Orange Pi5 kernel

Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards

3 Commits   0 Branches   0 Tags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   1) // SPDX-License-Identifier: GPL-2.0-or-later
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   3)  * Copyright (C) 2019 Linaro, Ltd. <ard.biesheuvel@linaro.org>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   4)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   5) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   6) #ifdef CONFIG_ARM64
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   7) #include <asm/neon-intrinsics.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   8) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   9) #define AES_ROUND	"aese %0.16b, %1.16b \n\t aesmc %0.16b, %0.16b"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  10) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  11) #include <arm_neon.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  12) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  13) #define AES_ROUND	"aese.8 %q0, %q1 \n\t aesmc.8 %q0, %q0"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  14) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  15) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  16) #define AEGIS_BLOCK_SIZE	16
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  17) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  18) #include <stddef.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  19) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  20) extern int aegis128_have_aes_insn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  21) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  22) void *memcpy(void *dest, const void *src, size_t n);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  23) void *memset(void *s, int c, size_t n);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  24) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  25) struct aegis128_state {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  26) 	uint8x16_t v[5];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  27) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  28) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  29) extern const uint8_t crypto_aes_sbox[];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  30) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  31) static struct aegis128_state aegis128_load_state_neon(const void *state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  32) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  33) 	return (struct aegis128_state){ {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  34) 		vld1q_u8(state),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  35) 		vld1q_u8(state + 16),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  36) 		vld1q_u8(state + 32),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  37) 		vld1q_u8(state + 48),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  38) 		vld1q_u8(state + 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  39) 	} };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  40) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  41) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  42) static void aegis128_save_state_neon(struct aegis128_state st, void *state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  43) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  44) 	vst1q_u8(state, st.v[0]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  45) 	vst1q_u8(state + 16, st.v[1]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  46) 	vst1q_u8(state + 32, st.v[2]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  47) 	vst1q_u8(state + 48, st.v[3]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  48) 	vst1q_u8(state + 64, st.v[4]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  49) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  50) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  51) static inline __attribute__((always_inline))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  52) uint8x16_t aegis_aes_round(uint8x16_t w)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  53) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  54) 	uint8x16_t z = {};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  55) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  56) #ifdef CONFIG_ARM64
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  57) 	if (!__builtin_expect(aegis128_have_aes_insn, 1)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  58) 		static const uint8_t shift_rows[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  59) 			0x0, 0x5, 0xa, 0xf, 0x4, 0x9, 0xe, 0x3,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  60) 			0x8, 0xd, 0x2, 0x7, 0xc, 0x1, 0x6, 0xb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  61) 		};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  62) 		static const uint8_t ror32by8[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  63) 			0x1, 0x2, 0x3, 0x0, 0x5, 0x6, 0x7, 0x4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  64) 			0x9, 0xa, 0xb, 0x8, 0xd, 0xe, 0xf, 0xc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  65) 		};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  66) 		uint8x16_t v;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  67) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  68) 		// shift rows
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  69) 		w = vqtbl1q_u8(w, vld1q_u8(shift_rows));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  70) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  71) 		// sub bytes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  72) #ifndef CONFIG_CC_IS_GCC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  73) 		v = vqtbl4q_u8(vld1q_u8_x4(crypto_aes_sbox), w);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  74) 		v = vqtbx4q_u8(v, vld1q_u8_x4(crypto_aes_sbox + 0x40), w - 0x40);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  75) 		v = vqtbx4q_u8(v, vld1q_u8_x4(crypto_aes_sbox + 0x80), w - 0x80);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  76) 		v = vqtbx4q_u8(v, vld1q_u8_x4(crypto_aes_sbox + 0xc0), w - 0xc0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  77) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  78) 		asm("tbl %0.16b, {v16.16b-v19.16b}, %1.16b" : "=w"(v) : "w"(w));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  79) 		w -= 0x40;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  80) 		asm("tbx %0.16b, {v20.16b-v23.16b}, %1.16b" : "+w"(v) : "w"(w));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  81) 		w -= 0x40;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  82) 		asm("tbx %0.16b, {v24.16b-v27.16b}, %1.16b" : "+w"(v) : "w"(w));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  83) 		w -= 0x40;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  84) 		asm("tbx %0.16b, {v28.16b-v31.16b}, %1.16b" : "+w"(v) : "w"(w));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  85) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  86) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  87) 		// mix columns
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  88) 		w = (v << 1) ^ (uint8x16_t)(((int8x16_t)v >> 7) & 0x1b);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  89) 		w ^= (uint8x16_t)vrev32q_u16((uint16x8_t)v);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  90) 		w ^= vqtbl1q_u8(v ^ w, vld1q_u8(ror32by8));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  91) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  92) 		return w;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  93) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  94) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  95) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  96) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  97) 	 * We use inline asm here instead of the vaeseq_u8/vaesmcq_u8 intrinsics
