From: Dmitry Dunaev dunaev@tecon.ru

--- common/image-fit.c | 7 ++ include/image.h | 12 +++ include/u-boot/xxhash.h | 143 ++++++++++++++++++++++++++ lib/Kconfig | 8 ++ lib/Makefile | 1 + lib/xxhash.c | 265 ++++++++++++++++++++++++++++++++++++++++++++++++ tools/Makefile | 2 + 7 files changed, 438 insertions(+) create mode 100644 include/u-boot/xxhash.h create mode 100644 lib/xxhash.c

diff --git a/common/image-fit.c b/common/image-fit.c index cb6a306..d7ea8bf 100644 --- a/common/image-fit.c +++ b/common/image-fit.c @@ -29,6 +29,7 @@ DECLARE_GLOBAL_DATA_PTR; #include <u-boot/md5.h> #include <u-boot/sha1.h> #include <u-boot/sha256.h> +#include <u-boot/xxhash.h>

/*****************************************************************************/ /* New uImage format routines */ @@ -1013,6 +1014,12 @@ int calculate_hash(const void *data, int data_len, const char *algo, } else if (IMAGE_ENABLE_MD5 && strcmp(algo, "md5") == 0) { md5_wd((unsigned char *)data, data_len, value, CHUNKSZ_MD5); *value_len = 16; + } else if (IMAGE_ENABLE_XXHASH && strcmp(algo, "xxh32") == 0) { + *((uint32_t *)value) = xxh32_wd(data, data_len, 0, CHUNKSZ_XXH); + *value_len = 4; + } else if (IMAGE_ENABLE_XXHASH && strcmp(algo, "xxh64") == 0) { + *((uint64_t *)value) = xxh64_wd(data, data_len, 0, CHUNKSZ_XXH); + *value_len = 8; } else { debug("Unsupported hash alogrithm\n"); return -1; diff --git a/include/image.h b/include/image.h index a41a836..57c1e38 100644 --- a/include/image.h +++ b/include/image.h @@ -65,11 +65,15 @@ struct lmb; # ifdef CONFIG_SPL_SHA1_SUPPORT # define IMAGE_ENABLE_SHA1 1 # endif +# ifdef CONFIG_SPL_XXHASH_SUPPORT +# define IMAGE_ENABLE_XXHASH 1 +# endif # else # define CONFIG_CRC32 /* FIT images need CRC32 support */ # define IMAGE_ENABLE_CRC32 1 # define IMAGE_ENABLE_MD5 1 # define IMAGE_ENABLE_SHA1 1 +# define IMAGE_ENABLE_XXHASH 1 # endif

#ifndef IMAGE_ENABLE_CRC32 @@ -84,6 +88,10 @@ struct lmb; #define IMAGE_ENABLE_SHA1 0 #endif

+#ifndef IMAGE_ENABLE_XXHASH +#define IMAGE_ENABLE_XXHASH 0 +#endif + #if defined(CONFIG_FIT_ENABLE_SHA256_SUPPORT) || \ defined(CONFIG_SPL_SHA256_SUPPORT) #define IMAGE_ENABLE_SHA256 1 @@ -419,6 +427,10 @@ extern bootm_headers_t images; #define CHUNKSZ_SHA1 (64 * 1024) #endif

+#ifndef CHUNKSZ_XXH +#define CHUNKSZ_XXH (64 * 1024) +#endif + #define uimage_to_cpu(x) be32_to_cpu(x) #define cpu_to_uimage(x) cpu_to_be32(x)

