commit 3b2ab9b73c2cb1dc289aeea9f1dc0bf71459f0a3
parent dd67eb7259d728d2f91021ec6155f60650f6266a
Author: Nihal Jere <nihal@nihaljere.xyz>
Date: Fri, 30 Jul 2021 10:12:44 -0500
swap out pbkdf2 for argon2
Diffstat:
16 files changed, 3107 insertions(+), 140 deletions(-)
diff --git a/argon2/Makefile b/argon2/Makefile
@@ -0,0 +1,16 @@
+dir := argon2
+
+ARGON2_SRC = argon2/argon2.c \
+ argon2/core.c \
+ argon2/ref.c \
+ argon2/thread.c \
+ argon2/blake2/blake2b.c
+
+ARGON2_OBJ = $(ARGON2_SRC:%.c=%.o)
+ARGON2_LIB = argon2/argon2.a
+
+$(ARGON2_LIB): $(ARGON2_OBJ)
+ $(AR) crs $@ $(ARGON2_OBJ)
+
+$(dir)-clean:
+ rm -f $(ARGON2_OBJ) $(ARGON2_LIB)
diff --git a/argon2/argon2.a b/argon2/argon2.a
Binary files differ.
diff --git a/argon2/argon2.c b/argon2/argon2.c
@@ -0,0 +1,354 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#include <string.h>
+#include <stdlib.h>
+#include <stdio.h>
+
+#include "argon2.h"
+#include "core.h"
+
+const char *argon2_type2string(argon2_type type, int uppercase) {
+ switch (type) {
+ case Argon2_d:
+ return uppercase ? "Argon2d" : "argon2d";
+ case Argon2_i:
+ return uppercase ? "Argon2i" : "argon2i";
+ case Argon2_id:
+ return uppercase ? "Argon2id" : "argon2id";
+ }
+
+ return NULL;
+}
+
+int argon2_ctx(argon2_context *context, argon2_type type) {
+ /* 1. Validate all inputs */
+ int result = validate_inputs(context);
+ uint32_t memory_blocks, segment_length;
+ argon2_instance_t instance;
+
+ if (ARGON2_OK != result) {
+ return result;
+ }
+
+ if (Argon2_d != type && Argon2_i != type && Argon2_id != type) {
+ return ARGON2_INCORRECT_TYPE;
+ }
+
+ /* 2. Align memory size */
+ /* Minimum memory_blocks = 8L blocks, where L is the number of lanes */
+ memory_blocks = context->m_cost;
+
+ if (memory_blocks < 2 * ARGON2_SYNC_POINTS * context->lanes) {
+ memory_blocks = 2 * ARGON2_SYNC_POINTS * context->lanes;
+ }
+
+ segment_length = memory_blocks / (context->lanes * ARGON2_SYNC_POINTS);
+ /* Ensure that all segments have equal length */
+ memory_blocks = segment_length * (context->lanes * ARGON2_SYNC_POINTS);
+
+ instance.version = context->version;
+ instance.memory = NULL;
+ instance.passes = context->t_cost;
+ instance.memory_blocks = memory_blocks;
+ instance.segment_length = segment_length;
+ instance.lane_length = segment_length * ARGON2_SYNC_POINTS;
+ instance.lanes = context->lanes;
+ instance.threads = context->threads;
+ instance.type = type;
+
+ if (instance.threads > instance.lanes) {
+ instance.threads = instance.lanes;
+ }
+
+ /* 3. Initialization: Hashing inputs, allocating memory, filling first
+ * blocks
+ */
+ result = initialize(&instance, context);
+
+ if (ARGON2_OK != result) {
+ return result;
+ }
+
+ /* 4. Filling memory */
+ result = fill_memory_blocks(&instance);
+
+ if (ARGON2_OK != result) {
+ return result;
+ }
+ /* 5. Finalization */
+ finalize(context, &instance);
+
+ return ARGON2_OK;
+}
+
+int argon2_hash(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt, const size_t saltlen,
+ void *hash, const size_t hashlen, char *encoded,
+ const size_t encodedlen, argon2_type type,
+ const uint32_t version){
+
+ argon2_context context;
+ int result;
+ uint8_t *out;
+
+ if (pwdlen > ARGON2_MAX_PWD_LENGTH) {
+ return ARGON2_PWD_TOO_LONG;
+ }
+
+ if (saltlen > ARGON2_MAX_SALT_LENGTH) {
+ return ARGON2_SALT_TOO_LONG;
+ }
+
+ if (hashlen > ARGON2_MAX_OUTLEN) {
+ return ARGON2_OUTPUT_TOO_LONG;
+ }
+
+ if (hashlen < ARGON2_MIN_OUTLEN) {
+ return ARGON2_OUTPUT_TOO_SHORT;
+ }
+
+ out = malloc(hashlen);
+ if (!out) {
+ return ARGON2_MEMORY_ALLOCATION_ERROR;
+ }
+
+ context.out = (uint8_t *)out;
+ context.outlen = (uint32_t)hashlen;
+ context.pwd = CONST_CAST(uint8_t *)pwd;
+ context.pwdlen = (uint32_t)pwdlen;
+ context.salt = CONST_CAST(uint8_t *)salt;
+ context.saltlen = (uint32_t)saltlen;
+ context.secret = NULL;
+ context.secretlen = 0;
+ context.ad = NULL;
+ context.adlen = 0;
+ context.t_cost = t_cost;
+ context.m_cost = m_cost;
+ context.lanes = parallelism;
+ context.threads = parallelism;
+ context.allocate_cbk = NULL;
+ context.free_cbk = NULL;
+ context.flags = ARGON2_DEFAULT_FLAGS;
+ context.version = version;
+
+ result = argon2_ctx(&context, type);
+
+ if (result != ARGON2_OK) {
+ clear_internal_memory(out, hashlen);
+ free(out);
+ return result;
+ }
+
+ /* if raw hash requested, write it */
+ if (hash) {
+ memcpy(hash, out, hashlen);
+ }
+
+ clear_internal_memory(out, hashlen);
+ free(out);
+
+ return ARGON2_OK;
+}
+
+int argon2i_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, const size_t hashlen,
+ char *encoded, const size_t encodedlen) {
+
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ NULL, hashlen, encoded, encodedlen, Argon2_i,
+ ARGON2_VERSION_NUMBER);
+}
+
+int argon2i_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash, const size_t hashlen) {
+
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ hash, hashlen, NULL, 0, Argon2_i, ARGON2_VERSION_NUMBER);
+}
+
+int argon2d_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, const size_t hashlen,
+ char *encoded, const size_t encodedlen) {
+
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ NULL, hashlen, encoded, encodedlen, Argon2_d,
+ ARGON2_VERSION_NUMBER);
+}
+
+int argon2d_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash, const size_t hashlen) {
+
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ hash, hashlen, NULL, 0, Argon2_d, ARGON2_VERSION_NUMBER);
+}
+
+int argon2id_hash_encoded(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, const size_t hashlen,
+ char *encoded, const size_t encodedlen) {
+
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ NULL, hashlen, encoded, encodedlen, Argon2_id,
+ ARGON2_VERSION_NUMBER);
+}
+
+int argon2id_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash, const size_t hashlen) {
+ return argon2_hash(t_cost, m_cost, parallelism, pwd, pwdlen, salt, saltlen,
+ hash, hashlen, NULL, 0, Argon2_id,
+ ARGON2_VERSION_NUMBER);
+}
+
+static int argon2_compare(const uint8_t *b1, const uint8_t *b2, size_t len) {
+ size_t i;
+ uint8_t d = 0U;
+
+ for (i = 0U; i < len; i++) {
+ d |= b1[i] ^ b2[i];
+ }
+ return (int)((1 & ((d - 1) >> 8)) - 1);
+}
+
+int argon2d_ctx(argon2_context *context) {
+ return argon2_ctx(context, Argon2_d);
+}
+
+int argon2i_ctx(argon2_context *context) {
+ return argon2_ctx(context, Argon2_i);
+}
+
+int argon2id_ctx(argon2_context *context) {
+ return argon2_ctx(context, Argon2_id);
+}
+
+int argon2_verify_ctx(argon2_context *context, const char *hash,
+ argon2_type type) {
+ int ret = argon2_ctx(context, type);
+ if (ret != ARGON2_OK) {
+ return ret;
+ }
+
+ if (argon2_compare((uint8_t *)hash, context->out, context->outlen)) {
+ return ARGON2_VERIFY_MISMATCH;
+ }
+
+ return ARGON2_OK;
+}
+
+int argon2d_verify_ctx(argon2_context *context, const char *hash) {
+ return argon2_verify_ctx(context, hash, Argon2_d);
+}
+
+int argon2i_verify_ctx(argon2_context *context, const char *hash) {
+ return argon2_verify_ctx(context, hash, Argon2_i);
+}
+
+int argon2id_verify_ctx(argon2_context *context, const char *hash) {
+ return argon2_verify_ctx(context, hash, Argon2_id);
+}
+
+const char *argon2_error_message(int error_code) {
+ switch (error_code) {
+ case ARGON2_OK:
+ return "OK";
+ case ARGON2_OUTPUT_PTR_NULL:
+ return "Output pointer is NULL";
+ case ARGON2_OUTPUT_TOO_SHORT:
+ return "Output is too short";
+ case ARGON2_OUTPUT_TOO_LONG:
+ return "Output is too long";
+ case ARGON2_PWD_TOO_SHORT:
+ return "Password is too short";
+ case ARGON2_PWD_TOO_LONG:
+ return "Password is too long";
+ case ARGON2_SALT_TOO_SHORT:
+ return "Salt is too short";
+ case ARGON2_SALT_TOO_LONG:
+ return "Salt is too long";
+ case ARGON2_AD_TOO_SHORT:
+ return "Associated data is too short";
+ case ARGON2_AD_TOO_LONG:
+ return "Associated data is too long";
+ case ARGON2_SECRET_TOO_SHORT:
+ return "Secret is too short";
+ case ARGON2_SECRET_TOO_LONG:
+ return "Secret is too long";
+ case ARGON2_TIME_TOO_SMALL:
+ return "Time cost is too small";
+ case ARGON2_TIME_TOO_LARGE:
+ return "Time cost is too large";
+ case ARGON2_MEMORY_TOO_LITTLE:
+ return "Memory cost is too small";
+ case ARGON2_MEMORY_TOO_MUCH:
+ return "Memory cost is too large";
+ case ARGON2_LANES_TOO_FEW:
+ return "Too few lanes";
+ case ARGON2_LANES_TOO_MANY:
+ return "Too many lanes";
+ case ARGON2_PWD_PTR_MISMATCH:
+ return "Password pointer is NULL, but password length is not 0";
+ case ARGON2_SALT_PTR_MISMATCH:
+ return "Salt pointer is NULL, but salt length is not 0";
+ case ARGON2_SECRET_PTR_MISMATCH:
+ return "Secret pointer is NULL, but secret length is not 0";
+ case ARGON2_AD_PTR_MISMATCH:
+ return "Associated data pointer is NULL, but ad length is not 0";
+ case ARGON2_MEMORY_ALLOCATION_ERROR:
+ return "Memory allocation error";
+ case ARGON2_FREE_MEMORY_CBK_NULL:
+ return "The free memory callback is NULL";
+ case ARGON2_ALLOCATE_MEMORY_CBK_NULL:
+ return "The allocate memory callback is NULL";
+ case ARGON2_INCORRECT_PARAMETER:
+ return "Argon2_Context context is NULL";
+ case ARGON2_INCORRECT_TYPE:
+ return "There is no such version of Argon2";
+ case ARGON2_OUT_PTR_MISMATCH:
+ return "Output pointer mismatch";
+ case ARGON2_THREADS_TOO_FEW:
+ return "Not enough threads";
+ case ARGON2_THREADS_TOO_MANY:
+ return "Too many threads";
+ case ARGON2_MISSING_ARGS:
+ return "Missing arguments";
+ case ARGON2_ENCODING_FAIL:
+ return "Encoding failed";
+ case ARGON2_DECODING_FAIL:
+ return "Decoding failed";
+ case ARGON2_THREAD_FAIL:
+ return "Threading failure";
+ case ARGON2_DECODING_LENGTH_FAIL:
+ return "Some of encoded parameters are too long or too short";
+ case ARGON2_VERIFY_MISMATCH:
+ return "The password does not match the supplied hash";
+ default:
+ return "Unknown error code";
+ }
+}
diff --git a/argon2/argon2.h b/argon2/argon2.h
@@ -0,0 +1,385 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef ARGON2_H
+#define ARGON2_H
+
+#include <stdint.h>
+#include <stddef.h>
+#include <limits.h>
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+/* Symbols visibility control */
+#ifdef A2_VISCTL
+#define ARGON2_PUBLIC __attribute__((visibility("default")))
+#define ARGON2_LOCAL __attribute__ ((visibility ("hidden")))
+#elif defined(_MSC_VER)
+#define ARGON2_PUBLIC __declspec(dllexport)
+#define ARGON2_LOCAL
+#else
+#define ARGON2_PUBLIC
+#define ARGON2_LOCAL
+#endif
+
+/*
+ * Argon2 input parameter restrictions
+ */
+
+/* Minimum and maximum number of lanes (degree of parallelism) */
+#define ARGON2_MIN_LANES UINT32_C(1)
+#define ARGON2_MAX_LANES UINT32_C(0xFFFFFF)
+
+/* Minimum and maximum number of threads */
+#define ARGON2_MIN_THREADS UINT32_C(1)
+#define ARGON2_MAX_THREADS UINT32_C(0xFFFFFF)
+
+/* Number of synchronization points between lanes per pass */
+#define ARGON2_SYNC_POINTS UINT32_C(4)
+
+/* Minimum and maximum digest size in bytes */
+#define ARGON2_MIN_OUTLEN UINT32_C(4)
+#define ARGON2_MAX_OUTLEN UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum number of memory blocks (each of BLOCK_SIZE bytes) */
+#define ARGON2_MIN_MEMORY (2 * ARGON2_SYNC_POINTS) /* 2 blocks per slice */
+
+#define ARGON2_MIN(a, b) ((a) < (b) ? (a) : (b))
