#include <linux/kernel.h>
#include <linux/string.h>

static const u32 r[] = {
	0xB7E15162, 0xBF715880, 0x38B4DA56, 0x324E7738,
	0xBB1185EB, 0x4F7C7B57, 0xCFBFA1C8, 0xC2B3293D
};

static inline void sparkle256(u32 *state)
{
	int i, j;
	u32 rc, tmpx, tmpy, x0, y0;

	for (i = 0; i < 6; i ++) {
		state[1] ^= r[i];
		state[3] ^= i;
		for (j = 0; j < 8; j += 2) {
			rc = r[j >> 1];
			state[j] += ror32(state[j + 1], 31);
			state[j + 1] ^= ror32(state[j], 24);
			state[j] ^= rc;
			state[j] += ror32(state[j + 1], 17);
			state[j + 1] ^= ror32(state[j], 17);
			state[j] ^= rc;
			state[j] += state[j+1];
			state[j + 1] ^= ror32(state[j], 31);
			state[j] ^= rc;
			state[j] += ror32(state[j + 1], 24);
			state[j + 1] ^= ror32(state[j], 16);
			state[j] ^= rc;
		}
		tmpx = x0 = state[0];
		tmpy = y0 = state[1];
		for (j = 2; j < 4; j += 2) {
			tmpx ^= state[j];
			tmpy ^= state[j + 1];
		}
		tmpx = ror32((tmpx ^ (tmpx << 16)), 16);
		tmpy = ror32((tmpy ^ (tmpy << 16)), 16);
		for (j = 2; j < 4; j += 2) {
			state[j - 2] = state[j + 4] ^ state[j] ^ tmpy;
			state[j + 4] = state[j];
			state[j - 1] = state[j+ 4 + 1] ^ state[j + 1] ^ tmpx;
			state[j + 4 + 1] = state[j + 1];
		}
		state[4 - 2] = state[4] ^ x0 ^ tmpy;
		state[4] = x0;
		state[4 - 1] = state[4 + 1] ^ y0 ^ tmpx;
		state[4 + 1] = y0;
	}
}

void mix_sparkle(u32 h[4], const u32 v[4])
{
	u32 s[8];
	memcpy(s, h, sizeof(u32) * 4);
	memcpy(s + 4, v, sizeof(u32) * 4);
	sparkle256(s);
	memcpy(h, s, sizeof(u32) * 4);
}