Public/VeraCrypt
1
0
Fork 0
mirror of https://github.com/veracrypt/VeraCrypt.git synced 2026-06-09 22:36:59 -05:00
Code Activity

Windows: Add support for Argon2id as an alternative to PBKDF2 key derivation

Browse Source
This commit is contained in:
Mounir IDRASSI
2025-06-25 15:44:31 +09:00
parent 228129362a
commit 3c17b8ced2
35 changed files with 4609 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/Crypto/Argon2/include/argon2.h
+354
View File
@@ -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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#ifndef ARGON2_H
#define ARGON2_H
#include "Common/Tcdefs.h"
#if defined(__cplusplus)
extern "C" {
#endif
#ifdef _MSC_VER
#define uint8_t uint8
#define uint16_t uint16
#define uint32_t uint32
#define uint64_t uint64
#define UINT32_C(x) (x ## U)
#define UINT64_C(x) (x ## ULL)
#define CHAR_BIT 8
#define UINT32_MAX 0xffffffffui32
#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(1) /* VeraCrypt: modified to 1 to silence error on driver build */
#define ARGON2_MAX_PWD_LENGTH UINT32_C(0xFFFFFFFF)
/* Minimum and maximum associated data length in bytes */
#define ARGON2_MIN_AD_LENGTH UINT32_C(1) /* VeraCrypt: modified to 1 to silence error on driver build */
#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(1) /* VeraCrypt: modified to 1 to silence error on driver build */
#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, argon2_type type,
const uint32_t version);
/**
* 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);
/**
* 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);
/*===============================================================================
*
* Selftest functions for Argon2d/2i/2id (Version 1.3).
* @return 0 = Selftest OK / no errors.
* >=1 = Selftest failed / number of failed test cases.
*/
ARGON2_PUBLIC int argon2i_selftest(void); /* 12 test cases, up to 256 MiB
plus 1 Argon2d test case */
ARGON2_PUBLIC int argon2id_selftest(void); /* 12 test cases, up to 256 MiB */
#if defined(__cplusplus)
}
#endif
#endif
src/Crypto/Argon2/src/argon2.c
+284
View File
@@ -0,0 +1,284 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#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, 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 = TCalloc(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);
TCfree(out);
return result;
}
/* if raw hash requested, write it */
if (hash) {
memcpy(hash, out, hashlen);
}
clear_internal_memory(out, hashlen);
TCfree(out);
return ARGON2_OK;
}
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, Argon2_i, 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, Argon2_d, 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, Argon2_id,
ARGON2_VERSION_NUMBER);
}
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);
}
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";
}
}
src/Crypto/Argon2/src/blake2/blake2-impl.h
+153
View File
@@ -0,0 +1,153 @@
/*
* 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
#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
src/Crypto/Argon2/src/blake2/blake2.h
+89
View File
@@ -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
src/Crypto/Argon2/src/blake2/blake2b.c
+386
View File
@@ -0,0 +1,386 @@
/*
* 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 "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 */
src/Crypto/Argon2/src/blake2/blamka-round-opt.h
+471
View File
@@ -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 */
src/Crypto/Argon2/src/blake2/blamka-round-ref.h
+56
View File
@@ -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
src/Crypto/Argon2/src/core.c
+595
View File
@@ -0,0 +1,595 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#include "core.h"
#include "blake2/blake2.h"
#include "blake2/blake2-impl.h"
#define secure_wipe_memory(v, n) burn((v), (n))
/***************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 = TCalloc(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 {
TCfree(memory);
}
}
/* 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;
}
src/Crypto/Argon2/src/core.h
+230
View File
@@ -0,0 +1,230 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#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
src/Crypto/Argon2/src/opt_avx2.asm
+633
View File
@@ -0,0 +1,633 @@
; opt_avx2.asm - Optimized Argon2 implementation using AVX2 instructions
;
; Generated from opt_avx2.c using GCC ""gcc -S -O2 -DARGON2_NO_THREADS -masm=intel -mavx2"
; Then converted manually to YASM syntax, after omitting SEH and GCC-specific directives
; Target: Win64 (x64), No SEH prologue/epilogue, No PTR, YASM section syntax
;
; This was needed because Windows driver doesn't allow using AVX2 C intrinsic in kernel mode
;
; For use in VeraCrypt.
; Copyright (c) 2025 Mounir IDRASSI <mounir.idrassi@amcrypto.jp>
section .text align=16
fill_block:
sub rsp, 1160
vmovups [rsp+1040], xmm6
vmovups [rsp+1056], xmm7
vmovups [rsp+1072], xmm8
vmovups [rsp+1088], xmm9
vmovups [rsp+1104], xmm10
vmovups [rsp+1120], xmm11
vmovups [rsp+1136], xmm12
xor eax, eax
mov r11, rdx
lea rdx, [rsp+31]
