3)) /* expanded words */
+
+   ZEROIZE(ESI)
+
+START_LOOP(.LOAD_INPUT)
+   ADD_IMM(ESI, 4)
+
+   ASSIGN(EAX, ARRAY4(EBP, 0))
+   ASSIGN(EBX, ARRAY4(EBP, 1))
+   ASSIGN(ECX, ARRAY4(EBP, 2))
+   ASSIGN(EDX, ARRAY4(EBP, 3))
+
+   ADD_IMM(EBP, 16)
+
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-4), EAX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-3), EBX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-2), ECX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-1), EDX)
+LOOP_UNTIL_EQ(ESI, 16, .LOAD_INPUT)
+
+   ASSIGN(EBP, ARG(1))
+   ASSIGN(EAX, ARRAY4(EBP, 0))
+   ASSIGN(EBX, ARRAY4(EBP, 1))
+   ASSIGN(ECX, ARRAY4(EBP, 2))
+   ASSIGN(EDX, ARRAY4(EBP, 3))
+
+#define MSG EDI
+#define T1 ESI
+#define T2 EBP
+
+#define FF(A, B, C, D, N, S)          \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, C)                    ; \
+   XOR(T2, D)                       ; \
+   AND(T2, B)                       ; \
+   XOR(T2, D)                       ; \
+   ADD(A, T1)                       ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ;
+
+#define GG(A, B, C, D, N, S)          \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, B)                    ; \
+   OR(T2, C)                        ; \
+   AND(T2, D)                       ; \
+   ADD3_IMM(A, T1, 0x5A827999)      ; \
+   ASSIGN(T1, B)                    ; \
+   AND(T1, C)                       ; \
+   OR(T2, T1)                       ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ;
+
+#define HH(A, B, C, D, N, S)          \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, B)                    ; \
+   XOR(T2, C)                       ; \
+   XOR(T2, D)                       ; \
+   ADD3_IMM(A, T1, 0x6ED9EBA1)      ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ;
+
+   FF(EAX,EBX,ECX,EDX, 0, 3);
+   FF(EDX,EAX,EBX,ECX, 1, 7);
+   FF(ECX,EDX,EAX,EBX, 2,11);
+   FF(EBX,ECX,EDX,EAX, 3,19);
+   FF(EAX,EBX,ECX,EDX, 4, 3);
+   FF(EDX,EAX,EBX,ECX, 5, 7);
+   FF(ECX,EDX,EAX,EBX, 6,11);
+   FF(EBX,ECX,EDX,EAX, 7,19);
+   FF(EAX,EBX,ECX,EDX, 8, 3);
+   FF(EDX,EAX,EBX,ECX, 9, 7);
+   FF(ECX,EDX,EAX,EBX,10,11);
+   FF(EBX,ECX,EDX,EAX,11,19);
+   FF(EAX,EBX,ECX,EDX,12, 3);
+   FF(EDX,EAX,EBX,ECX,13, 7);
+   FF(ECX,EDX,EAX,EBX,14,11);
+   FF(EBX,ECX,EDX,EAX,15,19);
+
+   GG(EAX,EBX,ECX,EDX, 0, 3);
+   GG(EDX,EAX,EBX,ECX, 4, 5);
+   GG(ECX,EDX,EAX,EBX, 8, 9);
+   GG(EBX,ECX,EDX,EAX,12,13);
+   GG(EAX,EBX,ECX,EDX, 1, 3);
+   GG(EDX,EAX,EBX,ECX, 5, 5);
+   GG(ECX,EDX,EAX,EBX, 9, 9);
+   GG(EBX,ECX,EDX,EAX,13,13);
+   GG(EAX,EBX,ECX,EDX, 2, 3);
+   GG(EDX,EAX,EBX,ECX, 6, 5);
+   GG(ECX,EDX,EAX,EBX,10, 9);
+   GG(EBX,ECX,EDX,EAX,14,13);
+   GG(EAX,EBX,ECX,EDX, 3, 3);
+   GG(EDX,EAX,EBX,ECX, 7, 5);
+   GG(ECX,EDX,EAX,EBX,11, 9);
+   GG(EBX,ECX,EDX,EAX,15,13);
+
+   HH(EAX,EBX,ECX,EDX, 0, 3);
+   HH(EDX,EAX,EBX,ECX, 8, 9);
+   HH(ECX,EDX,EAX,EBX, 4,11);
+   HH(EBX,ECX,EDX,EAX,12,15);
+   HH(EAX,EBX,ECX,EDX, 2, 3);
+   HH(EDX,EAX,EBX,ECX,10, 9);
+   HH(ECX,EDX,EAX,EBX, 6,11);
+   HH(EBX,ECX,EDX,EAX,14,15);
+   HH(EAX,EBX,ECX,EDX, 1, 3);
+   HH(EDX,EAX,EBX,ECX, 9, 9);
+   HH(ECX,EDX,EAX,EBX, 5,11);
+   HH(EBX,ECX,EDX,EAX,13,15);
+   HH(EAX,EBX,ECX,EDX, 3, 3);
+   HH(EDX,EAX,EBX,ECX,11, 9);
+   HH(ECX,EDX,EAX,EBX, 7,11);
+   HH(EBX,ECX,EDX,EAX,15,15);
+
+   ASSIGN(EBP, ARG(1))
+   ADD(ARRAY4(EBP, 0), EAX)
+   ADD(ARRAY4(EBP, 1), EBX)
+   ADD(ARRAY4(EBP, 2), ECX)
+   ADD(ARRAY4(EBP, 3), EDX)
