react-native-quick-crypto 1.0.0-beta.17 → 1.0.0-beta.18

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6/salsa20_xmm6.c

@@ -0,0 +1,31 @@
+
+#include <stdint.h>
+
+#include "utils.h"
+
+#include "../stream_salsa20.h"
+#include "salsa20_xmm6.h"
+
+#ifdef HAVE_AMD64_ASM
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+extern int stream_salsa20_xmm6(unsigned char *c, unsigned long long clen,
+                               const unsigned char *n, const unsigned char *k);
+
+extern int stream_salsa20_xmm6_xor_ic(unsigned char *c, const unsigned char *m,
+                                      unsigned long long mlen,
+                                      const unsigned char *n,
+                                      uint64_t ic, const unsigned char *k);
+#ifdef __cplusplus
+}
+#endif
+
+struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6_implementation = {
+        SODIUM_C99(.stream =) stream_salsa20_xmm6,
+        SODIUM_C99(.stream_xor_ic =) stream_salsa20_xmm6_xor_ic,
+    };
+
+#endif

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6/salsa20_xmm6.h

@@ -0,0 +1,8 @@
+
+#include <stdint.h>
+
+#include "../stream_salsa20.h"
+#include "crypto_stream_salsa20.h"
+
+extern struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6_implementation;

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6int/salsa20_xmm6int-avx2.c

@@ -0,0 +1,134 @@
+
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "crypto_stream_salsa20.h"
+#include "private/common.h"
+#include "utils.h"
+
+#if defined(HAVE_AVX2INTRIN_H) && defined(HAVE_EMMINTRIN_H) && \
+    defined(HAVE_TMMINTRIN_H) && defined(HAVE_SMMINTRIN_H)
+
+# ifdef __clang__
+#  pragma clang attribute push(__attribute__((target("sse2,ssse3,sse4.1,avx2"))), apply_to = function)
+# elif defined(__GNUC__)
+#  pragma GCC target("sse2,ssse3,sse4.1,avx2")
+# endif
+
+# include <emmintrin.h>
+# include <immintrin.h>
+# include <smmintrin.h>
+# include <tmmintrin.h>
+# include "private/sse2_64_32.h"
+
+# include "../stream_salsa20.h"
+# include "salsa20_xmm6int-avx2.h"
+
+# define ROUNDS 20
+
+typedef struct salsa_ctx {
+    uint32_t input[16];
+} salsa_ctx;
+
+static const int TR[16] = {
+    0, 5, 10, 15, 12, 1, 6, 11, 8, 13, 2, 7, 4, 9, 14, 3
+};
+
+static void
+salsa_keysetup(salsa_ctx *ctx, const uint8_t *k)
+{
+    ctx->input[TR[1]]  = LOAD32_LE(k + 0);
+    ctx->input[TR[2]]  = LOAD32_LE(k + 4);
+    ctx->input[TR[3]]  = LOAD32_LE(k + 8);