mov r10, rcx
mov rcx, rdx
and rcx, -32
test r9d, r9d
je .L5
align 64
align 16
align 8
.L3:
vmovdqu ymm0, [r11+rax]
vpxor ymm0, ymm0, [r10+rax]
add rcx, 32
vmovdqu [r10+rax], ymm0
vpxor ymm0, ymm0, [r8+rax]
add rax, 32
vmovdqu [rcx-32], ymm0
cmp rax, 1024
jne .L3
.L4:
vmovdqu ymm2, [rel LC0]
mov rax, r10
mov rcx, r10
vmovdqu ymm3, [rel LC1]
lea r9, [r10+1024]
.L6:
vmovdqu ymm0, [rcx+32]
vmovdqu ymm1, [rcx]
add rcx, 256
vmovdqu ymm7, [rcx-128]
vpmuludq ymm4, ymm1, ymm0
vpaddq ymm1, ymm1, ymm0
vpaddq ymm4, ymm4, ymm4
vpaddq ymm4, ymm4, ymm1
vpxor ymm5, ymm4, [rcx-160]
vmovdqu ymm1, [rcx-192]
vpshufd ymm5, ymm5, 177
vpmuludq ymm8, ymm1, ymm5
vpaddq ymm1, ymm1, ymm5
vpaddq ymm8, ymm8, ymm8
vpaddq ymm8, ymm8, ymm1
vmovdqu ymm1, [rcx-96]
vpxor ymm0, ymm8, ymm0
vpmuludq ymm6, ymm7, ymm1
vpshufb ymm0, ymm0, ymm2
vpaddq ymm7, ymm7, ymm1
vpmuludq ymm9, ymm4, ymm0
vpaddq ymm4, ymm4, ymm0
vpaddq ymm6, ymm6, ymm6
vpaddq ymm6, ymm6, ymm7
vpxor ymm12, ymm6, [rcx-32]
vmovdqu ymm7, [rcx-64]
vpaddq ymm9, ymm9, ymm9
vpaddq ymm9, ymm9, ymm4
vpshufd ymm12, ymm12, 177
vpxor ymm5, ymm9, ymm5
vpmuludq ymm10, ymm7, ymm12
vpshufb ymm5, ymm5, ymm3
vpaddq ymm7, ymm7, ymm12
vpaddq ymm10, ymm10, ymm10
vpaddq ymm10, ymm10, ymm7
vpmuludq ymm7, ymm8, ymm5
vpaddq ymm8, ymm8, ymm5
vpxor ymm1, ymm10, ymm1
vpermq ymm5, ymm5, 147
vpshufb ymm1, ymm1, ymm2
vpaddq ymm7, ymm7, ymm7
vpaddq ymm7, ymm7, ymm8
vpmuludq ymm8, ymm6, ymm1
vpaddq ymm6, ymm6, ymm1
vpxor ymm0, ymm7, ymm0
vpermq ymm7, ymm7, 78
vpsrlq ymm11, ymm0, 63
vpaddq ymm0, ymm0, ymm0
vpxor ymm0, ymm0, ymm11
vpermq ymm0, ymm0, 57
vpaddq ymm8, ymm8, ymm8
vpaddq ymm8, ymm8, ymm6
vpxor ymm4, ymm8, ymm12
vpmuludq ymm12, ymm9, ymm0
vpaddq ymm9, ymm0, ymm9
vpshufb ymm4, ymm4, ymm3
vpmuludq ymm6, ymm10, ymm4
vpaddq ymm10, ymm10, ymm4
vpermq ymm4, ymm4, 147
vpaddq ymm12, ymm12, ymm12
vpaddq ymm12, ymm12, ymm9
vpaddq ymm6, ymm6, ymm6
vpxor ymm5, ymm5, ymm12
vpaddq ymm6, ymm6, ymm10
vpshufd ymm5, ymm5, 177
vpxor ymm1, ymm6, ymm1
vpmuludq ymm11, ymm7, ymm5
vpaddq ymm7, ymm7, ymm5
vpsrlq ymm10, ymm1, 63
vpaddq ymm1, ymm1, ymm1
vpermq ymm6, ymm6, 78
vpxor ymm1, ymm1, ymm10
vpermq ymm1, ymm1, 57
vpmuludq ymm10, ymm8, ymm1
vpaddq ymm8, ymm1, ymm8
vpaddq ymm11, ymm11, ymm11
vpaddq ymm11, ymm11, ymm7
vpxor ymm0, ymm0, ymm11
vpshufb ymm0, ymm0, ymm2
vpaddq ymm10, ymm10, ymm10
vpaddq ymm10, ymm10, ymm8
vpxor ymm4, ymm4, ymm10
vpshufd ymm9, ymm4, 177
vpmuludq ymm4, ymm12, ymm0
vpaddq ymm12, ymm12, ymm0
vpmuludq ymm7, ymm6, ymm9
vpaddq ymm6, ymm6, ymm9
vpaddq ymm4, ymm4, ymm4
vpaddq ymm7, ymm7, ymm7
vpaddq ymm4, ymm4, ymm12
vpaddq ymm7, ymm7, ymm6
vmovdqu [rcx-256], ymm4
vpxor ymm4, ymm4, ymm5
vpxor ymm1, ymm1, ymm7
vpshufb ymm6, ymm4, ymm3
vpshufb ymm1, ymm1, ymm2
vpmuludq ymm8, ymm11, ymm6
vpaddq ymm11, ymm11, ymm6
vpmuludq ymm4, ymm10, ymm1
vpaddq ymm10, ymm10, ymm1
vpaddq ymm8, ymm8, ymm8
vpaddq ymm4, ymm4, ymm4
vpaddq ymm8, ymm8, ymm11
vpaddq ymm4, ymm4, ymm10
vpxor ymm0, ymm8, ymm0
vpermq ymm8, ymm8, 78
vmovdqu [rcx-128], ymm4
vpxor ymm4, ymm4, ymm9
vpsrlq ymm11, ymm0, 63
vpaddq ymm0, ymm0, ymm0
vpshufb ymm5, ymm4, ymm3
vpxor ymm0, ymm0, ymm11
vmovdqu [rcx-192], ymm8
vpmuludq ymm4, ymm7, ymm5
vpermq ymm0, ymm0, 147
vmovdqu [rcx-224], ymm0
vpaddq ymm9, ymm4, ymm4
vpaddq ymm4, ymm7, ymm5
vpaddq ymm7, ymm9, ymm4
vpermq ymm4, ymm6, 57
vpxor ymm1, ymm7, ymm1
vmovdqu [rcx-160], ymm4
vpermq ymm7, ymm7, 78
vpermq ymm4, ymm5, 57
vpsrlq ymm9, ymm1, 63
vpaddq ymm1, ymm1, ymm1
vmovdqu [rcx-64], ymm7
vpxor ymm1, ymm1, ymm9
vmovdqu [rcx-32], ymm4
vpermq ymm1, ymm1, 147
vmovdqu [rcx-96], ymm1
cmp r9, rcx
jne .L6
lea rcx, [r10+128]
.L7:
vmovdqu ymm1, [rax+256]
vmovdqu ymm0, [rax]
add rax, 32
vmovdqu ymm6, [rax+96]
vpmuludq ymm5, ymm0, ymm1
vpaddq ymm0, ymm0, ymm1
vpaddq ymm5, ymm5, ymm5
vpaddq ymm5, ymm5, ymm0
vpxor ymm9, ymm5, [rax+736]
vmovdqu ymm0, [rax+480]
vpshufd ymm9, ymm9, 177
vpmuludq ymm4, ymm0, ymm9
vpaddq ymm0, ymm0, ymm9
vpaddq ymm4, ymm4, ymm4
vpaddq ymm4, ymm4, ymm0
vmovdqu ymm0, [rax+352]
vpxor ymm1, ymm4, ymm1
vpmuludq ymm8, ymm6, ymm0
vpaddq ymm6, ymm6, ymm0
vpshufb ymm1, ymm1, ymm2
vpaddq ymm7, ymm5, ymm1
vpaddq ymm8, ymm8, ymm8
vpaddq ymm8, ymm8, ymm6
vpxor ymm12, ymm8, [rax+864]
vmovdqu ymm6, [rax+608]
vpshufd ymm12, ymm12, 177
vpmuludq ymm10, ymm6, ymm12
vpaddq ymm6, ymm6, ymm12
vpaddq ymm10, ymm10, ymm10
vpaddq ymm10, ymm10, ymm6
vpmuludq ymm6, ymm5, ymm1
vpxor ymm0, ymm10, ymm0
vpshufb ymm0, ymm0, ymm2
vpaddq ymm6, ymm6, ymm6
vpaddq ymm6, ymm6, ymm7
vpmuludq ymm7, ymm8, ymm0
vpxor ymm5, ymm6, ymm9
vpshufb ymm5, ymm5, ymm3
vpmuludq ymm9, ymm4, ymm5
vpaddq ymm4, ymm4, ymm5
vpaddq ymm7, ymm7, ymm7
vpaddq ymm9, ymm9, ymm9
vpaddq ymm9, ymm9, ymm4
vpaddq ymm4, ymm8, ymm0
vpaddq ymm7, ymm7, ymm4
vpxor ymm1, ymm9, ymm1
vpxor ymm8, ymm7, ymm12
vpsrlq ymm11, ymm1, 63
vpaddq ymm1, ymm1, ymm1
vpshufb ymm8, ymm8, ymm3
vpxor ymm1, ymm1, ymm11
vpmuludq ymm4, ymm10, ymm8
vpaddq ymm10, ymm10, ymm8
vpaddq ymm4, ymm4, ymm4
vpaddq ymm4, ymm4, ymm10
vpxor ymm0, ymm4, ymm0
vpsrlq ymm10, ymm0, 63
vpaddq ymm0, ymm0, ymm0
vpxor ymm0, ymm0, ymm10
