7 files changed, 1291 insertions, 0 deletions

diff --git a/crypto/scrypt/lib/README b/crypto/scrypt/lib/README
new file mode 100644
index 000000000..3bb211e84
--- /dev/null
+++ b/crypto/scrypt/lib/README
@@ -0,0 +1,6 @@
+The source code under this directory is taken from the client for the
+Tarsnap online backup system (and released under the 2-clause BSD license
+with permission of the author); keeping this code in sync with the Tarsnap
+code is highly desirable and explains why there is some functionality
+included here which is not actually used by the scrypt file encryption
+utility.
diff --git a/crypto/scrypt/lib/crypto/crypto_scrypt-neon-salsa208.h b/crypto/scrypt/lib/crypto/crypto_scrypt-neon-salsa208.h
new file mode 100644
index 000000000..a3b1019a7
--- /dev/null
+++ b/crypto/scrypt/lib/crypto/crypto_scrypt-neon-salsa208.h
@@ -0,0 +1,120 @@
+/*
+ * version 20110505
+ * D. J. Bernstein
+ * Public domain.
+ *
+ * Based on crypto_core/salsa208/armneon/core.c from SUPERCOP 20130419
+ */
+
+#define ROUNDS 8
+static void
+salsa20_8_intrinsic(void * input)
+{
+  int i;
+
+  const uint32x4_t abab = {-1,0,-1,0};
+
+  /*
+   * This is modified since we only have one argument. Usually you'd rearrange
+   * the constant, key, and input bytes, but we just have one linear array to
+   * rearrange which is a bit easier.
+   */
+
+  /*
+   * Change the input to be diagonals as if it's a 4x4 matrix of 32-bit values.
+   */
+  uint32x4_t x0x5x10x15;
+  uint32x4_t x12x1x6x11;
+  uint32x4_t x8x13x2x7;
+  uint32x4_t x4x9x14x3;
+
+  uint32x4_t x0x1x10x11;
+  uint32x4_t x12x13x6x7;
+  uint32x4_t x8x9x2x3;
+  uint32x4_t x4x5x14x15;
+
+  uint32x4_t x0x1x2x3;
+  uint32x4_t x4x5x6x7;
+  uint32x4_t x8x9x10x11;
+  uint32x4_t x12x13x14x15;
+
+  x0x1x2x3 = vld1q_u8((uint8_t *) input);
+  x4x5x6x7 = vld1q_u8(16 + (uint8_t *) input);
+  x8x9x10x11 = vld1q_u8(32 + (uint8_t *) input);
+  x12x13x14x15 = vld1q_u8(48 + (uint8_t *) input);
+
+  x0x1x10x11 = vcombine_u32(vget_low_u32(x0x1x2x3), vget_high_u32(x8x9x10x11));
+  x4x5x14x15 = vcombine_u32(vget_low_u32(x4x5x6x7), vget_high_u32(x12x13x14x15));
+  x8x9x2x3 = vcombine_u32(vget_low_u32(x8x9x10x11), vget_high_u32(x0x1x2x3));
+  x12x13x6x7 = vcombine_u32(vget_low_u32(x12x13x14x15), vget_high_u32(x4x5x6x7));
+
+  x0x5x10x15 = vbslq_u32(abab,x0x1x10x11,x4x5x14x15);
+  x8x13x2x7 = vbslq_u32(abab,x8x9x2x3,x12x13x6x7);
+  x4x9x14x3 = vbslq_u32(abab,x4x5x14x15,x8x9x2x3);
+  x12x1x6x11 = vbslq_u32(abab,x12x13x6x7,x0x1x10x11);
+
+  uint32x4_t start0 = x0x5x10x15;
+  uint32x4_t start1 = x12x1x6x11;
+  uint32x4_t start3 = x4x9x14x3;
+  uint32x4_t start2 = x8x13x2x7;
+
+  /* From here on this should be the same as the SUPERCOP version. */
+
+  uint32x4_t diag0 = start0;
+  uint32x4_t diag1 = start1;
+  uint32x4_t diag2 = start2;
+  uint32x4_t diag3 = start3;
+
+  uint32x4_t a0;
+  uint32x4_t a1;
+  uint32x4_t a2;
+  uint32x4_t a3;
+
+  for (i = ROUNDS;i > 0;i -= 2) {
+    a0 = diag1 + diag0;
+    diag3 ^= vsriq_n_u32(vshlq_n_u32(a0,7),a0,25);
+    a1 = diag0 + diag3;
+    diag2 ^= vsriq_n_u32(vshlq_n_u32(a1,9),a1,23);
+    a2 = diag3 + diag2;
+    diag1 ^= vsriq_n_u32(vshlq_n_u32(a2,13),a2,19);
+    a3 = diag2 + diag1;
+    diag0 ^= vsriq_n_u32(vshlq_n_u32(a3,18),a3,14);
+
+    diag3 = vextq_u32(diag3,diag3,3);
+    diag2 = vextq_u32(diag2,diag2,2);
+    diag1 = vextq_u32(diag1,diag1,1);
+
+    a0 = diag3 + diag0;
+    diag1 ^= vsriq_n_u32(vshlq_n_u32(a0,7),a0,25);
+    a1 = diag0 + diag1;
+    diag2 ^= vsriq_n_u32(vshlq_n_u32(a1,9),a1,23);
+    a2 = diag1 + diag2;
+    diag3 ^= vsriq_n_u32(vshlq_n_u32(a2,13),a2,19);
+    a3 = diag2 + diag3;
+    diag0 ^= vsriq_n_u32(vshlq_n_u32(a3,18),a3,14);
+
+    diag1 = vextq_u32(diag1,diag1,3);
+    diag2 = vextq_u32(diag2,diag2,2);
+    diag3 = vextq_u32(diag3,diag3,1);
+  }
+
+  x0x5x10x15 = diag0 + start0;
+  x12x1x6x11 = diag1 + start1;
+  x8x13x2x7 = diag2 + start2;