+/* Max memory size is addressing-space/2, topping at 2^32 blocks (4 TB) */
+#define ARGON2_MAX_MEMORY_BITS \
+ ARGON2_MIN(UINT32_C(32), (sizeof(void *) * CHAR_BIT - 10 - 1))
+#define ARGON2_MAX_MEMORY \
+ ARGON2_MIN(UINT32_C(0xFFFFFFFF), UINT64_C(1) << ARGON2_MAX_MEMORY_BITS)
+
+/* Minimum and maximum number of passes */
+#define ARGON2_MIN_TIME UINT32_C(1)
+#define ARGON2_MAX_TIME UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum password length in bytes */
+#define ARGON2_MIN_PWD_LENGTH UINT32_C(0)
+#define ARGON2_MAX_PWD_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum associated data length in bytes */
+#define ARGON2_MIN_AD_LENGTH UINT32_C(0)
+#define ARGON2_MAX_AD_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum salt length in bytes */
+#define ARGON2_MIN_SALT_LENGTH UINT32_C(8)
+#define ARGON2_MAX_SALT_LENGTH UINT32_C(0xFFFFFFFF)
+
+/* Minimum and maximum key length in bytes */
+#define ARGON2_MIN_SECRET UINT32_C(0)
+#define ARGON2_MAX_SECRET UINT32_C(0xFFFFFFFF)
+
+/* Flags to determine which fields are securely wiped (default = no wipe). */
+#define ARGON2_DEFAULT_FLAGS UINT32_C(0)
+#define ARGON2_FLAG_CLEAR_PASSWORD (UINT32_C(1) << 0)
+#define ARGON2_FLAG_CLEAR_SECRET (UINT32_C(1) << 1)
+
+/* Global flag to determine if we are wiping internal memory buffers. This flag
+ * is defined in core.c and defaults to 1 (wipe internal memory). */
+extern int FLAG_clear_internal_memory;
+
+/* Error codes */
+typedef enum Argon2_ErrorCodes {
+ ARGON2_OK = 0,
+
+ ARGON2_OUTPUT_PTR_NULL = -1,
+
+ ARGON2_OUTPUT_TOO_SHORT = -2,
+ ARGON2_OUTPUT_TOO_LONG = -3,
+
+ ARGON2_PWD_TOO_SHORT = -4,
+ ARGON2_PWD_TOO_LONG = -5,
+
+ ARGON2_SALT_TOO_SHORT = -6,
+ ARGON2_SALT_TOO_LONG = -7,
+
+ ARGON2_AD_TOO_SHORT = -8,
+ ARGON2_AD_TOO_LONG = -9,
+
+ ARGON2_SECRET_TOO_SHORT = -10,
+ ARGON2_SECRET_TOO_LONG = -11,
+
+ ARGON2_TIME_TOO_SMALL = -12,
+ ARGON2_TIME_TOO_LARGE = -13,
+
+ ARGON2_MEMORY_TOO_LITTLE = -14,
+ ARGON2_MEMORY_TOO_MUCH = -15,
+
+ ARGON2_LANES_TOO_FEW = -16,
+ ARGON2_LANES_TOO_MANY = -17,
+
+ ARGON2_PWD_PTR_MISMATCH = -18, /* NULL ptr with non-zero length */
+ ARGON2_SALT_PTR_MISMATCH = -19, /* NULL ptr with non-zero length */
+ ARGON2_SECRET_PTR_MISMATCH = -20, /* NULL ptr with non-zero length */
+ ARGON2_AD_PTR_MISMATCH = -21, /* NULL ptr with non-zero length */
+
+ ARGON2_MEMORY_ALLOCATION_ERROR = -22,
+
+ ARGON2_FREE_MEMORY_CBK_NULL = -23,
+ ARGON2_ALLOCATE_MEMORY_CBK_NULL = -24,
+
+ ARGON2_INCORRECT_PARAMETER = -25,
+ ARGON2_INCORRECT_TYPE = -26,
+
+ ARGON2_OUT_PTR_MISMATCH = -27,
+
+ ARGON2_THREADS_TOO_FEW = -28,
+ ARGON2_THREADS_TOO_MANY = -29,
+
+ ARGON2_MISSING_ARGS = -30,
+
+ ARGON2_ENCODING_FAIL = -31,
+
+ ARGON2_DECODING_FAIL = -32,
+
+ ARGON2_THREAD_FAIL = -33,
+
+ ARGON2_DECODING_LENGTH_FAIL = -34,
+
+ ARGON2_VERIFY_MISMATCH = -35
+} argon2_error_codes;
+
+/* Memory allocator types --- for external allocation */
+typedef int (*allocate_fptr)(uint8_t **memory, size_t bytes_to_allocate);
+typedef void (*deallocate_fptr)(uint8_t *memory, size_t bytes_to_allocate);
+
+/* Argon2 external data structures */
+
+/*
+ *****
+ * Context: structure to hold Argon2 inputs:
+ * output array and its length,
+ * password and its length,
+ * salt and its length,
+ * secret and its length,
+ * associated data and its length,
+ * number of passes, amount of used memory (in KBytes, can be rounded up a bit)
+ * number of parallel threads that will be run.
+ * All the parameters above affect the output hash value.
+ * Additionally, two function pointers can be provided to allocate and
+ * deallocate the memory (if NULL, memory will be allocated internally).
+ * Also, three flags indicate whether to erase password, secret as soon as they
+ * are pre-hashed (and thus not needed anymore), and the entire memory
+ *****
+ * Simplest situation: you have output array out[8], password is stored in
+ * pwd[32], salt is stored in salt[16], you do not have keys nor associated
+ * data. You need to spend 1 GB of RAM and you run 5 passes of Argon2d with
+ * 4 parallel lanes.
+ * You want to erase the password, but you're OK with last pass not being
+ * erased. You want to use the default memory allocator.
+ * Then you initialize:
+ Argon2_Context(out,8,pwd,32,salt,16,NULL,0,NULL,0,5,1<<20,4,4,NULL,NULL,true,false,false,false)
+ */
+typedef struct Argon2_Context {
+ uint8_t *out; /* output array */
+ uint32_t outlen; /* digest length */
+
+ uint8_t *pwd; /* password array */
+ uint32_t pwdlen; /* password length */
+
+ uint8_t *salt; /* salt array */
+ uint32_t saltlen; /* salt length */
+
+ uint8_t *secret; /* key array */
+ uint32_t secretlen; /* key length */
+
+ uint8_t *ad; /* associated data array */
+ uint32_t adlen; /* associated data length */
+
+ uint32_t t_cost; /* number of passes */
+ uint32_t m_cost; /* amount of memory requested (KB) */
+ uint32_t lanes; /* number of lanes */
+ uint32_t threads; /* maximum number of threads */
+
+ uint32_t version; /* version number */
+
+ allocate_fptr allocate_cbk; /* pointer to memory allocator */
+ deallocate_fptr free_cbk; /* pointer to memory deallocator */
+
+ uint32_t flags; /* array of bool options */
+} argon2_context;
+
+/* Argon2 primitive type */
+typedef enum Argon2_type {
+ Argon2_d = 0,
+ Argon2_i = 1,
+ Argon2_id = 2
+} argon2_type;
+
+/* Version of the algorithm */
+typedef enum Argon2_version {
+ ARGON2_VERSION_10 = 0x10,
+ ARGON2_VERSION_13 = 0x13,
+ ARGON2_VERSION_NUMBER = ARGON2_VERSION_13
+} argon2_version;
+
+/*
+ * Function that gives the string representation of an argon2_type.
+ * @param type The argon2_type that we want the string for
+ * @param uppercase Whether the string should have the first letter uppercase
+ * @return NULL if invalid type, otherwise the string representation.
+ */
+ARGON2_PUBLIC const char *argon2_type2string(argon2_type type, int uppercase);
+
+/*
+ * Function that performs memory-hard hashing with certain degree of parallelism
+ * @param context Pointer to the Argon2 internal structure
+ * @return Error code if smth is wrong, ARGON2_OK otherwise
+ */
+ARGON2_PUBLIC int argon2_ctx(argon2_context *context, argon2_type type);
+
+/**
+ * Hashes a password with Argon2i, producing a raw hash at @hash
+ * @param t_cost Number of iterations
+ * @param m_cost Sets memory usage to m_cost kibibytes
+ * @param parallelism Number of threads and compute lanes
+ * @param pwd Pointer to password
+ * @param pwdlen Password size in bytes
+ * @param salt Pointer to salt
+ * @param saltlen Salt size in bytes
+ * @param hash Buffer where to write the raw hash - updated by the function
+ * @param hashlen Desired length of the hash in bytes
+ * @pre Different parallelism levels will give different results
+ * @pre Returns ARGON2_OK if successful
+ */
+ARGON2_PUBLIC int argon2i_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash,
+ const size_t hashlen);
+
+
+ARGON2_PUBLIC int argon2d_hash_raw(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash,
+ const size_t hashlen);
+
+ARGON2_PUBLIC int argon2id_hash_raw(const uint32_t t_cost,
+ const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash,
+ const size_t hashlen);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_hash(const uint32_t t_cost, const uint32_t m_cost,
+ const uint32_t parallelism, const void *pwd,
+ const size_t pwdlen, const void *salt,
+ const size_t saltlen, void *hash,
+ const size_t hashlen, char *encoded,
+ const size_t encodedlen, argon2_type type,
+ const uint32_t version);
+
+/**
+ * Verifies a password against an encoded string
+ * Encoded string is restricted as in validate_inputs()
+ * @param encoded String encoding parameters, salt, hash
+ * @param pwd Pointer to password
+ * @pre Returns ARGON2_OK if successful
+ */
+ARGON2_PUBLIC int argon2i_verify(const char *encoded, const void *pwd,
+ const size_t pwdlen);
+
+ARGON2_PUBLIC int argon2d_verify(const char *encoded, const void *pwd,
+ const size_t pwdlen);
+
+ARGON2_PUBLIC int argon2id_verify(const char *encoded, const void *pwd,
+ const size_t pwdlen);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_verify(const char *encoded, const void *pwd,
+ const size_t pwdlen, argon2_type type);
+
+/**
+ * Argon2d: Version of Argon2 that picks memory blocks depending
+ * on the password and salt. Only for side-channel-free
+ * environment!!
+ *****
+ * @param context Pointer to current Argon2 context
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2d_ctx(argon2_context *context);
+
+/**
+ * Argon2i: Version of Argon2 that picks memory blocks
+ * independent on the password and salt. Good for side-channels,
+ * but worse w.r.t. tradeoff attacks if only one pass is used.
+ *****
+ * @param context Pointer to current Argon2 context
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2i_ctx(argon2_context *context);
+
+/**
+ * Argon2id: Version of Argon2 where the first half-pass over memory is
+ * password-independent, the rest are password-dependent (on the password and
+ * salt). OK against side channels (they reduce to 1/2-pass Argon2i), and
+ * better with w.r.t. tradeoff attacks (similar to Argon2d).
+ *****
+ * @param context Pointer to current Argon2 context
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2id_ctx(argon2_context *context);
+
+/**
+ * Verify if a given password is correct for Argon2d hashing
+ * @param context Pointer to current Argon2 context
+ * @param hash The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2d_verify_ctx(argon2_context *context, const char *hash);
+
+/**
+ * Verify if a given password is correct for Argon2i hashing
+ * @param context Pointer to current Argon2 context
+ * @param hash The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2i_verify_ctx(argon2_context *context, const char *hash);
+
+/**
+ * Verify if a given password is correct for Argon2id hashing
+ * @param context Pointer to current Argon2 context
+ * @param hash The password hash to verify. The length of the hash is
+ * specified by the context outlen member
+ * @return Zero if successful, a non zero error code otherwise
+ */
+ARGON2_PUBLIC int argon2id_verify_ctx(argon2_context *context,
+ const char *hash);
+
+/* generic function underlying the above ones */
+ARGON2_PUBLIC int argon2_verify_ctx(argon2_context *context, const char *hash,
+ argon2_type type);
+
+/**
+ * Get the associated error message for given error code
+ * @return The error message associated with the given error code
+ */
+ARGON2_PUBLIC const char *argon2_error_message(int error_code);
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif
diff --git a/argon2/blake2/blake2-impl.h b/argon2/blake2/blake2-impl.h
@@ -0,0 +1,156 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef PORTABLE_BLAKE2_IMPL_H
+#define PORTABLE_BLAKE2_IMPL_H
+
+#include <stdint.h>
+#include <string.h>
+
+#ifdef _WIN32
+#define BLAKE2_INLINE __inline
+#elif defined(__GNUC__) || defined(__clang__)
+#define BLAKE2_INLINE __inline__
+#else
+#define BLAKE2_INLINE
+#endif
+
+/* Argon2 Team - Begin Code */
+/*
+ Not an exhaustive list, but should cover the majority of modern platforms
+ Additionally, the code will always be correct---this is only a performance
+ tweak.
