summary refs log tree commit diff
path: root/src/liboutils/rmd160.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/liboutils/rmd160.c')
-rw-r--r--src/liboutils/rmd160.c369
1 files changed, 369 insertions, 0 deletions
diff --git a/src/liboutils/rmd160.c b/src/liboutils/rmd160.c
new file mode 100644
index 0000000..afd119e
--- /dev/null
+++ b/src/liboutils/rmd160.c
@@ -0,0 +1,369 @@
+/*
+ * Copyright (c) 2001 Markus Friedl.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+/*
+ * Preneel, Bosselaers, Dobbertin, "The Cryptographic Hash Function RIPEMD-160",
+ * RSA Laboratories, CryptoBytes, Volume 3, Number 2, Autumn 1997,
+ * ftp://ftp.rsasecurity.com/pub/cryptobytes/crypto3n2.pdf
+ */
+#include <sys/types.h>
+#include <endian.h>
+#include <string.h>
+#include <rmd160.h>
+
+#define PUT_64BIT_LE(cp, value) do {                                    \
+	(cp)[7] = (value) >> 56;                                        \
+	(cp)[6] = (value) >> 48;                                        \
+	(cp)[5] = (value) >> 40;                                        \
+	(cp)[4] = (value) >> 32;                                        \
+	(cp)[3] = (value) >> 24;                                        \
+	(cp)[2] = (value) >> 16;                                        \
+	(cp)[1] = (value) >> 8;                                         \
+	(cp)[0] = (value); } while (0)
+
+#define PUT_32BIT_LE(cp, value) do {                                    \
+	(cp)[3] = (value) >> 24;                                        \
+	(cp)[2] = (value) >> 16;                                        \
+	(cp)[1] = (value) >> 8;                                         \
+	(cp)[0] = (value); } while (0)
+
+#define	H0	0x67452301U
+#define	H1	0xEFCDAB89U
+#define	H2	0x98BADCFEU
+#define	H3	0x10325476U
+#define	H4	0xC3D2E1F0U
+
+#define	K0	0x00000000U
+#define	K1	0x5A827999U
+#define	K2	0x6ED9EBA1U
+#define	K3	0x8F1BBCDCU
+#define	K4	0xA953FD4EU
+
+#define	KK0	0x50A28BE6U
+#define	KK1	0x5C4DD124U
+#define	KK2	0x6D703EF3U
+#define	KK3	0x7A6D76E9U
+#define	KK4	0x00000000U
+
+/* rotate x left n bits.  */
+#define ROL(n, x) (((x) << (n)) | ((x) >> (32-(n))))
+
+#define F0(x, y, z) ((x) ^ (y) ^ (z))
+#define F1(x, y, z) (((x) & (y)) | ((~x) & (z)))
+#define F2(x, y, z) (((x) | (~y)) ^ (z))
+#define F3(x, y, z) (((x) & (z)) | ((y) & (~z)))
+#define F4(x, y, z) ((x) ^ ((y) | (~z)))
+
+#define R(a, b, c, d, e, Fj, Kj, sj, rj)                                \
+	do {                                                            \
+		a = ROL(sj, a + Fj(b,c,d) + X(rj) + Kj) + e;            \
+		c = ROL(10, c);                                         \
+	} while(0)
+
+#define X(i)	x[i]
+
+static u_int8_t PADDING[RMD160_BLOCK_LENGTH] = {
+	0x80, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+};
+
+void
+RMD160Init(RMD160_CTX *ctx)