+  x4x9x14x3 = diag3 + start3;
+
+  x0x1x10x11 = vbslq_u32(abab,x0x5x10x15,x12x1x6x11);
+  x12x13x6x7 = vbslq_u32(abab,x12x1x6x11,x8x13x2x7);
+  x8x9x2x3 = vbslq_u32(abab,x8x13x2x7,x4x9x14x3);
+  x4x5x14x15 = vbslq_u32(abab,x4x9x14x3,x0x5x10x15);
+
+  x0x1x2x3 = vcombine_u32(vget_low_u32(x0x1x10x11),vget_high_u32(x8x9x2x3));
+  x4x5x6x7 = vcombine_u32(vget_low_u32(x4x5x14x15),vget_high_u32(x12x13x6x7));
+  x8x9x10x11 = vcombine_u32(vget_low_u32(x8x9x2x3),vget_high_u32(x0x1x10x11));
+  x12x13x14x15 = vcombine_u32(vget_low_u32(x12x13x6x7),vget_high_u32(x4x5x14x15));
+
+  vst1q_u8((uint8_t *) input,(uint8x16_t) x0x1x2x3);
+  vst1q_u8(16 + (uint8_t *) input,(uint8x16_t) x4x5x6x7);
+  vst1q_u8(32 + (uint8_t *) input,(uint8x16_t) x8x9x10x11);
+  vst1q_u8(48 + (uint8_t *) input,(uint8x16_t) x12x13x14x15);
+}
diff --git a/crypto/scrypt/lib/crypto/crypto_scrypt-neon.c b/crypto/scrypt/lib/crypto/crypto_scrypt-neon.c
new file mode 100644
index 000000000..a3bf052b4
--- /dev/null
+++ b/crypto/scrypt/lib/crypto/crypto_scrypt-neon.c
@@ -0,0 +1,305 @@
+/*-
+ * Copyright 2009 Colin Percival
+ * 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 AND CONTRIBUTORS ``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 OR CONTRIBUTORS 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.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+#include "scrypt_platform.h"
+
+#include <machine/cpu-features.h>
+#include <arm_neon.h>
+
+#include <errno.h>
+#include <stdint.h>
+#include <limits.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef USE_OPENSSL_PBKDF2
+#include <openssl/evp.h>
+#else
+#include "sha256.h"
+#endif
+#include "sysendian.h"
+
+#include "crypto_scrypt.h"
+
+#include "crypto_scrypt-neon-salsa208.h"
+
+static void blkcpy(void *, void *, size_t);
+static void blkxor(void *, void *, size_t);
+void crypto_core_salsa208_armneon2(void *);
+static void blockmix_salsa8(uint8x16_t *, uint8x16_t *, uint8x16_t *, size_t);
+static uint64_t integerify(void *, size_t);
+static void smix(uint8_t *, size_t, uint64_t, void *, void *);
+
+static void
+blkcpy(void * dest, void * src, size_t len)
+{
+	uint8x16_t * D = dest;
+	uint8x16_t * S = src;
+	size_t L = len / 16;
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] = S[i];
+}
+
+static void
+blkxor(void * dest, void * src, size_t len)
+{
+	uint8x16_t * D = dest;
+	uint8x16_t * S = src;
+	size_t L = len / 16;
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] = veorq_u8(D[i], S[i]);
+}
+
+/**
+ * blockmix_salsa8(B, Y, r):
+ * Compute B = BlockMix_{salsa20/8, r}(B).  The input B must be 128r bytes in
+ * length; the temporary space Y must also be the same size.
+ */
+static void
+blockmix_salsa8(uint8x16_t * Bin, uint8x16_t * Bout, uint8x16_t * X, size_t r)
+{
+	size_t i;
+
+	/* 1: X <-- B_{2r - 1} */
+	blkcpy(X, &Bin[8 * r - 4], 64);
+
+	/* 2: for i = 0 to 2r - 1 do */
+	for (i = 0; i < r; i++) {
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 8], 64);
+                salsa20_8_intrinsic((void *) X);
+
+		/* 4: Y_i <-- X */
+		/* 6: B' <-- (Y_0, Y_2 ... Y_{2r-2}, Y_1, Y_3 ... Y_{2r-1}) */
+		blkcpy(&Bout[i * 4], X, 64);
+
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 8 + 4], 64);
+                salsa20_8_intrinsic((void *) X);
+
+		/* 4: Y_i <-- X */
+		/* 6: B' <-- (Y_0, Y_2 ... Y_{2r-2}, Y_1, Y_3 ... Y_{2r-1}) */
+		blkcpy(&Bout[(r + i) * 4], X, 64);
+	}
+}
+
+/**
+ * integerify(B, r):
+ * Return the result of parsing B_{2r-1} as a little-endian integer.
+ */
+static uint64_t
+integerify(void * B, size_t r)
+{
+	uint8_t * X = (void*)((uintptr_t)(B) + (2 * r - 1) * 64);
+
+	return (le64dec(X));
+}
+
+/**
+ * smix(B, r, N, V, XY):
+ * Compute B = SMix_r(B, N).  The input B must be 128r bytes in length; the
+ * temporary storage V must be 128rN bytes in length; the temporary storage
+ * XY must be 256r bytes in length.  The value N must be a power of 2.
