X-Git-Url: http://tinc-vpn.org/git/browse?a=blobdiff_plain;f=src%2Fchacha-poly1305%2Fpoly1305.c;fp=src%2Fchacha-poly1305%2Fpoly1305.c;h=0c90564cb0bba833886742713033b1fd2f43e560;hb=63df2139835bb532125562c73ec3efbc437a3634;hp=4d99b8c3ae15886d239c6f9206376ce5df5a8cc6;hpb=9ce00234190baec0eaa608a5093bb3823487ef0c;p=tinc

diff --git a/src/chacha-poly1305/poly1305.c b/src/chacha-poly1305/poly1305.c
index 4d99b8c3..0c90564c 100644
--- a/src/chacha-poly1305/poly1305.c
+++ b/src/chacha-poly1305/poly1305.c
@@ -1,205 +1,302 @@
 /*
- * Public Domain poly1305 from Andrew Moon
- * poly1305-donna-unrolled.c from https://github.com/floodyberry/poly1305-donna
- */
-
-#include "../system.h"
+poly1305 implementation using 32 bit * 32 bit = 64 bit multiplication and 64 bit addition
+public domain
+*/
 
 #include "poly1305.h"
 
-#define mul32x32_64(a,b) ((uint64_t)(a) * (b))
-
-#define U8TO32_LE(p) \
-	(((uint32_t)((p)[0])) | \
-	 ((uint32_t)((p)[1]) <<  8) | \
-	 ((uint32_t)((p)[2]) << 16) | \
-	 ((uint32_t)((p)[3]) << 24))
-
-#define U32TO8_LE(p, v) \
+#if (USE_UNALIGNED == 1)
+#define U8TO32(p) \
+	(*((uint32_t *)(p)))
+#define U32TO8(p, v) \
 	do { \
-		(p)[0] = (uint8_t)((v)); \
-		(p)[1] = (uint8_t)((v) >>  8); \
-		(p)[2] = (uint8_t)((v) >> 16); \
-		(p)[3] = (uint8_t)((v) >> 24); \
+		*((uint32_t *)(p)) = v; \
 	} while (0)
+#else
+/* interpret four 8 bit unsigned integers as a 32 bit unsigned integer in little endian */
+static uint32_t
+U8TO32(const unsigned char *p) {
+	return
+	        (((uint32_t)(p[0] & 0xff)) |
+	         ((uint32_t)(p[1] & 0xff) <<  8) |
+	         ((uint32_t)(p[2] & 0xff) << 16) |
+	         ((uint32_t)(p[3] & 0xff) << 24));
+}
+
+/* store a 32 bit unsigned integer as four 8 bit unsigned integers in little endian */
+static void
+U32TO8(unsigned char *p, uint32_t v) {
+	p[0] = (v) & 0xff;
+	p[1] = (v >>  8) & 0xff;
+	p[2] = (v >> 16) & 0xff;
+	p[3] = (v >> 24) & 0xff;
+}
+#endif
 