+*/
+#if (defined(__BYTE_ORDER__) && \
+ (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)) || \
+ defined(__LITTLE_ENDIAN__) || defined(__ARMEL__) || defined(__MIPSEL__) || \
+ defined(__AARCH64EL__) || defined(__amd64__) || defined(__i386__) || \
+ defined(_M_IX86) || defined(_M_X64) || defined(_M_AMD64) || \
+ defined(_M_ARM)
+#define NATIVE_LITTLE_ENDIAN
+#endif
+/* Argon2 Team - End Code */
+
+static BLAKE2_INLINE uint32_t load32(const void *src) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+ uint32_t w;
+ memcpy(&w, src, sizeof w);
+ return w;
+#else
+ const uint8_t *p = (const uint8_t *)src;
+ uint32_t w = *p++;
+ w |= (uint32_t)(*p++) << 8;
+ w |= (uint32_t)(*p++) << 16;
+ w |= (uint32_t)(*p++) << 24;
+ return w;
+#endif
+}
+
+static BLAKE2_INLINE uint64_t load64(const void *src) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+ uint64_t w;
+ memcpy(&w, src, sizeof w);
+ return w;
+#else
+ const uint8_t *p = (const uint8_t *)src;
+ uint64_t w = *p++;
+ w |= (uint64_t)(*p++) << 8;
+ w |= (uint64_t)(*p++) << 16;
+ w |= (uint64_t)(*p++) << 24;
+ w |= (uint64_t)(*p++) << 32;
+ w |= (uint64_t)(*p++) << 40;
+ w |= (uint64_t)(*p++) << 48;
+ w |= (uint64_t)(*p++) << 56;
+ return w;
+#endif
+}
+
+static BLAKE2_INLINE void store32(void *dst, uint32_t w) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+ memcpy(dst, &w, sizeof w);
+#else
+ uint8_t *p = (uint8_t *)dst;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+#endif
+}
+
+static BLAKE2_INLINE void store64(void *dst, uint64_t w) {
+#if defined(NATIVE_LITTLE_ENDIAN)
+ memcpy(dst, &w, sizeof w);
+#else
+ uint8_t *p = (uint8_t *)dst;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+#endif
+}
+
+static BLAKE2_INLINE uint64_t load48(const void *src) {
+ const uint8_t *p = (const uint8_t *)src;
+ uint64_t w = *p++;
+ w |= (uint64_t)(*p++) << 8;
+ w |= (uint64_t)(*p++) << 16;
+ w |= (uint64_t)(*p++) << 24;
+ w |= (uint64_t)(*p++) << 32;
+ w |= (uint64_t)(*p++) << 40;
+ return w;
+}
+
+static BLAKE2_INLINE void store48(void *dst, uint64_t w) {
+ uint8_t *p = (uint8_t *)dst;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+ w >>= 8;
+ *p++ = (uint8_t)w;
+}
+
+static BLAKE2_INLINE uint32_t rotr32(const uint32_t w, const unsigned c) {
+ return (w >> c) | (w << (32 - c));
+}
+
+static BLAKE2_INLINE uint64_t rotr64(const uint64_t w, const unsigned c) {
+ return (w >> c) | (w << (64 - c));
+}
+
+void clear_internal_memory(void *v, size_t n);
+
+#endif
diff --git a/argon2/blake2/blake2.h b/argon2/blake2/blake2.h
@@ -0,0 +1,89 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef PORTABLE_BLAKE2_H
+#define PORTABLE_BLAKE2_H
+
+#include "../argon2.h"
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+enum blake2b_constant {
+ BLAKE2B_BLOCKBYTES = 128,
+ BLAKE2B_OUTBYTES = 64,
+ BLAKE2B_KEYBYTES = 64,
+ BLAKE2B_SALTBYTES = 16,
+ BLAKE2B_PERSONALBYTES = 16
+};
+
+#pragma pack(push, 1)
+typedef struct __blake2b_param {
+ uint8_t digest_length; /* 1 */
+ uint8_t key_length; /* 2 */
+ uint8_t fanout; /* 3 */
+ uint8_t depth; /* 4 */
+ uint32_t leaf_length; /* 8 */
+ uint64_t node_offset; /* 16 */
+ uint8_t node_depth; /* 17 */
+ uint8_t inner_length; /* 18 */
+ uint8_t reserved[14]; /* 32 */
+ uint8_t salt[BLAKE2B_SALTBYTES]; /* 48 */
+ uint8_t personal[BLAKE2B_PERSONALBYTES]; /* 64 */
+} blake2b_param;
+#pragma pack(pop)
+
+typedef struct __blake2b_state {
+ uint64_t h[8];
+ uint64_t t[2];
+ uint64_t f[2];
+ uint8_t buf[BLAKE2B_BLOCKBYTES];
+ unsigned buflen;
+ unsigned outlen;
+ uint8_t last_node;
+} blake2b_state;
+
+/* Ensure param structs have not been wrongly padded */
+/* Poor man's static_assert */
+enum {
+ blake2_size_check_0 = 1 / !!(CHAR_BIT == 8),
+ blake2_size_check_2 =
+ 1 / !!(sizeof(blake2b_param) == sizeof(uint64_t) * CHAR_BIT)
+};
+
+/* Streaming API */
+ARGON2_LOCAL int blake2b_init(blake2b_state *S, size_t outlen);
+ARGON2_LOCAL int blake2b_init_key(blake2b_state *S, size_t outlen, const void *key,
+ size_t keylen);
+ARGON2_LOCAL int blake2b_init_param(blake2b_state *S, const blake2b_param *P);
+ARGON2_LOCAL int blake2b_update(blake2b_state *S, const void *in, size_t inlen);
+ARGON2_LOCAL int blake2b_final(blake2b_state *S, void *out, size_t outlen);
+
+/* Simple API */
+ARGON2_LOCAL int blake2b(void *out, size_t outlen, const void *in, size_t inlen,
+ const void *key, size_t keylen);
+
+/* Argon2 Team - Begin Code */
+ARGON2_LOCAL int blake2b_long(void *out, size_t outlen, const void *in, size_t inlen);
+/* Argon2 Team - End Code */
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif
diff --git a/argon2/blake2/blake2b.c b/argon2/blake2/blake2b.c
@@ -0,0 +1,390 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#include <stdint.h>
+#include <string.h>
+#include <stdio.h>
+
+#include "blake2.h"
+#include "blake2-impl.h"
+
+static const uint64_t blake2b_IV[8] = {
+ UINT64_C(0x6a09e667f3bcc908), UINT64_C(0xbb67ae8584caa73b),
+ UINT64_C(0x3c6ef372fe94f82b), UINT64_C(0xa54ff53a5f1d36f1),
+ UINT64_C(0x510e527fade682d1), UINT64_C(0x9b05688c2b3e6c1f),
+ UINT64_C(0x1f83d9abfb41bd6b), UINT64_C(0x5be0cd19137e2179)};
+
+static const unsigned int blake2b_sigma[12][16] = {
+ {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15},
+ {14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3},
+ {11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4},
+ {7, 9, 3, 1, 13, 12, 11, 14, 2, 6, 5, 10, 4, 0, 15, 8},
+ {9, 0, 5, 7, 2, 4, 10, 15, 14, 1, 11, 12, 6, 8, 3, 13},
+ {2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9},
+ {12, 5, 1, 15, 14, 13, 4, 10, 0, 7, 6, 3, 9, 2, 8, 11},
+ {13, 11, 7, 14, 12, 1, 3, 9, 5, 0, 15, 4, 8, 6, 2, 10},
+ {6, 15, 14, 9, 11, 3, 0, 8, 12, 2, 13, 7, 1, 4, 10, 5},
+ {10, 2, 8, 4, 7, 6, 1, 5, 15, 11, 9, 14, 3, 12, 13, 0},
+ {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15},
+ {14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3},
+};
+
+static BLAKE2_INLINE void blake2b_set_lastnode(blake2b_state *S) {
+ S->f[1] = (uint64_t)-1;
+}
+
+static BLAKE2_INLINE void blake2b_set_lastblock(blake2b_state *S) {
+ if (S->last_node) {
+ blake2b_set_lastnode(S);
+ }
+ S->f[0] = (uint64_t)-1;
+}
+
+static BLAKE2_INLINE void blake2b_increment_counter(blake2b_state *S,
+ uint64_t inc) {
+ S->t[0] += inc;
+ S->t[1] += (S->t[0] < inc);
+}
+
+static BLAKE2_INLINE void blake2b_invalidate_state(blake2b_state *S) {
+ clear_internal_memory(S, sizeof(*S)); /* wipe */
+ blake2b_set_lastblock(S); /* invalidate for further use */
+}
+
+static BLAKE2_INLINE void blake2b_init0(blake2b_state *S) {
+ memset(S, 0, sizeof(*S));
+ memcpy(S->h, blake2b_IV, sizeof(S->h));
+}
+
+int blake2b_init_param(blake2b_state *S, const blake2b_param *P) {
+ const unsigned char *p = (const unsigned char *)P;
+ unsigned int i;
+
+ if (NULL == P || NULL == S) {
+ return -1;
+ }
+
+ blake2b_init0(S);
+ /* IV XOR Parameter Block */
+ for (i = 0; i < 8; ++i) {
+ S->h[i] ^= load64(&p[i * sizeof(S->h[i])]);
+ }
+ S->outlen = P->digest_length;
+ return 0;
+}
+
+/* Sequential blake2b initialization */
+int blake2b_init(blake2b_state *S, size_t outlen) {
+ blake2b_param P;
+
+ if (S == NULL) {
+ return -1;
+ }
+
+ if ((outlen == 0) || (outlen > BLAKE2B_OUTBYTES)) {
+ blake2b_invalidate_state(S);
+ return -1;
+ }
+
+ /* Setup Parameter Block for unkeyed BLAKE2 */
+ P.digest_length = (uint8_t)outlen;
+ P.key_length = 0;
+ P.fanout = 1;
+ P.depth = 1;
+ P.leaf_length = 0;
+ P.node_offset = 0;
+ P.node_depth = 0;
+ P.inner_length = 0;
+ memset(P.reserved, 0, sizeof(P.reserved));
+ memset(P.salt, 0, sizeof(P.salt));
+ memset(P.personal, 0, sizeof(P.personal));
+
+ return blake2b_init_param(S, &P);
+}
+
+int blake2b_init_key(blake2b_state *S, size_t outlen, const void *key,
+ size_t keylen) {
+ blake2b_param P;
+
+ if (S == NULL) {
+ return -1;
+ }
+
+ if ((outlen == 0) || (outlen > BLAKE2B_OUTBYTES)) {
+ blake2b_invalidate_state(S);
+ return -1;
+ }
+
+ if ((key == 0) || (keylen == 0) || (keylen > BLAKE2B_KEYBYTES)) {
+ blake2b_invalidate_state(S);
+ return -1;
+ }
+
+ /* Setup Parameter Block for keyed BLAKE2 */
+ P.digest_length = (uint8_t)outlen;
+ P.key_length = (uint8_t)keylen;
+ P.fanout = 1;
+ P.depth = 1;
+ P.leaf_length = 0;
+ P.node_offset = 0;
+ P.node_depth = 0;
+ P.inner_length = 0;
+ memset(P.reserved, 0, sizeof(P.reserved));
+ memset(P.salt, 0, sizeof(P.salt));
+ memset(P.personal, 0, sizeof(P.personal));
+
+ if (blake2b_init_param(S, &P) < 0) {
+ blake2b_invalidate_state(S);
+ return -1;
+ }
+
+ {
+ uint8_t block[BLAKE2B_BLOCKBYTES];
+ memset(block, 0, BLAKE2B_BLOCKBYTES);
+ memcpy(block, key, keylen);
+ blake2b_update(S, block, BLAKE2B_BLOCKBYTES);
+ /* Burn the key from stack */
+ clear_internal_memory(block, BLAKE2B_BLOCKBYTES);
+ }
+ return 0;
+}
+
+static void blake2b_compress(blake2b_state *S, const uint8_t *block) {
+ uint64_t m[16];
+ uint64_t v[16];
+ unsigned int i, r;
+
+ for (i = 0; i < 16; ++i) {
+ m[i] = load64(block + i * sizeof(m[i]));
+ }
+
+ for (i = 0; i < 8; ++i) {
+ v[i] = S->h[i];
+ }
+
+ v[8] = blake2b_IV[0];
+ v[9] = blake2b_IV[1];
+ v[10] = blake2b_IV[2];
+ v[11] = blake2b_IV[3];
+ v[12] = blake2b_IV[4] ^ S->t[0];
+ v[13] = blake2b_IV[5] ^ S->t[1];
+ v[14] = blake2b_IV[6] ^ S->f[0];
+ v[15] = blake2b_IV[7] ^ S->f[1];
+
+#define G(r, i, a, b, c, d) \
+ do { \
+ a = a + b + m[blake2b_sigma[r][2 * i + 0]]; \
+ d = rotr64(d ^ a, 32); \
+ c = c + d; \
+ b = rotr64(b ^ c, 24); \
+ a = a + b + m[blake2b_sigma[r][2 * i + 1]]; \
+ d = rotr64(d ^ a, 16); \
+ c = c + d; \
+ b = rotr64(b ^ c, 63); \
+ } while ((void)0, 0)
+
+#define ROUND(r) \
+ do { \
+ G(r, 0, v[0], v[4], v[8], v[12]); \
+ G(r, 1, v[1], v[5], v[9], v[13]); \
+ G(r, 2, v[2], v[6], v[10], v[14]); \
+ G(r, 3, v[3], v[7], v[11], v[15]); \
+ G(r, 4, v[0], v[5], v[10], v[15]); \
+ G(r, 5, v[1], v[6], v[11], v[12]); \
+ G(r, 6, v[2], v[7], v[8], v[13]); \
+ G(r, 7, v[3], v[4], v[9], v[14]); \
+ } while ((void)0, 0)
+
+ for (r = 0; r < 12; ++r) {
+ ROUND(r);
+ }
+
+ for (i = 0; i < 8; ++i) {
+ S->h[i] = S->h[i] ^ v[i] ^ v[i + 8];
+ }
+
+#undef G
+#undef ROUND
+}
+
+int blake2b_update(blake2b_state *S, const void *in, size_t inlen) {
+ const uint8_t *pin = (const uint8_t *)in;
+
+ if (inlen == 0) {
+ return 0;
+ }
+
+ /* Sanity check */
+ if (S == NULL || in == NULL) {
+ return -1;
+ }
+
+ /* Is this a reused state? */
+ if (S->f[0] != 0) {
+ return -1;
+ }
+
+ if (S->buflen + inlen > BLAKE2B_BLOCKBYTES) {
+ /* Complete current block */
+ size_t left = S->buflen;
+ size_t fill = BLAKE2B_BLOCKBYTES - left;
+ memcpy(&S->buf[left], pin, fill);
+ blake2b_increment_counter(S, BLAKE2B_BLOCKBYTES);
+ blake2b_compress(S, S->buf);
+ S->buflen = 0;
+ inlen -= fill;
+ pin += fill;
+ /* Avoid buffer copies when possible */
+ while (inlen > BLAKE2B_BLOCKBYTES) {
+ blake2b_increment_counter(S, BLAKE2B_BLOCKBYTES);