+    ctx->input[TR[4]]  = LOAD32_LE(k + 12);
+    ctx->input[TR[11]] = LOAD32_LE(k + 16);
+    ctx->input[TR[12]] = LOAD32_LE(k + 20);
+    ctx->input[TR[13]] = LOAD32_LE(k + 24);
+    ctx->input[TR[14]] = LOAD32_LE(k + 28);
+    ctx->input[TR[0]]  = 0x61707865;
+    ctx->input[TR[5]]  = 0x3320646e;
+    ctx->input[TR[10]] = 0x79622d32;
+    ctx->input[TR[15]] = 0x6b206574;
+}
+
+static void
+salsa_ivsetup(salsa_ctx *ctx, const uint8_t *iv, const uint8_t *counter)
+{
+    ctx->input[TR[6]] = LOAD32_LE(iv + 0);
+    ctx->input[TR[7]] = LOAD32_LE(iv + 4);
+    ctx->input[TR[8]] = counter == NULL ? 0 : LOAD32_LE(counter + 0);
+    ctx->input[TR[9]] = counter == NULL ? 0 : LOAD32_LE(counter + 4);
+}
+
+static void
+salsa20_encrypt_bytes(salsa_ctx *ctx, const uint8_t *m, uint8_t *c,
+                      unsigned long long bytes)
+{
+    uint32_t * const x = &ctx->input[0];
+
+    if (!bytes) {
+        return; /* LCOV_EXCL_LINE */
+    }
+
+#include "u8.h"
+#include "u4.h"
+#include "u1.h"
+#include "u0.h"
+}
+
+static int
+stream_avx2(unsigned char *c, unsigned long long clen, const unsigned char *n,
+            const unsigned char *k)
+{
+    struct salsa_ctx ctx;
+
+    if (!clen) {
+        return 0;
+    }
+    COMPILER_ASSERT(crypto_stream_salsa20_KEYBYTES == 256 / 8);
+    salsa_keysetup(&ctx, k);
+    salsa_ivsetup(&ctx, n, NULL);
+    memset(c, 0, clen);
+    salsa20_encrypt_bytes(&ctx, c, c, clen);
+    sodium_memzero(&ctx, sizeof ctx);
+
+    return 0;
+}
+
+static int
+stream_avx2_xor_ic(unsigned char *c, const unsigned char *m,
+                   unsigned long long mlen, const unsigned char *n, uint64_t ic,
+                   const unsigned char *k)
+{
+    struct salsa_ctx ctx;
+    uint8_t          ic_bytes[8];
+    uint32_t         ic_high;
+    uint32_t         ic_low;
+
+    if (!mlen) {
+        return 0;
+    }
+    ic_high = (uint32_t) (ic >> 32);
+    ic_low  = (uint32_t) ic;
+    STORE32_LE(&ic_bytes[0], ic_low);
+    STORE32_LE(&ic_bytes[4], ic_high);
+    salsa_keysetup(&ctx, k);
+    salsa_ivsetup(&ctx, n, ic_bytes);
+    salsa20_encrypt_bytes(&ctx, m, c, mlen);
+    sodium_memzero(&ctx, sizeof ctx);
+
+    return 0;
+}
+
+struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6int_avx2_implementation = {
+        SODIUM_C99(.stream =) stream_avx2,
+        SODIUM_C99(.stream_xor_ic =) stream_avx2_xor_ic
+    };
+
+#ifdef __clang__
+# pragma clang attribute pop
+#endif
+
+#endif