+ */
+static void
+smix(uint8_t * B, size_t r, uint64_t N, void * V, void * XY)
+{
+	uint8x16_t * X = XY;
+	uint8x16_t * Y = (void *)((uintptr_t)(XY) + 128 * r);
+        uint8x16_t * Z = (void *)((uintptr_t)(XY) + 256 * r);
+        uint32_t * X32 = (void *)X;
+	uint64_t i, j;
+        size_t k;
+
+	/* 1: X <-- B */
+	blkcpy(X, B, 128 * r);
+
+	/* 2: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i += 2) {
+		/* 3: V_i <-- X */
+		blkcpy((void *)((uintptr_t)(V) + i * 128 * r), X, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(X, Y, Z, r);
+
+		/* 3: V_i <-- X */
+		blkcpy((void *)((uintptr_t)(V) + (i + 1) * 128 * r),
+		    Y, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(Y, X, Z, r);
+	}
+
+	/* 6: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i += 2) {
+		/* 7: j <-- Integerify(X) mod N */
+		j = integerify(X, r) & (N - 1);
+
+		/* 8: X <-- H(X \xor V_j) */
+		blkxor(X, (void *)((uintptr_t)(V) + j * 128 * r), 128 * r);
+		blockmix_salsa8(X, Y, Z, r);
+
+		/* 7: j <-- Integerify(X) mod N */
+		j = integerify(Y, r) & (N - 1);
+
+		/* 8: X <-- H(X \xor V_j) */
+		blkxor(Y, (void *)((uintptr_t)(V) + j * 128 * r), 128 * r);
+		blockmix_salsa8(Y, X, Z, r);
+	}
+
+	/* 10: B' <-- X */
+	blkcpy(B, X, 128 * r);
+}
+
+/**
+ * crypto_scrypt(passwd, passwdlen, salt, saltlen, N, r, p, buf, buflen):
+ * Compute scrypt(passwd[0 .. passwdlen - 1], salt[0 .. saltlen - 1], N, r,
+ * p, buflen) and write the result into buf.  The parameters r, p, and buflen
+ * must satisfy r * p < 2^30 and buflen <= (2^32 - 1) * 32.  The parameter N
+ * must be a power of 2.
+ *
+ * Return 0 on success; or -1 on error.
+ */
+int
+crypto_scrypt(const uint8_t * passwd, size_t passwdlen,
+    const uint8_t * salt, size_t saltlen, uint64_t N, uint32_t r, uint32_t p,
+    uint8_t * buf, size_t buflen)
+{
+	void * B0, * V0, * XY0;
+	uint8_t * B;
+	uint32_t * V;
+	uint32_t * XY;
+	uint32_t i;
+
+	/* Sanity-check parameters. */
+#if SIZE_MAX > UINT32_MAX
+	if (buflen > (((uint64_t)(1) << 32) - 1) * 32) {
+		errno = EFBIG;
+		goto err0;
+	}
+#endif
+	if ((uint64_t)(r) * (uint64_t)(p) >= (1 << 30)) {
+		errno = EFBIG;
+		goto err0;
+	}
+	if (((N & (N - 1)) != 0) || (N == 0)) {
+		errno = EINVAL;
+		goto err0;
+	}
+	if ((r > SIZE_MAX / 128 / p) ||
+#if SIZE_MAX / 256 <= UINT32_MAX
+	    (r > SIZE_MAX / 256) ||
+#endif
+	    (N > SIZE_MAX / 128 / r)) {
+		errno = ENOMEM;
+		goto err0;
+	}
+
+	/* Allocate memory. */
+#ifdef HAVE_POSIX_MEMALIGN
+	if ((errno = posix_memalign(&B0, 64, 128 * r * p)) != 0)
+		goto err0;
+	B = (uint8_t *)(B0);
+	if ((errno = posix_memalign(&XY0, 64, 256 * r + 64)) != 0)
+		goto err1;
+	XY = (uint32_t *)(XY0);
+#ifndef MAP_ANON
+	if ((errno = posix_memalign(&V0, 64, 128 * r * N)) != 0)
+		goto err2;
+	V = (uint32_t *)(V0);
+#endif
+#else
+	if ((B0 = malloc(128 * r * p + 63)) == NULL)
+		goto err0;
+	B = (uint8_t *)(((uintptr_t)(B0) + 63) & ~ (uintptr_t)(63));
+	if ((XY0 = malloc(256 * r + 64 + 63)) == NULL)
+		goto err1;
+	XY = (uint32_t *)(((uintptr_t)(XY0) + 63) & ~ (uintptr_t)(63));
+#ifndef MAP_ANON
+	if ((V0 = malloc(128 * r * N + 63)) == NULL)
+		goto err2;
+	V = (uint32_t *)(((uintptr_t)(V0) + 63) & ~ (uintptr_t)(63));
+#endif
+#endif
+#ifdef MAP_ANON
+	if ((V0 = mmap(NULL, 128 * r * N, PROT_READ | PROT_WRITE,
+#ifdef MAP_NOCORE
+	    MAP_ANON | MAP_PRIVATE | MAP_NOCORE,
+#else
+	    MAP_ANON | MAP_PRIVATE,
+#endif
+	    -1, 0)) == MAP_FAILED)
+		goto err2;
+	V = (uint32_t *)(V0);
+#endif
+
+	/* 1: (B_0 ... B_{p-1}) <-- PBKDF2(P, S, 1, p * MFLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, salt, saltlen, 1, EVP_sha256(), p * 128 * r, B);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, salt, saltlen, 1, B, p * 128 * r);
+#endif
+
+	/* 2: for i = 0 to p - 1 do */
+	for (i = 0; i < p; i++) {
+		/* 3: B_i <-- MF(B_i, N) */
+		smix(&B[i * 128 * r], r, N, V, XY);
+	}
+
+	/* 5: DK <-- PBKDF2(P, B, 1, dkLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, B, p * 128 * r, 1, EVP_sha256(), buflen, buf);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, B, p * 128 * r, 1, buf, buflen);
+#endif
+
+	/* Free memory. */
+#ifdef MAP_ANON
+	if (munmap(V0, 128 * r * N))
+		goto err2;
+#else
+	free(V0);
+#endif
+	free(XY0);
+	free(B0);
+
+	/* Success! */
+	return (0);
+
+err2:
+	free(XY0);
+err1:
+	free(B0);
+err0:
+	/* Failure! */
+	return (-1);
+}
diff --git a/crypto/scrypt/lib/crypto/crypto_scrypt-ref.c b/crypto/scrypt/lib/crypto/crypto_scrypt-ref.c
new file mode 100644
index 000000000..abe23eaa5
--- /dev/null
+++ b/crypto/scrypt/lib/crypto/crypto_scrypt-ref.c
@@ -0,0 +1,296 @@
+/*-
+ * Copyright 2009 Colin Percival
+ * 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 AND CONTRIBUTORS ``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 OR CONTRIBUTORS 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.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+#include "scrypt_platform.h"
+
+#include <errno.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef USE_OPENSSL_PBKDF2
+#include <openssl/evp.h>
+#else
+#include "sha256.h"
+#endif
+#include "sysendian.h"
+
+#include "crypto_scrypt.h"
+
+static void blkcpy(uint8_t *, uint8_t *, size_t);
+static void blkxor(uint8_t *, uint8_t *, size_t);
+static void salsa20_8(uint8_t[64]);
+static void blockmix_salsa8(uint8_t *, uint8_t *, size_t);
+static uint64_t integerify(uint8_t *, size_t);
+static void smix(uint8_t *, size_t, uint64_t, uint8_t *, uint8_t *);
+
+static void
+blkcpy(uint8_t * dest, uint8_t * src, size_t len)
+{
+	size_t i;
+
+	for (i = 0; i < len; i++)
+		dest[i] = src[i];
+}
+
+static void
+blkxor(uint8_t * dest, uint8_t * src, size_t len)
+{
+	size_t i;
+
+	for (i = 0; i < len; i++)
+		dest[i] ^= src[i];
+}
+
+/**
+ * salsa20_8(B):
+ * Apply the salsa20/8 core to the provided block.
