#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <math.h>
#include <limits.h>

#include "nmglobal.h"
#include "nbench1.h"


/********************
** IDEA Encryption **
*********************
** IDEA - International Data Encryption Algorithm.
** Based on code presented in Applied Cryptography by Bruce Schneier.
** Which was based on code developed by Xuejia Lai and James L. Massey.
** Other modifications made by Colin Plumb.
**
*/

/*
** DEFINES
*/
#define IDEAKEYSIZE 16
#define IDEABLOCKSIZE 8
#define ROUNDS 8
#define KEYLEN (6*ROUNDS+4)

/*
** MACROS
*/
#define low16(x) ((x) & 0x0FFFF)
#define MUL(x,y) (x=mul(low16(x),y))


typedef uint16_t IDEAkey[KEYLEN];

/*
** PROTOTYPES
*/
static unsigned long DoIDEAIteration(unsigned char *plain1,
	unsigned char *crypt1, unsigned char *plain2,
	unsigned long arraysize, unsigned long nloops,
	IDEAkey Z, IDEAkey DK);
static uint16_t mul(register uint16_t a, register uint16_t b);
static uint16_t inv(uint16_t x);
static void en_key_idea(uint16_t userkey[8], IDEAkey Z);
static void de_key_idea(IDEAkey Z, IDEAkey DK);
static void cipher_idea(uint16_t in[4], uint16_t out[4], IDEAkey Z);

/***********
** DoIDEA **
************
** Perform IDEA encryption.  Note that we time encryption & decryption
** time as being a single loop.
*/
void DoIDEA(void)
{
IDEAStruct *locideastruct;      /* Loc pointer to global structure */
int i;
IDEAkey Z,DK;
uint16_t userkey[8];
unsigned long accumtime;
double iterations;
char *context;
unsigned char *plain1 = NULL;               /* First plaintext buffer */
unsigned char *crypt1 = NULL;               /* Encryption buffer */
unsigned char *plain2 = NULL;               /* Second plaintext buffer */

/*
** Link to global data
*/
locideastruct=&global_ideastruct;

/*
** Set error context
*/
context="CPU:IDEA";

/*
** Re-init random-number generator.
*/
/* randnum(3L); */
randnum(3);

/*
** Build an encryption/decryption key
*/
for (i=0;i<8;i++)
        /* userkey[i]=(uint16_t)(abs_randwc(60000L) & 0xFFFF); */
	userkey[i]=(uint16_t)(abs_randwc((int32_t)60000) & 0xFFFF);
for(i=0;i<KEYLEN;i++)
	Z[i]=0;

/*
** Compute encryption/decryption subkeys
*/
en_key_idea(userkey,Z);
de_key_idea(Z,DK);

/*
** Allocate memory for buffers.  We'll make 3, called plain1,
** crypt1, and plain2.  It works like this:
**   plain1 >>encrypt>> crypt1 >>decrypt>> plain2.
** So, plain1 and plain2 should match.
** Also, fill up plain1 with sample text.
*/
plain1 = malloc(locideastruct->arraysize);
if (!plain1) {
        fprintf(stderr, "Error in %s, could not allocate memory. Exitting...\n", context);
        exit(1);
}

crypt1 = malloc(locideastruct->arraysize);
if (!crypt1) {
        fprintf(stderr, "Error in %s, could not allocate memory. Exitting...\n", context);
        free(plain1);
        exit(1);
}

plain2 = malloc(locideastruct->arraysize);
if (!plain2) {
        fprintf(stderr, "Error in %s, could not allocate memory. Exitting...\n", context);
        free(plain1);
        free(plain2);
        exit(1);
}
/*
** Note that we build the "plaintext" by simply loading
** the array up with random numbers.
*/
for(i=0;i<locideastruct->arraysize;i++)
	plain1[i]=(unsigned char)(abs_randwc(255) & 0xFF);

/*
** See if we need to perform self adjustment loop.
*/
if(locideastruct->adjust==0)
{
	/*
	** Do self-adjustment.  This involves initializing the
	** # of loops and increasing the loop count until we
	** get a number of loops that we can use.
	*/
	for(locideastruct->loops=100L;
	  locideastruct->loops<MAXIDEALOOPS;
	  locideastruct->loops+=10L)
		if(DoIDEAIteration(plain1,crypt1,plain2,
		  locideastruct->arraysize,
		  locideastruct->loops,
		  Z,DK)>global_min_ticks) break;
}

/*
** All's well if we get here.  Do the test.
*/
accumtime=0L;
iterations=(double)0.0;

do {
	accumtime+=DoIDEAIteration(plain1,crypt1,plain2,
		locideastruct->arraysize,
		locideastruct->loops,Z,DK);
	iterations+=(double)locideastruct->loops;
} while(TicksToSecs(accumtime)<locideastruct->request_secs);

/*
** Clean up, calculate results, and go home.  Be sure to
** show that we don't have to rerun adjustment code.
*/
free(plain1);
free(crypt1);
free(plain2);

locideastruct->iterspersec=iterations / TicksToFracSecs(accumtime);

if(locideastruct->adjust==0)
	locideastruct->adjust=1;

return;

}

/********************
** DoIDEAIteration **
*********************
** Execute a single iteration of the IDEA encryption algorithm.
** Actually, a single iteration is one encryption and one
** decryption.
*/
static unsigned long DoIDEAIteration(unsigned char *plain1,
			unsigned char *crypt1,
			unsigned char *plain2,
			unsigned long arraysize,
			unsigned long nloops,
			IDEAkey Z,
			IDEAkey DK)
{
register unsigned long i;
register unsigned long j;
unsigned long elapsed;
#ifdef DEBUG
int status=0;
#endif

