35 typedef unsigned vec __attribute__((vector_size(16)));  typedef
47 #define ONE (vec) vsetq_lane_u32(1, vdupq_n_u32(0), 0)
48 #define LOAD_ALIGNED(m) (vec)(*((vec *)(m)))
54     (*((vec *)(alignment_buffer))) = (r); \
57 #define ROTV1(x) (vec) vextq_u32((uint32x4_t)x, (uint32x4_t)x, 1)
58 #define ROTV2(x) (vec) vextq_u32((uint32x4_t)x, (uint32x4_t)x, 2)
59 #define ROTV3(x) (vec) vextq_u32((uint32x4_t)x, (uint32x4_t)x, 3)
60 #define ROTW16(x) (vec) vrev32q_u16((uint16x8_t)x)
62 #define ROTW7(x) (x << ((vec) {7, 7, 7, 7})) ^ (x >> ((vec) {25, 25, 25, 25}))
63 #define ROTW8(x) (x << ((vec) {8, 8, 8, 8})) ^ (x >> ((vec) {24, 24, 24, 24}))
65   (x << ((vec) {12, 12, 12, 12})) ^ (x >> ((vec) {20, 20, 20, 20}))
68   (vec) vsriq_n_u32(vshlq_n_u32((uint32x4_t)x, 7), (uint32x4_t)x, 25)
70   (vec) vsriq_n_u32(vshlq_n_u32((uint32x4_t)x, 8), (uint32x4_t)x, 24)
72   (vec) vsriq_n_u32(vshlq_n_u32((uint32x4_t)x, 12), (uint32x4_t)x, 20)
82 #define ONE (vec) _mm_set_epi32(0, 0, 0, 1)
83 #define LOAD(m) (vec) _mm_loadu_si128((__m128i *)(m))
84 #define LOAD_ALIGNED(m) (vec) _mm_load_si128((__m128i *)(m))
86 #define ROTV1(x) (vec) _mm_shuffle_epi32((__m128i)x, _MM_SHUFFLE(0, 3, 2, 1))
87 #define ROTV2(x) (vec) _mm_shuffle_epi32((__m128i)x, _MM_SHUFFLE(1, 0, 3, 2))
88 #define ROTV3(x) (vec) _mm_shuffle_epi32((__m128i)x, _MM_SHUFFLE(2, 1, 0, 3))
90   (vec)(_mm_slli_epi32((__m128i)x, 7) ^ _mm_srli_epi32((__m128i)x, 25))
92   (vec)(_mm_slli_epi32((__m128i)x, 12) ^ _mm_srli_epi32((__m128i)x, 20))
96   (vec) _mm_shuffle_epi8((__m128i)x, _mm_set_epi8(14, 13, 12, 15, 10, 9, 8, \
99   (vec) _mm_shuffle_epi8((__m128i)x, _mm_set_epi8(13, 12, 15, 14, 9, 8, 11, \
103   (vec)(_mm_slli_epi32((__m128i)x, 8) ^ _mm_srli_epi32((__m128i)x, 24))
105   (vec)(_mm_slli_epi32((__m128i)x, 16) ^ _mm_srli_epi32((__m128i)x, 16))
165 	vec s0, s1, s2, s3;  in CRYPTO_chacha_20_neon()
173 	s1 = LOAD(&((vec*)kp)[0]);  in CRYPTO_chacha_20_neon()
174 	s2 = LOAD(&((vec*)kp)[1]);  in CRYPTO_chacha_20_neon()
175 	s3 = (vec){  in CRYPTO_chacha_20_neon()
193 		vec v8,v9,v10,v11;  in CRYPTO_chacha_20_neon()
196 		vec v12,v13,v14,v15;  in CRYPTO_chacha_20_neon()
199 		vec v0,v1,v2,v3,v4,v5,v6,v7;  in CRYPTO_chacha_20_neon()
279 		vec v0 = s0, v1 = s1, v2 = s2, v3 = s3;  in CRYPTO_chacha_20_neon()
293 		__attribute__ ((aligned (16))) vec buf[4];  in CRYPTO_chacha_20_neon()
294 		vec v0,v1,v2,v3;  in CRYPTO_chacha_20_neon()