vpblendd ymm10, ymm5, ymm8, 204
vpblendd ymm5, ymm5, ymm8, 51
vpblendd ymm11, ymm1, ymm0, 204
vpblendd ymm1, ymm1, ymm0, 51
vpermq ymm10, ymm10, 177
vpermq ymm1, ymm1, 177
vpermq ymm11, ymm11, 177
vpermq ymm5, ymm5, 177
vpmuludq ymm0, ymm6, ymm1
vpaddq ymm6, ymm1, ymm6
vpaddq ymm0, ymm0, ymm0
vpaddq ymm6, ymm6, ymm0
vpxor ymm10, ymm10, ymm6
vpshufd ymm10, ymm10, 177
vpmuludq ymm8, ymm4, ymm10
vpaddq ymm4, ymm4, ymm10
vpaddq ymm8, ymm8, ymm8
vpaddq ymm8, ymm8, ymm4
vpmuludq ymm4, ymm7, ymm11
vpaddq ymm7, ymm11, ymm7
vpxor ymm1, ymm8, ymm1
vpshufb ymm0, ymm1, ymm2
vpaddq ymm4, ymm4, ymm4
vpaddq ymm4, ymm4, ymm7
vpxor ymm5, ymm5, ymm4
vpshufd ymm7, ymm5, 177
vpmuludq ymm1, ymm9, ymm7
vpaddq ymm5, ymm1, ymm1
vpaddq ymm1, ymm9, ymm7
vpaddq ymm9, ymm5, ymm1
vpmuludq ymm5, ymm6, ymm0
vpaddq ymm6, ymm6, ymm0
vpxor ymm1, ymm9, ymm11
vpshufb ymm1, ymm1, ymm2
vpaddq ymm5, ymm5, ymm5
vpaddq ymm5, ymm5, ymm6
vmovdqu [rax-32], ymm5
vpxor ymm5, ymm5, ymm10
vpmuludq ymm10, ymm4, ymm1
vpaddq ymm4, ymm4, ymm1
vpshufb ymm5, ymm5, ymm3
vpmuludq ymm6, ymm8, ymm5
vpaddq ymm8, ymm8, ymm5
vpaddq ymm10, ymm10, ymm10
vpaddq ymm4, ymm4, ymm10
vmovdqu [rax+96], ymm4
vpxor ymm4, ymm4, ymm7
vpaddq ymm6, ymm6, ymm6
vpshufb ymm7, ymm4, ymm3
vpaddq ymm6, ymm6, ymm8
vpmuludq ymm4, ymm9, ymm7
vpaddq ymm9, ymm9, ymm7
vpxor ymm0, ymm6, ymm0
vmovdqu [rax+608], ymm6
vpsrlq ymm8, ymm0, 63
vpaddq ymm0, ymm0, ymm0
vpxor ymm0, ymm0, ymm8
vpaddq ymm4, ymm4, ymm4
vpaddq ymm4, ymm4, ymm9
vpxor ymm1, ymm4, ymm1
vmovdqu [rax+480], ymm4
vpsrlq ymm9, ymm1, 63
vpaddq ymm1, ymm1, ymm1
vpxor ymm1, ymm1, ymm9
vpblendd ymm8, ymm0, ymm1, 204
vpblendd ymm0, ymm0, ymm1, 51
vpermq ymm0, ymm0, 177
vpermq ymm1, ymm8, 177
vmovdqu [rax+352], ymm0
vpblendd ymm0, ymm5, ymm7, 51
vpblendd ymm5, ymm5, ymm7, 204
vpermq ymm0, ymm0, 177
vpermq ymm5, ymm5, 177
vmovdqu [rax+224], ymm1
vmovdqu [rax+736], ymm0
vmovdqu [rax+864], ymm5
cmp rax, rcx
jne .L7
and rdx, -32
xor eax, eax
align 64
align 16
align 8
.L8:
vmovdqu ymm0, [r10+rax]
vpxor ymm0, ymm0, [rdx]
add rdx, 32
vmovdqu [r10+rax], ymm0
vmovdqu [r8+rax], ymm0
add rax, 32
cmp rax, 1024
jne .L8
vzeroupper
vmovups xmm6, [rsp+1040]
vmovups xmm7, [rsp+1056]
vmovups xmm8, [rsp+1072]
vmovups xmm9, [rsp+1088]
vmovups xmm10, [rsp+1104]
vmovups xmm11, [rsp+1120]
vmovups xmm12, [rsp+1136]
add rsp, 1160
ret
align 64
align 16
align 8
.L5:
vmovdqu ymm0, [r11+rax]
vpxor ymm0, ymm0, [r10+rax]
add rcx, 32
vmovdqu [r10+rax], ymm0
add rax, 32
vmovdqu [rcx-32], ymm0
cmp rax, 1024
jne .L5
jmp .L4
next_addresses:
push rdi
push rbx
sub rsp, 2104
xor eax, eax
xor r9d, r9d
add qword [rdx+48], 1
lea rbx, [rsp+63]
mov r8, rcx
mov ecx, 128
and rbx, -32
lea rdi, [rbx+1024]
rep stosq
mov rdi, rbx
mov ecx, 128
rep stosq
lea rcx, [rbx+1024]
call fill_block
xor r9d, r9d
mov rdx, r8
mov rcx, rbx
call fill_block
nop
add rsp, 2104
pop rbx
pop rdi
ret
global fill_segment_avx2
fill_segment_avx2:
push r15
push r14
push r13
push r12
push rbp
push rdi
push rsi
push rbx
sub rsp, 3160
vmovdqu xmm1, [rdx]
lea r14, [rsp+2143]
mov rbx, rcx
vmovdqu [rsp+48], xmm1
and r14, -32
test rcx, rcx
je .L44
mov edx, dword [rcx+36]
cmp edx, 1
je .L18
mov r12d, dword [rsp+48]
movzx eax, byte [rsp+56]
cmp edx, 2
je .L19
mov ebp, dword [rsp+52]
test r12d, r12d
jne .L37
xor r12d, r12d
test al, al
sete r12b
xor r15d, r15d
add r12d, r12d
.L20:
mov r8d, dword [rbx+24]
mov r9d, dword [rbx+20]
xor edx, edx
mov ecx, 128
mov rdi, r14
imul ebp, r8d
imul eax, r9d
add ebp, r12d
add ebp, eax
mov eax, ebp
lea r13d, [rbp-1]
div r8d
lea eax, [rbp+r8-1]
test edx, edx
cmove r13d, eax
mov esi, r13d
sal rsi, 10
add rsi, qword [rbx]
rep movsq
cmp r12d, r9d
jnb .L44
lea rax, [rsp+1088]
mov qword [rsp+32], rax
lea rax, [rsp+64]
mov qword [rsp+40], rax
jmp .L35
align 16
align 8
.L46:
cmp byte [rsp+56], 0
jne .L30
mov rdi, rcx
mov r9d, 1
.L31:
lea rdx, [rsp+48]
mov rcx, rbx
mov dword [rsp+60], r12d
call index_alpha
mov edx, dword [rbx+24]
mov r8, qword [rbx]
mov eax, eax
imul rdx, rdi
add rdx, rax
mov eax, ebp
sal rdx, 10
sal rax, 10
add rdx, r8
add r8, rax
cmp dword [rbx+8], 16
je .L32
mov eax, dword [rsp+48]
test eax, eax
je .L32
mov r9d, 1
mov rcx, r14
call fill_block
.L34:
add r12d, 1
cmp r12d, dword [rbx+20]
jnb .L44
mov r8d, dword [rbx+24]
add ebp, 1
.L35:
mov eax, ebp
xor edx, edx
div r8d
cmp edx, 1
je .L25
mov eax, r13d
add r13d, 1
.L26:
test r15d, r15d
je .L27
mov edi, r12d
and edi, 127
je .L45
.L28:
mov eax, edi
mov r8, qword [rsp+64+rax*8]
.L29:
mov edx, dword [rsp+48]
mov ecx, dword [rsp+52]
test edx, edx
je .L46
.L30:
mov edi, dword [rbx+28]
mov rax, r8
xor edx, edx
xor r9d, r9d
shr rax, 32
div rdi
cmp rdx, rcx
mov rdi, rdx
sete r9b
jmp .L31
align 16
align 8
.L44:
add rsp, 3160
pop rbx
pop rsi
pop rdi
pop rbp
pop r12
pop r13
pop r14
pop r15
ret
align 16
align 8
.L19:
test r12d, r12d
jne .L47
cmp al, 1
jbe .L18
mov ebp, dword [rsp+52]
xor r15d, r15d
jmp .L20
align 16
align 8
.L32:
xor r9d, r9d
mov rcx, r14
call fill_block
jmp .L34
align 16
align 8
.L27:
sal rax, 10
add rax, qword [rbx]
mov r8, qword [rax]
jmp .L29
align 16
align 8
.L25:
lea eax, [rbp-1]