/*
** Start the stopwatch.
*/
elapsed=StartStopwatch();

/*
** Do everything for nloops.
*/
for(i=0;i<nloops;i++)
{
	for(j=0;j<arraysize;j+=(sizeof(uint16_t)*4))
		cipher_idea((uint16_t *)(plain1+j),(uint16_t *)(crypt1+j),Z);       /* Encrypt */

	for(j=0;j<arraysize;j+=(sizeof(uint16_t)*4))
		cipher_idea((uint16_t *)(crypt1+j),(uint16_t *)(plain2+j),DK);      /* Decrypt */
}

#ifdef DEBUG
for(j=0;j<arraysize;j++)
	if(*(plain1+j)!=*(plain2+j)){
		printf("IDEA Error! \n");
                status=1;
                }
if (status==0) printf("IDEA: OK\n");
#endif

/*
** Get elapsed time.
*/
return(StopStopwatch(elapsed));
}

/********
** mul **
*********
** Performs multiplication, modulo (2**16)+1.  This code is structured
** on the assumption that untaken branches are cheaper than taken
** branches, and that the compiler doesn't schedule branches.
*/
static uint16_t mul(register uint16_t a, register uint16_t b)
{
register uint32_t p;
if(a)
{       if(b)
	{       p=(uint32_t)(a*b);
		b=low16(p);
		a=(uint16_t)(p>>16);
		return(b-a+(b<a));
	}
	else
		return(1-a);
}
else
	return(1-b);
}

/********
** inv **
*********
** Compute multiplicative inverse of x, modulo (2**16)+1
** using Euclid's GCD algorithm.  It is unrolled twice
** to avoid swapping the meaning of the registers.  And
** some subtracts are changed to adds.
*/
static uint16_t inv(uint16_t x)
{
uint16_t t0, t1;
uint16_t q, y;

if(x<=1)
	return(x);      /* 0 and 1 are self-inverse */
t1=0x10001 / x;
y=0x10001 % x;
if(y==1)
	return(low16(1-t1));
t0=1;
do {
	q=x/y;
	x=x%y;
	t0+=q*t1;
	if(x==1) return(t0);
	q=y/x;
	y=y%x;
	t1+=q*t0;
} while(y!=1);
return(low16(1-t1));
}

/****************
** en_key_idea **
*****************
** Compute IDEA encryption subkeys Z
*/
static void en_key_idea(uint16_t *userkey, uint16_t *Z)
{
int i,j;

/*
** shifts
*/
for(j=0;j<8;j++)
	Z[j]=*userkey++;
for(i=0;j<KEYLEN;j++)
{       i++;
	Z[i+7]=(Z[i&7]<<9)| (Z[(i+1) & 7] >> 7);
	Z+=i&8;
	i&=7;
}
return;
}

/****************
** de_key_idea **
*****************
** Compute IDEA decryption subkeys DK from encryption
** subkeys Z.
*/
static void de_key_idea(IDEAkey Z, IDEAkey DK)
{
IDEAkey TT;
int j;
uint16_t t1, t2, t3;
uint16_t *p;
p=(uint16_t *)(TT+KEYLEN);

t1=inv(*Z++);
t2=-*Z++;
t3=-*Z++;
*--p=inv(*Z++);
*--p=t3;
*--p=t2;
*--p=t1;

for(j=1;j<ROUNDS;j++)
{       t1=*Z++;
	*--p=*Z++;
	*--p=t1;
	t1=inv(*Z++);
	t2=-*Z++;
	t3=-*Z++;
	*--p=inv(*Z++);
	*--p=t2;
	*--p=t3;
	*--p=t1;
}
t1=*Z++;
*--p=*Z++;
*--p=t1;
t1=inv(*Z++);
t2=-*Z++;
t3=-*Z++;
*--p=inv(*Z++);
*--p=t3;
*--p=t2;
*--p=t1;
/*
** Copy and destroy temp copy
*/
for(j=0,p=TT;j<KEYLEN;j++)
{       *DK++=*p;
	*p++=0;
}

return;
}

/*
** MUL(x,y)
** This #define creates a macro that computes x=x*y modulo 0x10001.
** Requires temps t16 and t32.  Also requires y to be strictly 16
** bits.  Here, I am using the simplest form.  May not be the
** fastest. -- RG
*/
/* #define MUL(x,y) (x=mul(low16(x),y)) */

/****************
** cipher_idea **
*****************
** IDEA encryption/decryption algorithm.
*/
static void cipher_idea(uint16_t in[4],
		uint16_t out[4],
		register IDEAkey Z)
{
register uint16_t x1, x2, x3, x4, t1, t2;
/* register uint16_t t16;
register uint16_t t32; */
int r=ROUNDS;

x1=*in++;
x2=*in++;
x3=*in++;
x4=*in;

do {
	MUL(x1,*Z++);
	x2+=*Z++;
	x3+=*Z++;
	MUL(x4,*Z++);

	t2=x1^x3;
	MUL(t2,*Z++);
	t1=t2+(x2^x4);
	MUL(t1,*Z++);
	t2=t1+t2;

	x1^=t1;
	x4^=t2;

	t2^=x2;
	x2=x3^t1;
	x3=t2;
} while(--r);
MUL(x1,*Z++);
*out++=x1;
*out++=x3+*Z++;
*out++=x2+*Z++;
MUL(x4,*Z);
*out=x4;
return;
}