 void
-poly1305_auth(unsigned char out[POLY1305_TAGLEN], const unsigned char *m, size_t inlen, const unsigned char key[POLY1305_KEYLEN]) {
-	uint32_t t0, t1, t2, t3;
-	uint32_t h0, h1, h2, h3, h4;
+poly1305_init(struct poly1305_context *st, const unsigned char key[32]) {
+	/* r &= 0xffffffc0ffffffc0ffffffc0fffffff */
+	st->r[0] = (U8TO32(&key[ 0])) & 0x3ffffff;
+	st->r[1] = (U8TO32(&key[ 3]) >> 2) & 0x3ffff03;
+	st->r[2] = (U8TO32(&key[ 6]) >> 4) & 0x3ffc0ff;
+	st->r[3] = (U8TO32(&key[ 9]) >> 6) & 0x3f03fff;
+	st->r[4] = (U8TO32(&key[12]) >> 8) & 0x00fffff;
+
+	/* h = 0 */
+	st->h[0] = 0;
+	st->h[1] = 0;
+	st->h[2] = 0;
+	st->h[3] = 0;
+	st->h[4] = 0;
+
+	/* save pad for later */
+	st->pad[0] = U8TO32(&key[16]);
+	st->pad[1] = U8TO32(&key[20]);
+	st->pad[2] = U8TO32(&key[24]);
+	st->pad[3] = U8TO32(&key[28]);
+
+	st->leftover = 0;
+	st->final = 0;
+}
+
+static void
+poly1305_blocks(struct poly1305_context *st, const unsigned char *m, size_t bytes) {
+	const uint32_t hibit = (st->final) ? 0 : (1 << 24); /* 1 << 128 */
 	uint32_t r0, r1, r2, r3, r4;
 	uint32_t s1, s2, s3, s4;
-	uint32_t b, nb;
-	size_t j;
-	uint64_t t[5];
-	uint64_t f0, f1, f2, f3;
-	uint32_t g0, g1, g2, g3, g4;
-	uint64_t c;
-	unsigned char mp[16];
-
-	/* clamp key */
-	t0 = U8TO32_LE(key + 0);
-	t1 = U8TO32_LE(key + 4);
-	t2 = U8TO32_LE(key + 8);
-	t3 = U8TO32_LE(key + 12);
-
-	/* precompute multipliers */
-	r0 = t0 & 0x3ffffff;
-	t0 >>= 26;
-	t0 |= t1 << 6;
-	r1 = t0 & 0x3ffff03;
-	t1 >>= 20;
-	t1 |= t2 << 12;
-	r2 = t1 & 0x3ffc0ff;
-	t2 >>= 14;
-	t2 |= t3 << 18;
-	r3 = t2 & 0x3f03fff;
-	t3 >>= 8;
-	r4 = t3 & 0x00fffff;
+	uint32_t h0, h1, h2, h3, h4;
+	uint64_t d0, d1, d2, d3, d4;
+	uint32_t c;
+
+	r0 = st->r[0];
+	r1 = st->r[1];
+	r2 = st->r[2];
+	r3 = st->r[3];
+	r4 = st->r[4];
 
 	s1 = r1 * 5;
 	s2 = r2 * 5;
 	s3 = r3 * 5;
 	s4 = r4 * 5;
 
-	/* init state */
-	h0 = 0;
-	h1 = 0;
-	h2 = 0;
-	h3 = 0;
-	h4 = 0;
-
-	/* full blocks */
-	if(inlen < 16) {
-		goto poly1305_donna_atmost15bytes;
-	}
+	h0 = st->h[0];
+	h1 = st->h[1];
+	h2 = st->h[2];
+	h3 = st->h[3];
+	h4 = st->h[4];
 
-poly1305_donna_16bytes:
-	m += 16;
-	inlen -= 16;
-
-	t0 = U8TO32_LE(m - 16);
-	t1 = U8TO32_LE(m - 12);
-	t2 = U8TO32_LE(m - 8);
-	t3 = U8TO32_LE(m - 4);
-
-	h0 += t0 & 0x3ffffff;
-	h1 += ((((uint64_t) t1 << 32) | t0) >> 26) & 0x3ffffff;
-	h2 += ((((uint64_t) t2 << 32) | t1) >> 20) & 0x3ffffff;
-	h3 += ((((uint64_t) t3 << 32) | t2) >> 14) & 0x3ffffff;
-	h4 += (t3 >> 8) | (1 << 24);
-
-poly1305_donna_mul:
-	t[0] = mul32x32_64(h0, r0) + mul32x32_64(h1, s4) + mul32x32_64(h2, s3) + mul32x32_64(h3, s2) + mul32x32_64(h4, s1);
-	t[1] = mul32x32_64(h0, r1) + mul32x32_64(h1, r0) + mul32x32_64(h2, s4) + mul32x32_64(h3, s3) + mul32x32_64(h4, s2);
-	t[2] = mul32x32_64(h0, r2) + mul32x32_64(h1, r1) + mul32x32_64(h2, r0) + mul32x32_64(h3, s4) + mul32x32_64(h4, s3);
-	t[3] = mul32x32_64(h0, r3) + mul32x32_64(h1, r2) + mul32x32_64(h2, r1) + mul32x32_64(h3, r0) + mul32x32_64(h4, s4);
-	t[4] = mul32x32_64(h0, r4) + mul32x32_64(h1, r3) + mul32x32_64(h2, r2) + mul32x32_64(h3, r1) + mul32x32_64(h4, r0);
-
-	h0 = (uint32_t) t[0] & 0x3ffffff;
-	c = (t[0] >> 26);
-	t[1] += c;
-	h1 = (uint32_t) t[1] & 0x3ffffff;
-	b = (uint32_t)(t[1] >> 26);
-	t[2] += b;
-	h2 = (uint32_t) t[2] & 0x3ffffff;
-	b = (uint32_t)(t[2] >> 26);
-	t[3] += b;
-	h3 = (uint32_t) t[3] & 0x3ffffff;
-	b = (uint32_t)(t[3] >> 26);
-	t[4] += b;
-	h4 = (uint32_t) t[4] & 0x3ffffff;
-	b = (uint32_t)(t[4] >> 26);
-	h0 += b * 5;
-
-	if(inlen >= 16) {
-		goto poly1305_donna_16bytes;
-	}
+	while(bytes >= POLY1305_BLOCK_SIZE) {
+		/* h += m[i] */
+		h0 += (U8TO32(m + 0)) & 0x3ffffff;
+		h1 += (U8TO32(m + 3) >> 2) & 0x3ffffff;
+		h2 += (U8TO32(m + 6) >> 4) & 0x3ffffff;
+		h3 += (U8TO32(m + 9) >> 6) & 0x3ffffff;
+		h4 += (U8TO32(m + 12) >> 8) | hibit;
 