+ */
+static void
+salsa20_8(uint8_t B[64])
+{
+	uint32_t B32[16];
+	uint32_t x[16];
+	size_t i;
+
+	/* Convert little-endian values in. */
+	for (i = 0; i < 16; i++)
+		B32[i] = le32dec(&B[i * 4]);
+
+	/* Compute x = doubleround^4(B32). */
+	for (i = 0; i < 16; i++)
+		x[i] = B32[i];
+	for (i = 0; i < 8; i += 2) {
+#define R(a,b) (((a) << (b)) | ((a) >> (32 - (b))))
+		/* Operate on columns. */
+		x[ 4] ^= R(x[ 0]+x[12], 7);  x[ 8] ^= R(x[ 4]+x[ 0], 9);
+		x[12] ^= R(x[ 8]+x[ 4],13);  x[ 0] ^= R(x[12]+x[ 8],18);
+
+		x[ 9] ^= R(x[ 5]+x[ 1], 7);  x[13] ^= R(x[ 9]+x[ 5], 9);
+		x[ 1] ^= R(x[13]+x[ 9],13);  x[ 5] ^= R(x[ 1]+x[13],18);
+
+		x[14] ^= R(x[10]+x[ 6], 7);  x[ 2] ^= R(x[14]+x[10], 9);
+		x[ 6] ^= R(x[ 2]+x[14],13);  x[10] ^= R(x[ 6]+x[ 2],18);
+
+		x[ 3] ^= R(x[15]+x[11], 7);  x[ 7] ^= R(x[ 3]+x[15], 9);
+		x[11] ^= R(x[ 7]+x[ 3],13);  x[15] ^= R(x[11]+x[ 7],18);
+
+		/* Operate on rows. */
+		x[ 1] ^= R(x[ 0]+x[ 3], 7);  x[ 2] ^= R(x[ 1]+x[ 0], 9);
+		x[ 3] ^= R(x[ 2]+x[ 1],13);  x[ 0] ^= R(x[ 3]+x[ 2],18);
+
+		x[ 6] ^= R(x[ 5]+x[ 4], 7);  x[ 7] ^= R(x[ 6]+x[ 5], 9);
+		x[ 4] ^= R(x[ 7]+x[ 6],13);  x[ 5] ^= R(x[ 4]+x[ 7],18);
+
+		x[11] ^= R(x[10]+x[ 9], 7);  x[ 8] ^= R(x[11]+x[10], 9);
+		x[ 9] ^= R(x[ 8]+x[11],13);  x[10] ^= R(x[ 9]+x[ 8],18);
+
+		x[12] ^= R(x[15]+x[14], 7);  x[13] ^= R(x[12]+x[15], 9);
+		x[14] ^= R(x[13]+x[12],13);  x[15] ^= R(x[14]+x[13],18);
+#undef R
+	}
+
+	/* Compute B32 = B32 + x. */
+	for (i = 0; i < 16; i++)
+		B32[i] += x[i];
+
+	/* Convert little-endian values out. */
+	for (i = 0; i < 16; i++)
+		le32enc(&B[4 * i], B32[i]);
+}
+
+/**
+ * blockmix_salsa8(B, Y, r):
+ * Compute B = BlockMix_{salsa20/8, r}(B).  The input B must be 128r bytes in
+ * length; the temporary space Y must also be the same size.
+ */
+static void
+blockmix_salsa8(uint8_t * B, uint8_t * Y, size_t r)
+{
+	uint8_t X[64];
+	size_t i;
+
+	/* 1: X <-- B_{2r - 1} */
+	blkcpy(X, &B[(2 * r - 1) * 64], 64);
+
+	/* 2: for i = 0 to 2r - 1 do */
+	for (i = 0; i < 2 * r; i++) {
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &B[i * 64], 64);
+		salsa20_8(X);
+
+		/* 4: Y_i <-- X */
+		blkcpy(&Y[i * 64], X, 64);
+	}
+
+	/* 6: B' <-- (Y_0, Y_2 ... Y_{2r-2}, Y_1, Y_3 ... Y_{2r-1}) */
+	for (i = 0; i < r; i++)
+		blkcpy(&B[i * 64], &Y[(i * 2) * 64], 64);
+	for (i = 0; i < r; i++)
+		blkcpy(&B[(i + r) * 64], &Y[(i * 2 + 1) * 64], 64);
+}
+
+/**
+ * integerify(B, r):
+ * Return the result of parsing B_{2r-1} as a little-endian integer.