mov r13d, ebp
jmp .L26
align 16
align 8
.L45:
mov rdx, qword [rsp+32]
mov rcx, qword [rsp+40]
call next_addresses
jmp .L28
align 16
align 8
.L37:
xor r15d, r15d
xor r12d, r12d
jmp .L20
align 16
align 8
.L18:
xor edx, edx
lea rcx, [rsp+1088]
mov r15d, 1
call init_block_value
mov eax, dword [rsp+48]
vmovd xmm2, dword [rbx+12]
vpinsrd xmm0, xmm2, dword [rbx+36], 1
movzx edx, byte [rsp+56]
mov qword [rsp+1088], rax
mov ecx, dword [rbx+16]
mov r12, rax
mov eax, dword [rsp+52]
vpmovzxdq xmm0, xmm0
mov qword [rsp+1104], rdx
mov qword [rsp+1112], rcx
mov qword [rsp+1096], rax
mov rbp, rax
mov rax, rdx
vmovdqu [rsp+1120], xmm0
test r12d, r12d
jne .L38
test dl, dl
jne .L20
lea rcx, [rsp+64]
lea rdx, [rsp+1088]
mov r12d, 2
call next_addresses
mov ebp, dword [rsp+52]
movzx eax, byte [rsp+56]
jmp .L20
align 16
align 8
.L47:
mov ebp, dword [rsp+52]
xor r15d, r15d
xor r12d, r12d
jmp .L20
align 16
align 8
.L38:
xor r12d, r12d
jmp .L20
section .rdata align=32
LC0:
db 3,4,5,6,7,0,1,2,11,12,13,14,15,8,9,10
db 3,4,5,6,7,0,1,2,11,12,13,14,15,8,9,10
align 32
LC1:
db 2,3,4,5,6,7,0,1,10,11,12,13,14,15,8,9
db 2,3,4,5,6,7,0,1,10,11,12,13,14,15,8,9
; External functions
extern index_alpha
extern init_block_value
; end of file
src/Crypto/Argon2/src/opt_avx2.c
+200
View File
@@ -0,0 +1,200 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#include "argon2.h"
#include "core.h"
#include "blake2/blake2.h"
#include "blake2/blamka-round-opt.h"
/*
* Function fills a new memory block and optionally XORs the old block over the new one.
* Memory must be initialized.
* @param state Pointer to the just produced block. Content will be updated(!)
* @param ref_block Pointer to the reference block
* @param next_block Pointer to the block to be XORed over. May coincide with @ref_block
* @param with_xor Whether to XOR into the new block (1) or just overwrite (0)
* @pre all block pointers must be valid
*/
#if defined(__AVX2__)
static void fill_block(__m256i *state, const block *ref_block,
block *next_block, int with_xor) {
__m256i block_XY[ARGON2_HWORDS_IN_BLOCK];
unsigned int i;
if (with_xor) {
for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
state[i] = _mm256_xor_si256(
state[i], _mm256_loadu_si256((const __m256i *)ref_block->v + i));
block_XY[i] = _mm256_xor_si256(
state[i], _mm256_loadu_si256((const __m256i *)next_block->v + i));
}
} else {
for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
block_XY[i] = state[i] = _mm256_xor_si256(
state[i], _mm256_loadu_si256((const __m256i *)ref_block->v + i));
}
}
for (i = 0; i < 4; ++i) {
BLAKE2_ROUND_1(state[8 * i + 0], state[8 * i + 4], state[8 * i + 1], state[8 * i + 5],
state[8 * i + 2], state[8 * i + 6], state[8 * i + 3], state[8 * i + 7]);
}
for (i = 0; i < 4; ++i) {
BLAKE2_ROUND_2(state[ 0 + i], state[ 4 + i], state[ 8 + i], state[12 + i],
state[16 + i], state[20 + i], state[24 + i], state[28 + i]);
}
for (i = 0; i < ARGON2_HWORDS_IN_BLOCK; i++) {
state[i] = _mm256_xor_si256(state[i], block_XY[i]);
_mm256_storeu_si256((__m256i *)next_block->v + i, state[i]);
}
}
static void next_addresses(block *address_block, block *input_block) {
/*Temporary zero-initialized blocks*/
__m256i zero_block[ARGON2_HWORDS_IN_BLOCK];
__m256i zero2_block[ARGON2_HWORDS_IN_BLOCK];
memset(zero_block, 0, sizeof(zero_block));
memset(zero2_block, 0, sizeof(zero2_block));
/*Increasing index counter*/
input_block->v[6]++;
/*First iteration of G*/
fill_block(zero_block, input_block, address_block, 0);
/*Second iteration of G*/
fill_block(zero2_block, address_block, address_block, 0);
}
void fill_segment_avx2(const argon2_instance_t *instance,
argon2_position_t position) {
block *ref_block = NULL, *curr_block = NULL;
block address_block, input_block;
uint64_t pseudo_rand, ref_index, ref_lane;
uint32_t prev_offset, curr_offset;
uint32_t starting_index, i;
__m256i state[ARGON2_HWORDS_IN_BLOCK];
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(&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);
}
}
/* 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;
}
memcpy(state, ((instance->memory + prev_offset)->v), ARGON2_BLOCK_SIZE);
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);
}
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(state, ref_block, curr_block, 0);
} else {
if(0 == position.pass) {
fill_block(state, ref_block, curr_block, 0);
} else {
fill_block(state, ref_block, curr_block, 1);
}
}
}
}
#else
void fill_segment_avx2(const argon2_instance_t* instance,
argon2_position_t position) {
UNREFERENCED_PARAMETER(instance);
UNREFERENCED_PARAMETER(position);
}
#endif
src/Crypto/Argon2/src/opt_sse2.c
+202
View File
@@ -0,0 +1,202 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#include "argon2.h"
#include "core.h"
#include "blake2/blake2.h"
#include "blake2/blamka-round-opt.h"
#include "Crypto/config.h"
#if CRYPTOPP_BOOL_SSE2_INTRINSICS_AVAILABLE
/*
* Function fills a new memory block and optionally XORs the old block over the new one.
* Memory must be initialized.
* @param state Pointer to the just produced block. Content will be updated(!)