-	/* final bytes */
-poly1305_donna_atmost15bytes:
+		/* h *= r */
+		d0 = ((uint64_t)h0 * r0) + ((uint64_t)h1 * s4) + ((uint64_t)h2 * s3) + ((uint64_t)h3 * s2) + ((uint64_t)h4 * s1);
+		d1 = ((uint64_t)h0 * r1) + ((uint64_t)h1 * r0) + ((uint64_t)h2 * s4) + ((uint64_t)h3 * s3) + ((uint64_t)h4 * s2);
+		d2 = ((uint64_t)h0 * r2) + ((uint64_t)h1 * r1) + ((uint64_t)h2 * r0) + ((uint64_t)h3 * s4) + ((uint64_t)h4 * s3);
+		d3 = ((uint64_t)h0 * r3) + ((uint64_t)h1 * r2) + ((uint64_t)h2 * r1) + ((uint64_t)h3 * r0) + ((uint64_t)h4 * s4);
+		d4 = ((uint64_t)h0 * r4) + ((uint64_t)h1 * r3) + ((uint64_t)h2 * r2) + ((uint64_t)h3 * r1) + ((uint64_t)h4 * r0);
 
-	if(!inlen) {
-		goto poly1305_donna_finish;
-	}
+		/* (partial) h %= p */
+		c = (uint32_t)(d0 >> 26);
+		h0 = (uint32_t)d0 & 0x3ffffff;
+		d1 += c;
+		c = (uint32_t)(d1 >> 26);
+		h1 = (uint32_t)d1 & 0x3ffffff;
+		d2 += c;
+		c = (uint32_t)(d2 >> 26);
+		h2 = (uint32_t)d2 & 0x3ffffff;
+		d3 += c;
+		c = (uint32_t)(d3 >> 26);
+		h3 = (uint32_t)d3 & 0x3ffffff;
+		d4 += c;
+		c = (uint32_t)(d4 >> 26);
+		h4 = (uint32_t)d4 & 0x3ffffff;
+		h0 += c * 5;
+		c = (h0 >> 26);
+		h0 =           h0 & 0x3ffffff;
+		h1 += c;
 
-	for(j = 0; j < inlen; j++) {
-		mp[j] = m[j];
+		m += POLY1305_BLOCK_SIZE;
+		bytes -= POLY1305_BLOCK_SIZE;
 	}
 
-	mp[j++] = 1;
+	st->h[0] = h0;
+	st->h[1] = h1;
+	st->h[2] = h2;
+	st->h[3] = h3;
+	st->h[4] = h4;
+}
 