+ */
+static uint64_t
+integerify(uint8_t * B, size_t r)
+{
+	uint8_t * X = &B[(2 * r - 1) * 64];
+
+	return (le64dec(X));
+}
+
+/**
+ * smix(B, r, N, V, XY):
+ * Compute B = SMix_r(B, N).  The input B must be 128r bytes in length; the
+ * temporary storage V must be 128rN bytes in length; the temporary storage
+ * XY must be 256r bytes in length.  The value N must be a power of 2.
+ */
+static void
+smix(uint8_t * B, size_t r, uint64_t N, uint8_t * V, uint8_t * XY)
+{
+	uint8_t * X = XY;
+	uint8_t * Y = &XY[128 * r];
+	uint64_t i;
+	uint64_t j;
+
+	/* 1: X <-- B */
+	blkcpy(X, B, 128 * r);
+
+	/* 2: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i++) {
+		/* 3: V_i <-- X */
+		blkcpy(&V[i * (128 * r)], X, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(X, Y, r);
+	}
+
+	/* 6: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i++) {
+		/* 7: j <-- Integerify(X) mod N */
+		j = integerify(X, r) & (N - 1);
+
+		/* 8: X <-- H(X \xor V_j) */
+		blkxor(X, &V[j * (128 * r)], 128 * r);
+		blockmix_salsa8(X, Y, r);
+	}
+
+	/* 10: B' <-- X */
+	blkcpy(B, X, 128 * r);
+}
+
+/**
+ * crypto_scrypt(passwd, passwdlen, salt, saltlen, N, r, p, buf, buflen):
+ * Compute scrypt(passwd[0 .. passwdlen - 1], salt[0 .. saltlen - 1], N, r,
+ * p, buflen) and write the result into buf.  The parameters r, p, and buflen
+ * must satisfy r * p < 2^30 and buflen <= (2^32 - 1) * 32.  The parameter N
+ * must be a power of 2.
+ *
+ * Return 0 on success; or -1 on error.
+ */
+int
+crypto_scrypt(const uint8_t * passwd, size_t passwdlen,
+    const uint8_t * salt, size_t saltlen, uint64_t N, uint32_t r, uint32_t p,
+    uint8_t * buf, size_t buflen)
+{
+	uint8_t * B;
+	uint8_t * V;
+	uint8_t * XY;
+	uint32_t i;
+
+	/* Sanity-check parameters. */
+#if SIZE_MAX > UINT32_MAX
+	if (buflen > (((uint64_t)(1) << 32) - 1) * 32) {
+		errno = EFBIG;
+		goto err0;
+	}
+#endif
+	if ((uint64_t)(r) * (uint64_t)(p) >= (1 << 30)) {
+		errno = EFBIG;
+		goto err0;
+	}
+	if (((N & (N - 1)) != 0) || (N == 0)) {
+		errno = EINVAL;
+		goto err0;
+	}
+	if ((r > SIZE_MAX / 128 / p) ||
+#if SIZE_MAX / 256 <= UINT32_MAX
+	    (r > SIZE_MAX / 256) ||
+#endif
+	    (N > SIZE_MAX / 128 / r)) {
+		errno = ENOMEM;
+		goto err0;
+	}
+
+	/* Allocate memory. */
+	if ((B = malloc(128 * r * p)) == NULL)
+		goto err0;
+	if ((XY = malloc(256 * r)) == NULL)
+		goto err1;
+	if ((V = malloc(128 * r * N)) == NULL)
+		goto err2;
+
+	/* 1: (B_0 ... B_{p-1}) <-- PBKDF2(P, S, 1, p * MFLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, salt, saltlen, 1, EVP_sha256(), p * 128 * r, B);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, salt, saltlen, 1, B, p * 128 * r);
+#endif
+
+	/* 2: for i = 0 to p - 1 do */
+	for (i = 0; i < p; i++) {
+		/* 3: B_i <-- MF(B_i, N) */
+		smix(&B[i * 128 * r], r, N, V, XY);
+	}
+
+	/* 5: DK <-- PBKDF2(P, B, 1, dkLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, B, p * 128 * r, 1, EVP_sha256(), buflen, buf);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, B, p * 128 * r, 1, buf, buflen);
+#endif
+
+	/* Free memory. */
+	free(V);
+	free(XY);
+	free(B);
+
+	/* Success! */
+	return (0);
+
+err2:
+	free(XY);
+err1:
+	free(B);
+err0:
+	/* Failure! */
+	return (-1);
+}
diff --git a/crypto/scrypt/lib/crypto/crypto_scrypt-sse.c b/crypto/scrypt/lib/crypto/crypto_scrypt-sse.c
new file mode 100644
index 000000000..dd18f291c
--- /dev/null
+++ b/crypto/scrypt/lib/crypto/crypto_scrypt-sse.c
@@ -0,0 +1,378 @@
+/*-
+ * Copyright 2009 Colin Percival
+ * 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 AND CONTRIBUTORS ``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 OR CONTRIBUTORS 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.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+#include "scrypt_platform.h"
+
+#include <sys/types.h>
+#include <sys/mman.h>
+
+#include <emmintrin.h>
+#include <errno.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef USE_OPENSSL_PBKDF2
+#include <openssl/evp.h>
+#else
+#include "sha256.h"
+#endif
+#include "sysendian.h"
+
+#include "crypto_scrypt.h"
+
+static void blkcpy(void *, void *, size_t);
+static void blkxor(void *, void *, size_t);
+static void salsa20_8(__m128i *);
+static void blockmix_salsa8(__m128i *, __m128i *, __m128i *, size_t);
+static uint64_t integerify(void *, size_t);
+static void smix(uint8_t *, size_t, uint64_t, void *, void *);
+
+static void
+blkcpy(void * dest, void * src, size_t len)
+{
+	__m128i * D = dest;
+	__m128i * S = src;
+	size_t L = len / 16;
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] = S[i];
+}
+
+static void
+blkxor(void * dest, void * src, size_t len)
+{
+	__m128i * D = dest;
+	__m128i * S = src;
+	size_t L = len / 16;
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] = _mm_xor_si128(D[i], S[i]);
+}
+
+/**
+ * salsa20_8(B):
+ * Apply the salsa20/8 core to the provided block.