* @param ref_block Pointer to the reference block
* @param next_block Pointer to the block to be XORed over. May coincide with @ref_block
* @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(__m128i *state, const block *ref_block,
block *next_block, int with_xor) {
__m128i block_XY[ARGON2_OWORDS_IN_BLOCK];
unsigned int i;
if (with_xor) {
for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
state[i] = _mm_xor_si128(
state[i], _mm_loadu_si128((const __m128i *)ref_block->v + i));
block_XY[i] = _mm_xor_si128(
state[i], _mm_loadu_si128((const __m128i *)next_block->v + i));
}
} else {
for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
block_XY[i] = state[i] = _mm_xor_si128(
state[i], _mm_loadu_si128((const __m128i *)ref_block->v + i));
}
}
for (i = 0; i < 8; ++i) {
BLAKE2_ROUND(state[8 * i + 0], state[8 * i + 1], state[8 * i + 2],
state[8 * i + 3], state[8 * i + 4], state[8 * i + 5],
state[8 * i + 6], state[8 * i + 7]);
}
for (i = 0; i < 8; ++i) {
BLAKE2_ROUND(state[8 * 0 + i], state[8 * 1 + i], state[8 * 2 + i],
state[8 * 3 + i], state[8 * 4 + i], state[8 * 5 + i],
state[8 * 6 + i], state[8 * 7 + i]);
}
for (i = 0; i < ARGON2_OWORDS_IN_BLOCK; i++) {
state[i] = _mm_xor_si128(state[i], block_XY[i]);
_mm_storeu_si128((__m128i *)next_block->v + i, state[i]);
}
}
static void next_addresses(block *address_block, block *input_block) {
/*Temporary zero-initialized blocks*/
__m128i zero_block[ARGON2_OWORDS_IN_BLOCK];
__m128i zero2_block[ARGON2_OWORDS_IN_BLOCK];
memset(zero_block, 0, sizeof(zero_block));
memset(zero2_block, 0, sizeof(zero2_block));
/*Increasing index counter*/
input_block->v[6]++;
/*First iteration of G*/
fill_block(zero_block, input_block, address_block, 0);
/*Second iteration of G*/
fill_block(zero2_block, address_block, address_block, 0);
}
void fill_segment_sse2(const argon2_instance_t *instance,
argon2_position_t position) {
block *ref_block = NULL, *curr_block = NULL;
block address_block, input_block;
uint64_t pseudo_rand, ref_index, ref_lane;
uint32_t prev_offset, curr_offset;
uint32_t starting_index, i;
__m128i state[ARGON2_OWORDS_IN_BLOCK];
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(&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);
}
}
/* 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;
}
memcpy(state, ((instance->memory + prev_offset)->v), ARGON2_BLOCK_SIZE);
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);
}
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(state, ref_block, curr_block, 0);
} else {
if(0 == position.pass) {
fill_block(state, ref_block, curr_block, 0);
} else {
fill_block(state, ref_block, curr_block, 1);
}
}
}
}
#else
void fill_segment_sse2(const argon2_instance_t* instance,
argon2_position_t position) {
}
#endif
src/Crypto/Argon2/src/ref.c
+221
View File
@@ -0,0 +1,221 @@
/*
* 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.
*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#include "argon2.h"
#include "core.h"
#include "blake2/blamka-round-ref.h"
#include "blake2/blake2-impl.h"
#include "blake2/blake2.h"
#include "Crypto/config.h"
#include "Crypto/cpu.h"
#include "Crypto/misc.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_ref(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);
}
}
}
}
#if CRYPTOPP_BOOL_X64 || CRYPTOPP_BOOL_X86 || CRYPTOPP_BOOL_X32
extern void fill_segment_sse2(const argon2_instance_t* instance,
argon2_position_t position);
extern void fill_segment_avx2(const argon2_instance_t* instance,
argon2_position_t position);
#endif
void fill_segment(const argon2_instance_t* instance,
argon2_position_t position) {
#if CRYPTOPP_BOOL_X64 || CRYPTOPP_BOOL_X86 || CRYPTOPP_BOOL_X32
if (HasSAVX2())
{
fill_segment_avx2(instance, position);
}
else if (HasSSE2())
{
fill_segment_sse2(instance, position);
}
else
#endif
{
fill_segment_ref(instance, position);
}
}
src/Crypto/Argon2/src/selftest.c
+263
View File
@@ -0,0 +1,263 @@
/*
* 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.
*/
/*================================================================================*/
/* MODIFICATIONS by Alfred J. Baier (2018-2024) */
/* */
/* Selftests for Argon2d/2i/2id (Version 1.3) using test vectors from */
/* http://tools.ietf.org/id/draft-irtf-cfrg-argon2-03.html#section-6 */
/* Now: https://datatracker.ietf.org/doc/rfc9106/ Section 5. */
/* See also: */
/* https://github.com/RustCrypto/password-hashes/blob/master/argon2/tests/kat.rs */
/* */
/* Return values: 0 = Selftest OK / no error. */
/* >=1 = Selftest failed / number of failed test cases. */
/* */
/* Note: The selftest functions verify the correctness of Argon2 hashing on */
/* the current host processor (using the selected CPU extended features and */
/* optimizations) for various parameters and input values. */
/* They do, however, not provide a comprehensive test for all Argon2 functions, */
/* programme paths, and error cases. */
/*================================================================================*/
/* Modified for VeraCrypt integration - June 2025 by Mounir IDRASSI */
#include "argon2.h"
#undef TEST_LARGE_RAM /* Additional Argon2i test with 1 GiB */
/*==============================================================================
* Basic selftest for Argon2d/2i/2id (Version 1.3) using test vectors from
* http://tools.ietf.org/id/draft-irtf-cfrg-argon2-03.html#section-6
* Now: https://datatracker.ietf.org/doc/rfc9106/ Sections 5.1 - 5.3.
*
* Uses argon2_ctx() with memory cost of 32 Bytes.
* Runs 3 test cases.
* Returns number of failed test cases.
*/
static int argon2_selftest_ctx (void)
{
int i, err = 0;
argon2_context context;
uint8_t hash[32], pwd[32], salt[16], sec[8], ad[12];
for (i=0; i<32; i++) pwd[i] = 0x01; /* Fill test vectors */
for (i=0; i<16; i++) salt[i] = 0x02;
for (i=0; i< 8; i++) sec[i] = 0x03;
for (i=0; i<12; i++) ad[i] = 0x04;
static const uint8_t ref0d[32] = /* Reference hash results */
{0x51,0x2b,0x39,0x1b,0x6f,0x11,0x62,0x97,0x53,0x71,0xd3,0x09,0x19,0x73,0x42,0x94,
0xf8,0x68,0xe3,0xbe,0x39,0x84,0xf3,0xc1,0xa1,0x3a,0x4d,0xb9,0xfa,0xbe,0x4a,0xcb};
static const uint8_t ref0i[32] =
{0xc8,0x14,0xd9,0xd1,0xdc,0x7f,0x37,0xaa,0x13,0xf0,0xd7,0x7f,0x24,0x94,0xbd,0xa1,
0xc8,0xde,0x6b,0x01,0x6d,0xd3,0x88,0xd2,0x99,0x52,0xa4,0xc4,0x67,0x2b,0x6c,0xe8};
static const uint8_t ref0id[32] =
{0x0d,0x64,0x0d,0xf5,0x8d,0x78,0x76,0x6c,0x08,0xc0,0x37,0xa3,0x4a,0x8b,0x53,0xc9,
0xd0,0x1e,0xf0,0x45,0x2d,0x75,0xb6,0x5e,0xb5,0x25,0x20,0xe9,0x6b,0x01,0xe6,0x59};
context.out = (uint8_t*) hash; /* Init Argon2 context */
context.outlen = 32;
context.pwd = (uint8_t*) pwd;
context.pwdlen = 32;
context.salt = (uint8_t*) salt;
context.saltlen = 16;
context.secret = (uint8_t*) sec;
context.secretlen = 8;
context.ad = (uint8_t*) ad;
context.adlen = 12;
context.t_cost = 3;
context.m_cost = 32;
context.lanes = 4;
context.threads = 4;
context.allocate_cbk = NULL;
context.free_cbk = NULL;
context.flags = ARGON2_DEFAULT_FLAGS;
context.version = ARGON2_VERSION_13;
/* Test execution for Argon2d, Argon2i, Argon2id */
if (argon2_ctx (&context, Argon2_d) == ARGON2_OK &&
memcmp (hash, ref0d, 32) != 0) err++;
if (argon2_ctx (&context, Argon2_i) == ARGON2_OK &&
memcmp (hash, ref0i, 32) != 0) err++;
if (argon2_ctx (&context, Argon2_id) == ARGON2_OK &&
memcmp (hash, ref0id, 32) != 0) err++;
return err;
}
/*==============================================================================
* Hash test helper function for Argon2 (Version 1.3).
