scrypt 1.0.5 → 1.0.6

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    scrypt (1.0.5)
+    scrypt (1.0.6)
 
 GEM
   remote: http://rubygems.org/

data/ext/mri/Makefile

@@ -53,7 +53,7 @@ vendorarchdir = $(vendorlibdir)/$(sitearch)
 
 NULLCMD = :
 
-CC = /usr/bin/gcc-4.2 -Wall 
+CC = /usr/bin/gcc-4.2 -Wall -msse -msse2 
 CXX = g++-4.2
 LIBRUBY = $(LIBRUBY_SO)
 LIBRUBY_A = lib$(RUBY_SO_NAME)-static.a
@@ -115,8 +115,8 @@ extout_prefix =
 target_prefix = 
 LOCAL_LIBS = 
 LIBS = $(LIBRUBYARG_SHARED)  -lpthread -ldl -lobjc 
-SRCS = crypto_scrypt-ref.c memlimit.c scrypt_calibrate.c scrypt_ext.c scryptenc_cpuperf.c sha256.c
-OBJS = crypto_scrypt-ref.o memlimit.o scrypt_calibrate.o scrypt_ext.o scryptenc_cpuperf.o sha256.o
+SRCS = crypto_scrypt-sse.c memlimit.c scrypt_calibrate.c scrypt_ext.c scryptenc_cpuperf.c sha256.c
+OBJS = crypto_scrypt-sse.o memlimit.o scrypt_calibrate.o scrypt_ext.o scryptenc_cpuperf.o sha256.o
 TARGET = scrypt_ext
 DLLIB = $(TARGET).bundle
 EXTSTATIC = 

data/ext/mri/alt-impl/crypto_scrypt-nosse.c

@@ -0,0 +1,340 @@
+/*-
+ * 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>
+#ifndef __MINGW32__
+#include <sys/mman.h>
+#endif
+
+#include <errno.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sha256.h"
+#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(uint32_t[16]);
+static void blockmix_salsa8(uint32_t *, uint32_t *, uint32_t *, size_t);
+static uint64_t integerify(void *, size_t);
+static void smix(uint8_t *, size_t, uint64_t, uint32_t *, uint32_t *);
+
+static void
+blkcpy(void * dest, void * src, size_t len)
+{
+	size_t * D = dest;
+	size_t * S = src;
+	size_t L = len / sizeof(size_t);
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] = S[i];
+}
+
+static void
+blkxor(void * dest, void * src, size_t len)
+{
+	size_t * D = dest;
+	size_t * S = src;
+	size_t L = len / sizeof(size_t);
+	size_t i;
+
+	for (i = 0; i < L; i++)
+		D[i] ^= S[i];
+}
+
+/**
+ * salsa20_8(B):
+ * Apply the salsa20/8 core to the provided block.
+ */
+static void
+salsa20_8(uint32_t B[16])
+{
+	uint32_t x[16];
+	size_t i;
+
+	blkcpy(x, B, 64);
+	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
+	}
+	for (i = 0; i < 16; i++)
+		B[i] += x[i];
+}
+
+/**
+ * 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(uint32_t * Bin, uint32_t * Bout, uint32_t * X, size_t r)
+{
+	size_t i;
+
+	/* 1: X <-- B_{2r - 1} */
+	blkcpy(X, &Bin[(2 * r - 1) * 16], 64);
+
+	/* 2: for i = 0 to 2r - 1 do */
+	for (i = 0; i < 2 * r; i += 2) {
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 16], 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 * 8], X, 64);
+
+		/* 3: X <-- H(X \xor B_i) */
+		blkxor(X, &Bin[i * 16 + 16], 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 * 8 + r * 16], 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[1]) << 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, uint32_t * V, uint32_t * XY)
+{
+	uint32_t * X = XY;
+	uint32_t * Y = &XY[32 * r];
+	uint32_t * Z = &XY[64 * r];
+	uint64_t i;
+	uint64_t j;
+	size_t k;
+
+	/* 1: X <-- B */
+	for (k = 0; k < 32 * r; k++)
+		X[k] = le32dec(&B[4 * k]);
+
+	/* 2: for i = 0 to N - 1 do */
+	for (i = 0; i < N; i += 2) {
+		/* 3: V_i <-- X */
+		blkcpy(&V[i * (32 * r)], X, 128 * r);
+
+		/* 4: X <-- H(X) */
+		blockmix_salsa8(X, Y, Z, r);
+
+		/* 3: V_i <-- X */
+		blkcpy(&V[(i + 1) * (32 * 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, &V[j * (32 * 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, &V[j * (32 * r)], 128 * r);
+		blockmix_salsa8(Y, X, Z, r);
+	}
+
+	/* 10: B' <-- X */
+	for (k = 0; k < 32 * r; k++)
+		le32enc(&B[4 * k], X[k]);
+}
+
+/**
+ * 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 / 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) */
+	PBKDF2_scrypt_SHA256(passwd, passwdlen, salt, saltlen, 1, B, p * 128 * r);
+
+	/* 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) */
+	PBKDF2_scrypt_SHA256(passwd, passwdlen, B, p * 128 * r, 1, buf, buflen);
+
+	/* 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);
+}