diff --git a/include/u-boot/xxhash.h b/include/u-boot/xxhash.h new file mode 100644 index 0000000..ecbeb98 --- /dev/null +++ b/include/u-boot/xxhash.h @@ -0,0 +1,143 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Copyright (C) 2012-2016, Yann Collet. + * + * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * 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 COPYRIGHT HOLDERS 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 COPYRIGHT + * OWNER 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 program is free software; you can redistribute it and/or modify it under + * the terms of the GNU General Public License version 2 as published by the + * Free Software Foundation. This program is dual-licensed; you may select + * either version 2 of the GNU General Public License ("GPL") or BSD license + * ("BSD"). + * + * You can contact the author at: + * - xxHash homepage: http://cyan4973.github.io/xxHash/ + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/* + * Notice extracted from xxHash homepage: + * + * xxHash is an extremely fast Hash algorithm, running at RAM speed limits. + * It also successfully passes all tests from the SMHasher suite. + * + * Comparison (single thread, Windows Seven 32 bits, using SMHasher on a Core 2 + * Duo @3GHz) + * + * Name Speed Q.Score Author + * xxHash 5.4 GB/s 10 + * CrapWow 3.2 GB/s 2 Andrew + * MumurHash 3a 2.7 GB/s 10 Austin Appleby + * SpookyHash 2.0 GB/s 10 Bob Jenkins + * SBox 1.4 GB/s 9 Bret Mulvey + * Lookup3 1.2 GB/s 9 Bob Jenkins + * SuperFastHash 1.2 GB/s 1 Paul Hsieh + * CityHash64 1.05 GB/s 10 Pike & Alakuijala + * FNV 0.55 GB/s 5 Fowler, Noll, Vo + * CRC32 0.43 GB/s 9 + * MD5-32 0.33 GB/s 10 Ronald L. Rivest + * SHA1-32 0.28 GB/s 10 + * + * Q.Score is a measure of quality of the hash function. + * It depends on successfully passing SMHasher test set. + * 10 is a perfect score. + * + * A 64-bits version, named xxh64 offers much better speed, + * but for 64-bits applications only. + * Name Speed on 64 bits Speed on 32 bits + * xxh64 13.8 GB/s 1.9 GB/s + * xxh32 6.8 GB/s 6.0 GB/s + */ + +#ifndef XXHASH_H +#define XXHASH_H + +#include <linux/types.h> + +/* Reset watchdog each time we process this many bytes */ +#define CHUNKSZ_XXHASH (64 * 1024) + +/** + * xxh32() - calculate the 32-bit hash of the input with a given seed. + * + * @input: The data to hash. + * @length: The length of the data to hash. + * @seed: The seed can be used to alter the result predictably. + * + * Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s + * + * Return: The 32-bit hash of the data. + */ +uint32_t xxh32(const void *input, size_t length, uint32_t seed); + +/** + * xxh32_wd() - calculate the 32-bit hash of the input with a given seed. + * Triggering the watchdog every 'chunk_sz' bytes of input + * processed. + * + * @input: The data to hash. + * @length: The length of the data to hash. + * @seed: The seed can be used to alter the result predictably. + * @chunk_sz: The chunk size for watchdog triggering. + * + * Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s + * + * Return: The 32-bit hash of the data. + */ +uint32_t xxh32_wd(const void *input, const size_t len, const uint32_t seed, + const uint32_t chunk_sz); + +/** + * xxh64() - calculate the 64-bit hash of the input with a given seed. + * + * @input: The data to hash. + * @length: The length of the data to hash. + * @seed: The seed can be used to alter the result predictably. + * + * This function runs 2x faster on 64-bit systems, but slower on 32-bit systems. + * + * Return: The 64-bit hash of the data. + */ +uint64_t xxh64(const void *input, size_t length, uint64_t seed); + +/** + * xxh64_wd() - calculate the 64-bit hash of the input with a given seed. + * Triggering the watchdog every 'chunk_sz' bytes of input + * processed. + * + * @input: The data to hash. + * @length: The length of the data to hash. + * @seed: The seed can be used to alter the result predictably. + * @chunk_sz: The chunk size for watchdog triggering. + * + * This function runs 2x faster on 64-bit systems, but slower on 32-bit systems. + * + * Return: The 64-bit hash of the data. + */ +uint32_t xxh64_wd(const void *input, const size_t len, const uint64_t seed, + const uint32_t chunk_sz); + +#endif /* XXHASH_H */ diff --git a/lib/Kconfig b/lib/Kconfig index 00ac650..c279732 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -156,6 +156,14 @@ config SHA_PROG_HW_ACCEL Data can be streamed in a block at a time and the hashing is performed in hardware.

+config XXHASH + bool "Enable XXHASH support" + help + This option enables support of hashing using XXHASH algorithm. + The hash is calculated in software. + The SHA256 algorithm produces a 32-bit (4-byte) for xxh32 or + 64-bit (8-byte) for xxh64 hash value(digest). + config MD5 bool

diff --git a/lib/Makefile b/lib/Makefile index 8cd779f..f62716d 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -48,6 +48,7 @@ endif obj-$(CONFIG_RSA) += rsa/ obj-$(CONFIG_SHA1) += sha1.o obj-$(CONFIG_SHA256) += sha256.o +obj-$(CONFIG_XXHASH) += xxhash.o