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6int/salsa20_xmm6int-avx2.h

@@ -0,0 +1,8 @@
+
+#include <stdint.h>
+
+#include "../stream_salsa20.h"
+#include "crypto_stream_salsa20.h"
+
+extern struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6int_avx2_implementation;

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6int/salsa20_xmm6int-sse2.c

@@ -0,0 +1,128 @@
+
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "crypto_stream_salsa20.h"
+#include "private/common.h"
+#include "utils.h"
+
+#ifdef HAVE_EMMINTRIN_H
+
+# ifdef __clang__
+#  pragma clang attribute push(__attribute__((target("sse2"))), apply_to = function)
+# elif defined(__GNUC__)
+#  pragma GCC target("sse2")
+# endif
+# include <emmintrin.h>
+# include "private/sse2_64_32.h"
+
+# include "../stream_salsa20.h"
+# include "salsa20_xmm6int-sse2.h"
+
+# define ROUNDS 20
+
+typedef struct salsa_ctx {
+    uint32_t input[16];
+} salsa_ctx;
+
+static const int TR[16] = {
+    0, 5, 10, 15, 12, 1, 6, 11, 8, 13, 2, 7, 4, 9, 14, 3
+};
+
+static void
+salsa_keysetup(salsa_ctx *ctx, const uint8_t *k)
+{
+    ctx->input[TR[1]]  = LOAD32_LE(k + 0);
+    ctx->input[TR[2]]  = LOAD32_LE(k + 4);
+    ctx->input[TR[3]]  = LOAD32_LE(k + 8);
+    ctx->input[TR[4]]  = LOAD32_LE(k + 12);
+    ctx->input[TR[11]] = LOAD32_LE(k + 16);
+    ctx->input[TR[12]] = LOAD32_LE(k + 20);
+    ctx->input[TR[13]] = LOAD32_LE(k + 24);
+    ctx->input[TR[14]] = LOAD32_LE(k + 28);
+    ctx->input[TR[0]]  = 0x61707865;
+    ctx->input[TR[5]]  = 0x3320646e;
+    ctx->input[TR[10]] = 0x79622d32;
+    ctx->input[TR[15]] = 0x6b206574;
+}
+
+static void
+salsa_ivsetup(salsa_ctx *ctx, const uint8_t *iv, const uint8_t *counter)
+{
+    ctx->input[TR[6]] = LOAD32_LE(iv + 0);
+    ctx->input[TR[7]] = LOAD32_LE(iv + 4);
+    ctx->input[TR[8]] = counter == NULL ? 0 : LOAD32_LE(counter + 0);
+    ctx->input[TR[9]] = counter == NULL ? 0 : LOAD32_LE(counter + 4);
+}
+
+static void
+salsa20_encrypt_bytes(salsa_ctx *ctx, const uint8_t *m, uint8_t *c,
+                      unsigned long long bytes)
+{
+    uint32_t * const x = &ctx->input[0];
+
+    if (!bytes) {
+        return; /* LCOV_EXCL_LINE */
+    }
+
+#include "u4.h"
+#include "u1.h"
+#include "u0.h"
+}
+
+static int
+stream_sse2(unsigned char *c, unsigned long long clen, const unsigned char *n,
+            const unsigned char *k)
+{
+    struct salsa_ctx ctx;
+
+    if (!clen) {
+        return 0;
+    }
+    COMPILER_ASSERT(crypto_stream_salsa20_KEYBYTES == 256 / 8);
+    salsa_keysetup(&ctx, k);
+    salsa_ivsetup(&ctx, n, NULL);
+    memset(c, 0, clen);
+    salsa20_encrypt_bytes(&ctx, c, c, clen);
+    sodium_memzero(&ctx, sizeof ctx);
+
+    return 0;
+}
+
+static int
+stream_sse2_xor_ic(unsigned char *c, const unsigned char *m,
+                   unsigned long long mlen, const unsigned char *n, uint64_t ic,
+                   const unsigned char *k)
+{
+    struct salsa_ctx ctx;
+    uint8_t          ic_bytes[8];
+    uint32_t         ic_high;
+    uint32_t         ic_low;
+
+    if (!mlen) {
+        return 0;
+    }
+    ic_high = (uint32_t) (ic >> 32);
+    ic_low  = (uint32_t) (ic);
+    STORE32_LE(&ic_bytes[0], ic_low);
+    STORE32_LE(&ic_bytes[4], ic_high);
+    salsa_keysetup(&ctx, k);
+    salsa_ivsetup(&ctx, n, ic_bytes);
+    salsa20_encrypt_bytes(&ctx, m, c, mlen);
+    sodium_memzero(&ctx, sizeof ctx);
+
+    return 0;
+}
+
+struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6int_sse2_implementation = {
+        SODIUM_C99(.stream =) stream_sse2,
+        SODIUM_C99(.stream_xor_ic =) stream_sse2_xor_ic
+    };
+
+#ifdef __clang__
+# pragma clang attribute pop
+#endif
+
+#endif

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6int/salsa20_xmm6int-sse2.h

@@ -0,0 +1,8 @@
+
+#include <stdint.h>
+
+#include "../stream_salsa20.h"
+#include "crypto_stream_salsa20.h"
+
+extern struct crypto_stream_salsa20_implementation
+    crypto_stream_salsa20_xmm6int_sse2_implementation;