+ */
+static void
+salsa20_8(__m128i B[4])
+{
+	__m128i X0, X1, X2, X3;
+	__m128i T;
+	size_t i;
+
+	X0 = B[0];
+	X1 = B[1];
+	X2 = B[2];
+	X3 = B[3];
+
+	for (i = 0; i < 8; i += 2) {
+		/* Operate on "columns". */
+		T = _mm_add_epi32(X0, X3);
+		X1 = _mm_xor_si128(X1, _mm_slli_epi32(T, 7));
+		X1 = _mm_xor_si128(X1, _mm_srli_epi32(T, 25));
+		T = _mm_add_epi32(X1, X0);
+		X2 = _mm_xor_si128(X2, _mm_slli_epi32(T, 9));
+		X2 = _mm_xor_si128(X2, _mm_srli_epi32(T, 23));
+		T = _mm_add_epi32(X2, X1);
+		X3 = _mm_xor_si128(X3, _mm_slli_epi32(T, 13));
+		X3 = _mm_xor_si128(X3, _mm_srli_epi32(T, 19));
+		T = _mm_add_epi32(X3, X2);
+		X0 = _mm_xor_si128(X0, _mm_slli_epi32(T, 18));
+		X0 = _mm_xor_si128(X0, _mm_srli_epi32(T, 14));
+
+		/* Rearrange data. */
+		X1 = _mm_shuffle_epi32(X1, 0x93);
+		X2 = _mm_shuffle_epi32(X2, 0x4E);
+		X3 = _mm_shuffle_epi32(X3, 0x39);
+
+		/* Operate on "rows". */
+		T = _mm_add_epi32(X0, X1);
+		X3 = _mm_xor_si128(X3, _mm_slli_epi32(T, 7));
+		X3 = _mm_xor_si128(X3, _mm_srli_epi32(T, 25));
+		T = _mm_add_epi32(X3, X0);
+		X2 = _mm_xor_si128(X2, _mm_slli_epi32(T, 9));
+		X2 = _mm_xor_si128(X2, _mm_srli_epi32(T, 23));
+		T = _mm_add_epi32(X2, X3);
+		X1 = _mm_xor_si128(X1, _mm_slli_epi32(T, 13));
+		X1 = _mm_xor_si128(X1, _mm_srli_epi32(T, 19));
+		T = _mm_add_epi32(X1, X2);
+		X0 = _mm_xor_si128(X0, _mm_slli_epi32(T, 18));
+		X0 = _mm_xor_si128(X0, _mm_srli_epi32(T, 14));
+
+		/* Rearrange data. */
+		X1 = _mm_shuffle_epi32(X1, 0x39);
+		X2 = _mm_shuffle_epi32(X2, 0x4E);
+		X3 = _mm_shuffle_epi32(X3, 0x93);
+	}
+
+	B[0] = _mm_add_epi32(B[0], X0);
+	B[1] = _mm_add_epi32(B[1], X1);
+	B[2] = _mm_add_epi32(B[2], X2);
+	B[3] = _mm_add_epi32(B[3], X3);
+}
+
+/**
+ * blockmix_salsa8(Bin, Bout, X, r):
+ * Compute Bout = BlockMix_{salsa20/8, r}(Bin).  The input Bin must be 128r
+ * bytes in length; the output Bout must also be the same size.  The
+ * temporary space X must be 64 bytes.
+ */
+static void
+blockmix_salsa8(__m128i * Bin, __m128i * Bout, __m128i * X, size_t r)
+{
+	size_t i;
+
+	/* 1: X <-- B_{2r - 1} */
+	blkcpy(X, &Bin[8 * r - 4], 64);
+
+	/* 2: for i = 0 to 2r - 1 do */
+	for (i = 0; i < r; i++) {
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 8], 64);
+		salsa20_8(X);
+
+		/* 4: Y_i <-- X */
+		/* 6: B' <-- (Y_0, Y_2 ... Y_{2r-2}, Y_1, Y_3 ... Y_{2r-1}) */
+		blkcpy(&Bout[i * 4], X, 64);
+
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 8 + 4], 64);
+		salsa20_8(X);
+
+		/* 4: Y_i <-- X */
+		/* 6: B' <-- (Y_0, Y_2 ... Y_{2r-2}, Y_1, Y_3 ... Y_{2r-1}) */
+		blkcpy(&Bout[(r + i) * 4], X, 64);
+	}
+}
+
+/**
+ * integerify(B, r):
+ * Return the result of parsing B_{2r-1} as a little-endian integer.
+ */
+static uint64_t
+integerify(void * B, size_t r)
+{
+	uint32_t * X = (void *)((uintptr_t)(B) + (2 * r - 1) * 64);
+
+	return (((uint64_t)(X[13]) << 32) + X[0]);
+}
+
+/**
+ * smix(B, r, N, V, XY):
+ * Compute B = SMix_r(B, N).  The input B must be 128r bytes in length;
+ * the temporary storage V must be 128rN bytes in length; the temporary
+ * storage XY must be 256r + 64 bytes in length.  The value N must be a
+ * power of 2 greater than 1.  The arrays B, V, and XY must be aligned to a
+ * multiple of 64 bytes.