+ blake2b_compress(S, pin);
+ inlen -= BLAKE2B_BLOCKBYTES;
+ pin += BLAKE2B_BLOCKBYTES;
+ }
+ }
+ memcpy(&S->buf[S->buflen], pin, inlen);
+ S->buflen += (unsigned int)inlen;
+ return 0;
+}
+
+int blake2b_final(blake2b_state *S, void *out, size_t outlen) {
+ uint8_t buffer[BLAKE2B_OUTBYTES] = {0};
+ unsigned int i;
+
+ /* Sanity checks */
+ if (S == NULL || out == NULL || outlen < S->outlen) {
+ return -1;
+ }
+
+ /* Is this a reused state? */
+ if (S->f[0] != 0) {
+ return -1;
+ }
+
+ blake2b_increment_counter(S, S->buflen);
+ blake2b_set_lastblock(S);
+ memset(&S->buf[S->buflen], 0, BLAKE2B_BLOCKBYTES - S->buflen); /* Padding */
+ blake2b_compress(S, S->buf);
+
+ for (i = 0; i < 8; ++i) { /* Output full hash to temp buffer */
+ store64(buffer + sizeof(S->h[i]) * i, S->h[i]);
+ }
+
+ memcpy(out, buffer, S->outlen);
+ clear_internal_memory(buffer, sizeof(buffer));
+ clear_internal_memory(S->buf, sizeof(S->buf));
+ clear_internal_memory(S->h, sizeof(S->h));
+ return 0;
+}
+
+int blake2b(void *out, size_t outlen, const void *in, size_t inlen,
+ const void *key, size_t keylen) {
+ blake2b_state S;
+ int ret = -1;
+
+ /* Verify parameters */
+ if (NULL == in && inlen > 0) {
+ goto fail;
+ }
+
+ if (NULL == out || outlen == 0 || outlen > BLAKE2B_OUTBYTES) {
+ goto fail;
+ }
+
+ if ((NULL == key && keylen > 0) || keylen > BLAKE2B_KEYBYTES) {
+ goto fail;
+ }
+
+ if (keylen > 0) {
+ if (blake2b_init_key(&S, outlen, key, keylen) < 0) {
+ goto fail;
+ }
+ } else {
+ if (blake2b_init(&S, outlen) < 0) {
+ goto fail;
+ }
+ }
+
+ if (blake2b_update(&S, in, inlen) < 0) {
+ goto fail;
+ }
+ ret = blake2b_final(&S, out, outlen);
+
+fail:
+ clear_internal_memory(&S, sizeof(S));
+ return ret;
+}
+
+/* Argon2 Team - Begin Code */
+int blake2b_long(void *pout, size_t outlen, const void *in, size_t inlen) {
+ uint8_t *out = (uint8_t *)pout;
+ blake2b_state blake_state;
+ uint8_t outlen_bytes[sizeof(uint32_t)] = {0};
+ int ret = -1;
+
+ if (outlen > UINT32_MAX) {
+ goto fail;
+ }
+
+ /* Ensure little-endian byte order! */
+ store32(outlen_bytes, (uint32_t)outlen);
+
+#define TRY(statement) \
+ do { \
+ ret = statement; \
+ if (ret < 0) { \
+ goto fail; \
+ } \
+ } while ((void)0, 0)
+
+ if (outlen <= BLAKE2B_OUTBYTES) {
+ TRY(blake2b_init(&blake_state, outlen));
+ TRY(blake2b_update(&blake_state, outlen_bytes, sizeof(outlen_bytes)));
+ TRY(blake2b_update(&blake_state, in, inlen));
+ TRY(blake2b_final(&blake_state, out, outlen));
+ } else {
+ uint32_t toproduce;
+ uint8_t out_buffer[BLAKE2B_OUTBYTES];
+ uint8_t in_buffer[BLAKE2B_OUTBYTES];
+ TRY(blake2b_init(&blake_state, BLAKE2B_OUTBYTES));
+ TRY(blake2b_update(&blake_state, outlen_bytes, sizeof(outlen_bytes)));
+ TRY(blake2b_update(&blake_state, in, inlen));
+ TRY(blake2b_final(&blake_state, out_buffer, BLAKE2B_OUTBYTES));
+ memcpy(out, out_buffer, BLAKE2B_OUTBYTES / 2);
+ out += BLAKE2B_OUTBYTES / 2;
+ toproduce = (uint32_t)outlen - BLAKE2B_OUTBYTES / 2;
+
+ while (toproduce > BLAKE2B_OUTBYTES) {
+ memcpy(in_buffer, out_buffer, BLAKE2B_OUTBYTES);
+ TRY(blake2b(out_buffer, BLAKE2B_OUTBYTES, in_buffer,
+ BLAKE2B_OUTBYTES, NULL, 0));
+ memcpy(out, out_buffer, BLAKE2B_OUTBYTES / 2);
+ out += BLAKE2B_OUTBYTES / 2;
+ toproduce -= BLAKE2B_OUTBYTES / 2;
+ }
+
+ memcpy(in_buffer, out_buffer, BLAKE2B_OUTBYTES);
+ TRY(blake2b(out_buffer, toproduce, in_buffer, BLAKE2B_OUTBYTES, NULL,
+ 0));
+ memcpy(out, out_buffer, toproduce);
+ }
+fail:
+ clear_internal_memory(&blake_state, sizeof(blake_state));
+ return ret;
+#undef TRY
+}
+/* Argon2 Team - End Code */
diff --git a/argon2/blake2/blamka-round-opt.h b/argon2/blake2/blamka-round-opt.h
@@ -0,0 +1,471 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef BLAKE_ROUND_MKA_OPT_H
+#define BLAKE_ROUND_MKA_OPT_H
+
+#include "blake2-impl.h"
+
+#include <emmintrin.h>
+#if defined(__SSSE3__)
+#include <tmmintrin.h> /* for _mm_shuffle_epi8 and _mm_alignr_epi8 */
+#endif
+
+#if defined(__XOP__) && (defined(__GNUC__) || defined(__clang__))
+#include <x86intrin.h>
+#endif
+
+#if !defined(__AVX512F__)
+#if !defined(__AVX2__)
+#if !defined(__XOP__)
+#if defined(__SSSE3__)
+#define r16 \
+ (_mm_setr_epi8(2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9))
+#define r24 \
+ (_mm_setr_epi8(3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10))
+#define _mm_roti_epi64(x, c) \
+ (-(c) == 32) \
+ ? _mm_shuffle_epi32((x), _MM_SHUFFLE(2, 3, 0, 1)) \
+ : (-(c) == 24) \
+ ? _mm_shuffle_epi8((x), r24) \
+ : (-(c) == 16) \
+ ? _mm_shuffle_epi8((x), r16) \
+ : (-(c) == 63) \
+ ? _mm_xor_si128(_mm_srli_epi64((x), -(c)), \
+ _mm_add_epi64((x), (x))) \
+ : _mm_xor_si128(_mm_srli_epi64((x), -(c)), \
+ _mm_slli_epi64((x), 64 - (-(c))))
+#else /* defined(__SSE2__) */
+#define _mm_roti_epi64(r, c) \
+ _mm_xor_si128(_mm_srli_epi64((r), -(c)), _mm_slli_epi64((r), 64 - (-(c))))
+#endif
+#else
+#endif
+
+static BLAKE2_INLINE __m128i fBlaMka(__m128i x, __m128i y) {
+ const __m128i z = _mm_mul_epu32(x, y);
+ return _mm_add_epi64(_mm_add_epi64(x, y), _mm_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ A0 = fBlaMka(A0, B0); \
+ A1 = fBlaMka(A1, B1); \
+ \
+ D0 = _mm_xor_si128(D0, A0); \
+ D1 = _mm_xor_si128(D1, A1); \
+ \
+ D0 = _mm_roti_epi64(D0, -32); \
+ D1 = _mm_roti_epi64(D1, -32); \
+ \
+ C0 = fBlaMka(C0, D0); \
+ C1 = fBlaMka(C1, D1); \
+ \
+ B0 = _mm_xor_si128(B0, C0); \
+ B1 = _mm_xor_si128(B1, C1); \
+ \
+ B0 = _mm_roti_epi64(B0, -24); \
+ B1 = _mm_roti_epi64(B1, -24); \
+ } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ A0 = fBlaMka(A0, B0); \
+ A1 = fBlaMka(A1, B1); \
+ \
+ D0 = _mm_xor_si128(D0, A0); \
+ D1 = _mm_xor_si128(D1, A1); \
+ \
+ D0 = _mm_roti_epi64(D0, -16); \
+ D1 = _mm_roti_epi64(D1, -16); \
+ \
+ C0 = fBlaMka(C0, D0); \
+ C1 = fBlaMka(C1, D1); \
+ \
+ B0 = _mm_xor_si128(B0, C0); \
+ B1 = _mm_xor_si128(B1, C1); \
+ \
+ B0 = _mm_roti_epi64(B0, -63); \
+ B1 = _mm_roti_epi64(B1, -63); \
+ } while ((void)0, 0)
+
+#if defined(__SSSE3__)
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ __m128i t0 = _mm_alignr_epi8(B1, B0, 8); \
+ __m128i t1 = _mm_alignr_epi8(B0, B1, 8); \
+ B0 = t0; \
+ B1 = t1; \
+ \
+ t0 = C0; \
+ C0 = C1; \
+ C1 = t0; \
+ \
+ t0 = _mm_alignr_epi8(D1, D0, 8); \
+ t1 = _mm_alignr_epi8(D0, D1, 8); \
+ D0 = t1; \
+ D1 = t0; \
+ } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ __m128i t0 = _mm_alignr_epi8(B0, B1, 8); \
+ __m128i t1 = _mm_alignr_epi8(B1, B0, 8); \
+ B0 = t0; \
+ B1 = t1; \
+ \
+ t0 = C0; \
+ C0 = C1; \
+ C1 = t0; \
+ \
+ t0 = _mm_alignr_epi8(D0, D1, 8); \
+ t1 = _mm_alignr_epi8(D1, D0, 8); \
+ D0 = t1; \
+ D1 = t0; \
+ } while ((void)0, 0)
+#else /* SSE2 */
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ __m128i t0 = D0; \
+ __m128i t1 = B0; \
+ D0 = C0; \
+ C0 = C1; \
+ C1 = D0; \
+ D0 = _mm_unpackhi_epi64(D1, _mm_unpacklo_epi64(t0, t0)); \
+ D1 = _mm_unpackhi_epi64(t0, _mm_unpacklo_epi64(D1, D1)); \
+ B0 = _mm_unpackhi_epi64(B0, _mm_unpacklo_epi64(B1, B1)); \
+ B1 = _mm_unpackhi_epi64(B1, _mm_unpacklo_epi64(t1, t1)); \
+ } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ __m128i t0, t1; \
+ t0 = C0; \
+ C0 = C1; \
+ C1 = t0; \
+ t0 = B0; \
+ t1 = D0; \
+ B0 = _mm_unpackhi_epi64(B1, _mm_unpacklo_epi64(B0, B0)); \
+ B1 = _mm_unpackhi_epi64(t0, _mm_unpacklo_epi64(B1, B1)); \
+ D0 = _mm_unpackhi_epi64(D0, _mm_unpacklo_epi64(D1, D1)); \
+ D1 = _mm_unpackhi_epi64(D1, _mm_unpacklo_epi64(t1, t1)); \
+ } while ((void)0, 0)
+#endif
+
+#define BLAKE2_ROUND(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+ G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+ \
+ DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+ \
+ G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+ G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+ \
+ UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+ } while ((void)0, 0)
+#else /* __AVX2__ */
+
+#include <immintrin.h>
+
+#define rotr32(x) _mm256_shuffle_epi32(x, _MM_SHUFFLE(2, 3, 0, 1))
+#define rotr24(x) _mm256_shuffle_epi8(x, _mm256_setr_epi8(3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10, 3, 4, 5, 6, 7, 0, 1, 2, 11, 12, 13, 14, 15, 8, 9, 10))
+#define rotr16(x) _mm256_shuffle_epi8(x, _mm256_setr_epi8(2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9, 2, 3, 4, 5, 6, 7, 0, 1, 10, 11, 12, 13, 14, 15, 8, 9))
+#define rotr63(x) _mm256_xor_si256(_mm256_srli_epi64((x), 63), _mm256_add_epi64((x), (x)))
+
+#define G1_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ __m256i ml = _mm256_mul_epu32(A0, B0); \
+ ml = _mm256_add_epi64(ml, ml); \
+ A0 = _mm256_add_epi64(A0, _mm256_add_epi64(B0, ml)); \
+ D0 = _mm256_xor_si256(D0, A0); \
+ D0 = rotr32(D0); \
+ \
+ ml = _mm256_mul_epu32(C0, D0); \
+ ml = _mm256_add_epi64(ml, ml); \
+ C0 = _mm256_add_epi64(C0, _mm256_add_epi64(D0, ml)); \
+ \
+ B0 = _mm256_xor_si256(B0, C0); \
+ B0 = rotr24(B0); \
+ \
+ ml = _mm256_mul_epu32(A1, B1); \
+ ml = _mm256_add_epi64(ml, ml); \
+ A1 = _mm256_add_epi64(A1, _mm256_add_epi64(B1, ml)); \
+ D1 = _mm256_xor_si256(D1, A1); \
+ D1 = rotr32(D1); \
+ \
+ ml = _mm256_mul_epu32(C1, D1); \
+ ml = _mm256_add_epi64(ml, ml); \
+ C1 = _mm256_add_epi64(C1, _mm256_add_epi64(D1, ml)); \
+ \
+ B1 = _mm256_xor_si256(B1, C1); \
+ B1 = rotr24(B1); \
+ } while((void)0, 0);
+
+#define G2_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ __m256i ml = _mm256_mul_epu32(A0, B0); \
+ ml = _mm256_add_epi64(ml, ml); \
+ A0 = _mm256_add_epi64(A0, _mm256_add_epi64(B0, ml)); \
+ D0 = _mm256_xor_si256(D0, A0); \
+ D0 = rotr16(D0); \
+ \
+ ml = _mm256_mul_epu32(C0, D0); \
+ ml = _mm256_add_epi64(ml, ml); \
+ C0 = _mm256_add_epi64(C0, _mm256_add_epi64(D0, ml)); \
+ B0 = _mm256_xor_si256(B0, C0); \
+ B0 = rotr63(B0); \
+ \
+ ml = _mm256_mul_epu32(A1, B1); \
+ ml = _mm256_add_epi64(ml, ml); \
+ A1 = _mm256_add_epi64(A1, _mm256_add_epi64(B1, ml)); \
+ D1 = _mm256_xor_si256(D1, A1); \
+ D1 = rotr16(D1); \
+ \
+ ml = _mm256_mul_epu32(C1, D1); \
+ ml = _mm256_add_epi64(ml, ml); \
+ C1 = _mm256_add_epi64(C1, _mm256_add_epi64(D1, ml)); \
+ B1 = _mm256_xor_si256(B1, C1); \
+ B1 = rotr63(B1); \
+ } while((void)0, 0);
+
+#define DIAGONALIZE_1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ B0 = _mm256_permute4x64_epi64(B0, _MM_SHUFFLE(0, 3, 2, 1)); \
+ C0 = _mm256_permute4x64_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+ D0 = _mm256_permute4x64_epi64(D0, _MM_SHUFFLE(2, 1, 0, 3)); \
+ \
+ B1 = _mm256_permute4x64_epi64(B1, _MM_SHUFFLE(0, 3, 2, 1)); \
+ C1 = _mm256_permute4x64_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+ D1 = _mm256_permute4x64_epi64(D1, _MM_SHUFFLE(2, 1, 0, 3)); \
+ } while((void)0, 0);
+
+#define DIAGONALIZE_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ __m256i tmp1 = _mm256_blend_epi32(B0, B1, 0xCC); \
+ __m256i tmp2 = _mm256_blend_epi32(B0, B1, 0x33); \
+ B1 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+ B0 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+ \