obj-$(CONFIG_$(SPL_)ZLIB) += zlib/ obj-$(CONFIG_$(SPL_)GZIP) += gunzip.o diff --git a/lib/xxhash.c b/lib/xxhash.c new file mode 100644 index 0000000..c9fe476 --- /dev/null +++ b/lib/xxhash.c @@ -0,0 +1,265 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Copyright (C) 2012-2016, Yann Collet. + * + * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * 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 COPYRIGHT HOLDERS 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 COPYRIGHT + * OWNER 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 program is free software; you can redistribute it and/or modify it under + * the terms of the GNU General Public License version 2 as published by the + * Free Software Foundation. This program is dual-licensed; you may select + * either version 2 of the GNU General Public License ("GPL") or BSD license + * ("BSD"). + * + * You can contact the author at: + * - xxHash homepage: http://cyan4973.github.io/xxHash/ + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +#include "compiler.h" + +#ifndef USE_HOSTCC +#include <common.h> +#include <asm/types.h> +#include <asm/unaligned.h> +#endif /* USE_HOSTCC */ +#include <watchdog.h> +#include <u-boot/xxhash.h> + +/* Macros */ +#define xxh_rotl32(x, r) ((x << r) | (x >> (32 - r))) +#define xxh_rotl64(x, r) ((x << r) | (x >> (64 - r))) + +#ifdef USE_HOSTCC +#define get_unaligned_le32(x) cpu_to_le32(*((uint32_t *)(x))) +#define get_unaligned_le64(x) cpu_to_le64(*((uint64_t *)(x))) +#endif + +/* Constants */ +static const uint32_t PRIME32_1 = 2654435761U; +static const uint32_t PRIME32_2 = 2246822519U; +static const uint32_t PRIME32_3 = 3266489917U; +static const uint32_t PRIME32_4 = 668265263U; +static const uint32_t PRIME32_5 = 374761393U; + +static const uint64_t PRIME64_1 = 11400714785074694791ULL; +static const uint64_t PRIME64_2 = 14029467366897019727ULL; +static const uint64_t PRIME64_3 = 1609587929392839161ULL; +static const uint64_t PRIME64_4 = 9650029242287828579ULL; +static const uint64_t PRIME64_5 = 2870177450012600261ULL; + +/* Simple Hash Functions */ +static uint32_t xxh32_round(uint32_t seed, const uint32_t input) +{ + seed += input * PRIME32_2; + seed = xxh_rotl32(seed, 13); + seed *= PRIME32_1; + return seed; +} + +uint32_t xxh32(const void *input, const size_t len, const uint32_t seed) +{ + const uint8_t *p = (const uint8_t *)input; + const uint8_t *b_end = p + len; + uint32_t h32; + + if (len >= 16) { + const uint8_t *const limit = b_end - 16; + uint32_t v1 = seed + PRIME32_1 + PRIME32_2; + uint32_t v2 = seed + PRIME32_2; + uint32_t v3 = seed + 0; + uint32_t v4 = seed - PRIME32_1; + + do { + v1 = xxh32_round(v1, get_unaligned_le32(p)); + p += 4; + v2 = xxh32_round(v2, get_unaligned_le32(p)); + p += 4; + v3 = xxh32_round(v3, get_unaligned_le32(p)); + p += 4; + v4 = xxh32_round(v4, get_unaligned_le32(p)); + p += 4; + } while (p <= limit); + + h32 = xxh_rotl32(v1, 1) + xxh_rotl32(v2, 7) + + xxh_rotl32(v3, 12) + xxh_rotl32(v4, 18); + } else { + h32 = seed + PRIME32_5; + } + + h32 += (uint32_t)len; + + while (p + 4 <= b_end) { + h32 += get_unaligned_le32(p) * PRIME32_3; + h32 = xxh_rotl32(h32, 17) * PRIME32_4; + p += 4; + } + + while (p < b_end) { + h32 += (*p) * PRIME32_5; + h32 = xxh_rotl32(h32, 11) * PRIME32_1; + p++; + } + + h32 ^= h32 >> 15; + h32 *= PRIME32_2; + h32 ^= h32 >> 13; + h32 *= PRIME32_3; + h32 ^= h32 >> 16; + + return