data/ext/mri/crypto_scrypt-sse.c

@@ -0,0 +1,368 @@
+/*-
+ * 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>
+#ifndef __MINGW32__
+#include <sys/mman.h>
+#endif
+
+#include <emmintrin.h>
+#include <errno.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sha256.h"
+#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) */
+	PBKDF2_scrypt_SHA256(passwd, passwdlen, salt, saltlen, 1, B, p * 128 * r);
+
+	/* 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) */
+	PBKDF2_scrypt_SHA256(passwd, passwdlen, B, p * 128 * r, 1, buf, buflen);
+
+	/* 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);
+}

data/ext/mri/crypto_scrypt.h

@@ -40,7 +40,6 @@
  *
  * 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);
+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_ */

data/ext/mri/extconf.rb

@@ -1,4 +1,5 @@
 require "mkmf"
 dir_config("scrypt_ext")
-CONFIG['CC'] << " -Wall "
+CONFIG['CC'] << " -Wall -msse -msse2 "
+CONFIG['CC'] << " -D_GNU_SOURCE=1 " if CONFIG["target_os"]["mingw"]
 create_makefile("scrypt_ext")

data/ext/mri/memlimit.c

@@ -29,7 +29,9 @@
 #include "scrypt_platform.h"
 
 #include <sys/types.h>
+#ifndef __MINGW32__
 #include <sys/resource.h>
+#endif
 
 #ifdef HAVE_SYS_PARAM_H
 #include <sys/param.h>
@@ -141,6 +143,7 @@ memlimit_sysinfo(size_t * memlimit)
 static int
 memlimit_rlimit(size_t * memlimit)
 {
+#ifndef __MINGW32__
 	struct rlimit rl;
 	uint64_t memrlimit;
 
@@ -181,7 +184,9 @@ memlimit_rlimit(size_t * memlimit)
 #else
 	*memlimit = memrlimit;
 #endif
-
+#else
+	*memlimit = SIZE_MAX;
+#endif
 	/* Success! */
 	return (0);
 }