+
+   RESTORE_REGS()
+END_FUNCTION(md4_core)
diff --git a/src/hash/md5_ia32/md5.cpp b/src/hash/md5_ia32/md5.cpp
new file mode 100644
index 000000000..cfe48e7e9
--- /dev/null
+++ b/src/hash/md5_ia32/md5.cpp
@@ -0,0 +1,43 @@
+/*************************************************
+* MD5 Source File                                *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/md5.h>
+#include <botan/loadstor.h>
+
+namespace Botan {
+
+extern "C" void md5_core(u32bit[4], const byte[64], u32bit[16]);
+
+/*************************************************
+* MD5 Compression Function                       *
+*************************************************/
+void MD5::hash(const byte input[])
+   {
+   md5_core(digest, input, M);
+   }
+
+/*************************************************
+* Copy out the digest                            *
+*************************************************/
+void MD5::copy_out(byte output[])
+   {
+   for(u32bit j = 0; j != OUTPUT_LENGTH; ++j)
+      output[j] = get_byte(3 - (j % 4), digest[j/4]);
+   }
+
+/*************************************************
+* Clear memory of sensitive data                 *
+*************************************************/
+void MD5::clear() throw()
+   {
+   MDx_HashFunction::clear();
+   M.clear();
+   digest[0] = 0x67452301;
+   digest[1] = 0xEFCDAB89;
+   digest[2] = 0x98BADCFE;
+   digest[3] = 0x10325476;
+   }
+
+}
diff --git a/src/hash/md5_ia32/md5core.S b/src/hash/md5_ia32/md5core.S
new file mode 100644
index 000000000..8ebe469f3
--- /dev/null
+++ b/src/hash/md5_ia32/md5core.S
@@ -0,0 +1,164 @@
+/*************************************************
+* MD5 Source File                                *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/asm_macr.h>
+
+START_LISTING(md5core.S)
+
+START_FUNCTION(md5_core)
+   SPILL_REGS()
+
+#define PUSHED 4
+
+   ASSIGN(EBP, ARG(2)) /* input block */
+   ASSIGN(EDI, ARG(3)) /* expanded words */
+
+   ZEROIZE(ESI)
+
+START_LOOP(.LOAD_INPUT)
+   ADD_IMM(ESI, 4)
+
+   ASSIGN(EAX, ARRAY4(EBP, 0))
+   ASSIGN(EBX, ARRAY4(EBP, 1))
+   ASSIGN(ECX, ARRAY4(EBP, 2))
+   ASSIGN(EDX, ARRAY4(EBP, 3))
+
+   ADD_IMM(EBP, 16)
+
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-4), EAX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-3), EBX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-2), ECX)
+   ASSIGN(ARRAY4_INDIRECT(EDI,ESI,-1), EDX)
+LOOP_UNTIL_EQ(ESI, 16, .LOAD_INPUT)
+
+   ASSIGN(EBP, ARG(1))
+   ASSIGN(EAX, ARRAY4(EBP, 0))
+   ASSIGN(EBX, ARRAY4(EBP, 1))
+   ASSIGN(ECX, ARRAY4(EBP, 2))
+   ASSIGN(EDX, ARRAY4(EBP, 3))
+
+#define MSG EDI
+#define T1 ESI
+#define T2 EBP
+
+#define FF(A, B, C, D, N, S, MAGIC)   \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, C)                    ; \
+   XOR(T2, D)                       ; \
+   AND(T2, B)                       ; \
+   XOR(T2, D)                       ; \
+   ADD3_IMM(A, T1, MAGIC)           ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ; \
+   ADD(A, B)                        ;
+
+#define GG(A, B, C, D, N, S, MAGIC)   \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, B)                    ; \
+   XOR(T2, C)                       ; \
+   AND(T2, D)                       ; \
+   XOR(T2, C)                       ; \
+   ADD3_IMM(A, T1, MAGIC)           ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ; \
+   ADD(A, B)                        ;
+