package/ios/libsodium-stable/src/libsodium/crypto_stream/salsa20/xmm6int/u0.h

@@ -0,0 +1,195 @@
+if (bytes > 0) {
+    __m128i diag0 = _mm_loadu_si128((const __m128i *) (x + 0));
+    __m128i diag1 = _mm_loadu_si128((const __m128i *) (x + 4));
+    __m128i diag2 = _mm_loadu_si128((const __m128i *) (x + 8));
+    __m128i diag3 = _mm_loadu_si128((const __m128i *) (x + 12));
+    __m128i a0, a1, a2, a3, a4, a5, a6, a7;
+    __m128i b0, b1, b2, b3, b4, b5, b6, b7;
+    uint8_t partialblock[64] = { 0 };
+
+    unsigned int i;
+
+    a0 = diag1;
+    for (i = 0; i < ROUNDS; i += 4) {
+        a0    = _mm_add_epi32(a0, diag0);
+        a1    = diag0;
+        b0    = a0;
+        a0    = _mm_slli_epi32(a0, 7);
+        b0    = _mm_srli_epi32(b0, 25);
+        diag3 = _mm_xor_si128(diag3, a0);
+
+        diag3 = _mm_xor_si128(diag3, b0);
+
+        a1    = _mm_add_epi32(a1, diag3);
+        a2    = diag3;
+        b1    = a1;
+        a1    = _mm_slli_epi32(a1, 9);
+        b1    = _mm_srli_epi32(b1, 23);
+        diag2 = _mm_xor_si128(diag2, a1);
+        diag3 = _mm_shuffle_epi32(diag3, 0x93);
+        diag2 = _mm_xor_si128(diag2, b1);
+
+        a2    = _mm_add_epi32(a2, diag2);
+        a3    = diag2;
+        b2    = a2;
+        a2    = _mm_slli_epi32(a2, 13);
+        b2    = _mm_srli_epi32(b2, 19);
+        diag1 = _mm_xor_si128(diag1, a2);
+        diag2 = _mm_shuffle_epi32(diag2, 0x4e);
+        diag1 = _mm_xor_si128(diag1, b2);
+
+        a3    = _mm_add_epi32(a3, diag1);
+        a4    = diag3;
+        b3    = a3;
+        a3    = _mm_slli_epi32(a3, 18);
+        b3    = _mm_srli_epi32(b3, 14);
+        diag0 = _mm_xor_si128(diag0, a3);
+        diag1 = _mm_shuffle_epi32(diag1, 0x39);
+        diag0 = _mm_xor_si128(diag0, b3);
+
+        a4    = _mm_add_epi32(a4, diag0);
+        a5    = diag0;
+        b4    = a4;
+        a4    = _mm_slli_epi32(a4, 7);
+        b4    = _mm_srli_epi32(b4, 25);
+        diag1 = _mm_xor_si128(diag1, a4);
+
+        diag1 = _mm_xor_si128(diag1, b4);
+
+        a5    = _mm_add_epi32(a5, diag1);
+        a6    = diag1;
+        b5    = a5;
+        a5    = _mm_slli_epi32(a5, 9);
+        b5    = _mm_srli_epi32(b5, 23);
+        diag2 = _mm_xor_si128(diag2, a5);
+        diag1 = _mm_shuffle_epi32(diag1, 0x93);
+        diag2 = _mm_xor_si128(diag2, b5);
+
+        a6    = _mm_add_epi32(a6, diag2);
+        a7    = diag2;
+        b6    = a6;
+        a6    = _mm_slli_epi32(a6, 13);
+        b6    = _mm_srli_epi32(b6, 19);
+        diag3 = _mm_xor_si128(diag3, a6);
+        diag2 = _mm_shuffle_epi32(diag2, 0x4e);
+        diag3 = _mm_xor_si128(diag3, b6);
+
+        a7    = _mm_add_epi32(a7, diag3);
+        a0    = diag1;
+        b7    = a7;
+        a7    = _mm_slli_epi32(a7, 18);
+        b7    = _mm_srli_epi32(b7, 14);
+        diag0 = _mm_xor_si128(diag0, a7);
+        diag3 = _mm_shuffle_epi32(diag3, 0x39);
+        diag0 = _mm_xor_si128(diag0, b7);
+
+        a0    = _mm_add_epi32(a0, diag0);
+        a1    = diag0;
+        b0    = a0;
+        a0    = _mm_slli_epi32(a0, 7);
+        b0    = _mm_srli_epi32(b0, 25);
+        diag3 = _mm_xor_si128(diag3, a0);
+
+        diag3 = _mm_xor_si128(diag3, b0);
+
+        a1    = _mm_add_epi32(a1, diag3);
+        a2    = diag3;
+        b1    = a1;
+        a1    = _mm_slli_epi32(a1, 9);
+        b1    = _mm_srli_epi32(b1, 23);
+        diag2 = _mm_xor_si128(diag2, a1);
+        diag3 = _mm_shuffle_epi32(diag3, 0x93);
+        diag2 = _mm_xor_si128(diag2, b1);
+
+        a2    = _mm_add_epi32(a2, diag2);
+        a3    = diag2;
+        b2    = a2;