* Calls argon2_hash().
*/
static int hashtest (uint32_t t, uint32_t m, uint32_t p, const char *pwd,
const char *salt, argon2_type type, const uint8_t *ref)
{
uint8_t out[32];
return (argon2_hash(t, 1 << m, p, pwd, strlen (pwd), salt, strlen (salt),
out, 32, type, ARGON2_VERSION_NUMBER) == ARGON2_OK &&
memcmp (out, ref, 32) == 0) ? 0 : 1;
}
/*==============================================================================
* Selftest function for Argon2i (Version 1.3). Includes 1 Argon2d test case.
* Using test vectors from Argon2 reference source code package ("test.c")
* and from https://tools.ietf.org/id/draft-irtf-cfrg-argon2-03.html#section-5.6
* See also:
* https://github.com/RustCrypto/password-hashes/blob/master/argon2/tests/kat.rs
*
* Uses memory costs of 32 Bytes, 256 KiB, 4 MiB, 64 MiB, 256 MiB.
* Runs 13 test cases in total + 1 optional large memory test (1 GiB).
* Returns number of failed test cases.
*/
int argon2i_selftest (void)
{
int err = argon2_selftest_ctx (); /* Basic test using argon2_ctx() */
static const uint8_t ref1i[32] =
{0x89,0xe9,0x02,0x9f,0x46,0x37,0xb2,0x95,0xbe,0xb0,0x27,0x05,0x6a,0x73,0x36,0xc4,
0x14,0xfa,0xdd,0x43,0xf6,0xb2,0x08,0x64,0x52,0x81,0xcb,0x21,0x4a,0x56,0x45,0x2f};
err += hashtest(2, 8, 1, "password", "somesalt", Argon2_i, ref1i);
static const uint8_t ref2i[32] =
{0x4f,0xf5,0xce,0x27,0x69,0xa1,0xd7,0xf4,0xc8,0xa4,0x91,0xdf,0x09,0xd4,0x1a,0x9f,
0xbe,0x90,0xe5,0xeb,0x02,0x15,0x5a,0x13,0xe4,0xc0,0x1e,0x20,0xcd,0x4e,0xab,0x61};
err += hashtest(2, 8, 2, "password", "somesalt", Argon2_i, ref2i);
static const uint8_t ref3i[32] =
{0x95,0x7f,0xc0,0x72,0x7d,0x83,0xf4,0x06,0x0b,0xb0,0xf1,0x07,0x1e,0xb5,0x90,0xa1,
0x9a,0x8c,0x44,0x8f,0xc0,0x20,0x94,0x97,0xee,0x4f,0x54,0xca,0x24,0x1f,0x3c,0x90};
err += hashtest (3, 12, 1, "pasword", "somesalt", Argon2_i, ref3i);
static const uint8_t ref4d[32] =
{0x0b,0x3f,0x09,0xe7,0xb8,0xd0,0x36,0xe5,0x8c,0xcd,0x08,0xf0,0x8c,0xb6,0xba,0xbf,
0x7e,0x5e,0x24,0x63,0xc2,0x6b,0xcf,0x2a,0x9e,0x4e,0xa7,0x0d,0x74,0x7c,0x40,0x98};
err += hashtest (3, 12, 1, "pasword", "somesalt", Argon2_d, ref4d);
static const uint8_t ref5i[32] =
{0xd1,0x68,0x07,0x5c,0x4d,0x98,0x5e,0x13,0xeb,0xea,0xe5,0x60,0xcf,0x8b,0x94,0xc3,
0xb5,0xd8,0xa1,0x6c,0x51,0x91,0x6b,0x6f,0x4a,0xc2,0xda,0x3a,0xc1,0x1b,0xbe,0xcf};
err += hashtest (1, 16, 1, "password", "somesalt", Argon2_i, ref5i);
static const uint8_t ref6i[32] =
{0xc1,0x62,0x88,0x32,0x14,0x7d,0x97,0x20,0xc5,0xbd,0x1c,0xfd,0x61,0x36,0x70,0x78,
0x72,0x9f,0x6d,0xfb,0x6f,0x8f,0xea,0x9f,0xf9,0x81,0x58,0xe0,0xd7,0x81,0x6e,0xd0};
err += hashtest (2, 16, 1, "password", "somesalt", Argon2_i, ref6i);
static const uint8_t ref7i[32] =
{0x14,0xae,0x8d,0xa0,0x1a,0xfe,0xa8,0x70,0x0c,0x23,0x58,0xdc,0xef,0x7c,0x53,0x58,
0xd9,0x02,0x12,0x82,0xbd,0x88,0x66,0x3a,0x45,0x62,0xf5,0x9f,0xb7,0x4d,0x22,0xee};
err += hashtest (2, 16, 1, "differentpassword", "somesalt", Argon2_i, ref7i);
static const uint8_t ref8i[32] =
{0xb0,0x35,0x7c,0xcc,0xfb,0xef,0x91,0xf3,0x86,0x0b,0x0d,0xba,0x44,0x7b,0x23,0x48,
0xcb,0xef,0xec,0xad,0xaf,0x99,0x0a,0xbf,0xe9,0xcc,0x40,0x72,0x6c,0x52,0x12,0x71};
err += hashtest (2, 16, 1, "password", "diffsalt", Argon2_i, ref8i);
static const uint8_t ref9i[32] =
{0xaa,0xa9,0x53,0xd5,0x8a,0xf3,0x70,0x6c,0xe3,0xdf,0x1a,0xef,0xd4,0xa6,0x4a,0x84,
0xe3,0x1d,0x7f,0x54,0x17,0x52,0x31,0xf1,0x28,0x52,0x59,0xf8,0x81,0x74,0xce,0x5b};
err += hashtest (4, 16, 1, "password", "somesalt", Argon2_i, ref9i);
static const uint8_t ref10i[32] =
{0x29,0x6d,0xba,0xe8,0x0b,0x80,0x7c,0xdc,0xea,0xad,0x44,0xae,0x74,0x1b,0x50,0x6f,
0x14,0xdb,0x09,0x59,0x26,0x7b,0x18,0x3b,0x11,0x8f,0x9b,0x24,0x22,0x9b,0xc7,0xcb};
err += hashtest (2, 18, 1, "password", "somesalt", Argon2_i, ref10i);
#ifdef TEST_LARGE_RAM /* Argon2i: 1 GiB Test */
static const uint8_t ref11i[32] =
{0xd1,0x58,0x7a,0xca,0x09,0x22,0xc3,0xb5,0xd6,0xa8,0x3e,0xda,0xb3,0x1b,0xee,0x3c,
0x4e,0xba,0xef,0x34,0x2e,0xd6,0x12,0x7a,0x55,0xd1,0x9b,0x23,0x51,0xad,0x1f,0x41};
err += hashtest (2, 20, 1, "password", "somesalt", Argon2_i, ref11i);
#endif
return err;
}
/*==============================================================================
* Selftest function for Argon2id (Version 1.3).