+#define HH(A, B, C, D, N, S, MAGIC)   \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, B)                    ; \
+   XOR(T2, C)                       ; \
+   XOR(T2, D)                       ; \
+   ADD3_IMM(A, T1, MAGIC)           ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ; \
+   ADD(A, B)                        ;
+
+#define II(A, B, C, D, N, S, MAGIC)   \
+   ASSIGN(T1, ARRAY4(MSG, N))       ; \
+   ASSIGN(T2, D)                    ; \
+   NOT(T2)                          ; \
+   OR(T2, B)                        ; \
+   XOR(T2, C)                       ; \
+   ADD3_IMM(A, T1, MAGIC)           ; \
+   ADD(A, T2)                       ; \
+   ROTL_IMM(A, S)                   ; \
+   ADD(A, B)                        ;
+
+   FF(EAX,EBX,ECX,EDX, 0, 7,0xD76AA478);
+   FF(EDX,EAX,EBX,ECX, 1,12,0xE8C7B756);
+   FF(ECX,EDX,EAX,EBX, 2,17,0x242070DB);
+   FF(EBX,ECX,EDX,EAX, 3,22,0xC1BDCEEE);
+   FF(EAX,EBX,ECX,EDX, 4, 7,0xF57C0FAF);
+   FF(EDX,EAX,EBX,ECX, 5,12,0x4787C62A);
+   FF(ECX,EDX,EAX,EBX, 6,17,0xA8304613);
+   FF(EBX,ECX,EDX,EAX, 7,22,0xFD469501);
+   FF(EAX,EBX,ECX,EDX, 8, 7,0x698098D8);
+   FF(EDX,EAX,EBX,ECX, 9,12,0x8B44F7AF);
+   FF(ECX,EDX,EAX,EBX,10,17,0xFFFF5BB1);
+   FF(EBX,ECX,EDX,EAX,11,22,0x895CD7BE);
+   FF(EAX,EBX,ECX,EDX,12, 7,0x6B901122);
+   FF(EDX,EAX,EBX,ECX,13,12,0xFD987193);
+   FF(ECX,EDX,EAX,EBX,14,17,0xA679438E);
+   FF(EBX,ECX,EDX,EAX,15,22,0x49B40821);
+
+   GG(EAX,EBX,ECX,EDX, 1, 5,0xF61E2562);
+   GG(EDX,EAX,EBX,ECX, 6, 9,0xC040B340);
+   GG(ECX,EDX,EAX,EBX,11,14,0x265E5A51);
+   GG(EBX,ECX,EDX,EAX, 0,20,0xE9B6C7AA);
+   GG(EAX,EBX,ECX,EDX, 5, 5,0xD62F105D);
+   GG(EDX,EAX,EBX,ECX,10, 9,0x02441453);
+   GG(ECX,EDX,EAX,EBX,15,14,0xD8A1E681);
+   GG(EBX,ECX,EDX,EAX, 4,20,0xE7D3FBC8);
+   GG(EAX,EBX,ECX,EDX, 9, 5,0x21E1CDE6);
+   GG(EDX,EAX,EBX,ECX,14, 9,0xC33707D6);
+   GG(ECX,EDX,EAX,EBX, 3,14,0xF4D50D87);
+   GG(EBX,ECX,EDX,EAX, 8,20,0x455A14ED);
+   GG(EAX,EBX,ECX,EDX,13, 5,0xA9E3E905);
+   GG(EDX,EAX,EBX,ECX, 2, 9,0xFCEFA3F8);
+   GG(ECX,EDX,EAX,EBX, 7,14,0x676F02D9);
+   GG(EBX,ECX,EDX,EAX,12,20,0x8D2A4C8A);
+
+   HH(EAX,EBX,ECX,EDX, 5, 4,0xFFFA3942);
+   HH(EDX,EAX,EBX,ECX, 8,11,0x8771F681);
+   HH(ECX,EDX,EAX,EBX,11,16,0x6D9D6122);
+   HH(EBX,ECX,EDX,EAX,14,23,0xFDE5380C);
+   HH(EAX,EBX,ECX,EDX, 1, 4,0xA4BEEA44);
+   HH(EDX,EAX,EBX,ECX, 4,11,0x4BDECFA9);
+   HH(ECX,EDX,EAX,EBX, 7,16,0xF6BB4B60);
+   HH(EBX,ECX,EDX,EAX,10,23,0xBEBFBC70);
+   HH(EAX,EBX,ECX,EDX,13, 4,0x289B7EC6);
+   HH(EDX,EAX,EBX,ECX, 0,11,0xEAA127FA);
+   HH(ECX,EDX,EAX,EBX, 3,16,0xD4EF3085);
+   HH(EBX,ECX,EDX,EAX, 6,23,0x04881D05);
+   HH(EAX,EBX,ECX,EDX, 9, 4,0xD9D4D039);
+   HH(EDX,EAX,EBX,ECX,12,11,0xE6DB99E5);
+   HH(ECX,EDX,EAX,EBX,15,16,0x1FA27CF8);
+   HH(EBX,ECX,EDX,EAX, 2,23,0xC4AC5665);
+
+   II(EAX,EBX,ECX,EDX, 0, 6,0xF4292244);
+   II(EDX,EAX,EBX,ECX, 7,10,0x432AFF97);
+   II(ECX,EDX,EAX,EBX,14,15,0xAB9423A7);
+   II(EBX,ECX,EDX,EAX, 5,21,0xFC93A039);
+   II(EAX,EBX,ECX,EDX,12, 6,0x655B59C3);
+   II(EDX,EAX,EBX,ECX, 3,10,0x8F0CCC92);
+   II(ECX,EDX,EAX,EBX,10,15,0xFFEFF47D);
+   II(EBX,ECX,EDX,EAX, 1,21,0x85845DD1);
+   II(EAX,EBX,ECX,EDX, 8, 6,0x6FA87E4F);
+   II(EDX,EAX,EBX,ECX,15,10,0xFE2CE6E0);
+   II(ECX,EDX,EAX,EBX, 6,15,0xA3014314);
+   II(EBX,ECX,EDX,EAX,13,21,0x4E0811A1);
+   II(EAX,EBX,ECX,EDX, 4, 6,0xF7537E82);
+   II(EDX,EAX,EBX,ECX,11,10,0xBD3AF235);
+   II(ECX,EDX,EAX,EBX, 2,15,0x2AD7D2BB);
+   II(EBX,ECX,EDX,EAX, 9,21,0xEB86D391);