+        a2    = _mm_slli_epi32(a2, 13);
+        b2    = _mm_srli_epi32(b2, 19);
+        diag1 = _mm_xor_si128(diag1, a2);
+        diag2 = _mm_shuffle_epi32(diag2, 0x4e);
+        diag1 = _mm_xor_si128(diag1, b2);
+
+        a3    = _mm_add_epi32(a3, diag1);
+        a4    = diag3;
+        b3    = a3;
+        a3    = _mm_slli_epi32(a3, 18);
+        b3    = _mm_srli_epi32(b3, 14);
+        diag0 = _mm_xor_si128(diag0, a3);
+        diag1 = _mm_shuffle_epi32(diag1, 0x39);
+        diag0 = _mm_xor_si128(diag0, b3);
+
+        a4    = _mm_add_epi32(a4, diag0);
+        a5    = diag0;
+        b4    = a4;
+        a4    = _mm_slli_epi32(a4, 7);
+        b4    = _mm_srli_epi32(b4, 25);
+        diag1 = _mm_xor_si128(diag1, a4);
+
+        diag1 = _mm_xor_si128(diag1, b4);
+
+        a5    = _mm_add_epi32(a5, diag1);
+        a6    = diag1;
+        b5    = a5;
+        a5    = _mm_slli_epi32(a5, 9);
+        b5    = _mm_srli_epi32(b5, 23);
+        diag2 = _mm_xor_si128(diag2, a5);
+        diag1 = _mm_shuffle_epi32(diag1, 0x93);
+        diag2 = _mm_xor_si128(diag2, b5);
+
+        a6    = _mm_add_epi32(a6, diag2);
+        a7    = diag2;
+        b6    = a6;
+        a6    = _mm_slli_epi32(a6, 13);
+        b6    = _mm_srli_epi32(b6, 19);
+        diag3 = _mm_xor_si128(diag3, a6);
+        diag2 = _mm_shuffle_epi32(diag2, 0x4e);
+        diag3 = _mm_xor_si128(diag3, b6);
+
+        a7    = _mm_add_epi32(a7, diag3);
+        a0    = diag1;
+        b7    = a7;
+        a7    = _mm_slli_epi32(a7, 18);
+        b7    = _mm_srli_epi32(b7, 14);
+        diag0 = _mm_xor_si128(diag0, a7);
+        diag3 = _mm_shuffle_epi32(diag3, 0x39);
+        diag0 = _mm_xor_si128(diag0, b7);
+    }
+
+    diag0 = _mm_add_epi32(diag0, _mm_loadu_si128((const __m128i *) (x + 0)));
+    diag1 = _mm_add_epi32(diag1, _mm_loadu_si128((const __m128i *) (x + 4)));
+    diag2 = _mm_add_epi32(diag2, _mm_loadu_si128((const __m128i *) (x + 8)));
+    diag3 = _mm_add_epi32(diag3, _mm_loadu_si128((const __m128i *) (x + 12)));
+
+#define ONEQUAD_SHUFFLE(A, B, C, D)                                              \
+    do {                                                                         \
+        uint32_t in##A                         = _mm_cvtsi128_si32(diag0);       \
+        uint32_t in##B                         = _mm_cvtsi128_si32(diag1);       \
+        uint32_t in##C                         = _mm_cvtsi128_si32(diag2);       \
+        uint32_t in##D                         = _mm_cvtsi128_si32(diag3);       \
+        diag0                                  = _mm_shuffle_epi32(diag0, 0x39); \
+        diag1                                  = _mm_shuffle_epi32(diag1, 0x39); \
+        diag2                                  = _mm_shuffle_epi32(diag2, 0x39); \
+        diag3                                  = _mm_shuffle_epi32(diag3, 0x39); \
+        *(uint32_t *) (partialblock + (A * 4)) = in##A;                          \
+        *(uint32_t *) (partialblock + (B * 4)) = in##B;                          \
+        *(uint32_t *) (partialblock + (C * 4)) = in##C;                          \
+        *(uint32_t *) (partialblock + (D * 4)) = in##D;                          \
+    } while (0)
+
+#define ONEQUAD(A, B, C, D) ONEQUAD_SHUFFLE(A, B, C, D)
+
+    ONEQUAD(0, 12, 8, 4);
+    ONEQUAD(5, 1, 13, 9);
+    ONEQUAD(10, 6, 2, 14);
+    ONEQUAD(15, 11, 7, 3);
+
+#undef ONEQUAD
+#undef ONEQUAD_SHUFFLE
+
+    for (i = 0; i < bytes; i++) {
+        c[i] = m[i] ^ partialblock[i];
+    }
+
+    sodium_memzero(partialblock, sizeof partialblock);
+}