+ tmp1 = C0; \
+ C0 = C1; \
+ C1 = tmp1; \
+ \
+ tmp1 = _mm256_blend_epi32(D0, D1, 0xCC); \
+ tmp2 = _mm256_blend_epi32(D0, D1, 0x33); \
+ D0 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+ D1 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+ } while(0);
+
+#define UNDIAGONALIZE_1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ B0 = _mm256_permute4x64_epi64(B0, _MM_SHUFFLE(2, 1, 0, 3)); \
+ C0 = _mm256_permute4x64_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+ D0 = _mm256_permute4x64_epi64(D0, _MM_SHUFFLE(0, 3, 2, 1)); \
+ \
+ B1 = _mm256_permute4x64_epi64(B1, _MM_SHUFFLE(2, 1, 0, 3)); \
+ C1 = _mm256_permute4x64_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+ D1 = _mm256_permute4x64_epi64(D1, _MM_SHUFFLE(0, 3, 2, 1)); \
+ } while((void)0, 0);
+
+#define UNDIAGONALIZE_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ __m256i tmp1 = _mm256_blend_epi32(B0, B1, 0xCC); \
+ __m256i tmp2 = _mm256_blend_epi32(B0, B1, 0x33); \
+ B0 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+ B1 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+ \
+ tmp1 = C0; \
+ C0 = C1; \
+ C1 = tmp1; \
+ \
+ tmp1 = _mm256_blend_epi32(D0, D1, 0x33); \
+ tmp2 = _mm256_blend_epi32(D0, D1, 0xCC); \
+ D0 = _mm256_permute4x64_epi64(tmp1, _MM_SHUFFLE(2,3,0,1)); \
+ D1 = _mm256_permute4x64_epi64(tmp2, _MM_SHUFFLE(2,3,0,1)); \
+ } while((void)0, 0);
+
+#define BLAKE2_ROUND_1(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do{ \
+ G1_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ G2_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ \
+ DIAGONALIZE_1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ \
+ G1_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ G2_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ \
+ UNDIAGONALIZE_1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ } while((void)0, 0);
+
+#define BLAKE2_ROUND_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do{ \
+ G1_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ G2_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ \
+ DIAGONALIZE_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ \
+ G1_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ G2_AVX2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ \
+ UNDIAGONALIZE_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ } while((void)0, 0);
+
+#endif /* __AVX2__ */
+
+#else /* __AVX512F__ */
+
+#include <immintrin.h>
+
+#define ror64(x, n) _mm512_ror_epi64((x), (n))
+
+static __m512i muladd(__m512i x, __m512i y)
+{
+ __m512i z = _mm512_mul_epu32(x, y);
+ return _mm512_add_epi64(_mm512_add_epi64(x, y), _mm512_add_epi64(z, z));
+}
+
+#define G1(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ A0 = muladd(A0, B0); \
+ A1 = muladd(A1, B1); \
+\
+ D0 = _mm512_xor_si512(D0, A0); \
+ D1 = _mm512_xor_si512(D1, A1); \
+\
+ D0 = ror64(D0, 32); \
+ D1 = ror64(D1, 32); \
+\
+ C0 = muladd(C0, D0); \
+ C1 = muladd(C1, D1); \
+\
+ B0 = _mm512_xor_si512(B0, C0); \
+ B1 = _mm512_xor_si512(B1, C1); \
+\
+ B0 = ror64(B0, 24); \
+ B1 = ror64(B1, 24); \
+ } while ((void)0, 0)
+
+#define G2(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ A0 = muladd(A0, B0); \
+ A1 = muladd(A1, B1); \
+\
+ D0 = _mm512_xor_si512(D0, A0); \
+ D1 = _mm512_xor_si512(D1, A1); \
+\
+ D0 = ror64(D0, 16); \
+ D1 = ror64(D1, 16); \
+\
+ C0 = muladd(C0, D0); \
+ C1 = muladd(C1, D1); \
+\
+ B0 = _mm512_xor_si512(B0, C0); \
+ B1 = _mm512_xor_si512(B1, C1); \
+\
+ B0 = ror64(B0, 63); \
+ B1 = ror64(B1, 63); \
+ } while ((void)0, 0)
+
+#define DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ B0 = _mm512_permutex_epi64(B0, _MM_SHUFFLE(0, 3, 2, 1)); \
+ B1 = _mm512_permutex_epi64(B1, _MM_SHUFFLE(0, 3, 2, 1)); \
+\
+ C0 = _mm512_permutex_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+ C1 = _mm512_permutex_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+ D0 = _mm512_permutex_epi64(D0, _MM_SHUFFLE(2, 1, 0, 3)); \
+ D1 = _mm512_permutex_epi64(D1, _MM_SHUFFLE(2, 1, 0, 3)); \
+ } while ((void)0, 0)
+
+#define UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ B0 = _mm512_permutex_epi64(B0, _MM_SHUFFLE(2, 1, 0, 3)); \
+ B1 = _mm512_permutex_epi64(B1, _MM_SHUFFLE(2, 1, 0, 3)); \
+\
+ C0 = _mm512_permutex_epi64(C0, _MM_SHUFFLE(1, 0, 3, 2)); \
+ C1 = _mm512_permutex_epi64(C1, _MM_SHUFFLE(1, 0, 3, 2)); \
+\
+ D0 = _mm512_permutex_epi64(D0, _MM_SHUFFLE(0, 3, 2, 1)); \
+ D1 = _mm512_permutex_epi64(D1, _MM_SHUFFLE(0, 3, 2, 1)); \
+ } while ((void)0, 0)
+
+#define BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1) \
+ do { \
+ G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+ G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+ DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+ G1(A0, B0, C0, D0, A1, B1, C1, D1); \
+ G2(A0, B0, C0, D0, A1, B1, C1, D1); \
+\
+ UNDIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+ } while ((void)0, 0)
+
+#define SWAP_HALVES(A0, A1) \
+ do { \
+ __m512i t0, t1; \
+ t0 = _mm512_shuffle_i64x2(A0, A1, _MM_SHUFFLE(1, 0, 1, 0)); \
+ t1 = _mm512_shuffle_i64x2(A0, A1, _MM_SHUFFLE(3, 2, 3, 2)); \
+ A0 = t0; \
+ A1 = t1; \
+ } while((void)0, 0)
+
+#define SWAP_QUARTERS(A0, A1) \
+ do { \
+ SWAP_HALVES(A0, A1); \
+ A0 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A0); \
+ A1 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A1); \
+ } while((void)0, 0)
+
+#define UNSWAP_QUARTERS(A0, A1) \
+ do { \
+ A0 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A0); \
+ A1 = _mm512_permutexvar_epi64(_mm512_setr_epi64(0, 1, 4, 5, 2, 3, 6, 7), A1); \
+ SWAP_HALVES(A0, A1); \
+ } while((void)0, 0)
+
+#define BLAKE2_ROUND_1(A0, C0, B0, D0, A1, C1, B1, D1) \
+ do { \
+ SWAP_HALVES(A0, B0); \
+ SWAP_HALVES(C0, D0); \
+ SWAP_HALVES(A1, B1); \
+ SWAP_HALVES(C1, D1); \
+ BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1); \
+ SWAP_HALVES(A0, B0); \
+ SWAP_HALVES(C0, D0); \
+ SWAP_HALVES(A1, B1); \
+ SWAP_HALVES(C1, D1); \
+ } while ((void)0, 0)
+
+#define BLAKE2_ROUND_2(A0, A1, B0, B1, C0, C1, D0, D1) \
+ do { \
+ SWAP_QUARTERS(A0, A1); \
+ SWAP_QUARTERS(B0, B1); \
+ SWAP_QUARTERS(C0, C1); \
+ SWAP_QUARTERS(D0, D1); \
+ BLAKE2_ROUND(A0, B0, C0, D0, A1, B1, C1, D1); \
+ UNSWAP_QUARTERS(A0, A1); \
+ UNSWAP_QUARTERS(B0, B1); \
+ UNSWAP_QUARTERS(C0, C1); \
+ UNSWAP_QUARTERS(D0, D1); \
+ } while ((void)0, 0)
+
+#endif /* __AVX512F__ */
+#endif /* BLAKE_ROUND_MKA_OPT_H */
diff --git a/argon2/blake2/blamka-round-ref.h b/argon2/blake2/blamka-round-ref.h
@@ -0,0 +1,56 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef BLAKE_ROUND_MKA_H
+#define BLAKE_ROUND_MKA_H
+
+#include "blake2.h"
+#include "blake2-impl.h"
+
+/* designed by the Lyra PHC team */
+static BLAKE2_INLINE uint64_t fBlaMka(uint64_t x, uint64_t y) {
+ const uint64_t m = UINT64_C(0xFFFFFFFF);
+ const uint64_t xy = (x & m) * (y & m);
+ return x + y + 2 * xy;
+}
+
+#define G(a, b, c, d) \
+ do { \
+ a = fBlaMka(a, b); \
+ d = rotr64(d ^ a, 32); \
+ c = fBlaMka(c, d); \
+ b = rotr64(b ^ c, 24); \
+ a = fBlaMka(a, b); \
+ d = rotr64(d ^ a, 16); \
+ c = fBlaMka(c, d); \
+ b = rotr64(b ^ c, 63); \
+ } while ((void)0, 0)
+
+#define BLAKE2_ROUND_NOMSG(v0, v1, v2, v3, v4, v5, v6, v7, v8, v9, v10, v11, \
+ v12, v13, v14, v15) \
+ do { \
+ G(v0, v4, v8, v12); \
+ G(v1, v5, v9, v13); \
+ G(v2, v6, v10, v14); \
+ G(v3, v7, v11, v15); \
+ G(v0, v5, v10, v15); \
+ G(v1, v6, v11, v12); \
+ G(v2, v7, v8, v13); \
+ G(v3, v4, v9, v14); \
+ } while ((void)0, 0)
+
+#endif
diff --git a/argon2/core.c b/argon2/core.c
@@ -0,0 +1,644 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+/*For memory wiping*/
+#ifdef _WIN32
+#include <windows.h>
+#include <winbase.h> /* For SecureZeroMemory */
+#endif
+#if defined __STDC_LIB_EXT1__
+#define __STDC_WANT_LIB_EXT1__ 1
+#endif
+#define VC_GE_2005(version) (version >= 1400)
+
+/* for explicit_bzero() on glibc */
+#define _DEFAULT_SOURCE
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "core.h"
+#include "thread.h"
+#include "blake2/blake2.h"
+#include "blake2/blake2-impl.h"
+
+#if defined(__clang__)
+#if __has_attribute(optnone)
+#define NOT_OPTIMIZED __attribute__((optnone))
+#endif
+#elif defined(__GNUC__)
+#define GCC_VERSION \
+ (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
+#if GCC_VERSION >= 40400
+#define NOT_OPTIMIZED __attribute__((optimize("O0")))
+#endif
+#endif
+#ifndef NOT_OPTIMIZED
+#define NOT_OPTIMIZED
+#endif
+
+/***************Instance and Position constructors**********/
+void init_block_value(block *b, uint8_t in) { memset(b->v, in, sizeof(b->v)); }
+
+void copy_block(block *dst, const block *src) {
+ memcpy(dst->v, src->v, sizeof(uint64_t) * ARGON2_QWORDS_IN_BLOCK);
+}
+
+void xor_block(block *dst, const block *src) {
+ int i;
+ for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+ dst->v[i] ^= src->v[i];
+ }
+}
+
+static void load_block(block *dst, const void *input) {
+ unsigned i;
+ for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+ dst->v[i] = load64((const uint8_t *)input + i * sizeof(dst->v[i]));
+ }
+}
+
+static void store_block(void *output, const block *src) {
+ unsigned i;
+ for (i = 0; i < ARGON2_QWORDS_IN_BLOCK; ++i) {
+ store64((uint8_t *)output + i * sizeof(src->v[i]), src->v[i]);
+ }
+}
+
+/***************Memory functions*****************/
+
+int allocate_memory(const argon2_context *context, uint8_t **memory,
+ size_t num, size_t size) {
+ size_t memory_size = num*size;
+ if (memory == NULL) {
+ return ARGON2_MEMORY_ALLOCATION_ERROR;
+ }
+
+ /* 1. Check for multiplication overflow */
+ if (size != 0 && memory_size / size != num) {
+ return ARGON2_MEMORY_ALLOCATION_ERROR;
+ }
+
+ /* 2. Try to allocate with appropriate allocator */
+ if (context->allocate_cbk) {
+ (context->allocate_cbk)(memory, memory_size);
+ } else {
+ *memory = malloc(memory_size);
+ }
+
+ if (*memory == NULL) {
+ return ARGON2_MEMORY_ALLOCATION_ERROR;
+ }
+
+ return ARGON2_OK;
+}
+
+void free_memory(const argon2_context *context, uint8_t *memory,
+ size_t num, size_t size) {
+ size_t memory_size = num*size;
+ clear_internal_memory(memory, memory_size);
+ if (context->free_cbk) {
+ (context->free_cbk)(memory, memory_size);
+ } else {
+ free(memory);
+ }
+}
+
+#if defined(__OpenBSD__)
+#define HAVE_EXPLICIT_BZERO 1
+#elif defined(__GLIBC__) && defined(__GLIBC_PREREQ)
+#if __GLIBC_PREREQ(2,25)
+#define HAVE_EXPLICIT_BZERO 1
+#endif
+#endif
+
+void NOT_OPTIMIZED secure_wipe_memory(void *v, size_t n) {
+#if defined(_MSC_VER) && VC_GE_2005(_MSC_VER) || defined(__MINGW32__)
+ SecureZeroMemory(v, n);
+#elif defined memset_s
+ memset_s(v, n, 0, n);
+#elif defined(HAVE_EXPLICIT_BZERO)
+ explicit_bzero(v, n);
+#else
+ static void *(*const volatile memset_sec)(void *, int, size_t) = &memset;
+ memset_sec(v, 0, n);
+#endif
+}
+