+
+   ASSIGN(EBP, ARG(1))
+   ADD(ARRAY4(EBP, 0), EAX)
+   ADD(ARRAY4(EBP, 1), EBX)
+   ADD(ARRAY4(EBP, 2), ECX)
+   ADD(ARRAY4(EBP, 3), EDX)
+
+   RESTORE_REGS()
+END_FUNCTION(md5_core)
diff --git a/src/hash/sha1_amd64/sha160.cpp b/src/hash/sha1_amd64/sha160.cpp
new file mode 100644
index 000000000..cfac02f45
--- /dev/null
+++ b/src/hash/sha1_amd64/sha160.cpp
@@ -0,0 +1,52 @@
+/*************************************************
+* SHA-160 Source File                            *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/sha160.h>
+#include <botan/loadstor.h>
+
+namespace Botan {
+
+extern "C" void botan_sha160_asm_amd64(u32bit[5], const byte[64], u32bit[80]);
+
+/*************************************************
+* SHA-160 Compression Function                   *
+*************************************************/
+void SHA_160::hash(const byte input[])
+   {
+   botan_sha160_asm_amd64(digest, input, W);
+   }
+
+/*************************************************
+* Copy out the digest                            *
+*************************************************/
+void SHA_160::copy_out(byte output[])
+   {
+   for(u32bit j = 0; j != OUTPUT_LENGTH; ++j)
+      output[j] = get_byte(j % 4, digest[j/4]);
+   }
+
+/*************************************************
+* Clear memory of sensitive data                 *
+*************************************************/
+void SHA_160::clear() throw()
+   {
+   MDx_HashFunction::clear();
+   W.clear();
+   digest[0] = 0x67452301;
+   digest[1] = 0xEFCDAB89;
+   digest[2] = 0x98BADCFE;
+   digest[3] = 0x10325476;
+   digest[4] = 0xC3D2E1F0;
+   }
+
+/*************************************************
+* SHA_160 Constructor                            *
+*************************************************/
+SHA_160::SHA_160() : MDx_HashFunction(20, 64, true, true), W(80)
+   {
+   clear();
+   }
+
+}
diff --git a/src/hash/sha1_amd64/sha1_asm.S b/src/hash/sha1_amd64/sha1_asm.S
new file mode 100644
index 000000000..ecf4a18ce
--- /dev/null
+++ b/src/hash/sha1_amd64/sha1_asm.S
@@ -0,0 +1,258 @@
+/*************************************************
+* SHA-160 Source File                            *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/asm_macr.h>
+
+START_LISTING(sha1_asm.S)
+
+START_FUNCTION(botan_sha160_asm_amd64)
+
+#define DIGEST_ARR %rdi
+#define INPUT %rsi
+#define W %rdx
+#define LOOP_CTR %eax
+
+#define A %r8d
+#define B %r9d
+#define C %r10d
+#define D %r11d
+#define E %ecx
+
+   ZEROIZE(LOOP_CTR)
+
+ALIGN;
+.LOOP_LOAD_INPUT:
+   addl $8, %eax
+
+   movq ARRAY8(INPUT, 0), %r8
+   movq ARRAY8(INPUT, 1), %r9
+   movq ARRAY8(INPUT, 2), %r10
+   movq ARRAY8(INPUT, 3), %r11
+
+   bswap %r8
+   bswap %r9
+   bswap %r10
+   bswap %r11
+
+   rolq $32, %r8
+   rolq $32, %r9
+   rolq $32, %r10
+   rolq $32, %r11
+
+   movq %r8, ARRAY8(W, 0)
+   movq %r9, ARRAY8(W, 1)
+   movq %r10, ARRAY8(W, 2)
+   movq %r11, ARRAY8(W, 3)
+
+   addq $32, W
+   addq $32, INPUT
+
+   cmp IMM(16), LOOP_CTR
+   jne .LOOP_LOAD_INPUT
+
+/*
+#define A %r8d
+#define B %r9d
+#define C %r10d
+#define D %r11d
+#define E %ecx
+*/
+
+ALIGN;
+.LOOP_EXPANSION:
+   addl $4, LOOP_CTR
+
+   ZEROIZE(A)
+   ASSIGN(B, ARRAY4(W, -1))
+   ASSIGN(C, ARRAY4(W, -2))
+   ASSIGN(D, ARRAY4(W, -3))
+
+   XOR(A, ARRAY4(W, -5))
+   XOR(B, ARRAY4(W, -6))
+   XOR(C, ARRAY4(W, -7))
+   XOR(D, ARRAY4(W, -8))
+
+   XOR(A, ARRAY4(W, -11))
