/*
--------------------------------------------------------------------
lookup.c, by Bob Jenkins, 1996, Public Domain.
hash(), hash2(), hash3, and mix() are externally useful functions.
Routines to test the hash are included if SELF_TEST is defined.
You can use this free for any purpose.  It has no warranty.

Obsolete.  Use lookup3.c instead, it is faster and more thorough.
--------------------------------------------------------------------
*/
#define SELF_TEST

#include <stdio.h>
#include <stddef.h>
#include <stdlib.h>
typedef  unsigned long  int  ub4;   /* unsigned 4-byte quantities */
typedef  unsigned       char ub1;

#define hashsize(n) ((ub4)1<<(n))
#define hashmask(n) (hashsize(n)-1)

/*
--------------------------------------------------------------------
mix -- mix 4 32-bit values reversibly.
Changing any input bit will usually change at least 32 output bits,
  whether the hash is run forward or in reverse.
Changing any input bit will change each of the 32 output bits in d
  about half the time when inputs a,b,c,d are uniformly distributed.
mix() takes 32 machine cycles.  No 16 or 24 cycle mixer works; this
  was confirmed by brute-force search.  This hash is the best of
  about 500,000 32-cycle hashes examined.
--------------------------------------------------------------------
*/
#define mix(a,b,c,d) \
{ \
  a += d; d += a; a ^= (a>>7);  \
  b += a; a += b; b ^= (b<<13);  \
  c += b; b += c; c ^= (c>>17); \
  d += c; c += d; d ^= (d<<9); \
  a += d; d += a; a ^= (a>>3);  \
  b += a; a += b; b ^= (b<<7);  \
  c += b; b += c; c ^= (c>>15); \
  d += c; c += d; d ^= (d<<11); \
}

/*
--------------------------------------------------------------------
hash() -- hash a variable-length key into a 32-bit value
  k     : the key (the unaligned variable-length array of bytes)
  len   : the length of the key, counting by bytes
  level : can be any 4-byte value
Returns a 32-bit value.  Every bit of the key affects every bit of
the return value.  There are no funnels smaller than 32 bits.
About 34+5len instructions.

The best hash table sizes are powers of 2.  There is no need to do
mod a prime (mod is sooo slow!).  If you need less than 32 bits,
use a bitmask.  For example, if you need only 10 bits, do
  h = (h & hashmask(10));
In which case, the hash table should have hashsize(10) elements.

If you are hashing n strings (ub1 **)k, do it like this:
  for (i=0, h=0; i<n; ++i) h = hash( k[i], len[i], h);

By Bob Jenkins, 1996.  74512.261@compuserve.com.  You may use this
code any way you wish, private, educational, or commercial.

See http://ourworld.compuserve.com/homepages/bob_jenkins/evahash.htm
Use for hash table lookup, or anything where one collision in 2^32 is
acceptable.  Do NOT use for cryptographic purposes.
--------------------------------------------------------------------
*/

ub4 hash( k, length, level)
register ub1 *k;        /* the key */
register ub4  length;   /* the length of the key */
register ub4  level;    /* the previous hash, or an arbitrary value */
{
   register ub4 a,b,c,d,len;

   /* Set up the internal state */
   len = length;
   a = b = c = 0x9e3779b9;  /* the golden ratio; an arbitrary value */
   d = level;               /* the previous hash value */

   /*---------------------------------------- handle most of the key */
   while (len >= 16)
   {
      a += (k[0] +((ub4)k[1]<<8) +((ub4)k[2]<<16) +((ub4)k[3]<<24));
      b += (k[4] +((ub4)k[5]<<8) +((ub4)k[6]<<16) +((ub4)k[7]<<24));
      c += (k[8] +((ub4)k[9]<<8) +((ub4)k[10]<<16)+((ub4)k[11]<<24));
      d += (k[12]+((ub4)k[13]<<8)+((ub4)k[14]<<16)+((ub4)k[15]<<24));
      mix(a,b,c,d);
      k += 16; len -= 16;
   }