+/* Memory clear flag defaults to true. */
+int FLAG_clear_internal_memory = 1;
+void clear_internal_memory(void *v, size_t n) {
+ if (FLAG_clear_internal_memory && v) {
+ secure_wipe_memory(v, n);
+ }
+}
+
+void finalize(const argon2_context *context, argon2_instance_t *instance) {
+ if (context != NULL && instance != NULL) {
+ block blockhash;
+ uint32_t l;
+
+ copy_block(&blockhash, instance->memory + instance->lane_length - 1);
+
+ /* XOR the last blocks */
+ for (l = 1; l < instance->lanes; ++l) {
+ uint32_t last_block_in_lane =
+ l * instance->lane_length + (instance->lane_length - 1);
+ xor_block(&blockhash, instance->memory + last_block_in_lane);
+ }
+
+ /* Hash the result */
+ {
+ uint8_t blockhash_bytes[ARGON2_BLOCK_SIZE];
+ store_block(blockhash_bytes, &blockhash);
+ blake2b_long(context->out, context->outlen, blockhash_bytes,
+ ARGON2_BLOCK_SIZE);
+ /* clear blockhash and blockhash_bytes */
+ clear_internal_memory(blockhash.v, ARGON2_BLOCK_SIZE);
+ clear_internal_memory(blockhash_bytes, ARGON2_BLOCK_SIZE);
+ }
+
+#ifdef GENKAT
+ print_tag(context->out, context->outlen);
+#endif
+
+ free_memory(context, (uint8_t *)instance->memory,
+ instance->memory_blocks, sizeof(block));
+ }
+}
+
+uint32_t index_alpha(const argon2_instance_t *instance,
+ const argon2_position_t *position, uint32_t pseudo_rand,
+ int same_lane) {
+ /*
+ * Pass 0:
+ * This lane : all already finished segments plus already constructed
+ * blocks in this segment
+ * Other lanes : all already finished segments
+ * Pass 1+:
+ * This lane : (SYNC_POINTS - 1) last segments plus already constructed
+ * blocks in this segment
+ * Other lanes : (SYNC_POINTS - 1) last segments
+ */
+ uint32_t reference_area_size;
+ uint64_t relative_position;
+ uint32_t start_position, absolute_position;
+
+ if (0 == position->pass) {
+ /* First pass */
+ if (0 == position->slice) {
+ /* First slice */
+ reference_area_size =
+ position->index - 1; /* all but the previous */
+ } else {
+ if (same_lane) {
+ /* The same lane => add current segment */
+ reference_area_size =
+ position->slice * instance->segment_length +
+ position->index - 1;
+ } else {
+ reference_area_size =
+ position->slice * instance->segment_length +
+ ((position->index == 0) ? (-1) : 0);
+ }
+ }
+ } else {
+ /* Second pass */
+ if (same_lane) {
+ reference_area_size = instance->lane_length -
+ instance->segment_length + position->index -
+ 1;
+ } else {
+ reference_area_size = instance->lane_length -
+ instance->segment_length +
+ ((position->index == 0) ? (-1) : 0);
+ }
+ }
+
+ /* 1.2.4. Mapping pseudo_rand to 0..<reference_area_size-1> and produce
+ * relative position */
+ relative_position = pseudo_rand;
+ relative_position = relative_position * relative_position >> 32;
+ relative_position = reference_area_size - 1 -
+ (reference_area_size * relative_position >> 32);
+
+ /* 1.2.5 Computing starting position */
+ start_position = 0;
+
+ if (0 != position->pass) {
+ start_position = (position->slice == ARGON2_SYNC_POINTS - 1)
+ ? 0
+ : (position->slice + 1) * instance->segment_length;
+ }
+
+ /* 1.2.6. Computing absolute position */
+ absolute_position = (start_position + relative_position) %
+ instance->lane_length; /* absolute position */
+ return absolute_position;
+}
+
+/* Single-threaded version for p=1 case */
+static int fill_memory_blocks_st(argon2_instance_t *instance) {
+ uint32_t r, s, l;
+
+ for (r = 0; r < instance->passes; ++r) {
+ for (s = 0; s < ARGON2_SYNC_POINTS; ++s) {
+ for (l = 0; l < instance->lanes; ++l) {
+ argon2_position_t position = {r, l, (uint8_t)s, 0};
+ fill_segment(instance, position);
+ }
+ }
+#ifdef GENKAT
+ internal_kat(instance, r); /* Print all memory blocks */
+#endif
+ }
+ return ARGON2_OK;
+}
+
+#if !defined(ARGON2_NO_THREADS)
+
+#ifdef _WIN32
+static unsigned __stdcall fill_segment_thr(void *thread_data)
+#else
+static void *fill_segment_thr(void *thread_data)
+#endif
+{
+ argon2_thread_data *my_data = thread_data;
+ fill_segment(my_data->instance_ptr, my_data->pos);
+ argon2_thread_exit();
+ return 0;
+}
+
+/* Multi-threaded version for p > 1 case */
+static int fill_memory_blocks_mt(argon2_instance_t *instance) {
+ uint32_t r, s;
+ argon2_thread_handle_t *thread = NULL;
+ argon2_thread_data *thr_data = NULL;
+ int rc = ARGON2_OK;
+
+ /* 1. Allocating space for threads */
+ thread = calloc(instance->lanes, sizeof(argon2_thread_handle_t));
+ if (thread == NULL) {
+ rc = ARGON2_MEMORY_ALLOCATION_ERROR;
+ goto fail;
+ }
+
+ thr_data = calloc(instance->lanes, sizeof(argon2_thread_data));
+ if (thr_data == NULL) {
+ rc = ARGON2_MEMORY_ALLOCATION_ERROR;
+ goto fail;
+ }
+
+ for (r = 0; r < instance->passes; ++r) {
+ for (s = 0; s < ARGON2_SYNC_POINTS; ++s) {
+ uint32_t l, ll;
+
+ /* 2. Calling threads */
+ for (l = 0; l < instance->lanes; ++l) {
+ argon2_position_t position;
+
+ /* 2.1 Join a thread if limit is exceeded */
+ if (l >= instance->threads) {
+ if (argon2_thread_join(thread[l - instance->threads])) {
+ rc = ARGON2_THREAD_FAIL;
+ goto fail;
+ }
+ }
+
+ /* 2.2 Create thread */
+ position.pass = r;
+ position.lane = l;
+ position.slice = (uint8_t)s;
+ position.index = 0;
+ thr_data[l].instance_ptr =
+ instance; /* preparing the thread input */
+ memcpy(&(thr_data[l].pos), &position,
+ sizeof(argon2_position_t));
+ if (argon2_thread_create(&thread[l], &fill_segment_thr,
+ (void *)&thr_data[l])) {
+ /* Wait for already running threads */
+ for (ll = 0; ll < l; ++ll)
+ argon2_thread_join(thread[ll]);
+ rc = ARGON2_THREAD_FAIL;
+ goto fail;
+ }
+
+ /* fill_segment(instance, position); */
+ /*Non-thread equivalent of the lines above */
+ }
+
+ /* 3. Joining remaining threads */
+ for (l = instance->lanes - instance->threads; l < instance->lanes;
+ ++l) {
+ if (argon2_thread_join(thread[l])) {
+ rc = ARGON2_THREAD_FAIL;
+ goto fail;
+ }
+ }
+ }
+
+#ifdef GENKAT
+ internal_kat(instance, r); /* Print all memory blocks */
+#endif
+ }
+
+fail:
+ if (thread != NULL) {
+ free(thread);
+ }
+ if (thr_data != NULL) {
+ free(thr_data);
+ }
+ return rc;
+}
+
+#endif /* ARGON2_NO_THREADS */
+
+int fill_memory_blocks(argon2_instance_t *instance) {
+ if (instance == NULL || instance->lanes == 0) {
+ return ARGON2_INCORRECT_PARAMETER;
+ }
+#if defined(ARGON2_NO_THREADS)
+ return fill_memory_blocks_st(instance);
+#else
+ return instance->threads == 1 ?
+ fill_memory_blocks_st(instance) : fill_memory_blocks_mt(instance);
+#endif
+}
+
+int validate_inputs(const argon2_context *context) {
+ if (NULL == context) {
+ return ARGON2_INCORRECT_PARAMETER;
+ }
+
+ if (NULL == context->out) {
+ return ARGON2_OUTPUT_PTR_NULL;
+ }
+
+ /* Validate output length */
+ if (ARGON2_MIN_OUTLEN > context->outlen) {
+ return ARGON2_OUTPUT_TOO_SHORT;
+ }
+
+ if (ARGON2_MAX_OUTLEN < context->outlen) {
+ return ARGON2_OUTPUT_TOO_LONG;
+ }
+
+ /* Validate password (required param) */
+ if (NULL == context->pwd) {
+ if (0 != context->pwdlen) {
+ return ARGON2_PWD_PTR_MISMATCH;
+ }
+ }
+
+ if (ARGON2_MIN_PWD_LENGTH > context->pwdlen) {
+ return ARGON2_PWD_TOO_SHORT;
+ }
+
+ if (ARGON2_MAX_PWD_LENGTH < context->pwdlen) {
+ return ARGON2_PWD_TOO_LONG;
+ }
+
+ /* Validate salt (required param) */
+ if (NULL == context->salt) {
+ if (0 != context->saltlen) {
+ return ARGON2_SALT_PTR_MISMATCH;
+ }
+ }
+
+ if (ARGON2_MIN_SALT_LENGTH > context->saltlen) {
+ return ARGON2_SALT_TOO_SHORT;
+ }
+
+ if (ARGON2_MAX_SALT_LENGTH < context->saltlen) {
+ return ARGON2_SALT_TOO_LONG;
+ }
+
+ /* Validate secret (optional param) */
+ if (NULL == context->secret) {
+ if (0 != context->secretlen) {
+ return ARGON2_SECRET_PTR_MISMATCH;
+ }
+ } else {
+ if (ARGON2_MIN_SECRET > context->secretlen) {
+ return ARGON2_SECRET_TOO_SHORT;
+ }
+ if (ARGON2_MAX_SECRET < context->secretlen) {
+ return ARGON2_SECRET_TOO_LONG;
+ }
+ }
+
+ /* Validate associated data (optional param) */
+ if (NULL == context->ad) {
+ if (0 != context->adlen) {
+ return ARGON2_AD_PTR_MISMATCH;
+ }
+ } else {
+ if (ARGON2_MIN_AD_LENGTH > context->adlen) {
+ return ARGON2_AD_TOO_SHORT;
+ }
+ if (ARGON2_MAX_AD_LENGTH < context->adlen) {
+ return ARGON2_AD_TOO_LONG;
+ }
+ }
+
+ /* Validate memory cost */
+ if (ARGON2_MIN_MEMORY > context->m_cost) {
+ return ARGON2_MEMORY_TOO_LITTLE;
+ }
+
+ if (ARGON2_MAX_MEMORY < context->m_cost) {
+ return ARGON2_MEMORY_TOO_MUCH;
+ }
+
+ if (context->m_cost < 8 * context->lanes) {
+ return ARGON2_MEMORY_TOO_LITTLE;
+ }
+
+ /* Validate time cost */
+ if (ARGON2_MIN_TIME > context->t_cost) {
+ return ARGON2_TIME_TOO_SMALL;
+ }
+
+ if (ARGON2_MAX_TIME < context->t_cost) {
+ return ARGON2_TIME_TOO_LARGE;
+ }
+
+ /* Validate lanes */
+ if (ARGON2_MIN_LANES > context->lanes) {
+ return ARGON2_LANES_TOO_FEW;
+ }
+
+ if (ARGON2_MAX_LANES < context->lanes) {
+ return ARGON2_LANES_TOO_MANY;
+ }
+
+ /* Validate threads */
+ if (ARGON2_MIN_THREADS > context->threads) {
+ return ARGON2_THREADS_TOO_FEW;
+ }
+
+ if (ARGON2_MAX_THREADS < context->threads) {
+ return ARGON2_THREADS_TOO_MANY;
+ }
+
+ if (NULL != context->allocate_cbk && NULL == context->free_cbk) {
+ return ARGON2_FREE_MEMORY_CBK_NULL;
+ }
+
+ if (NULL == context->allocate_cbk && NULL != context->free_cbk) {
+ return ARGON2_ALLOCATE_MEMORY_CBK_NULL;
+ }
+
+ return ARGON2_OK;
+}
+
+void fill_first_blocks(uint8_t *blockhash, const argon2_instance_t *instance) {
+ uint32_t l;
+ /* Make the first and second block in each lane as G(H0||0||i) or
+ G(H0||1||i) */
+ uint8_t blockhash_bytes[ARGON2_BLOCK_SIZE];
+ for (l = 0; l < instance->lanes; ++l) {
+
+ store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH, 0);
+ store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH + 4, l);
+ blake2b_long(blockhash_bytes, ARGON2_BLOCK_SIZE, blockhash,
+ ARGON2_PREHASH_SEED_LENGTH);
+ load_block(&instance->memory[l * instance->lane_length + 0],
+ blockhash_bytes);
+
+ store32(blockhash + ARGON2_PREHASH_DIGEST_LENGTH, 1);
+ blake2b_long(blockhash_bytes, ARGON2_BLOCK_SIZE, blockhash,
+ ARGON2_PREHASH_SEED_LENGTH);
+ load_block(&instance->memory[l * instance->lane_length + 1],
+ blockhash_bytes);
+ }
+ clear_internal_memory(blockhash_bytes, ARGON2_BLOCK_SIZE);
+}
+
+void initial_hash(uint8_t *blockhash, argon2_context *context,
+ argon2_type type) {
+ blake2b_state BlakeHash;
+ uint8_t value[sizeof(uint32_t)];
+
+ if (NULL == context || NULL == blockhash) {
+ return;
+ }
+
+ blake2b_init(&BlakeHash, ARGON2_PREHASH_DIGEST_LENGTH);
+
+ store32(&value, context->lanes);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, context->outlen);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, context->m_cost);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, context->t_cost);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, context->version);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, (uint32_t)type);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ store32(&value, context->pwdlen);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ if (context->pwd != NULL) {