+   XOR(B, ARRAY4(W, -12))
+   XOR(C, ARRAY4(W, -13))
+   XOR(D, ARRAY4(W, -14))
+
+   XOR(A, ARRAY4(W, -13))
+   XOR(B, ARRAY4(W, -14))
+   XOR(C, ARRAY4(W, -15))
+   XOR(D, ARRAY4(W, -16))
+
+   ROTL_IMM(D, 1)
+   ROTL_IMM(C, 1)
+   ROTL_IMM(B, 1)
+   XOR(A, D)
+   ROTL_IMM(A, 1)
+
+   ASSIGN(ARRAY4(W, 0), D)
+   ASSIGN(ARRAY4(W, 1), C)
+   ASSIGN(ARRAY4(W, 2), B)
+   ASSIGN(ARRAY4(W, 3), A)
+
+   addq $16, W
+   cmp IMM(80), LOOP_CTR
+   jne .LOOP_EXPANSION
+
+   subq $320, W
+
+#define MAGIC1 0x5A827999
+#define MAGIC2 0x6ED9EBA1
+#define MAGIC3 0x8F1BBCDC
+#define MAGIC4 0xCA62C1D6
+
+#define T %esi
+#define T2 %eax
+
+#define F1(A, B, C, D, E, F, N)       \
+   ASSIGN(T2, ARRAY4(W, N))         ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, C)                     ; \
+   XOR(E, D)                        ; \
+   ADD3_IMM(F, T2, MAGIC1)          ; \
+   AND(E, B)                        ; \
+   XOR(E, D)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F2_4(A, B, C, D, E, F, N, MAGIC) \
+   ASSIGN(T2, ARRAY4(W, N))         ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, B)                     ; \
+   XOR(E, C)                        ; \
+   ADD3_IMM(F, T2, MAGIC)           ; \
+   XOR(E, D)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F3(A, B, C, D, E, F, N)       \
+   ASSIGN(T2, ARRAY4(W, N))         ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, B)                     ; \
+   OR(E, C)                         ; \
+   AND(E, D)                        ; \
+   ADD3_IMM(F, T2, MAGIC3)          ; \
+   ASSIGN(T2, B)                    ; \
+   AND(T2, C)                       ; \
+   OR(E, T2)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F2(A, B, C, D, E, F, W) \
+   F2_4(A, B, C, D, E, F, W, MAGIC2)
+
+#define F4(A, B, C, D, E, F, W) \
+   F2_4(A, B, C, D, E, F, W, MAGIC4)
+
+   ASSIGN(T, ARRAY4(DIGEST_ARR, 0))
+   ASSIGN(B, ARRAY4(DIGEST_ARR, 1))
+   ASSIGN(C, ARRAY4(DIGEST_ARR, 2))
+   ASSIGN(D, ARRAY4(DIGEST_ARR, 3))
+   ASSIGN(E, ARRAY4(DIGEST_ARR, 4))
+
+   /* First Round */
+   F1(A, B, C, D, E, T, 0)
+   F1(T, A, B, C, D, E, 1)
+   F1(E, T, A, B, C, D, 2)
+   F1(D, E, T, A, B, C, 3)
+   F1(C, D, E, T, A, B, 4)
+   F1(B, C, D, E, T, A, 5)
+   F1(A, B, C, D, E, T, 6)
+   F1(T, A, B, C, D, E, 7)
+   F1(E, T, A, B, C, D, 8)
+   F1(D, E, T, A, B, C, 9)
+   F1(C, D, E, T, A, B, 10)
+   F1(B, C, D, E, T, A, 11)
+   F1(A, B, C, D, E, T, 12)
+   F1(T, A, B, C, D, E, 13)
+   F1(E, T, A, B, C, D, 14)
+   F1(D, E, T, A, B, C, 15)
+   F1(C, D, E, T, A, B, 16)
+   F1(B, C, D, E, T, A, 17)
+   F1(A, B, C, D, E, T, 18)
+   F1(T, A, B, C, D, E, 19)
+
+   /* Second Round */
+   F2(E, T, A, B, C, D, 20)
+   F2(D, E, T, A, B, C, 21)
+   F2(C, D, E, T, A, B, 22)
+   F2(B, C, D, E, T, A, 23)
+   F2(A, B, C, D, E, T, 24)
+   F2(T, A, B, C, D, E, 25)
+   F2(E, T, A, B, C, D, 26)
+   F2(D, E, T, A, B, C, 27)
+   F2(C, D, E, T, A, B, 28)
+   F2(B, C, D, E, T, A, 29)
+   F2(A, B, C, D, E, T, 30)
+   F2(T, A, B, C, D, E, 31)
+   F2(E, T, A, B, C, D, 32)
+   F2(D, E, T, A, B, C, 33)
+   F2(C, D, E, T, A, B, 34)
+   F2(B, C, D, E, T, A, 35)
+   F2(A, B, C, D, E, T, 36)
+   F2(T, A, B, C, D, E, 37)
+   F2(E, T, A, B, C, D, 38)
+   F2(D, E, T, A, B, C, 39)
+
+   /* Third Round */
+   F3(C, D, E, T, A, B, 40)
+   F3(B, C, D, E, T, A, 41)
+   F3(A, B, C, D, E, T, 42)
+   F3(T, A, B, C, D, E, 43)
+   F3(E, T, A, B, C, D, 44)
+   F3(D, E, T, A, B, C, 45)