   /*------------------------------------- handle the last 15 bytes */
   d += length;
   switch(len)              /* all the case statements fall through */
   {
   case 15: d+=((ub4)k[14]<<24);
   case 14: d+=((ub4)k[13]<<16);
   case 13: d+=((ub4)k[12]<<8);
      /* the first byte of d is reserved for the length */
   case 12: c+=((ub4)k[11]<<24);
   case 11: c+=((ub4)k[10]<<16);
   case 10: c+=((ub4)k[9]<<8);
   case 9 : c+=k[8];
   case 8 : b+=((ub4)k[7]<<24);
   case 7 : b+=((ub4)k[6]<<16);
   case 6 : b+=((ub4)k[5]<<8);
   case 5 : b+=k[4];
   case 4 : a+=((ub4)k[3]<<24);
   case 3 : a+=((ub4)k[2]<<16);
   case 2 : a+=((ub4)k[1]<<8);
   case 1 : a+=k[0];
     /* case 0: nothing left to add */
   }
   mix(a,b,c,d);
   /*-------------------------------------------- report the result */
   return d;
}


/*
--------------------------------------------------------------------
 This works on all machines, is identical to hash() on little-endian 
 machines, and it is much faster than hash(), but it requires
 -- that buffers be aligned, that is, ASSERT(((ub4)k)&3 == 0), and
 -- that all your machines have the same endianness.
--------------------------------------------------------------------
*/
ub4 hash2( k, length, level)
register ub1 *k;        /* the key */
register ub4  length;   /* the length of the key */
register ub4  level;    /* the previous hash, or an arbitrary value */
{
   register ub4 a,b,c,d,len;

   /* Set up the internal state */
   len = length;
   a = b = c = 0x9e3779b9;  /* the golden ratio; an arbitrary value */
   d = level;               /* the previous hash value */

   /*---------------------------------------- handle most of the key */
   while (len >= 16)
   {
      a += *(ub4 *)(k+0);
      b += *(ub4 *)(k+4);
      c += *(ub4 *)(k+8);
      d += *(ub4 *)(k+12);
      mix(a,b,c,d);
      k += 16; len -= 16;
   }

   /*------------------------------------- handle the last 15 bytes */
   d += length;
   switch(len)              /* all the case statements fall through */
   {
   case 15: d+=((ub4)k[14]<<24);
   case 14: d+=((ub4)k[13]<<16);
   case 13: d+=((ub4)k[12]<<8);
      /* the first byte of d is reserved for the length */
   case 12: c+=((ub4)k[11]<<24);
   case 11: c+=((ub4)k[10]<<16);
   case 10: c+=((ub4)k[9]<<8);
   case 9 : c+=k[8];
   case 8 : b+=((ub4)k[7]<<24);
   case 7 : b+=((ub4)k[6]<<16);
   case 6 : b+=((ub4)k[5]<<8);
   case 5 : b+=k[4];
   case 4 : a+=((ub4)k[3]<<24);
   case 3 : a+=((ub4)k[2]<<16);
   case 2 : a+=((ub4)k[1]<<8);
   case 1 : a+=k[0];
     /* case 0: nothing left to add */
   }
   mix(a,b,c,d);
   /*-------------------------------------------- report the result */
   return d;
}

/*
--------------------------------------------------------------------
 This is identical to hash() on little-endian machines, and it is much
 faster than hash(), but a little slower than hash2(), and it requires
 -- that all your machines be little-endian, for example all Intel x86
    chips or all VAXen.  It gives wrong results on big-endian machines.
--------------------------------------------------------------------
*/

ub4 hash3( k, length, level)
register ub1 *k;        /* the key */
register ub4  length;   /* the length of the key */
register ub4  level;    /* the previous hash, or an arbitrary value */
{
   register ub4 a,b,c,d,len;

   /* Set up the internal state */
   len = length;
   a = b = c = 0x9e3779b9;  /* the golden ratio; an arbitrary value */
   d = level;               /* the previous hash value */