h32; +} + +uint32_t xxh32_wd(const void *input, const size_t len, const uint32_t seed, + const uint32_t chunk_sz) +{ +#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG) + uint8_t *end, *curr; + uint32_t hash = (uint32_t)seed; + int chunk; + + curr = input; + end = input + len; + while (curr < end) { + chunk = end - curr; + if (chunk > chunk_sz) + chunk = chunk_sz; + hash = xxh32(curr, chunk, hash); + curr += chunk; + WATCHDOG_RESET(); + } + + return hash; +#else + return xxh32(input, len, 0); +#endif +} + +static uint64_t xxh64_round(uint64_t acc, const uint64_t input) +{ + acc += input * PRIME64_2; + acc = xxh_rotl64(acc, 31); + acc *= PRIME64_1; + return acc; +} + +static uint64_t xxh64_merge_round(uint64_t acc, uint64_t val) +{ + val = xxh64_round(0, val); + acc ^= val; + acc = acc * PRIME64_1 + PRIME64_4; + return acc; +} + +uint64_t xxh64(const void *input, const size_t len, const uint64_t seed) +{ + const uint8_t *p = (const uint8_t *)input; + const uint8_t *const b_end = p + len; + uint64_t h64; + + if (len >= 32) { + const uint8_t *const limit = b_end - 32; + uint64_t v1 = seed + PRIME64_1 + PRIME64_2; + uint64_t v2 = seed + PRIME64_2; + uint64_t v3 = seed + 0; + uint64_t v4 = seed - PRIME64_1; + + do { + v1 = xxh64_round(v1, get_unaligned_le64(p)); + p += 8; + v2 = xxh64_round(v2, get_unaligned_le64(p)); + p += 8; + v3 = xxh64_round(v3, get_unaligned_le64(p)); + p += 8; + v4 = xxh64_round(v4, get_unaligned_le64(p)); + p += 8; + } while (p <= limit); + + h64 = xxh_rotl64(v1, 1) + xxh_rotl64(v2, 7) + + xxh_rotl64(v3, 12) + xxh_rotl64(v4, 18); + h64 = xxh64_merge_round(h64, v1); + h64 = xxh64_merge_round(h64, v2); + h64 = xxh64_merge_round(h64, v3); + h64 = xxh64_merge_round(h64, v4); + + } else { + h64 = seed + PRIME64_5; + } + + h64 += (uint64_t)len; + + while (p + 8 <= b_end) { + const uint64_t k1 = xxh64_round(0, get_unaligned_le64(p)); + + h64 ^= k1; + h64 = xxh_rotl64(h64, 27) * PRIME64_1 + PRIME64_4; + p += 8; + } + + if (p + 4 <= b_end) { + h64 ^= (uint64_t)(get_unaligned_le32(p)) * PRIME64_1; + h64 = xxh_rotl64(h64, 23) * PRIME64_2 + PRIME64_3; + p += 4; + } + + while (p < b_end) { + h64 ^= (*p) * PRIME64_5; + h64 = xxh_rotl64(h64, 11) * PRIME64_1; + p++; + } + + h64 ^= h64 >> 33; + h64 *= PRIME64_2; + h64 ^= h64 >> 29; + h64 *= PRIME64_3; + h64 ^= h64 >> 32; + + return h64; +} + +uint32_t xxh64_wd(const void *input, const size_t len, const uint64_t seed, + const uint32_t chunk_sz) +{ +#if defined(CONFIG_HW_WATCHDOG) || defined(CONFIG_WATCHDOG) + uint8_t *end, *curr; + uint64_t hash = (uint64_t)seed; + int chunk; + + curr = input; + end = input + len; + while (curr < end) { + chunk = end - curr; + if (chunk > chunk_sz) + chunk = chunk_sz; + hash = xxh64(curr, chunk, hash); + curr += chunk; + WATCHDOG_RESET(); + } + + return hash; +#else + return xxh64(input, len, 0); +#endif +} diff --git a/tools/Makefile b/tools/Makefile index 428cb8a..cdcaf89 100644 --- a/tools/Makefile +++ b/tools/Makefile @@ -105,6 +105,7 @@ dumpimage-mkimage-objs := aisimage.o \ socfpgaimage.o \ lib/sha1.o \ lib/sha256.o \ + lib/xxhash.o \ common/hash.o \ ublimage.o \ zynqimage.o \ @@ -200,6 +201,7 @@ HOSTCFLAGS_crc8.o := -pedantic HOSTCFLAGS_md5.o := -pedantic HOSTCFLAGS_sha1.o := -pedantic HOSTCFLAGS_sha256.o := -pedantic +HOSTCFLAGS_xxhash.o := -pedantic

quiet_cmd_wrap = WRAP $@ cmd_wrap = echo "#include <../$(patsubst $(obj)/%,%,$@)>" >$@