laag-libsodium 1.0.16.0

data/vendor/github.com/jedisct1/libsodium/src/libsodium/crypto_pwhash/argon2/blamka-round-avx2.h

@@ -0,0 +1,150 @@
+#ifndef blamka_round_avx2_H
+#define blamka_round_avx2_H
+
+#include "private/common.h"
+#include "private/sse2_64_32.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

data/vendor/github.com/jedisct1/libsodium/src/libsodium/crypto_pwhash/argon2/blamka-round-avx512f.h

@@ -0,0 +1,145 @@
+#ifndef blamka_round_avx512f_H
+#define blamka_round_avx512f_H
+
+#include "private/common.h"
+#include "private/sse2_64_32.h"
+
+#define ror64(x, n) _mm512_ror_epi64((x), (n))
+
+static inline __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_AVX512F(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_AVX512F(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_AVX512F(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2_AVX512F(A0, B0, C0, D0, A1, B1, C1, D1); \
+        \
+        DIAGONALIZE(A0, B0, C0, D0, A1, B1, C1, D1); \
+        \
+        G1_AVX512F(A0, B0, C0, D0, A1, B1, C1, D1); \
+        G2_AVX512F(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

data/vendor/github.com/jedisct1/libsodium/src/libsodium/crypto_pwhash/argon2/blamka-round-ref.h

@@ -0,0 +1,40 @@
+#ifndef blamka_round_ref_H
+#define blamka_round_ref_H
+
+#include "private/common.h"
+
+/*designed by the Lyra PHC team */
+static 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

data/vendor/github.com/jedisct1/libsodium/src/libsodium/crypto_pwhash/argon2/blamka-round-ssse3.h

@@ -0,0 +1,120 @@
+#ifndef blamka_round_ssse3_H
+#define blamka_round_ssse3_H
+
+#include "private/common.h"
+#include "private/sse2_64_32.h"
+
+#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))))
+
+static 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)
+
+#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)
+
+#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)
+
+#endif