/* mmh3.c
 */
#include <string.h>
#include <stdint.h>

#include "mmh3.h"


static uint64_t
load64_le(const uint8_t *p)
{
    return ((uint64_t)p[0])       |
           ((uint64_t)p[1] << 8)  |
           ((uint64_t)p[2] << 16) |
           ((uint64_t)p[3] << 24) |
           ((uint64_t)p[4] << 32) |
           ((uint64_t)p[5] << 40) |
           ((uint64_t)p[6] << 48) |
           ((uint64_t)p[7] << 56);
}

/* mmh3_128() -- calculate 128 bit mmh3 hash.
 *
 * Args:
 *     key
 *     len
 *     seed
 *     out
 *
 * Returns:
 *     Nothing. Output is written to 'out'.
 */
void mmh3_128(const void *key, const size_t len, const uint64_t seed, uint64_t *out) {
    const uint64_t  c1 = 0x87c37b91114253d5;
    const uint64_t  c2 = 0x4cf5ad432745937f;
    const uint8_t   *data = (const uint8_t *)key;
    const size_t    nblocks = len / 16;
    uint64_t        h1 = seed;
    uint64_t        h2 = seed;

    // Process the data in 128-bit chunks
    for (size_t i = 0; i < nblocks; i++) {
        const uint8_t *block = data + i * 16;

	uint64_t k1 = load64_le(block);
	uint64_t k2 = load64_le(block + 8);

        k1 *= c1; k1 = (k1 << 31) | (k1 >> 33); k1 *= c2; h1 ^= k1;
        h1 = (h1 << 27) | (h1 >> 37); h1 += h2; h1 = h1 * 5 + 0x52dce729;

        k2 *= c2; k2 = (k2 << 33) | (k2 >> 31); k2 *= c1; h2 ^= k2;
        h2 = (h2 << 31) | (h2 >> 33); h2 += h1; h2 = h2 * 5 + 0x38495ab5;
    }

    // Process the remaining bytes
    const uint8_t *tail = (const uint8_t *)(data + nblocks * 16);
    uint64_t k1 = 0;
    uint64_t k2 = 0;

    switch (len & 15) {
    case 15: k2 ^= ((uint64_t)tail[14]) << 48; /* fall through */
    case 14: k2 ^= ((uint64_t)tail[13]) << 40;  /* fall through */
    case 13: k2 ^= ((uint64_t)tail[12]) << 32;  /* fall through */
    case 12: k2 ^= ((uint64_t)tail[11]) << 24;  /* fall through */
    case 11: k2 ^= ((uint64_t)tail[10]) << 16; /* fall through */
    case 10: k2 ^= ((uint64_t)tail[9]) << 8; /* fall through */
    case  9: k2 ^= ((uint64_t)tail[8]) << 0; /* fall through */
      k2 *= c2; k2 = (k2 << 33) | (k2 >> 31); k2 *= c1; h2 ^= k2;
      /* fall through */
    case  8: k1 ^= ((uint64_t)tail[7]) << 56; /* fall through */
    case  7: k1 ^= ((uint64_t)tail[6]) << 48; /* fall through */
    case  6: k1 ^= ((uint64_t)tail[5]) << 40; /* fall through */
    case  5: k1 ^= ((uint64_t)tail[4]) << 32; /* fall through */
    case  4: k1 ^= ((uint64_t)tail[3]) << 24; /* fall through */
    case  3: k1 ^= ((uint64_t)tail[2]) << 16; /* fall through */
    case  2: k1 ^= ((uint64_t)tail[1]) << 8; /* fall through */
    case  1: k1 ^= ((uint64_t)tail[0]) << 0; /* fall through */
      k1 *= c1; k1 = (k1 << 31) | (k1 >> 33); k1 *= c2; h1 ^= k1;
    }

    // Finalization
    h1 ^= len;
    h2 ^= len;

    h1 += h2;
    h2 += h1;

    h1 ^= h1 >> 33;
    h1 *= 0xff51afd7ed558ccd;
    h1 ^= h1 >> 33;
    h1 *= 0xc4ceb9fe1a85ec53;
    h1 ^= h1 >> 33;

    h2 ^= h2 >> 33;
    h2 *= 0xff51afd7ed558ccd;
    h2 ^= h2 >> 33;
    h2 *= 0xc4ceb9fe1a85ec53;
    h2 ^= h2 >> 33;

    h1 += h2;
    h2 += h1;

    out[0] = h1;
    out[1] = h2;
}