data/ext/mri/scrypt_calibrate.c

@@ -24,31 +24,30 @@
 
 
 static int
-pickparams(size_t maxmem, double maxmemfrac, double maxtime,
-		   int * logN, uint32_t * r, uint32_t * p)
+pickparams(size_t maxmem, double maxmemfrac, double maxtime, int * logN, uint32_t * r, uint32_t * p)
 {
 	size_t memlimit;
 	double opps;
 	double opslimit;
 	double maxN, maxrp;
 	int rc;
-	
+
 	/* Figure out how much memory to use. */
 	if (memtouse(maxmem, maxmemfrac, &memlimit))
 		return (1);
-	
+
 	/* Figure out how fast the CPU is. */
 	if ((rc = scryptenc_cpuperf(&opps)) != 0)
 		return (rc);
 	opslimit = opps * maxtime;
-	
+
 	/* Allow a minimum of 2^15 salsa20/8 cores. */
 	if (opslimit < 32768)
 		opslimit = 32768;
-	
+
 	/* Fix r = 8 for now. */
 	*r = 8;
-	
+
 	/*
 	 * The memory limit requires that 128Nr <= memlimit, while the CPU
 	 * limit requires that 4Nrp <= opslimit.  If opslimit < memlimit/32,
@@ -73,32 +72,31 @@ pickparams(size_t maxmem, double maxmemfrac, double maxtime,
 			if ((uint64_t)(1) << *logN > maxN / 2)
 				break;
 		}
-		
+
 		/* Choose p based on the CPU limit. */
 		maxrp = (opslimit / 4) / ((uint64_t)(1) << *logN);
 		if (maxrp > 0x3fffffff)
 			maxrp = 0x3fffffff;
 		*p = (uint32_t)(maxrp) / *r;
 	}
-	
+
 #ifdef DEBUG
 	fprintf(stderr, "N = %zu r = %d p = %d\n",
 			(size_t)(1) << *logN, (int)(*r), (int)(*p));
 #endif
-	
+
 	/* Success! */
 	return (0);
 }
 
 int
-calibrate(size_t maxmem, double maxmemfrac, double maxtime,
-		  uint64_t * n, uint32_t * r, uint32_t * p)
+calibrate(size_t maxmem, double maxmemfrac, double maxtime, uint64_t * n, uint32_t * r, uint32_t * p)
 {
 	int logN = 0;
 	int result = pickparams( maxmem, maxmemfrac, maxtime, & logN, r, p );
 	if (result == 0)
 	{
 		*n = (uint64_t)(1) << logN;
-	}	
+	}
 	return result;
 }

data/ext/mri/scrypt_ext.c

@@ -26,7 +26,11 @@ static VALUE sc_calibrate( VALUE self, VALUE maxmem, VALUE maxmemfrac, VALUE max
 	{
 		char cost_str[33];
 		memset( cost_str, '\0', 33 );
+		#ifdef __MINGW32__
+		sprintf( cost_str, "%lx$%x$%x$", (long unsigned int)n, (unsigned int)r, (unsigned int)p );
+		#else
 		sprintf( cost_str, "%Lx$%x$%x$", n, r, p );
+		#endif
 		return rb_str_new2( cost_str );
 	}
 
@@ -36,6 +40,11 @@ static VALUE sc_calibrate( VALUE self, VALUE maxmem, VALUE maxmemfrac, VALUE max
 
 static VALUE sc_crypt( VALUE self, VALUE key, VALUE salt, VALUE cost )
 {
+	uint64_t n = 0;
+	uint32_t r = 0;
+	uint32_t p = 0;
+	int result;
+
 	const char * safe_key = RSTRING_PTR(key) ? RSTRING_PTR(key) : "";
 	const char * safe_salt = RSTRING_PTR(salt) ? RSTRING_PTR(salt) : "";
 
@@ -48,12 +57,13 @@ static VALUE sc_crypt( VALUE self, VALUE key, VALUE salt, VALUE cost )
 		return Qnil;
 	}
 
-	uint64_t n = 0;
-	uint32_t r = 0;
-	uint32_t p = 0;
+	#ifdef __MINGW32__
+	sscanf( RSTRING_PTR( cost ), "%lx$%x$%x$", (long unsigned int*)& n, (unsigned int*)& r, (unsigned int*)& p );
+	#else
 	sscanf( RSTRING_PTR( cost ), "%Lx$%x$%x$", & n, & r, & p );
+	#endif
 
-	int result = crypto_scrypt(
+	result = crypto_scrypt(
 		(uint8_t *) safe_key, strlen(safe_key),
 		(uint8_t *) safe_salt, strlen(safe_salt),
 		n, r, p,

data/ext/mri/scryptenc_cpuperf.c

@@ -38,6 +38,13 @@
 
 #include "scryptenc_cpuperf.h"
 