+ blake2b_update(&BlakeHash, (const uint8_t *)context->pwd,
+ context->pwdlen);
+
+ if (context->flags & ARGON2_FLAG_CLEAR_PASSWORD) {
+ secure_wipe_memory(context->pwd, context->pwdlen);
+ context->pwdlen = 0;
+ }
+ }
+
+ store32(&value, context->saltlen);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ if (context->salt != NULL) {
+ blake2b_update(&BlakeHash, (const uint8_t *)context->salt,
+ context->saltlen);
+ }
+
+ store32(&value, context->secretlen);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ if (context->secret != NULL) {
+ blake2b_update(&BlakeHash, (const uint8_t *)context->secret,
+ context->secretlen);
+
+ if (context->flags & ARGON2_FLAG_CLEAR_SECRET) {
+ secure_wipe_memory(context->secret, context->secretlen);
+ context->secretlen = 0;
+ }
+ }
+
+ store32(&value, context->adlen);
+ blake2b_update(&BlakeHash, (const uint8_t *)&value, sizeof(value));
+
+ if (context->ad != NULL) {
+ blake2b_update(&BlakeHash, (const uint8_t *)context->ad,
+ context->adlen);
+ }
+
+ blake2b_final(&BlakeHash, blockhash, ARGON2_PREHASH_DIGEST_LENGTH);
+}
+
+int initialize(argon2_instance_t *instance, argon2_context *context) {
+ uint8_t blockhash[ARGON2_PREHASH_SEED_LENGTH];
+ int result = ARGON2_OK;
+
+ if (instance == NULL || context == NULL)
+ return ARGON2_INCORRECT_PARAMETER;
+ instance->context_ptr = context;
+
+ /* 1. Memory allocation */
+ result = allocate_memory(context, (uint8_t **)&(instance->memory),
+ instance->memory_blocks, sizeof(block));
+ if (result != ARGON2_OK) {
+ return result;
+ }
+
+ /* 2. Initial hashing */
+ /* H_0 + 8 extra bytes to produce the first blocks */
+ /* uint8_t blockhash[ARGON2_PREHASH_SEED_LENGTH]; */
+ /* Hashing all inputs */
+ initial_hash(blockhash, context, instance->type);
+ /* Zeroing 8 extra bytes */
+ clear_internal_memory(blockhash + ARGON2_PREHASH_DIGEST_LENGTH,
+ ARGON2_PREHASH_SEED_LENGTH -
+ ARGON2_PREHASH_DIGEST_LENGTH);
+
+#ifdef GENKAT
+ initial_kat(blockhash, context, instance->type);
+#endif
+
+ /* 3. Creating first blocks, we always have at least two blocks in a slice
+ */
+ fill_first_blocks(blockhash, instance);
+ /* Clearing the hash */
+ clear_internal_memory(blockhash, ARGON2_PREHASH_SEED_LENGTH);
+
+ return ARGON2_OK;
+}
diff --git a/argon2/core.h b/argon2/core.h
@@ -0,0 +1,228 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef ARGON2_CORE_H
+#define ARGON2_CORE_H
+
+#include "argon2.h"
+
+#define CONST_CAST(x) (x)(uintptr_t)
+
+/**********************Argon2 internal constants*******************************/
+
+enum argon2_core_constants {
+ /* Memory block size in bytes */
+ ARGON2_BLOCK_SIZE = 1024,
+ ARGON2_QWORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 8,
+ ARGON2_OWORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 16,
+ ARGON2_HWORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 32,
+ ARGON2_512BIT_WORDS_IN_BLOCK = ARGON2_BLOCK_SIZE / 64,
+
+ /* Number of pseudo-random values generated by one call to Blake in Argon2i
+ to
+ generate reference block positions */
+ ARGON2_ADDRESSES_IN_BLOCK = 128,
+
+ /* Pre-hashing digest length and its extension*/
+ ARGON2_PREHASH_DIGEST_LENGTH = 64,
+ ARGON2_PREHASH_SEED_LENGTH = 72
+};
+
+/*************************Argon2 internal data types***********************/
+
+/*
+ * Structure for the (1KB) memory block implemented as 128 64-bit words.
+ * Memory blocks can be copied, XORed. Internal words can be accessed by [] (no
+ * bounds checking).
+ */
+typedef struct block_ { uint64_t v[ARGON2_QWORDS_IN_BLOCK]; } block;
+
+/*****************Functions that work with the block******************/
+
+/* Initialize each byte of the block with @in */
+void init_block_value(block *b, uint8_t in);
+
+/* Copy block @src to block @dst */
+void copy_block(block *dst, const block *src);
+
+/* XOR @src onto @dst bytewise */
+void xor_block(block *dst, const block *src);
+
+/*
+ * Argon2 instance: memory pointer, number of passes, amount of memory, type,
+ * and derived values.
+ * Used to evaluate the number and location of blocks to construct in each
+ * thread
+ */
+typedef struct Argon2_instance_t {
+ block *memory; /* Memory pointer */
+ uint32_t version;
+ uint32_t passes; /* Number of passes */
+ uint32_t memory_blocks; /* Number of blocks in memory */
+ uint32_t segment_length;
+ uint32_t lane_length;
+ uint32_t lanes;
+ uint32_t threads;
+ argon2_type type;
+ int print_internals; /* whether to print the memory blocks */
+ argon2_context *context_ptr; /* points back to original context */
+} argon2_instance_t;
+
+/*
+ * Argon2 position: where we construct the block right now. Used to distribute
+ * work between threads.
+ */
+typedef struct Argon2_position_t {
+ uint32_t pass;
+ uint32_t lane;
+ uint8_t slice;
+ uint32_t index;
+} argon2_position_t;
+
+/*Struct that holds the inputs for thread handling FillSegment*/
+typedef struct Argon2_thread_data {
+ argon2_instance_t *instance_ptr;
+ argon2_position_t pos;
+} argon2_thread_data;
+
+/*************************Argon2 core functions********************************/
+
+/* Allocates memory to the given pointer, uses the appropriate allocator as
+ * specified in the context. Total allocated memory is num*size.
+ * @param context argon2_context which specifies the allocator
+ * @param memory pointer to the pointer to the memory
+ * @param size the size in bytes for each element to be allocated
+ * @param num the number of elements to be allocated
+ * @return ARGON2_OK if @memory is a valid pointer and memory is allocated
+ */
+int allocate_memory(const argon2_context *context, uint8_t **memory,
+ size_t num, size_t size);
+
+/*
+ * Frees memory at the given pointer, uses the appropriate deallocator as
+ * specified in the context. Also cleans the memory using clear_internal_memory.
+ * @param context argon2_context which specifies the deallocator
+ * @param memory pointer to buffer to be freed
+ * @param size the size in bytes for each element to be deallocated
+ * @param num the number of elements to be deallocated
+ */
+void free_memory(const argon2_context *context, uint8_t *memory,
+ size_t num, size_t size);
+
+/* Function that securely cleans the memory. This ignores any flags set
+ * regarding clearing memory. Usually one just calls clear_internal_memory.
+ * @param mem Pointer to the memory
+ * @param s Memory size in bytes
+ */
+void secure_wipe_memory(void *v, size_t n);
+
+/* Function that securely clears the memory if FLAG_clear_internal_memory is
+ * set. If the flag isn't set, this function does nothing.
+ * @param mem Pointer to the memory
+ * @param s Memory size in bytes
+ */
+void clear_internal_memory(void *v, size_t n);
+
+/*
+ * Computes absolute position of reference block in the lane following a skewed
+ * distribution and using a pseudo-random value as input
+ * @param instance Pointer to the current instance
+ * @param position Pointer to the current position
+ * @param pseudo_rand 32-bit pseudo-random value used to determine the position
+ * @param same_lane Indicates if the block will be taken from the current lane.
+ * If so we can reference the current segment
+ * @pre All pointers must be valid
+ */
+uint32_t index_alpha(const argon2_instance_t *instance,
+ const argon2_position_t *position, uint32_t pseudo_rand,
+ int same_lane);
+
+/*
+ * Function that validates all inputs against predefined restrictions and return
+ * an error code
+ * @param context Pointer to current Argon2 context
+ * @return ARGON2_OK if everything is all right, otherwise one of error codes
+ * (all defined in <argon2.h>
+ */
+int validate_inputs(const argon2_context *context);
+
+/*
+ * Hashes all the inputs into @a blockhash[PREHASH_DIGEST_LENGTH], clears
+ * password and secret if needed
+ * @param context Pointer to the Argon2 internal structure containing memory
+ * pointer, and parameters for time and space requirements.
+ * @param blockhash Buffer for pre-hashing digest
+ * @param type Argon2 type
+ * @pre @a blockhash must have at least @a PREHASH_DIGEST_LENGTH bytes
+ * allocated
+ */
+void initial_hash(uint8_t *blockhash, argon2_context *context,
+ argon2_type type);
+
+/*
+ * Function creates first 2 blocks per lane
+ * @param instance Pointer to the current instance
+ * @param blockhash Pointer to the pre-hashing digest
+ * @pre blockhash must point to @a PREHASH_SEED_LENGTH allocated values
+ */
+void fill_first_blocks(uint8_t *blockhash, const argon2_instance_t *instance);
+
+/*
+ * Function allocates memory, hashes the inputs with Blake, and creates first
+ * two blocks. Returns the pointer to the main memory with 2 blocks per lane
+ * initialized
+ * @param context Pointer to the Argon2 internal structure containing memory
+ * pointer, and parameters for time and space requirements.
+ * @param instance Current Argon2 instance
+ * @return Zero if successful, -1 if memory failed to allocate. @context->state
+ * will be modified if successful.
+ */
+int initialize(argon2_instance_t *instance, argon2_context *context);
+
+/*
+ * XORing the last block of each lane, hashing it, making the tag. Deallocates
+ * the memory.
+ * @param context Pointer to current Argon2 context (use only the out parameters
+ * from it)
+ * @param instance Pointer to current instance of Argon2
+ * @pre instance->state must point to necessary amount of memory
+ * @pre context->out must point to outlen bytes of memory
+ * @pre if context->free_cbk is not NULL, it should point to a function that
+ * deallocates memory
+ */
+void finalize(const argon2_context *context, argon2_instance_t *instance);
+
+/*
+ * Function that fills the segment using previous segments also from other
+ * threads
+ * @param context current context
+ * @param instance Pointer to the current instance
+ * @param position Current position
+ * @pre all block pointers must be valid
+ */
+void fill_segment(const argon2_instance_t *instance,
+ argon2_position_t position);
+
+/*
+ * Function that fills the entire memory t_cost times based on the first two
+ * blocks in each lane
+ * @param instance Pointer to the current instance
+ * @return ARGON2_OK if successful, @context->state
+ */
+int fill_memory_blocks(argon2_instance_t *instance);
+
+#endif
diff --git a/argon2/ref.c b/argon2/ref.c
@@ -0,0 +1,194 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#include <stdint.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include "argon2.h"
+#include "core.h"
+
+#include "blake2/blamka-round-ref.h"
+#include "blake2/blake2-impl.h"
+#include "blake2/blake2.h"
+
+
+/*
+ * Function fills a new memory block and optionally XORs the old block over the new one.
+ * @next_block must be initialized.
