1 files changed, 183 insertions, 0 deletions
diff --git a/src/math/mp/mp_asm.cpp b/src/math/mp/mp_asm.cpp
new file mode 100644
index 000000000..4fcdee7a4
--- /dev/null
+++ b/src/math/mp/mp_asm.cpp
@@ -0,0 +1,183 @@
+/*
+* Lowest Level MPI Algorithms
+* (C) 1999-2010 Jack Lloyd
+*     2006 Luca Piccarreta
+*
+* Distributed under the terms of the Botan license
+*/
+
+#include <botan/internal/mp_asm.h>
+#include <botan/internal/mp_asmi.h>
+#include <botan/internal/mp_core.h>
+#include <botan/exceptn.h>
+#include <botan/mem_ops.h>
+
+namespace Botan {
+
+extern "C" {
+
+/*
+* Two Operand Addition, No Carry
+*/
+word bigint_add2_nc(word x[], u32bit x_size, const word y[], u32bit y_size)
+   {
+   word carry = 0;
+
+   const u32bit blocks = y_size - (y_size % 8);
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      carry = word8_add2(x + i, y + i, carry);
+
+   for(u32bit i = blocks; i != y_size; ++i)
+      x[i] = word_add(x[i], y[i], &carry);
+
+   for(u32bit i = y_size; i != x_size; ++i)
+      x[i] = word_add(x[i], 0, &carry);
+
+   return carry;
+   }
+
+/*
+* Three Operand Addition, No Carry
+*/
+word bigint_add3_nc(word z[], const word x[], u32bit x_size,
+                              const word y[], u32bit y_size)
+   {
+   if(x_size < y_size)
+      { return bigint_add3_nc(z, y, y_size, x, x_size); }
+
+   word carry = 0;
+
+   const u32bit blocks = y_size - (y_size % 8);
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      carry = word8_add3(z + i, x + i, y + i, carry);
+
+   for(u32bit i = blocks; i != y_size; ++i)
+      z[i] = word_add(x[i], y[i], &carry);
+
+   for(u32bit i = y_size; i != x_size; ++i)
+      z[i] = word_add(x[i], 0, &carry);
+
+   return carry;
+   }
+
+/*
+* Two Operand Addition
+*/
+void bigint_add2(word x[], u32bit x_size, const word y[], u32bit y_size)
+   {
+   x[x_size] += bigint_add2_nc(x, x_size, y, y_size);
+   }
+
+/*
+* Three Operand Addition
+*/
+void bigint_add3(word z[], const word x[], u32bit x_size,
+                           const word y[], u32bit y_size)
+   {
+   z[(x_size > y_size ? x_size : y_size)] +=
+      bigint_add3_nc(z, x, x_size, y, y_size);
+   }
+
+/*
+* Two Operand Subtraction
+*/
+word bigint_sub2(word x[], u32bit x_size, const word y[], u32bit y_size)
+   {
+   word borrow = 0;
+
+   const u32bit blocks = y_size - (y_size % 8);
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      borrow = word8_sub2(x + i, y + i, borrow);
+
+   for(u32bit i = blocks; i != y_size; ++i)
+      x[i] = word_sub(x[i], y[i], &borrow);
+
+   for(u32bit i = y_size; i != x_size; ++i)
+      x[i] = word_sub(x[i], 0, &borrow);
+
+   return borrow;
+   }
+
+/*
+* Two Operand Subtraction x = y - x
+*/
+void bigint_sub2_rev(word x[],  const word y[], u32bit y_size)
+   {
+   word borrow = 0;
+
+   const u32bit blocks = y_size - (y_size % 8);
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      borrow = word8_sub2_rev(x + i, y + i, borrow);
+
+   for(u32bit i = blocks; i != y_size; ++i)
+      x[i] = word_sub(y[i], x[i], &borrow);
+
+   if(borrow)
+      throw Internal_Error("bigint_sub2_rev: x >= y");
+   }
+
+/*
+* Three Operand Subtraction
+*/
+word bigint_sub3(word z[], const word x[], u32bit x_size,
+                           const word y[], u32bit y_size)
+   {
+   word borrow = 0;
+
+   const u32bit blocks = y_size - (y_size % 8);
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      borrow = word8_sub3(z + i, x + i, y + i, borrow);
+
+   for(u32bit i = blocks; i != y_size; ++i)
+      z[i] = word_sub(x[i], y[i], &borrow);
+
+   for(u32bit i = y_size; i != x_size; ++i)
+      z[i] = word_sub(x[i], 0, &borrow);
+
+   return borrow;
+   }
+
+/*
+* Two Operand Linear Multiply
+*/
+void bigint_linmul2(word x[], u32bit x_size, word y)
+   {
+   const u32bit blocks = x_size - (x_size % 8);
+
+   word carry = 0;
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      carry = word8_linmul2(x + i, y, carry);
+
+   for(u32bit i = blocks; i != x_size; ++i)
+      x[i] = word_madd2(x[i], y, &carry);
+
+   x[x_size] = carry;
+   }
+
+/*
+* Three Operand Linear Multiply
+*/
+void bigint_linmul3(word z[], const word x[], u32bit x_size, word y)
+   {
+   const u32bit blocks = x_size - (x_size % 8);
+
+   word carry = 0;
+
+   for(u32bit i = 0; i != blocks; i += 8)
+      carry = word8_linmul3(z + i, x + i, y, carry);
+
+   for(u32bit i = blocks; i != x_size; ++i)
+      z[i] = word_madd2(x[i], y, &carry);
+
+   z[x_size] = carry;
+   }
+
+}
+
+}