+   F3(C, D, E, T, A, B, 46)
+   F3(B, C, D, E, T, A, 47)
+   F3(A, B, C, D, E, T, 48)
+   F3(T, A, B, C, D, E, 49)
+   F3(E, T, A, B, C, D, 50)
+   F3(D, E, T, A, B, C, 51)
+   F3(C, D, E, T, A, B, 52)
+   F3(B, C, D, E, T, A, 53)
+   F3(A, B, C, D, E, T, 54)
+   F3(T, A, B, C, D, E, 55)
+   F3(E, T, A, B, C, D, 56)
+   F3(D, E, T, A, B, C, 57)
+   F3(C, D, E, T, A, B, 58)
+   F3(B, C, D, E, T, A, 59)
+
+   /* Fourth Round */
+   F4(A, B, C, D, E, T, 60)
+   F4(T, A, B, C, D, E, 61)
+   F4(E, T, A, B, C, D, 62)
+   F4(D, E, T, A, B, C, 63)
+   F4(C, D, E, T, A, B, 64)
+   F4(B, C, D, E, T, A, 65)
+   F4(A, B, C, D, E, T, 66)
+   F4(T, A, B, C, D, E, 67)
+   F4(E, T, A, B, C, D, 68)
+   F4(D, E, T, A, B, C, 69)
+   F4(C, D, E, T, A, B, 70)
+   F4(B, C, D, E, T, A, 71)
+   F4(A, B, C, D, E, T, 72)
+   F4(T, A, B, C, D, E, 73)
+   F4(E, T, A, B, C, D, 74)
+   F4(D, E, T, A, B, C, 75)
+   F4(C, D, E, T, A, B, 76)
+   F4(B, C, D, E, T, A, 77)
+   F4(A, B, C, D, E, T, 78)
+   F4(T, A, B, C, D, E, 79)
+
+   ADD(ARRAY4(DIGEST_ARR, 0), D)
+   ADD(ARRAY4(DIGEST_ARR, 1), T)
+   ADD(ARRAY4(DIGEST_ARR, 2), A)
+   ADD(ARRAY4(DIGEST_ARR, 3), B)
+   ADD(ARRAY4(DIGEST_ARR, 4), C)
+
+END_FUNCTION(botan_sha160_asm_amd64)
diff --git a/src/hash/sha1_ia32/sha160.cpp b/src/hash/sha1_ia32/sha160.cpp
new file mode 100644
index 000000000..7725541d5
--- /dev/null
+++ b/src/hash/sha1_ia32/sha160.cpp
@@ -0,0 +1,52 @@
+/*************************************************
+* SHA-160 Source File                            *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/sha160.h>
+#include <botan/loadstor.h>
+
+namespace Botan {
+
+extern "C" void botan_sha160_asm_ia32(u32bit[5], const byte[64], u32bit[81]);
+
+/*************************************************
+* SHA-160 Compression Function                   *
+*************************************************/
+void SHA_160::hash(const byte input[])
+   {
+   botan_sha160_asm_ia32(digest, input, W);
+   }
+
+/*************************************************
+* Copy out the digest                            *
+*************************************************/
+void SHA_160::copy_out(byte output[])
+   {
+   for(u32bit j = 0; j != OUTPUT_LENGTH; ++j)
+      output[j] = get_byte(j % 4, digest[j/4]);
+   }
+
+/*************************************************
+* Clear memory of sensitive data                 *
+*************************************************/
+void SHA_160::clear() throw()
+   {
+   MDx_HashFunction::clear();
+   W.clear();
+   digest[0] = 0x67452301;
+   digest[1] = 0xEFCDAB89;
+   digest[2] = 0x98BADCFE;
+   digest[3] = 0x10325476;
+   digest[4] = 0xC3D2E1F0;
+   }
+
+/*************************************************
+* SHA_160 Constructor                            *
+*************************************************/
+SHA_160::SHA_160() : MDx_HashFunction(20, 64, true, true), W(81)
+   {
+   clear();
+   }
+
+}
diff --git a/src/hash/sha1_ia32/sha1_asm.S b/src/hash/sha1_ia32/sha1_asm.S
new file mode 100644
index 000000000..85bc9dc2c
--- /dev/null
+++ b/src/hash/sha1_ia32/sha1_asm.S
@@ -0,0 +1,242 @@
+/*************************************************
+* SHA-160 Source File                            *
+* (C) 1999-2007 Jack Lloyd                       *
+*************************************************/
+
+#include <botan/asm_macr.h>
+
+START_LISTING(sha1_asm.S)
+
+START_FUNCTION(botan_sha160_asm_ia32)