+{
+	ctx->count = 0;
+	ctx->state[0] = H0;
+	ctx->state[1] = H1;
+	ctx->state[2] = H2;
+	ctx->state[3] = H3;
+	ctx->state[4] = H4;
+}
+
+void
+RMD160Update(RMD160_CTX *ctx, const u_int8_t *input, size_t len)
+{
+	size_t have, off, need;
+
+	have = (ctx->count / 8) % RMD160_BLOCK_LENGTH;
+	need = RMD160_BLOCK_LENGTH - have;
+	ctx->count += 8 * len;
+	off = 0;
+
+	if (len >= need) {
+		if (have) {
+			memcpy(ctx->buffer + have, input, need);
+			RMD160Transform(ctx->state, ctx->buffer);
+			off = need;
+			have = 0;
+		}
+		/* now the buffer is empty */
+		while (off + RMD160_BLOCK_LENGTH <= len) {
+			RMD160Transform(ctx->state, input+off);
+			off += RMD160_BLOCK_LENGTH;
+		}
+	}
+	if (off < len)
+		memcpy(ctx->buffer + have, input+off, len-off);
+}
+
+void
+RMD160Pad(RMD160_CTX *ctx)
+{
+	u_int8_t size[8];
+	size_t padlen;
+
+	PUT_64BIT_LE(size, ctx->count);
+
+	/*
+	 * pad to RMD160_BLOCK_LENGTH byte blocks, at least one byte from
+	 * PADDING plus 8 bytes for the size
+	 */
+	padlen = RMD160_BLOCK_LENGTH - ((ctx->count / 8) % RMD160_BLOCK_LENGTH);
+	if (padlen < 1 + 8)
+		padlen += RMD160_BLOCK_LENGTH;
+	RMD160Update(ctx, PADDING, padlen - 8);		/* padlen - 8 <= 64 */
+	RMD160Update(ctx, size, 8);
+}
+
+void
+RMD160Final(u_int8_t digest[RMD160_DIGEST_LENGTH], RMD160_CTX *ctx)
+{
+	int i;
+
+	RMD160Pad(ctx);
+	for (i = 0; i < 5; i++)
+		PUT_32BIT_LE(digest + i*4, ctx->state[i]);
+	memset(ctx, 0, sizeof (*ctx));
+}
+
+void
+RMD160Transform(u_int32_t state[5], const u_int8_t block[RMD160_BLOCK_LENGTH])
+{
+	u_int32_t a, b, c, d, e, aa, bb, cc, dd, ee, t, x[16];
+
+#if BYTE_ORDER == LITTLE_ENDIAN
+	memcpy(x, block, RMD160_BLOCK_LENGTH);
+#else
+	int i;
+
+	for (i = 0; i < 16; i++)
+		x[i] = (u_int32_t)(
+		    (u_int32_t)(block[i*4 + 0]) |
+		    (u_int32_t)(block[i*4 + 1]) <<  8 |
+		    (u_int32_t)(block[i*4 + 2]) << 16 |
+		    (u_int32_t)(block[i*4 + 3]) << 24);
+#endif
+
+	a = state[0];
+	b = state[1];
+	c = state[2];
+	d = state[3];
+	e = state[4];
+
+	/* Round 1 */
+	R(a, b, c, d, e, F0, K0, 11,  0);
+	R(e, a, b, c, d, F0, K0, 14,  1);
+	R(d, e, a, b, c, F0, K0, 15,  2);
+	R(c, d, e, a, b, F0, K0, 12,  3);
+	R(b, c, d, e, a, F0, K0,  5,  4);
+	R(a, b, c, d, e, F0, K0,  8,  5);
+	R(e, a, b, c, d, F0, K0,  7,  6);
+	R(d, e, a, b, c, F0, K0,  9,  7);
+	R(c, d, e, a, b, F0, K0, 11,  8);
+	R(b, c, d, e, a, F0, K0, 13,  9);
+	R(a, b, c, d, e, F0, K0, 14, 10);
+	R(e, a, b, c, d, F0, K0, 15, 11);
+	R(d, e, a, b, c, F0, K0,  6, 12);
+	R(c, d, e, a, b, F0, K0,  7, 13);
+	R(b, c, d, e, a, F0, K0,  9, 14);
+	R(a, b, c, d, e, F0, K0,  8, 15); /* #15 */
+	/* Round 2 */
+	R(e, a, b, c, d, F1, K1,  7,  7);
+	R(d, e, a, b, c, F1, K1,  6,  4);
+	R(c, d, e, a, b, F1, K1,  8, 13);
+	R(b, c, d, e, a, F1, K1, 13,  1);
+	R(a, b, c, d, e, F1, K1, 11, 10);
+	R(e, a, b, c, d, F1, K1,  9,  6);