+ */
+static void
+smix(uint8_t * B, size_t r, uint64_t N, void * V, void * XY)
+{
+	__m128i * X = XY;
+	__m128i * Y = (void *)((uintptr_t)(XY) + 128 * r);
+	__m128i * Z = (void *)((uintptr_t)(XY) + 256 * r);
+	uint32_t * X32 = (void *)X;
+	uint64_t i, j;
+	size_t k;
+
+	/* 1: X <-- B */
+	for (k = 0; k < 2 * r; k++) {
+		for (i = 0; i < 16; i++) {
+			X32[k * 16 + i] =
+			    le32dec(&B[(k * 16 + (i * 5 % 16)) * 4]);
+		}
+	}
+
+	/* 2: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i += 2) {
+		/* 3: V_i <-- X */
+		blkcpy((void *)((uintptr_t)(V) + i * 128 * r), X, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(X, Y, Z, r);
+
+		/* 3: V_i <-- X */
+		blkcpy((void *)((uintptr_t)(V) + (i + 1) * 128 * r),
+		    Y, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(Y, X, Z, r);
+	}
+
+	/* 6: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i += 2) {
+		/* 7: j <-- Integerify(X) mod N */
+		j = integerify(X, r) & (N - 1);
+
+		/* 8: X <-- H(X \xor V_j) */
+		blkxor(X, (void *)((uintptr_t)(V) + j * 128 * r), 128 * r);
+		blockmix_salsa8(X, Y, Z, r);
+
+		/* 7: j <-- Integerify(X) mod N */
+		j = integerify(Y, r) & (N - 1);
+
+		/* 8: X <-- H(X \xor V_j) */
+		blkxor(Y, (void *)((uintptr_t)(V) + j * 128 * r), 128 * r);
+		blockmix_salsa8(Y, X, Z, r);
+	}
+
+	/* 10: B' <-- X */
+	for (k = 0; k < 2 * r; k++) {
+		for (i = 0; i < 16; i++) {
+			le32enc(&B[(k * 16 + (i * 5 % 16)) * 4],
+			    X32[k * 16 + i]);
+		}
+	}
+}
+
+/**
+ * crypto_scrypt(passwd, passwdlen, salt, saltlen, N, r, p, buf, buflen):
+ * Compute scrypt(passwd[0 .. passwdlen - 1], salt[0 .. saltlen - 1], N, r,
+ * p, buflen) and write the result into buf.  The parameters r, p, and buflen
+ * must satisfy r * p < 2^30 and buflen <= (2^32 - 1) * 32.  The parameter N
+ * must be a power of 2 greater than 1.
+ *
+ * Return 0 on success; or -1 on error.
+ */
+int
+crypto_scrypt(const uint8_t * passwd, size_t passwdlen,
+    const uint8_t * salt, size_t saltlen, uint64_t N, uint32_t r, uint32_t p,
+    uint8_t * buf, size_t buflen)
+{
+	void * B0, * V0, * XY0;
+	uint8_t * B;
+	uint32_t * V;
+	uint32_t * XY;
+	uint32_t i;
+
+	/* Sanity-check parameters. */
+#if SIZE_MAX > UINT32_MAX
+	if (buflen > (((uint64_t)(1) << 32) - 1) * 32) {
+		errno = EFBIG;
+		goto err0;
+	}
+#endif
+	if ((uint64_t)(r) * (uint64_t)(p) >= (1 << 30)) {
+		errno = EFBIG;
+		goto err0;
+	}
+	if (((N & (N - 1)) != 0) || (N == 0)) {
+		errno = EINVAL;
+		goto err0;
+	}
+	if ((r > SIZE_MAX / 128 / p) ||
+#if SIZE_MAX / 256 <= UINT32_MAX
+	    (r > (SIZE_MAX - 64) / 256) ||
+#endif
+	    (N > SIZE_MAX / 128 / r)) {
+		errno = ENOMEM;
+		goto err0;
+	}
+
+	/* Allocate memory. */
+#ifdef HAVE_POSIX_MEMALIGN
+	if ((errno = posix_memalign(&B0, 64, 128 * r * p)) != 0)
+		goto err0;
+	B = (uint8_t *)(B0);
+	if ((errno = posix_memalign(&XY0, 64, 256 * r + 64)) != 0)
+		goto err1;
+	XY = (uint32_t *)(XY0);
+#ifndef MAP_ANON
+	if ((errno = posix_memalign(&V0, 64, 128 * r * N)) != 0)
+		goto err2;
+	V = (uint32_t *)(V0);
+#endif
+#else
+	if ((B0 = malloc(128 * r * p + 63)) == NULL)
+		goto err0;
+	B = (uint8_t *)(((uintptr_t)(B0) + 63) & ~ (uintptr_t)(63));
+	if ((XY0 = malloc(256 * r + 64 + 63)) == NULL)
+		goto err1;
+	XY = (uint32_t *)(((uintptr_t)(XY0) + 63) & ~ (uintptr_t)(63));
+#ifndef MAP_ANON
+	if ((V0 = malloc(128 * r * N + 63)) == NULL)
+		goto err2;
+	V = (uint32_t *)(((uintptr_t)(V0) + 63) & ~ (uintptr_t)(63));
+#endif
+#endif
+#ifdef MAP_ANON
+	if ((V0 = mmap(NULL, 128 * r * N, PROT_READ | PROT_WRITE,
+#ifdef MAP_NOCORE
+	    MAP_ANON | MAP_PRIVATE | MAP_NOCORE,
+#else
+	    MAP_ANON | MAP_PRIVATE,
+#endif
+	    -1, 0)) == MAP_FAILED)
+		goto err2;
+	V = (uint32_t *)(V0);
+#endif
+
+	/* 1: (B_0 ... B_{p-1}) <-- PBKDF2(P, S, 1, p * MFLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, salt, saltlen, 1, EVP_sha256(), p * 128 * r, B);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, salt, saltlen, 1, B, p * 128 * r);
+#endif
+
+	/* 2: for i = 0 to p - 1 do */
+	for (i = 0; i < p; i++) {
+		/* 3: B_i <-- MF(B_i, N) */
+		smix(&B[i * 128 * r], r, N, V, XY);
+	}
+
+	/* 5: DK <-- PBKDF2(P, B, 1, dkLen) */
+#ifdef USE_OPENSSL_PBKDF2
+	PKCS5_PBKDF2_HMAC((const char *)passwd, passwdlen, B, p * 128 * r, 1, EVP_sha256(), buflen, buf);
+#else
+	PBKDF2_SHA256(passwd, passwdlen, B, p * 128 * r, 1, buf, buflen);
+#endif
+
+	/* Free memory. */
+#ifdef MAP_ANON
+	if (munmap(V0, 128 * r * N))
+		goto err2;
+#else
+	free(V0);
+#endif
+	free(XY0);
+	free(B0);
+
+	/* Success! */
+	return (0);
+
+err2:
+	free(XY0);
+err1:
+	free(B0);
+err0:
+	/* Failure! */
+	return (-1);
+}
diff --git a/crypto/scrypt/lib/crypto/crypto_scrypt.h b/crypto/scrypt/lib/crypto/crypto_scrypt.h
new file mode 100644
index 000000000..f72e1f4b0
--- /dev/null
+++ b/crypto/scrypt/lib/crypto/crypto_scrypt.h
@@ -0,0 +1,46 @@
+/*-
+ * Copyright 2009 Colin Percival
+ * 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 AND CONTRIBUTORS ``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 OR CONTRIBUTORS 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.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+#ifndef _CRYPTO_SCRYPT_H_
+#define _CRYPTO_SCRYPT_H_
+
+#include <stdint.h>
+
+/**
+ * crypto_scrypt(passwd, passwdlen, salt, saltlen, N, r, p, buf, buflen):
+ * Compute scrypt(passwd[0 .. passwdlen - 1], salt[0 .. saltlen - 1], N, r,
+ * p, buflen) and write the result into buf.  The parameters r, p, and buflen
+ * must satisfy r * p < 2^30 and buflen <= (2^32 - 1) * 32.  The parameter N
+ * must be a power of 2 greater than 1.