+   SPILL_REGS()
+
+#define PUSHED 4
+
+   ASSIGN(EDI, ARG(2))
+   ASSIGN(EBP, ARG(3))
+
+   ZEROIZE(ESI)
+
+START_LOOP(.LOAD_INPUT)
+   ADD_IMM(ESI, 4)
+
+   ASSIGN(EAX, ARRAY4(EDI, 0))
+   ASSIGN(EBX, ARRAY4(EDI, 1))
+   ASSIGN(ECX, ARRAY4(EDI, 2))
+   ASSIGN(EDX, ARRAY4(EDI, 3))
+
+   ADD_IMM(EDI, 16)
+
+   BSWAP(EAX)
+   BSWAP(EBX)
+   BSWAP(ECX)
+   BSWAP(EDX)
+
+   ASSIGN(ARRAY4_INDIRECT(EBP,ESI,-4), EAX)
+   ASSIGN(ARRAY4_INDIRECT(EBP,ESI,-3), EBX)
+   ASSIGN(ARRAY4_INDIRECT(EBP,ESI,-2), ECX)
+   ASSIGN(ARRAY4_INDIRECT(EBP,ESI,-1), EDX)
+LOOP_UNTIL_EQ(ESI, 16, .LOAD_INPUT)
+
+   ADD2_IMM(EDI, EBP, 64)
+
+START_LOOP(.EXPANSION)
+   ADD_IMM(ESI, 4)
+
+   ZEROIZE(EAX)
+   ASSIGN(EBX, ARRAY4(EDI, -1))
+   ASSIGN(ECX, ARRAY4(EDI, -2))
+   ASSIGN(EDX, ARRAY4(EDI, -3))
+
+   XOR(EAX, ARRAY4(EDI, -5))
+   XOR(EBX, ARRAY4(EDI, -6))
+   XOR(ECX, ARRAY4(EDI, -7))
+   XOR(EDX, ARRAY4(EDI, -8))
+
+   XOR(EAX, ARRAY4(EDI, -11))
+   XOR(EBX, ARRAY4(EDI, -12))
+   XOR(ECX, ARRAY4(EDI, -13))
+   XOR(EDX, ARRAY4(EDI, -14))
+
+   XOR(EAX, ARRAY4(EDI, -13))
+   XOR(EBX, ARRAY4(EDI, -14))
+   XOR(ECX, ARRAY4(EDI, -15))
+   XOR(EDX, ARRAY4(EDI, -16))
+
+   ROTL_IMM(EDX, 1)
+   ROTL_IMM(ECX, 1)
+   ROTL_IMM(EBX, 1)
+   XOR(EAX, EDX)
+   ROTL_IMM(EAX, 1)
+
+   ASSIGN(ARRAY4(EDI, 0), EDX)
+   ASSIGN(ARRAY4(EDI, 1), ECX)
+   ASSIGN(ARRAY4(EDI, 2), EBX)
+   ASSIGN(ARRAY4(EDI, 3), EAX)
+
+   ADD_IMM(EDI, 16)
+LOOP_UNTIL_EQ(ESI, 80, .EXPANSION)
+
+#define MAGIC1 0x5A827999
+#define MAGIC2 0x6ED9EBA1
+#define MAGIC3 0x8F1BBCDC
+#define MAGIC4 0xCA62C1D6
+
+#define MSG ESP
+#define T2 EBP
+
+#define F1(A, B, C, D, E, F, N)       \
+   ASSIGN(T2, ARRAY4(MSG, N))       ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, C)                     ; \
+   XOR(E, D)                        ; \
+   ADD3_IMM(F, T2, MAGIC1)          ; \
+   AND(E, B)                        ; \
+   XOR(E, D)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F2_4(A, B, C, D, E, F, N, MAGIC) \
+   ASSIGN(T2, ARRAY4(MSG, N))       ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, B)                     ; \
+   XOR(E, C)                        ; \
+   ADD3_IMM(F, T2, MAGIC)           ; \
+   XOR(E, D)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F3(A, B, C, D, E, F, N)       \
+   ASSIGN(T2, ARRAY4(MSG, N))       ; \
+   ASSIGN(A, F)                     ; \
+   ROTL_IMM(F, 5)                   ; \
+   ADD(F, E)                        ; \
+   ASSIGN(E, B)                     ; \
+   OR(E, C)                         ; \
+   AND(E, D)                        ; \
+   ADD3_IMM(F, T2, MAGIC3)          ; \
+   ASSIGN(T2, B)                    ; \
+   AND(T2, C)                       ; \
+   OR(E, T2)                        ; \
+   ROTR_IMM(B, 2)                   ; \
+   ADD(E, F)                        ;
+
+#define F2(A, B, C, D, E, F, MSG) \
+   F2_4(A, B, C, D, E, F, MSG, MAGIC2)
+
+#define F4(A, B, C, D, E, F, MSG) \
+   F2_4(A, B, C, D, E, F, MSG, MAGIC4)
+
+   ASSIGN(EAX, ARG(1))
+   ASSIGN(EDI, ARRAY4(EAX, 0))
+   ASSIGN(EBX, ARRAY4(EAX, 1))
+   ASSIGN(ECX, ARRAY4(EAX, 2))
+   ASSIGN(EDX, ARRAY4(EAX, 3))
+   ASSIGN(ESI, ARRAY4(EAX, 4))
+
+   ASSIGN(ARRAY4(EBP, 80), ESP)
+   ASSIGN(ESP, EBP)
+
+   /* First Round */
+   F1(EAX, EBX, ECX, EDX, ESI, EDI, 0)
+   F1(EDI, EAX, EBX, ECX, EDX, ESI, 1)
+   F1(ESI, EDI, EAX, EBX, ECX, EDX, 2)
+   F1(EDX, ESI, EDI, EAX, EBX, ECX, 3)
+   F1(ECX, EDX, ESI, EDI, EAX, EBX, 4)