+	R(d, e, a, b, c, F1, K1,  7, 15);
+	R(c, d, e, a, b, F1, K1, 15,  3);
+	R(b, c, d, e, a, F1, K1,  7, 12);
+	R(a, b, c, d, e, F1, K1, 12,  0);
+	R(e, a, b, c, d, F1, K1, 15,  9);
+	R(d, e, a, b, c, F1, K1,  9,  5);
+	R(c, d, e, a, b, F1, K1, 11,  2);
+	R(b, c, d, e, a, F1, K1,  7, 14);
+	R(a, b, c, d, e, F1, K1, 13, 11);
+	R(e, a, b, c, d, F1, K1, 12,  8); /* #31 */
+	/* Round 3 */
+	R(d, e, a, b, c, F2, K2, 11,  3);
+	R(c, d, e, a, b, F2, K2, 13, 10);
+	R(b, c, d, e, a, F2, K2,  6, 14);
+	R(a, b, c, d, e, F2, K2,  7,  4);
+	R(e, a, b, c, d, F2, K2, 14,  9);
+	R(d, e, a, b, c, F2, K2,  9, 15);
+	R(c, d, e, a, b, F2, K2, 13,  8);
+	R(b, c, d, e, a, F2, K2, 15,  1);
+	R(a, b, c, d, e, F2, K2, 14,  2);
+	R(e, a, b, c, d, F2, K2,  8,  7);
+	R(d, e, a, b, c, F2, K2, 13,  0);
+	R(c, d, e, a, b, F2, K2,  6,  6);
+	R(b, c, d, e, a, F2, K2,  5, 13);
+	R(a, b, c, d, e, F2, K2, 12, 11);
+	R(e, a, b, c, d, F2, K2,  7,  5);
+	R(d, e, a, b, c, F2, K2,  5, 12); /* #47 */
+	/* Round 4 */
+	R(c, d, e, a, b, F3, K3, 11,  1);
+	R(b, c, d, e, a, F3, K3, 12,  9);
+	R(a, b, c, d, e, F3, K3, 14, 11);
+	R(e, a, b, c, d, F3, K3, 15, 10);
+	R(d, e, a, b, c, F3, K3, 14,  0);
+	R(c, d, e, a, b, F3, K3, 15,  8);
+	R(b, c, d, e, a, F3, K3,  9, 12);
+	R(a, b, c, d, e, F3, K3,  8,  4);
+	R(e, a, b, c, d, F3, K3,  9, 13);
+	R(d, e, a, b, c, F3, K3, 14,  3);
+	R(c, d, e, a, b, F3, K3,  5,  7);
+	R(b, c, d, e, a, F3, K3,  6, 15);
+	R(a, b, c, d, e, F3, K3,  8, 14);
+	R(e, a, b, c, d, F3, K3,  6,  5);
+	R(d, e, a, b, c, F3, K3,  5,  6);
+	R(c, d, e, a, b, F3, K3, 12,  2); /* #63 */
+	/* Round 5 */
+	R(b, c, d, e, a, F4, K4,  9,  4);
+	R(a, b, c, d, e, F4, K4, 15,  0);
+	R(e, a, b, c, d, F4, K4,  5,  5);
+	R(d, e, a, b, c, F4, K4, 11,  9);
+	R(c, d, e, a, b, F4, K4,  6,  7);
+	R(b, c, d, e, a, F4, K4,  8, 12);
+	R(a, b, c, d, e, F4, K4, 13,  2);
+	R(e, a, b, c, d, F4, K4, 12, 10);
+	R(d, e, a, b, c, F4, K4,  5, 14);
+	R(c, d, e, a, b, F4, K4, 12,  1);
+	R(b, c, d, e, a, F4, K4, 13,  3);
+	R(a, b, c, d, e, F4, K4, 14,  8);
+	R(e, a, b, c, d, F4, K4, 11, 11);
+	R(d, e, a, b, c, F4, K4,  8,  6);
+	R(c, d, e, a, b, F4, K4,  5, 15);
+	R(b, c, d, e, a, F4, K4,  6, 13); /* #79 */
+
+	aa = a ; bb = b; cc = c; dd = d; ee = e;
+
+	a = state[0];
+	b = state[1];
+	c = state[2];
+	d = state[3];
+	e = state[4];
+
+	/* Parallel round 1 */
+	R(a, b, c, d, e, F4, KK0,  8,  5);
+	R(e, a, b, c, d, F4, KK0,  9, 14);
+	R(d, e, a, b, c, F4, KK0,  9,  7);
+	R(c, d, e, a, b, F4, KK0, 11,  0);
+	R(b, c, d, e, a, F4, KK0, 13,  9);
+	R(a, b, c, d, e, F4, KK0, 15,  2);
+	R(e, a, b, c, d, F4, KK0, 15, 11);
+	R(d, e, a, b, c, F4, KK0,  5,  4);
+	R(c, d, e, a, b, F4, KK0,  7, 13);
+	R(b, c, d, e, a, F4, KK0,  7,  6);
+	R(a, b, c, d, e, F4, KK0,  8, 15);
+	R(e, a, b, c, d, F4, KK0, 11,  8);
+	R(d, e, a, b, c, F4, KK0, 14,  1);