+ *
+ * Return 0 on success; or -1 on error.
+ */
+int crypto_scrypt(const uint8_t *, size_t, const uint8_t *, size_t, uint64_t,
+    uint32_t, uint32_t, uint8_t *, size_t);
+
+#endif /* !_CRYPTO_SCRYPT_H_ */
diff --git a/crypto/scrypt/lib/util/sysendian.h b/crypto/scrypt/lib/util/sysendian.h
new file mode 100644
index 000000000..62ef31a42
--- /dev/null
+++ b/crypto/scrypt/lib/util/sysendian.h
@@ -0,0 +1,140 @@
+/*-
+ * Copyright 2007-2009 Colin Percival
+ * 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 AND CONTRIBUTORS ``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 OR CONTRIBUTORS 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.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+#ifndef _SYSENDIAN_H_
+#define _SYSENDIAN_H_
+
+#include "scrypt_platform.h"
+
+/* If we don't have be64enc, the <sys/endian.h> we have isn't usable. */
+#if !HAVE_DECL_BE64ENC
+#undef HAVE_SYS_ENDIAN_H
+#endif
+
+#ifdef HAVE_SYS_ENDIAN_H
+
+#include <sys/endian.h>
+
+#else
+
+#include <stdint.h>
+
+static inline uint32_t
+be32dec(const void *pp)
+{
+	const uint8_t *p = (uint8_t const *)pp;
+
+	return ((uint32_t)(p[3]) + ((uint32_t)(p[2]) << 8) +
+	    ((uint32_t)(p[1]) << 16) + ((uint32_t)(p[0]) << 24));
+}
+
+static inline void
+be32enc(void *pp, uint32_t x)
+{
+	uint8_t * p = (uint8_t *)pp;
+
+	p[3] = x & 0xff;
+	p[2] = (x >> 8) & 0xff;
+	p[1] = (x >> 16) & 0xff;
+	p[0] = (x >> 24) & 0xff;
+}
+
+static inline uint64_t
+be64dec(const void *pp)
+{
+	const uint8_t *p = (uint8_t const *)pp;
+
+	return ((uint64_t)(p[7]) + ((uint64_t)(p[6]) << 8) +
+	    ((uint64_t)(p[5]) << 16) + ((uint64_t)(p[4]) << 24) +
+	    ((uint64_t)(p[3]) << 32) + ((uint64_t)(p[2]) << 40) +
+	    ((uint64_t)(p[1]) << 48) + ((uint64_t)(p[0]) << 56));
+}
+
+static inline void
+be64enc(void *pp, uint64_t x)
+{
+	uint8_t * p = (uint8_t *)pp;
+
+	p[7] = x & 0xff;
+	p[6] = (x >> 8) & 0xff;
+	p[5] = (x >> 16) & 0xff;
+	p[4] = (x >> 24) & 0xff;
+	p[3] = (x >> 32) & 0xff;
+	p[2] = (x >> 40) & 0xff;
+	p[1] = (x >> 48) & 0xff;
+	p[0] = (x >> 56) & 0xff;
+}
+
+static inline uint32_t
+le32dec(const void *pp)
+{
+	const uint8_t *p = (uint8_t const *)pp;
+
+	return ((uint32_t)(p[0]) + ((uint32_t)(p[1]) << 8) +
+	    ((uint32_t)(p[2]) << 16) + ((uint32_t)(p[3]) << 24));
+}
+
+static inline void
+le32enc(void *pp, uint32_t x)
+{
+	uint8_t * p = (uint8_t *)pp;
+
+	p[0] = x & 0xff;
+	p[1] = (x >> 8) & 0xff;
+	p[2] = (x >> 16) & 0xff;
+	p[3] = (x >> 24) & 0xff;
+}
+
+static inline uint64_t
+le64dec(const void *pp)
+{
+	const uint8_t *p = (uint8_t const *)pp;
+
+	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));
+}
+
+static inline void
+le64enc(void *pp, uint64_t x)
+{
+	uint8_t * p = (uint8_t *)pp;
+
+	p[0] = x & 0xff;
+	p[1] = (x >> 8) & 0xff;
+	p[2] = (x >> 16) & 0xff;
+	p[3] = (x >> 24) & 0xff;
+	p[4] = (x >> 32) & 0xff;
+	p[5] = (x >> 40) & 0xff;
+	p[6] = (x >> 48) & 0xff;
+	p[7] = (x >> 56) & 0xff;
+}
+#endif /* !HAVE_SYS_ENDIAN_H */
+
+#endif /* !_SYSENDIAN_H_ */