+   F1(EBX, ECX, EDX, ESI, EDI, EAX, 5)
+   F1(EAX, EBX, ECX, EDX, ESI, EDI, 6)
+   F1(EDI, EAX, EBX, ECX, EDX, ESI, 7)
+   F1(ESI, EDI, EAX, EBX, ECX, EDX, 8)
+   F1(EDX, ESI, EDI, EAX, EBX, ECX, 9)
+   F1(ECX, EDX, ESI, EDI, EAX, EBX, 10)
+   F1(EBX, ECX, EDX, ESI, EDI, EAX, 11)
+   F1(EAX, EBX, ECX, EDX, ESI, EDI, 12)
+   F1(EDI, EAX, EBX, ECX, EDX, ESI, 13)
+   F1(ESI, EDI, EAX, EBX, ECX, EDX, 14)
+   F1(EDX, ESI, EDI, EAX, EBX, ECX, 15)
+   F1(ECX, EDX, ESI, EDI, EAX, EBX, 16)
+   F1(EBX, ECX, EDX, ESI, EDI, EAX, 17)
+   F1(EAX, EBX, ECX, EDX, ESI, EDI, 18)
+   F1(EDI, EAX, EBX, ECX, EDX, ESI, 19)
+
+   /* Second Round */
+   F2(ESI, EDI, EAX, EBX, ECX, EDX, 20)
+   F2(EDX, ESI, EDI, EAX, EBX, ECX, 21)
+   F2(ECX, EDX, ESI, EDI, EAX, EBX, 22)
+   F2(EBX, ECX, EDX, ESI, EDI, EAX, 23)
+   F2(EAX, EBX, ECX, EDX, ESI, EDI, 24)
+   F2(EDI, EAX, EBX, ECX, EDX, ESI, 25)
+   F2(ESI, EDI, EAX, EBX, ECX, EDX, 26)
+   F2(EDX, ESI, EDI, EAX, EBX, ECX, 27)
+   F2(ECX, EDX, ESI, EDI, EAX, EBX, 28)
+   F2(EBX, ECX, EDX, ESI, EDI, EAX, 29)
+   F2(EAX, EBX, ECX, EDX, ESI, EDI, 30)
+   F2(EDI, EAX, EBX, ECX, EDX, ESI, 31)
+   F2(ESI, EDI, EAX, EBX, ECX, EDX, 32)
+   F2(EDX, ESI, EDI, EAX, EBX, ECX, 33)
+   F2(ECX, EDX, ESI, EDI, EAX, EBX, 34)
+   F2(EBX, ECX, EDX, ESI, EDI, EAX, 35)
+   F2(EAX, EBX, ECX, EDX, ESI, EDI, 36)
+   F2(EDI, EAX, EBX, ECX, EDX, ESI, 37)
+   F2(ESI, EDI, EAX, EBX, ECX, EDX, 38)
+   F2(EDX, ESI, EDI, EAX, EBX, ECX, 39)
+
+   /* Third Round */
+   F3(ECX, EDX, ESI, EDI, EAX, EBX, 40)
+   F3(EBX, ECX, EDX, ESI, EDI, EAX, 41)
+   F3(EAX, EBX, ECX, EDX, ESI, EDI, 42)
+   F3(EDI, EAX, EBX, ECX, EDX, ESI, 43)
+   F3(ESI, EDI, EAX, EBX, ECX, EDX, 44)
+   F3(EDX, ESI, EDI, EAX, EBX, ECX, 45)
+   F3(ECX, EDX, ESI, EDI, EAX, EBX, 46)
+   F3(EBX, ECX, EDX, ESI, EDI, EAX, 47)
+   F3(EAX, EBX, ECX, EDX, ESI, EDI, 48)
+   F3(EDI, EAX, EBX, ECX, EDX, ESI, 49)
+   F3(ESI, EDI, EAX, EBX, ECX, EDX, 50)
+   F3(EDX, ESI, EDI, EAX, EBX, ECX, 51)
+   F3(ECX, EDX, ESI, EDI, EAX, EBX, 52)
+   F3(EBX, ECX, EDX, ESI, EDI, EAX, 53)
+   F3(EAX, EBX, ECX, EDX, ESI, EDI, 54)
+   F3(EDI, EAX, EBX, ECX, EDX, ESI, 55)
+   F3(ESI, EDI, EAX, EBX, ECX, EDX, 56)
+   F3(EDX, ESI, EDI, EAX, EBX, ECX, 57)
+   F3(ECX, EDX, ESI, EDI, EAX, EBX, 58)
+   F3(EBX, ECX, EDX, ESI, EDI, EAX, 59)
+
+   /* Fourth Round */
+   F4(EAX, EBX, ECX, EDX, ESI, EDI, 60)
+   F4(EDI, EAX, EBX, ECX, EDX, ESI, 61)
+   F4(ESI, EDI, EAX, EBX, ECX, EDX, 62)
+   F4(EDX, ESI, EDI, EAX, EBX, ECX, 63)
+   F4(ECX, EDX, ESI, EDI, EAX, EBX, 64)
+   F4(EBX, ECX, EDX, ESI, EDI, EAX, 65)
+   F4(EAX, EBX, ECX, EDX, ESI, EDI, 66)
+   F4(EDI, EAX, EBX, ECX, EDX, ESI, 67)
+   F4(ESI, EDI, EAX, EBX, ECX, EDX, 68)
+   F4(EDX, ESI, EDI, EAX, EBX, ECX, 69)
+   F4(ECX, EDX, ESI, EDI, EAX, EBX, 70)
+   F4(EBX, ECX, EDX, ESI, EDI, EAX, 71)
+   F4(EAX, EBX, ECX, EDX, ESI, EDI, 72)
+   F4(EDI, EAX, EBX, ECX, EDX, ESI, 73)
+   F4(ESI, EDI, EAX, EBX, ECX, EDX, 74)
+   F4(EDX, ESI, EDI, EAX, EBX, ECX, 75)
+   F4(ECX, EDX, ESI, EDI, EAX, EBX, 76)
+   F4(EBX, ECX, EDX, ESI, EDI, EAX, 77)
+   F4(EAX, EBX, ECX, EDX, ESI, EDI, 78)
+   F4(EDI, EAX, EBX, ECX, EDX, ESI, 79)
+
+   ASSIGN(ESP, ARRAY4(ESP, 80))
+
+   ASSIGN(EBP, ARG(1))
+   ADD(ARRAY4(EBP, 0), EDX)
+   ADD(ARRAY4(EBP, 1), EDI)
+   ADD(ARRAY4(EBP, 2), EAX)
+   ADD(ARRAY4(EBP, 3), EBX)
+   ADD(ARRAY4(EBP, 4), ECX)
+
+   RESTORE_REGS()
+END_FUNCTION(botan_sha160_asm_ia32)