+ * @param prev_block Pointer to the previous block
+ * @param ref_block Pointer to the reference block
+ * @param next_block Pointer to the block to be constructed
+ * @param with_xor Whether to XOR into the new block (1) or just overwrite (0)
+ * @pre all block pointers must be valid
+ */
+static void fill_block(const block *prev_block, const block *ref_block,
+ block *next_block, int with_xor) {
+ block blockR, block_tmp;
+ unsigned i;
+
+ copy_block(&blockR, ref_block);
+ xor_block(&blockR, prev_block);
+ copy_block(&block_tmp, &blockR);
+ /* Now blockR = ref_block + prev_block and block_tmp = ref_block + prev_block */
+ if (with_xor) {
+ /* Saving the next block contents for XOR over: */
+ xor_block(&block_tmp, next_block);
+ /* Now blockR = ref_block + prev_block and
+ block_tmp = ref_block + prev_block + next_block */
+ }
+
+ /* Apply Blake2 on columns of 64-bit words: (0,1,...,15) , then
+ (16,17,..31)... finally (112,113,...127) */
+ for (i = 0; i < 8; ++i) {
+ BLAKE2_ROUND_NOMSG(
+ blockR.v[16 * i], blockR.v[16 * i + 1], blockR.v[16 * i + 2],
+ blockR.v[16 * i + 3], blockR.v[16 * i + 4], blockR.v[16 * i + 5],
+ blockR.v[16 * i + 6], blockR.v[16 * i + 7], blockR.v[16 * i + 8],
+ blockR.v[16 * i + 9], blockR.v[16 * i + 10], blockR.v[16 * i + 11],
+ blockR.v[16 * i + 12], blockR.v[16 * i + 13], blockR.v[16 * i + 14],
+ blockR.v[16 * i + 15]);
+ }
+
+ /* Apply Blake2 on rows of 64-bit words: (0,1,16,17,...112,113), then
+ (2,3,18,19,...,114,115).. finally (14,15,30,31,...,126,127) */
+ for (i = 0; i < 8; i++) {
+ BLAKE2_ROUND_NOMSG(
+ blockR.v[2 * i], blockR.v[2 * i + 1], blockR.v[2 * i + 16],
+ blockR.v[2 * i + 17], blockR.v[2 * i + 32], blockR.v[2 * i + 33],
+ blockR.v[2 * i + 48], blockR.v[2 * i + 49], blockR.v[2 * i + 64],
+ blockR.v[2 * i + 65], blockR.v[2 * i + 80], blockR.v[2 * i + 81],
+ blockR.v[2 * i + 96], blockR.v[2 * i + 97], blockR.v[2 * i + 112],
+ blockR.v[2 * i + 113]);
+ }
+
+ copy_block(next_block, &block_tmp);
+ xor_block(next_block, &blockR);
+}
+
+static void next_addresses(block *address_block, block *input_block,
+ const block *zero_block) {
+ input_block->v[6]++;
+ fill_block(zero_block, input_block, address_block, 0);
+ fill_block(zero_block, address_block, address_block, 0);
+}
+
+void fill_segment(const argon2_instance_t *instance,
+ argon2_position_t position) {
+ block *ref_block = NULL, *curr_block = NULL;
+ block address_block, input_block, zero_block;
+ uint64_t pseudo_rand, ref_index, ref_lane;
+ uint32_t prev_offset, curr_offset;
+ uint32_t starting_index;
+ uint32_t i;
+ int data_independent_addressing;
+
+ if (instance == NULL) {
+ return;
+ }
+
+ data_independent_addressing =
+ (instance->type == Argon2_i) ||
+ (instance->type == Argon2_id && (position.pass == 0) &&
+ (position.slice < ARGON2_SYNC_POINTS / 2));
+
+ if (data_independent_addressing) {
+ init_block_value(&zero_block, 0);
+ init_block_value(&input_block, 0);
+
+ input_block.v[0] = position.pass;
+ input_block.v[1] = position.lane;
+ input_block.v[2] = position.slice;
+ input_block.v[3] = instance->memory_blocks;
+ input_block.v[4] = instance->passes;
+ input_block.v[5] = instance->type;
+ }
+
+ starting_index = 0;
+
+ if ((0 == position.pass) && (0 == position.slice)) {
+ starting_index = 2; /* we have already generated the first two blocks */
+
+ /* Don't forget to generate the first block of addresses: */
+ if (data_independent_addressing) {
+ next_addresses(&address_block, &input_block, &zero_block);
+ }
+ }
+
+ /* Offset of the current block */
+ curr_offset = position.lane * instance->lane_length +
+ position.slice * instance->segment_length + starting_index;
+
+ if (0 == curr_offset % instance->lane_length) {
+ /* Last block in this lane */
+ prev_offset = curr_offset + instance->lane_length - 1;
+ } else {
+ /* Previous block */
+ prev_offset = curr_offset - 1;
+ }
+
+ for (i = starting_index; i < instance->segment_length;
+ ++i, ++curr_offset, ++prev_offset) {
+ /*1.1 Rotating prev_offset if needed */
+ if (curr_offset % instance->lane_length == 1) {
+ prev_offset = curr_offset - 1;
+ }
+
+ /* 1.2 Computing the index of the reference block */
+ /* 1.2.1 Taking pseudo-random value from the previous block */
+ if (data_independent_addressing) {
+ if (i % ARGON2_ADDRESSES_IN_BLOCK == 0) {
+ next_addresses(&address_block, &input_block, &zero_block);
+ }
+ pseudo_rand = address_block.v[i % ARGON2_ADDRESSES_IN_BLOCK];
+ } else {
+ pseudo_rand = instance->memory[prev_offset].v[0];
+ }
+
+ /* 1.2.2 Computing the lane of the reference block */
+ ref_lane = ((pseudo_rand >> 32)) % instance->lanes;
+
+ if ((position.pass == 0) && (position.slice == 0)) {
+ /* Can not reference other lanes yet */
+ ref_lane = position.lane;
+ }
+
+ /* 1.2.3 Computing the number of possible reference block within the
+ * lane.
+ */
+ position.index = i;
+ ref_index = index_alpha(instance, &position, pseudo_rand & 0xFFFFFFFF,
+ ref_lane == position.lane);
+
+ /* 2 Creating a new block */
+ ref_block =
+ instance->memory + instance->lane_length * ref_lane + ref_index;
+ curr_block = instance->memory + curr_offset;
+ if (ARGON2_VERSION_10 == instance->version) {
+ /* version 1.2.1 and earlier: overwrite, not XOR */
+ fill_block(instance->memory + prev_offset, ref_block, curr_block, 0);
+ } else {
+ if(0 == position.pass) {
+ fill_block(instance->memory + prev_offset, ref_block,
+ curr_block, 0);
+ } else {
+ fill_block(instance->memory + prev_offset, ref_block,
+ curr_block, 1);
+ }
+ }
+ }
+}
diff --git a/argon2/thread.c b/argon2/thread.c
@@ -0,0 +1,57 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#if !defined(ARGON2_NO_THREADS)
+
+#include "thread.h"
+#if defined(_WIN32)
+#include <windows.h>
+#endif
+
+int argon2_thread_create(argon2_thread_handle_t *handle,
+ argon2_thread_func_t func, void *args) {
+ if (NULL == handle || func == NULL) {
+ return -1;
+ }
+#if defined(_WIN32)
+ *handle = _beginthreadex(NULL, 0, func, args, 0, NULL);
+ return *handle != 0 ? 0 : -1;
+#else
+ return pthread_create(handle, NULL, func, args);
+#endif
+}
+
+int argon2_thread_join(argon2_thread_handle_t handle) {
+#if defined(_WIN32)
+ if (WaitForSingleObject((HANDLE)handle, INFINITE) == WAIT_OBJECT_0) {
+ return CloseHandle((HANDLE)handle) != 0 ? 0 : -1;
+ }
+ return -1;
+#else
+ return pthread_join(handle, NULL);
+#endif
+}
+
+void argon2_thread_exit(void) {
+#if defined(_WIN32)
+ _endthreadex(0);
+#else
+ pthread_exit(NULL);
+#endif
+}
+
+#endif /* ARGON2_NO_THREADS */
diff --git a/argon2/thread.h b/argon2/thread.h
@@ -0,0 +1,67 @@
+/*
+ * Argon2 reference source code package - reference C implementations
+ *
+ * Copyright 2015
+ * Daniel Dinu, Dmitry Khovratovich, Jean-Philippe Aumasson, and Samuel Neves
+ *
+ * You may use this work under the terms of a Creative Commons CC0 1.0
+ * License/Waiver or the Apache Public License 2.0, at your option. The terms of
+ * these licenses can be found at:
+ *
+ * - CC0 1.0 Universal : https://creativecommons.org/publicdomain/zero/1.0
+ * - Apache 2.0 : https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * You should have received a copy of both of these licenses along with this
+ * software. If not, they may be obtained at the above URLs.
+ */
+
+#ifndef ARGON2_THREAD_H
+#define ARGON2_THREAD_H
+
+#if !defined(ARGON2_NO_THREADS)
+
+/*
+ Here we implement an abstraction layer for the simpĺe requirements
+ of the Argon2 code. We only require 3 primitives---thread creation,
+ joining, and termination---so full emulation of the pthreads API
+ is unwarranted. Currently we wrap pthreads and Win32 threads.
+
+ The API defines 2 types: the function pointer type,
+ argon2_thread_func_t,
+ and the type of the thread handle---argon2_thread_handle_t.
+*/
+#if defined(_WIN32)
+#include <process.h>
+typedef unsigned(__stdcall *argon2_thread_func_t)(void *);
+typedef uintptr_t argon2_thread_handle_t;
+#else
+#include <pthread.h>
+typedef void *(*argon2_thread_func_t)(void *);
+typedef pthread_t argon2_thread_handle_t;
+#endif
+
+/* Creates a thread
+ * @param handle pointer to a thread handle, which is the output of this
+ * function. Must not be NULL.
+ * @param func A function pointer for the thread's entry point. Must not be
+ * NULL.
+ * @param args Pointer that is passed as an argument to @func. May be NULL.
+ * @return 0 if @handle and @func are valid pointers and a thread is successfully
+ * created.
+ */
+int argon2_thread_create(argon2_thread_handle_t *handle,
+ argon2_thread_func_t func, void *args);
+
+/* Waits for a thread to terminate
+ * @param handle Handle to a thread created with argon2_thread_create.
+ * @return 0 if @handle is a valid handle, and joining completed successfully.
+*/
+int argon2_thread_join(argon2_thread_handle_t handle);
+
+/* Terminate the current thread. Must be run inside a thread created by
+ * argon2_thread_create.
+*/
+void argon2_thread_exit(void);
+
+#endif /* ARGON2_NO_THREADS */
+#endif
diff --git a/pkcs5_pbkdf2.c b/pkcs5_pbkdf2.c
@@ -1,122 +0,0 @@
-/*-
- * Copyright (c) 2008 Damien Bergamini <damien.bergamini@free.fr>
- *
- * Permission to use, copy, modify, and distribute this software for any
- * purpose with or without fee is hereby granted, provided that the above
- * copyright notice and this permission notice appear in all copies.
- *
- * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
- * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
- * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
- * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
- * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
- */
-
-#include <sys/random.h>
-#include <sys/types.h>
-
-#include <string.h>
-#include <stdint.h>
-#include <stdlib.h>
-
-#include "common.h"
-#include "pkcs5_pbkdf2.h"
-#include "sha1.h"
-
-#define MINIMUM(a,b) (((a) < (b)) ? (a) : (b))
-
-/*
- * HMAC-SHA-1 (from RFC 2202).
- */
-static void
-hmac_sha1(const u_int8_t *text, size_t text_len, const u_int8_t *key,
- size_t key_len, u_int8_t digest[SHA1_DIGEST_LENGTH])
-{
- struct sha1 ctx;
- u_int8_t k_pad[SHA1_BLOCK_LENGTH];
- u_int8_t tk[SHA1_DIGEST_LENGTH];
- int i;
-
- if (key_len > SHA1_BLOCK_LENGTH) {
- sha1_init(&ctx);
- sha1_update(&ctx, key, key_len);
- sha1_sum(&ctx, tk);
-
- key = tk;
- key_len = SHA1_DIGEST_LENGTH;
- }
-
- bzero(k_pad, sizeof k_pad);
- bcopy(key, k_pad, key_len);
- for (i = 0; i < SHA1_BLOCK_LENGTH; i++)
- k_pad[i] ^= 0x36;
-
- sha1_init(&ctx);
- sha1_update(&ctx, k_pad, SHA1_BLOCK_LENGTH);
- sha1_update(&ctx, text, text_len);
- sha1_sum(&ctx, digest);
-
- bzero(k_pad, sizeof k_pad);
- bcopy(key, k_pad, key_len);
- for (i = 0; i < SHA1_BLOCK_LENGTH; i++)
- k_pad[i] ^= 0x5c;
-
- sha1_init(&ctx);
- sha1_update(&ctx, k_pad, SHA1_BLOCK_LENGTH);
- sha1_update(&ctx, digest, SHA1_DIGEST_LENGTH);
- sha1_sum(&ctx, digest);
-}
-
-/*
- * Password-Based Key Derivation Function 2 (PKCS #5 v2.0).
- * Code based on IEEE Std 802.11-2007, Annex H.4.2.
- */
-int
-pkcs5_pbkdf2(const char *pass, size_t pass_len, const uint8_t *salt,
- size_t salt_len, uint8_t *key, size_t key_len, unsigned int rounds)
-{
- uint8_t asalt[SALT_LEN + 4], obuf[SHA1_DIGEST_LENGTH];
- uint8_t d1[SHA1_DIGEST_LENGTH], d2[SHA1_DIGEST_LENGTH];
- unsigned int i, j;
- unsigned int count;
- size_t r;
-
- if (rounds < 1 || key_len == 0)
- goto bad;
-
- memcpy(asalt, salt, salt_len);
-
- for (count = 1; key_len > 0; count++) {
- asalt[salt_len + 0] = (count >> 24) & 0xff;
- asalt[salt_len + 1] = (count >> 16) & 0xff;
- asalt[salt_len + 2] = (count >> 8) & 0xff;
- asalt[salt_len + 3] = count & 0xff;
- hmac_sha1(asalt, salt_len + 4, pass, pass_len, d1);
- memcpy(obuf, d1, sizeof(obuf));
-
- for (i = 1; i < rounds; i++) {
- hmac_sha1(d1, sizeof(d1), pass, pass_len, d2);
- memcpy(d1, d2, sizeof(d1));
- for (j = 0; j < sizeof(obuf); j++)
- obuf[j] ^= d1[j];
- }
-
- r = MINIMUM(key_len, SHA1_DIGEST_LENGTH);
- memcpy(key, obuf, r);
- key += r;
- key_len -= r;
- };
- explicit_bzero(asalt, salt_len + 4);
- explicit_bzero(d1, sizeof(d1));
- explicit_bzero(d2, sizeof(d2));
- explicit_bzero(obuf, sizeof(obuf));
-
- return 0;
-
-bad:
- /* overwrite with random in case caller doesn't check return code */
- getrandom(key, key_len, 0);
- return -1;
-}
diff --git a/pkcs5_pbkdf2.h b/pkcs5_pbkdf2.h
@@ -1,18 +0,0 @@
-/*-
- * Copyright (c) 2008 Damien Bergamini <damien.bergamini@free.fr>
- *
- * Permission to use, copy, modify, and distribute this software for any
- * purpose with or without fee is hereby granted, provided that the above
- * copyright notice and this permission notice appear in all copies.
- *
- * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
- * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
- * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
- * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
- * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
- */
-
-int pkcs5_pbkdf2(const char *pass, size_t pass_len, const uint8_t *salt, size_t
- salt_len, uint8_t *key, size_t key_len, unsigned int rounds);