   /*---------------------------------------- handle most of the key */
   if (((ub4)k)&3)
   {
      while (len >= 16)    /* unaligned */
      {
         a += (k[0] +((ub4)k[1]<<8) +((ub4)k[2]<<16) +((ub4)k[3]<<24));
         b += (k[4] +((ub4)k[5]<<8) +((ub4)k[6]<<16) +((ub4)k[7]<<24));
         c += (k[8] +((ub4)k[9]<<8) +((ub4)k[10]<<16)+((ub4)k[11]<<24));
         d += (k[12]+((ub4)k[13]<<8)+((ub4)k[14]<<16)+((ub4)k[15]<<24));
         mix(a,b,c,d);
         k += 16; len -= 16;
      }
   }
   else
   {
      while (len >= 16)    /* aligned */
      {
         a += *(ub4 *)(k+0);
         b += *(ub4 *)(k+4);
         c += *(ub4 *)(k+8);
         d += *(ub4 *)(k+12);
         mix(a,b,c,d);
         k += 16; len -= 16;
      }
   }

   /*------------------------------------- handle the last 15 bytes */
   d += length;
   switch(len)              /* all the case statements fall through */
   {
   case 15: d+=((ub4)k[14]<<24);
   case 14: d+=((ub4)k[13]<<16);
   case 13: d+=((ub4)k[12]<<8);
      /* the first byte of d is reserved for the length */
   case 12: c+=((ub4)k[11]<<24);
   case 11: c+=((ub4)k[10]<<16);
   case 10: c+=((ub4)k[9]<<8);
   case 9 : c+=k[8];
   case 8 : b+=((ub4)k[7]<<24);
   case 7 : b+=((ub4)k[6]<<16);
   case 6 : b+=((ub4)k[5]<<8);
   case 5 : b+=k[4];
   case 4 : a+=((ub4)k[3]<<24);
   case 3 : a+=((ub4)k[2]<<16);
   case 2 : a+=((ub4)k[1]<<8);
   case 1 : a+=k[0];
     /* case 0: nothing left to add */
   }
   mix(a,b,c,d);
   /*-------------------------------------------- report the result */
   return d;
}



#ifdef SELF_TEST

/* used for timings */
void driver1()
{
  ub4 buf[256];
  ub4 i;
  ub4 h=0;

  for (i=0; i<256; ++i) 
  {
    h = hash(buf,i,h);
  }
}

/* check that every input bit changes every output bit half the time */
#define HASHSTATE 1
#define HASHLEN   1
#define MAXPAIR 80
#define MAXLEN 70
void driver2()
{
  ub1 qa[MAXLEN+1], qb[MAXLEN+2], *a = &qa[0], *b = &qb[1];
  ub4 c[HASHSTATE], d[HASHSTATE], i, j=0, k, l, m, z;
  ub4 e[HASHSTATE],f[HASHSTATE],g[HASHSTATE],h[HASHSTATE];
  ub4 x[HASHSTATE],y[HASHSTATE];
  ub4 hlen;

  printf("No more than %d trials should ever be needed \n",MAXPAIR/2);
  for (hlen=0; hlen < MAXLEN; ++hlen)
  {
    z=0;
    for (i=0; i<hlen; ++i)  /*----------------------- for each byte, */
    {
      for (j=0; j<8; ++j)   /*------------------------ for each bit, */
      {
	for (m=1; m<8; ++m) /*-------- for serveral possible levels, */
	{
	  for (l=0; l<HASHSTATE; ++l) e[l]=f[l]=g[l]=h[l]=x[l]=y[l]=~((ub4)0);