+	R(c, d, e, a, b, F4, KK0, 14, 10);
+	R(b, c, d, e, a, F4, KK0, 12,  3);
+	R(a, b, c, d, e, F4, KK0,  6, 12); /* #15 */
+	/* Parallel round 2 */
+	R(e, a, b, c, d, F3, KK1,  9,  6);
+	R(d, e, a, b, c, F3, KK1, 13, 11);
+	R(c, d, e, a, b, F3, KK1, 15,  3);
+	R(b, c, d, e, a, F3, KK1,  7,  7);
+	R(a, b, c, d, e, F3, KK1, 12,  0);
+	R(e, a, b, c, d, F3, KK1,  8, 13);
+	R(d, e, a, b, c, F3, KK1,  9,  5);
+	R(c, d, e, a, b, F3, KK1, 11, 10);
+	R(b, c, d, e, a, F3, KK1,  7, 14);
+	R(a, b, c, d, e, F3, KK1,  7, 15);
+	R(e, a, b, c, d, F3, KK1, 12,  8);
+	R(d, e, a, b, c, F3, KK1,  7, 12);
+	R(c, d, e, a, b, F3, KK1,  6,  4);
+	R(b, c, d, e, a, F3, KK1, 15,  9);
+	R(a, b, c, d, e, F3, KK1, 13,  1);
+	R(e, a, b, c, d, F3, KK1, 11,  2); /* #31 */
+	/* Parallel round 3 */
+	R(d, e, a, b, c, F2, KK2,  9, 15);
+	R(c, d, e, a, b, F2, KK2,  7,  5);
+	R(b, c, d, e, a, F2, KK2, 15,  1);
+	R(a, b, c, d, e, F2, KK2, 11,  3);
+	R(e, a, b, c, d, F2, KK2,  8,  7);
+	R(d, e, a, b, c, F2, KK2,  6, 14);
+	R(c, d, e, a, b, F2, KK2,  6,  6);
+	R(b, c, d, e, a, F2, KK2, 14,  9);
+	R(a, b, c, d, e, F2, KK2, 12, 11);
+	R(e, a, b, c, d, F2, KK2, 13,  8);
+	R(d, e, a, b, c, F2, KK2,  5, 12);
+	R(c, d, e, a, b, F2, KK2, 14,  2);
+	R(b, c, d, e, a, F2, KK2, 13, 10);
+	R(a, b, c, d, e, F2, KK2, 13,  0);
+	R(e, a, b, c, d, F2, KK2,  7,  4);
+	R(d, e, a, b, c, F2, KK2,  5, 13); /* #47 */
+	/* Parallel round 4 */
+	R(c, d, e, a, b, F1, KK3, 15,  8);
+	R(b, c, d, e, a, F1, KK3,  5,  6);
+	R(a, b, c, d, e, F1, KK3,  8,  4);
+	R(e, a, b, c, d, F1, KK3, 11,  1);
+	R(d, e, a, b, c, F1, KK3, 14,  3);
+	R(c, d, e, a, b, F1, KK3, 14, 11);
+	R(b, c, d, e, a, F1, KK3,  6, 15);
+	R(a, b, c, d, e, F1, KK3, 14,  0);
+	R(e, a, b, c, d, F1, KK3,  6,  5);
+	R(d, e, a, b, c, F1, KK3,  9, 12);
+	R(c, d, e, a, b, F1, KK3, 12,  2);
+	R(b, c, d, e, a, F1, KK3,  9, 13);
+	R(a, b, c, d, e, F1, KK3, 12,  9);
+	R(e, a, b, c, d, F1, KK3,  5,  7);
+	R(d, e, a, b, c, F1, KK3, 15, 10);
+	R(c, d, e, a, b, F1, KK3,  8, 14); /* #63 */
+	/* Parallel round 5 */
+	R(b, c, d, e, a, F0, KK4,  8, 12);
+	R(a, b, c, d, e, F0, KK4,  5, 15);
+	R(e, a, b, c, d, F0, KK4, 12, 10);
+	R(d, e, a, b, c, F0, KK4,  9,  4);
+	R(c, d, e, a, b, F0, KK4, 12,  1);
+	R(b, c, d, e, a, F0, KK4,  5,  5);
+	R(a, b, c, d, e, F0, KK4, 14,  8);
+	R(e, a, b, c, d, F0, KK4,  6,  7);
+	R(d, e, a, b, c, F0, KK4,  8,  6);
+	R(c, d, e, a, b, F0, KK4, 13,  2);
+	R(b, c, d, e, a, F0, KK4,  6, 13);
+	R(a, b, c, d, e, F0, KK4,  5, 14);
+	R(e, a, b, c, d, F0, KK4, 15,  0);
+	R(d, e, a, b, c, F0, KK4, 13,  3);
+	R(c, d, e, a, b, F0, KK4, 11,  9);
+	R(b, c, d, e, a, F0, KK4, 11, 11); /* #79 */
+
+	t =        state[1] + cc + d;
+	state[1] = state[2] + dd + e;
+	state[2] = state[3] + ee + a;
+	state[3] = state[4] + aa + b;
+	state[4] = state[0] + bb + c;
+	state[0] = t;
+}