+#ifdef __MINGW32__
+struct timespec {
+	long tv_sec;
+	long tv_nsec;
+};
+#endif
+
 #ifdef HAVE_CLOCK_GETTIME
 
 static clock_t clocktouse;
@@ -116,7 +123,7 @@ getclockdiff(struct timespec * st, double * diffd)
 	if (getclocktime(&en))
 		return (1);
 	*diffd = (en.tv_nsec - st->tv_nsec) * 0.000000001 +
-	    (en.tv_sec - st->tv_sec);
+			(en.tv_sec - st->tv_sec);
 
 	return (0);
 }
@@ -176,7 +183,7 @@ scryptenc_cpuperf(double * opps)
 
 #ifdef DEBUG
 	fprintf(stderr, "%ju salsa20/8 cores performed in %f seconds\n",
-	    (uintmax_t)i, diffd);
+			(uintmax_t)i, diffd);
 #endif
 
 	/* We can do approximately i salsa20/8 cores per diffd seconds. */

data/ext/mri/sha256.c

@@ -80,10 +80,10 @@ be32dec_vect(uint32_t *dst, const unsigned char *src, size_t len)
 /* Adjusted round function for rotating state */
 #define RNDr(S, W, i, k)			\
 	RND(S[(64 - i) % 8], S[(65 - i) % 8],	\
-	    S[(66 - i) % 8], S[(67 - i) % 8],	\
-	    S[(68 - i) % 8], S[(69 - i) % 8],	\
-	    S[(70 - i) % 8], S[(71 - i) % 8],	\
-	    W[i] + k)
+			S[(66 - i) % 8], S[(67 - i) % 8],	\
+			S[(68 - i) % 8], S[(69 - i) % 8],	\
+			S[(70 - i) % 8], S[(71 - i) % 8],	\
+			W[i] + k)
 
 /*
  * scrypt_SHA256 block compression function.  The 256-bit state is transformed via

data/ext/mri/sha256.h

@@ -56,7 +56,6 @@ void	HMAC_scrypt_SHA256_Final(unsigned char [32], HMAC_scrypt_SHA256_CTX *);
  * Compute PBKDF2(passwd, salt, c, dkLen) using HMAC-scrypt_SHA256 as the PRF, and
  * write the output to buf.  The value dkLen must be at most 32 * (2^32 - 1).
  */
-void	PBKDF2_scrypt_SHA256(const uint8_t *, size_t, const uint8_t *, size_t,
-    uint64_t, uint8_t *, size_t);
+void	PBKDF2_scrypt_SHA256(const uint8_t *, size_t, const uint8_t *, size_t, uint64_t, uint8_t *, size_t);
 
 #endif /* !_scrypt_SHA256_H_ */

data/ext/mri/sysendian.h

@@ -50,7 +50,7 @@ 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));
+		((uint32_t)(p[1]) << 16) + ((uint32_t)(p[0]) << 24));
 }
 
 static inline void
@@ -70,9 +70,9 @@ 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));
+		((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
@@ -116,9 +116,9 @@ 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));
+		((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

data/lib/scrypt/version.rb

@@ -1,3 +1,3 @@
 module SCrypt
-  VERSION = "1.0.5"
+  VERSION = "1.0.6"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: scrypt
 version: !ruby/object:Gem::Version
-  version: 1.0.5
+  version: 1.0.6
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-04-11 00:00:00.000000000 Z
+date: 2012-05-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -80,7 +80,9 @@ files:
 - Rakefile
 - autotest/discover.rb
 - ext/mri/Makefile
-- ext/mri/crypto_scrypt-ref.c
+- ext/mri/alt-impl/crypto_scrypt-nosse.c
+- ext/mri/alt-impl/crypto_scrypt-ref.c
+- ext/mri/crypto_scrypt-sse.c
 - ext/mri/crypto_scrypt.h
 - ext/mri/extconf.rb
 - ext/mri/memlimit.c
@@ -115,7 +117,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2356724981074653609
+      hash: 4242054223667576181
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -124,7 +126,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2356724981074653609
+      hash: 4242054223667576181
 requirements: []
 rubyforge_project: scrypt
 rubygems_version: 1.8.21