-	for(; j < 16; j++) {
-		mp[j] = 0;
-	}
+void
+poly1305_finish(struct poly1305_context *st, unsigned char mac[16]) {
+	uint32_t h0, h1, h2, h3, h4, c;
+	uint32_t g0, g1, g2, g3, g4;
+	uint64_t f;
+	uint32_t mask;
 
-	inlen = 0;
+	/* process the remaining block */
+	if(st->leftover) {
+		size_t i = st->leftover;
+		st->buffer[i++] = 1;
 
-	t0 = U8TO32_LE(mp + 0);
-	t1 = U8TO32_LE(mp + 4);
-	t2 = U8TO32_LE(mp + 8);
-	t3 = U8TO32_LE(mp + 12);
+		for(; i < POLY1305_BLOCK_SIZE; i++) {
+			st->buffer[i] = 0;
+		}
 
-	h0 += t0 & 0x3ffffff;
-	h1 += ((((uint64_t) t1 << 32) | t0) >> 26) & 0x3ffffff;
-	h2 += ((((uint64_t) t2 << 32) | t1) >> 20) & 0x3ffffff;
-	h3 += ((((uint64_t) t3 << 32) | t2) >> 14) & 0x3ffffff;
-	h4 += (t3 >> 8);
+		st->final = 1;
+		poly1305_blocks(st, st->buffer, POLY1305_BLOCK_SIZE);
+	}
 
-	goto poly1305_donna_mul;
+	/* fully carry h */
+	h0 = st->h[0];
+	h1 = st->h[1];
+	h2 = st->h[2];
+	h3 = st->h[3];
+	h4 = st->h[4];
 
-poly1305_donna_finish:
-	b = h0 >> 26;
-	h0 = h0 & 0x3ffffff;
-	h1 += b;
-	b = h1 >> 26;
+	c = h1 >> 26;
 	h1 = h1 & 0x3ffffff;
-	h2 += b;
-	b = h2 >> 26;
+	h2 +=     c;
+	c = h2 >> 26;
 	h2 = h2 & 0x3ffffff;
-	h3 += b;
-	b = h3 >> 26;
+	h3 +=     c;
+	c = h3 >> 26;
 	h3 = h3 & 0x3ffffff;
-	h4 += b;
-	b = h4 >> 26;
+	h4 +=     c;
+	c = h4 >> 26;
 	h4 = h4 & 0x3ffffff;
-	h0 += b * 5;
-	b = h0 >> 26;
+	h0 += c * 5;
+	c = h0 >> 26;
 	h0 = h0 & 0x3ffffff;
-	h1 += b;
+	h1 +=     c;
 