      	  /*---- check that every input bit affects every output bit */
	  for (k=0; k<MAXPAIR; k+=2)
	  { 
	    ub4 finished=1;
	    /* keys have one bit different */
	    for (l=0; l<hlen+1; ++l) {a[l] = b[l] = (ub1)0;}
	    /* have a and b be two keys differing in only one bit */
	    a[i] ^= (k<<j);
	    a[i] ^= (k>>(8-j));
	     c[0] = hash(a, hlen, m);
	    b[i] ^= ((k+1)<<j);
	    b[i] ^= ((k+1)>>(8-j));
	     d[0] = hash(b, hlen, m);
	    /* check every bit is 1, 0, set, and not set at least once */
	    for (l=0; l<HASHSTATE; ++l)
	    {
	      e[l] &= (c[l]^d[l]);
	      f[l] &= ~(c[l]^d[l]);
	      g[l] &= c[l];
	      h[l] &= ~c[l];
	      x[l] &= d[l];
	      y[l] &= ~d[l];
	      if (e[l]|f[l]|g[l]|h[l]|x[l]|y[l]) finished=0;
	    }
	    if (finished) break;
	  }
	  if (k>z) z=k;
	  if (k==MAXPAIR) 
	  {
	     printf("Some bit didn't change: ");
	     printf("%.8lx %.8lx %.8lx %.8lx %.8lx %.8lx  ",
	            e[0],f[0],g[0],h[0],x[0],y[0]);
	     printf("i %ld j %ld len %ld\n",i,j,hlen);
	  }
	  if (z==MAXPAIR) goto done;
	}
      }
    }
   done:
    if (z < MAXPAIR)
    {
      printf("Mix success  %2ld bytes  %2ld levels  ",i,j);
      printf("required  %ld  trials\n",z/2);
    }
  }
  printf("\n");
}

/* Check for reading beyond the end of the buffer and alignment problems */
void driver3()
{
  ub1 buf[MAXLEN+20], *b;
  ub4 len;
  ub1 q[] = "This is the time for all good men to come to the aid of their country";
  ub1 qq[] = "xThis is the time for all good men to come to the aid of their country";
  ub1 qqq[] = "xxThis is the time for all good men to come to the aid of their country";
  ub1 qqqq[] = "xxxThis is the time for all good men to come to the aid of their country";
  ub4 h,i,j,ref,x,y;

  printf("Endianness.  These should all be the same:\n");
  printf("%.8lx\n", hash(q, sizeof(q)-1, (ub4)0));
  printf("%.8lx\n", hash(qq+1, sizeof(q)-1, (ub4)0));
  printf("%.8lx\n", hash(qqq+2, sizeof(q)-1, (ub4)0));
  printf("%.8lx\n", hash(qqqq+3, sizeof(q)-1, (ub4)0));
  printf("\n");
  for (h=0, b=buf+1; h<8; ++h, ++b)
  {
    for (i=0; i<MAXLEN; ++i)
    {
      len = i;
      for (j=0; j<i; ++j) *(b+j)=0;

      /* these should all be equal */
      ref = hash(b, len, (ub4)1);
      *(b+i)=(ub1)~0;
      *(b-1)=(ub1)~0;
      x = hash(b, len, (ub4)1);
      y = hash(b, len, (ub4)1);
      if ((ref != x) || (ref != y)) 
      {
	printf("alignment error: %.8lx %.8lx %.8lx %ld %ld\n",ref,x,y,h,i);
      }
    }
  }
}

/* check for problems with nulls */
 void driver4()
{
  ub1 buf[1];
  ub4 h,i,state[HASHSTATE];


  buf[0] = ~0;
  for (i=0; i<HASHSTATE; ++i) state[i] = 1;
  printf("These should all be different\n");
  for (i=0, h=0; i<8; ++i)
  {
    h = hash(buf, (ub4)0, h);
    printf("%2ld  0-byte strings, hash is  %.8lx\n", i, h);
  }
}


int main()
{
  driver1();   /* test that the key is hashed: used for timings */
  driver2();   /* test that whole key is hashed thoroughly */
  driver3();   /* test that nothing but the key is hashed */
  driver4();   /* test hashing multiple buffers (all buffers are null) */
  return 1;
}

#endif  /* SELF_TEST */