* Using test vectors from Argon2 reference source code package ("test.c")
* and from https://tools.ietf.org/id/draft-irtf-cfrg-argon2-03.html#section-5.6
* See also:
* https://github.com/RustCrypto/password-hashes/blob/master/argon2/tests/kat.rs
*
* Uses memory costs of 32 Bytes, 256 KiB, 4 MiB, 64 MiB, 256 MiB.
* Runs 12 test cases in total.
* Returns number of failed test cases.
*/
int argon2id_selftest (void)
{
int err = argon2_selftest_ctx (); /* Basic test using argon2_ctx() */
static const uint8_t ref1id[32] =
{0x9d,0xfe,0xb9,0x10,0xe8,0x0b,0xad,0x03,0x11,0xfe,0xe2,0x0f,0x9c,0x0e,0x2b,0x12,
0xc1,0x79,0x87,0xb4,0xca,0xc9,0x0c,0x2e,0xf5,0x4d,0x5b,0x30,0x21,0xc6,0x8b,0xfe};
err += hashtest (2, 8, 1, "password", "somesalt", Argon2_id, ref1id);
static const uint8_t ref2id[32] =
{0x6d,0x09,0x3c,0x50,0x1f,0xd5,0x99,0x96,0x45,0xe0,0xea,0x3b,0xf6,0x20,0xd7,0xb8,
0xbe,0x7f,0xd2,0xdb,0x59,0xc2,0x0d,0x9f,0xff,0x95,0x39,0xda,0x2b,0xf5,0x70,0x37};
err += hashtest (2, 8, 2, "password", "somesalt", Argon2_id, ref2id);
static const uint8_t ref3id[32] =
{0xf5,0x55,0x35,0xbf,0xe9,0x48,0x71,0x00,0x51,0x42,0x4c,0x74,0x24,0xb1,0x1b,0xa9,
0xa1,0x3a,0x50,0x23,0x9b,0x04,0x59,0xf5,0x6c,0xa6,0x95,0xea,0x14,0xbc,0x19,0x5e};
err += hashtest (3, 12, 1, "pasword", "somesalt", Argon2_id, ref3id);
static const uint8_t ref4id[32] =
{0xf6,0xa5,0xad,0xc1,0xba,0x72,0x3d,0xdd,0xef,0x9b,0x5a,0xc1,0xd4,0x64,0xe1,0x80,
0xfc,0xd9,0xdf,0xfc,0x9d,0x1c,0xbf,0x76,0xcc,0xa2,0xfe,0xd7,0x95,0xd9,0xca,0x98};
err += hashtest (1, 16, 1, "password", "somesalt", Argon2_id, ref4id);
/*
static const uint8_t ref5id[32] =
{0x09,0x31,0x61,0x15,0xd5,0xcf,0x24,0xed,0x5a,0x15,0xa3,0x1a,0x3b,0xa3,0x26,0xe5,
0xcf,0x32,0xed,0xc2,0x47,0x02,0x98,0x7c,0x02,0xb6,0x56,0x6f,0x61,0x91,0x3c,0xf7};
err += hashtest (2, 16, 1, "password", "somesalt", Argon2_id, ref5id);
static const uint8_t ref6id[32] =
{0x0b,0x84,0xd6,0x52,0xcf,0x6b,0x0c,0x4b,0xea,0xef,0x0d,0xfe,0x27,0x8b,0xa6,0xa8,
0x0d,0xf6,0x69,0x62,0x81,0xd7,0xe0,0xd2,0x89,0x1b,0x81,0x7d,0x8c,0x45,0x8f,0xde};
err += hashtest (2, 16, 1, "differentpassword", "somesalt", Argon2_id, ref6id);
static const uint8_t ref7id[32] =
{0xbd,0xf3,0x2b,0x05,0xcc,0xc4,0x2e,0xb1,0x5d,0x58,0xfd,0x19,0xb1,0xf8,0x56,0xb1,
0x13,0xda,0x1e,0x9a,0x58,0x74,0xfd,0xcc,0x54,0x43,0x08,0x56,0x5a,0xa8,0x14,0x1c};
err += hashtest (2, 16, 1, "password", "diffsalt", Argon2_id, ref7id);
static const uint8_t ref8id[32] =
{0x90,0x25,0xd4,0x8e,0x68,0xef,0x73,0x95,0xcc,0xa9,0x07,0x9d,0xa4,0xc4,0xec,0x3a,
0xff,0xb3,0xc8,0x91,0x1f,0xe4,0xf8,0x6d,0x1a,0x25,0x20,0x85,0x6f,0x63,0x17,0x2c};
err += hashtest (4, 16, 1, "password", "somesalt", Argon2_id, ref8id);
static const uint8_t ref9id[32] =
{0x78,0xfe,0x1e,0xc9,0x1f,0xb3,0xaa,0x56,0x57,0xd7,0x2e,0x71,0x08,0x54,0xe4,0xc3,
0xd9,0xb9,0x19,0x8c,0x74,0x2f,0x96,0x16,0xc2,0xf0,0x85,0xbe,0xd9,0x5b,0x2e,0x8c};
err += hashtest (2, 18, 1, "password", "somesalt", Argon2_id, ref9id);
*/
return err;
}
src/Crypto/Crypto.vcxproj
+42 -8
View File
@@ -85,8 +85,8 @@
</Midl>
<ClCompile>
<Optimization>Disabled</Optimization>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;DEBUG;_DEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;$(ProjectDir)\Argon2\include;$(ProjectDir)\Argon2\src;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;ARGON2_NO_THREADS;DEBUG;_DEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<MinimalRebuild>true</MinimalRebuild>
<BasicRuntimeChecks>Default</BasicRuntimeChecks>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
@@ -105,8 +105,8 @@
<Midl />
<ClCompile>
<Optimization>Disabled</Optimization>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;DEBUG;_DEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;$(ProjectDir)\Argon2\include;$(ProjectDir)\Argon2\src;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;ARGON2_NO_THREADS;DEBUG;_DEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<MinimalRebuild>true</MinimalRebuild>
<BasicRuntimeChecks>Default</BasicRuntimeChecks>
<RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
@@ -127,8 +127,8 @@
</Midl>
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;NDEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;$(ProjectDir)\Argon2\include;$(ProjectDir)\Argon2\src;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;ARGON2_NO_THREADS;NDEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
<BufferSecurityCheck>true</BufferSecurityCheck>
<PrecompiledHeader>
@@ -149,8 +149,8 @@
<Midl />
<ClCompile>
<Optimization>MaxSpeed</Optimization>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;NDEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<AdditionalIncludeDirectories>$(ProjectDir)\..;$(ProjectDir)\..\Common;$(ProjectDir)\Argon2\include;$(ProjectDir)\Argon2\src;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
<PreprocessorDefinitions>WIN32;ARGON2_NO_THREADS;NDEBUG;_LIB;_CRT_SECURE_NO_DEPRECATE;_CRT_NON_CONFORMING_SWPRINTFS;%(PreprocessorDefinitions)</PreprocessorDefinitions>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
<BufferSecurityCheck>true</BufferSecurityCheck>
<PrecompiledHeader>
@@ -230,6 +230,21 @@
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">true</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Release|x64'">true</ExcludedFromBuild>
</ClCompile>
<ClCompile Include="Argon2\src\argon2.c" />
<ClCompile Include="Argon2\src\blake2\blake2b.c" />
<ClCompile Include="Argon2\src\core.c" />
<ClCompile Include="Argon2\src\opt_avx2.c">
<EnableEnhancedInstructionSet Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">AdvancedVectorExtensions2</EnableEnhancedInstructionSet>