+	/* compute h + -p */
 	g0 = h0 + 5;
-	b = g0 >> 26;
+	c = g0 >> 26;
 	g0 &= 0x3ffffff;
-	g1 = h1 + b;
-	b = g1 >> 26;
+	g1 = h1 + c;
+	c = g1 >> 26;
 	g1 &= 0x3ffffff;
-	g2 = h2 + b;
-	b = g2 >> 26;
+	g2 = h2 + c;
+	c = g2 >> 26;
 	g2 &= 0x3ffffff;
-	g3 = h3 + b;
-	b = g3 >> 26;
+	g3 = h3 + c;
+	c = g3 >> 26;
 	g3 &= 0x3ffffff;
-	g4 = h4 + b - (1 << 26);
-
-	b = (g4 >> 31) - 1;
-	nb = ~b;
-	h0 = (h0 & nb) | (g0 & b);
-	h1 = (h1 & nb) | (g1 & b);
-	h2 = (h2 & nb) | (g2 & b);
-	h3 = (h3 & nb) | (g3 & b);
-	h4 = (h4 & nb) | (g4 & b);
-
-	f0 = ((h0) | (h1 << 26)) + (uint64_t) U8TO32_LE(&key[16]);
-	f1 = ((h1 >> 6) | (h2 << 20)) + (uint64_t) U8TO32_LE(&key[20]);
-	f2 = ((h2 >> 12) | (h3 << 14)) + (uint64_t) U8TO32_LE(&key[24]);
-	f3 = ((h3 >> 18) | (h4 << 8)) + (uint64_t) U8TO32_LE(&key[28]);
-
-	U32TO8_LE(&out[0], f0);
-	f1 += (f0 >> 32);
-	U32TO8_LE(&out[4], f1);
-	f2 += (f1 >> 32);
-	U32TO8_LE(&out[8], f2);
-	f3 += (f2 >> 32);
-	U32TO8_LE(&out[12], f3);
+	g4 = h4 + c - (1 << 26);
+
+	/* select h if h < p, or h + -p if h >= p */
+	mask = (g4 >> ((sizeof(uint32_t) * 8) - 1)) - 1;
+	g0 &= mask;
+	g1 &= mask;
+	g2 &= mask;
+	g3 &= mask;
+	g4 &= mask;
+	mask = ~mask;
+	h0 = (h0 & mask) | g0;
+	h1 = (h1 & mask) | g1;
+	h2 = (h2 & mask) | g2;
+	h3 = (h3 & mask) | g3;
+	h4 = (h4 & mask) | g4;
+
+	/* h = h % (2^128) */
+	h0 = ((h0) | (h1 << 26)) & 0xffffffff;
+	h1 = ((h1 >>  6) | (h2 << 20)) & 0xffffffff;
+	h2 = ((h2 >> 12) | (h3 << 14)) & 0xffffffff;
+	h3 = ((h3 >> 18) | (h4 <<  8)) & 0xffffffff;
+
+	/* mac = (h + pad) % (2^128) */
+	f = (uint64_t)h0 + st->pad[0]            ;
+	h0 = (uint32_t)f;
+	f = (uint64_t)h1 + st->pad[1] + (f >> 32);
+	h1 = (uint32_t)f;
+	f = (uint64_t)h2 + st->pad[2] + (f >> 32);
+	h2 = (uint32_t)f;
+	f = (uint64_t)h3 + st->pad[3] + (f >> 32);
+	h3 = (uint32_t)f;
+
+	U32TO8(mac +  0, h0);
+	U32TO8(mac +  4, h1);
+	U32TO8(mac +  8, h2);
+	U32TO8(mac + 12, h3);
+
+	/* zero out the state */
+	st->h[0] = 0;
+	st->h[1] = 0;
+	st->h[2] = 0;
+	st->h[3] = 0;
+	st->h[4] = 0;
+	st->r[0] = 0;
+	st->r[1] = 0;
+	st->r[2] = 0;
+	st->r[3] = 0;
+	st->r[4] = 0;
+	st->pad[0] = 0;
+	st->pad[1] = 0;
+	st->pad[2] = 0;
+	st->pad[3] = 0;
+}
+
+
+void
+poly1305_update(struct poly1305_context *st, const unsigned char *m, size_t bytes) {
+	size_t i;
+
+	/* handle leftover */
+	if(st->leftover) {
+		size_t want = (POLY1305_BLOCK_SIZE - st->leftover);
+
+		if(want > bytes) {
+			want = bytes;
+		}
+
+		for(i = 0; i < want; i++) {
+			st->buffer[st->leftover + i] = m[i];
+		}
+
+		bytes -= want;
+		m += want;
+		st->leftover += want;
+
+		if(st->leftover < POLY1305_BLOCK_SIZE) {
+			return;
+		}
+
+		poly1305_blocks(st, st->buffer, POLY1305_BLOCK_SIZE);
+		st->leftover = 0;
+	}
+
+	/* process full blocks */
+	if(bytes >= POLY1305_BLOCK_SIZE) {
+		size_t want = (bytes & ~(POLY1305_BLOCK_SIZE - 1));
+		poly1305_blocks(st, m, want);
+		m += want;
+		bytes -= want;
+	}
+
+	/* store leftover */
+	if(bytes) {
+#if (USE_MEMCPY == 1)
+		memcpy(st->buffer + st->leftover, m, bytes);
+#else
+
+		for(i = 0; i < bytes; i++) {
+			st->buffer[st->leftover + i] = m[i];
+		}
+
+#endif
+		st->leftover += bytes;
+	}
+}
+
+void
+poly1305_auth(unsigned char mac[16], const unsigned char *m, size_t bytes, const unsigned char key[32]) {
+	struct poly1305_context ctx;
+	poly1305_init(&ctx, key);
+	poly1305_update(&ctx, m, bytes);
+	poly1305_finish(&ctx, mac);
 }