<EnableEnhancedInstructionSet Condition="'$(Configuration)|$(Platform)'=='Release|x64'">AdvancedVectorExtensions2</EnableEnhancedInstructionSet>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">true</ExcludedFromBuild>
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Release|x64'">true</ExcludedFromBuild>
</ClCompile>
<ClCompile Include="Argon2\src\opt_sse2.c">
<EnableEnhancedInstructionSet Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">StreamingSIMDExtensions2</EnableEnhancedInstructionSet>
<EnableEnhancedInstructionSet Condition="'$(Configuration)|$(Platform)'=='Release|x64'">StreamingSIMDExtensions2</EnableEnhancedInstructionSet>
</ClCompile>
<ClCompile Include="Argon2\src\ref.c" />
<ClCompile Include="Argon2\src\selftest.c" />
<ClCompile Include="blake2s.c" />
<ClCompile Include="blake2s_SSE2.c" />
<ClCompile Include="blake2s_SSE41.c" />
@@ -272,6 +287,12 @@
<ClInclude Include="Aes_hw_cpu.h" />
<ClInclude Include="Aesopt.h" />
<ClInclude Include="Aestab.h" />
<ClInclude Include="Argon2\include\argon2.h" />
<ClInclude Include="Argon2\src\blake2\blake2-impl.h" />
<ClInclude Include="Argon2\src\blake2\blake2.h" />
<ClInclude Include="Argon2\src\blake2\blamka-round-opt.h" />
<ClInclude Include="Argon2\src\blake2\blamka-round-ref.h" />
<ClInclude Include="Argon2\src\core.h" />
<ClInclude Include="Camellia.h" />
<ClInclude Include="chacha256.h" />
<ClInclude Include="chachaRng.h" />
@@ -480,6 +501,19 @@
<ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Release|ARM64'">true</ExcludedFromBuild>
</CustomBuild>
</ItemGroup>
<ItemGroup>
<CustomBuild Include="Argon2\src\opt_avx2.asm">
<FileType>Document</FileType>
<Command Condition="'$(Configuration)|$(Platform)'=='Debug|ARM64'">echo %(Filename)%(Extension) &amp; yasm.exe -D WINABI -D __YASM__ -f win64 -o "$(TargetDir)\%(Filename).obj" -l "$(TargetDir)\%(Filename).lst" "%(FullPath)"</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">echo %(Filename)%(Extension) &amp; yasm.exe -D WINABI -D __YASM__ -f win64 -o "$(TargetDir)\%(Filename).obj" -l "$(TargetDir)\%(Filename).lst" "%(FullPath)"</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='Release|ARM64'">echo %(Filename)%(Extension) &amp; yasm.exe -D WINABI -D __YASM__ -f win64 -o "$(TargetDir)\%(Filename).obj" -l "$(TargetDir)\%(Filename).lst" "%(FullPath)"</Command>
<Command Condition="'$(Configuration)|$(Platform)'=='Release|x64'">echo %(Filename)%(Extension) &amp; yasm.exe -D WINABI -D __YASM__ -f win64 -o "$(TargetDir)\%(Filename).obj" -l "$(TargetDir)\%(Filename).lst" "%(FullPath)"</Command>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Debug|ARM64'">$(TargetDir)\%(Filename).obj;%(Outputs)</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">$(TargetDir)\%(Filename).obj;%(Outputs)</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|ARM64'">$(TargetDir)\%(Filename).obj;%(Outputs)</Outputs>
<Outputs Condition="'$(Configuration)|$(Platform)'=='Release|x64'">$(TargetDir)\%(Filename).obj;%(Outputs)</Outputs>
</CustomBuild>
</ItemGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
<ImportGroup Label="ExtensionTargets">
</ImportGroup>
src/Crypto/Crypto.vcxproj.filters
+51
View File
@@ -13,6 +13,15 @@
<UniqueIdentifier>{67DA6AB6-F800-4c08-8B7A-83BB121AAD01}</UniqueIdentifier>
<Extensions>rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx</Extensions>
</Filter>
<Filter Include="Source Files\Argon2">
<UniqueIdentifier>{c436eacb-1512-4925-8060-274b3cc4149d}</UniqueIdentifier>
</Filter>
<Filter Include="Source Files\Argon2\blake2">
<UniqueIdentifier>{8d89b4d3-54fe-42f8-9abd-92ec0b4a914c}</UniqueIdentifier>
</Filter>
<Filter Include="Header Files\Argon2">
<UniqueIdentifier>{a64d03f7-3bd3-4dd3-98ca-6a851b689d09}</UniqueIdentifier>
</Filter>
</ItemGroup>
<ItemGroup>
<ClCompile Include="Aeskey.c">
@@ -99,6 +108,27 @@
<ClCompile Include="sha256_armv8.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\argon2.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\core.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\ref.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\blake2\blake2b.c">
<Filter>Source Files\Argon2\blake2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\opt_avx2.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\opt_sse2.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
<ClCompile Include="Argon2\src\selftest.c">
<Filter>Source Files\Argon2</Filter>
</ClCompile>
</ItemGroup>
<ItemGroup>
<ClInclude Include="Aes.h">
@@ -176,6 +206,24 @@
<ClInclude Include="t1ha_selfcheck.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="Argon2\src\core.h">
<Filter>Source Files\Argon2</Filter>
</ClInclude>
<ClInclude Include="Argon2\src\blake2\blake2.h">
<Filter>Source Files\Argon2\blake2</Filter>
</ClInclude>
<ClInclude Include="Argon2\src\blake2\blake2-impl.h">
<Filter>Source Files\Argon2\blake2</Filter>
</ClInclude>
<ClInclude Include="Argon2\src\blake2\blamka-round-opt.h">
<Filter>Source Files\Argon2\blake2</Filter>
</ClInclude>
<ClInclude Include="Argon2\src\blake2\blamka-round-ref.h">
<Filter>Source Files\Argon2\blake2</Filter>
</ClInclude>
<ClInclude Include="Argon2\include\argon2.h">
<Filter>Header Files\Argon2</Filter>
</ClInclude>
</ItemGroup>
<ItemGroup>
<CustomBuild Include="Aes_hw_cpu.asm">
@@ -229,5 +277,8 @@
<CustomBuild Include="rdseed_ml.asm">
<Filter>Source Files</Filter>
</CustomBuild>
<CustomBuild Include="Argon2\src\opt_avx2.asm">
<Filter>Source Files\Argon2</Filter>
</CustomBuild>
</ItemGroup>
</Project>