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  98) 	 * to force the compiler to issue the aese/aesmc instructions in pairs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  99) 	 * This is much faster on many cores, where the instruction pair can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) 	 * execute in a single cycle.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) 	asm(AES_ROUND : "+w"(w) : "w"(z));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) 	return w;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) static inline __attribute__((always_inline))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) struct aegis128_state aegis128_update_neon(struct aegis128_state st,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) 					   uint8x16_t m)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) 	m       ^= aegis_aes_round(st.v[4]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) 	st.v[4] ^= aegis_aes_round(st.v[3]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) 	st.v[3] ^= aegis_aes_round(st.v[2]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) 	st.v[2] ^= aegis_aes_round(st.v[1]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) 	st.v[1] ^= aegis_aes_round(st.v[0]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) 	st.v[0] ^= m;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) 	return st;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) static inline __attribute__((always_inline))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) void preload_sbox(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) 	if (!IS_ENABLED(CONFIG_ARM64) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) 	    !IS_ENABLED(CONFIG_CC_IS_GCC) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) 	    __builtin_expect(aegis128_have_aes_insn, 1))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) 		return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) 	asm("ld1	{v16.16b-v19.16b}, [%0], #64	\n\t"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) 	    "ld1	{v20.16b-v23.16b}, [%0], #64	\n\t"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) 	    "ld1	{v24.16b-v27.16b}, [%0], #64	\n\t"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) 	    "ld1	{v28.16b-v31.16b}, [%0]		\n\t"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) 	    :: "r"(crypto_aes_sbox));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) void crypto_aegis128_init_neon(void *state, const void *key, const void *iv)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) 	static const uint8_t const0[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) 		0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) 		0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) 	};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) 	static const uint8_t const1[] = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) 		0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) 		0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) 	};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) 	uint8x16_t k = vld1q_u8(key);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) 	uint8x16_t kiv = k ^ vld1q_u8(iv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) 	struct aegis128_state st = {{
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) 		kiv,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) 		vld1q_u8(const1),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) 		vld1q_u8(const0),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) 		k ^ vld1q_u8(const0),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) 		k ^ vld1q_u8(const1),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) 	}};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) 	preload_sbox();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) 	for (i = 0; i < 5; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) 		st = aegis128_update_neon(st, k);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) 		st = aegis128_update_neon(st, kiv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) 	aegis128_save_state_neon(st, state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) void crypto_aegis128_update_neon(void *state, const void *msg)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) 	struct aegis128_state st = aegis128_load_state_neon(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) 	preload_sbox();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) 	st = aegis128_update_neon(st, vld1q_u8(msg));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) 	aegis128_save_state_neon(st, state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) void crypto_aegis128_encrypt_chunk_neon(void *state, void *dst, const void *src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) 					unsigned int size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) 	struct aegis128_state st = aegis128_load_state_neon(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) 	uint8x16_t msg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) 	preload_sbox();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) 	while (size >= AEGIS_BLOCK_SIZE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) 		uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) 		msg = vld1q_u8(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) 		st = aegis128_update_neon(st, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) 		vst1q_u8(dst, msg ^ s);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 		size -= AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 		src += AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) 		dst += AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) 	if (size > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) 		uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) 		uint8_t buf[AEGIS_BLOCK_SIZE] = {};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) 		memcpy(buf, src, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) 		msg = vld1q_u8(buf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) 		st = aegis128_update_neon(st, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) 		vst1q_u8(buf, msg ^ s);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) 		memcpy(dst, buf, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) 	aegis128_save_state_neon(st, state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) void crypto_aegis128_decrypt_chunk_neon(void *state, void *dst, const void *src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) 					unsigned int size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) 	struct aegis128_state st = aegis128_load_state_neon(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) 	uint8x16_t msg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) 	preload_sbox();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) 	while (size >= AEGIS_BLOCK_SIZE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) 		msg = vld1q_u8(src) ^ st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) 		st = aegis128_update_neon(st, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) 		vst1q_u8(dst, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) 		size -= AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) 		src += AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) 		dst += AEGIS_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) 	if (size > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) 		uint8x16_t s = st.v[1] ^ (st.v[2] & st.v[3]) ^ st.v[4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) 		uint8_t buf[AEGIS_BLOCK_SIZE];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) 		vst1q_u8(buf, s);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) 		memcpy(buf, src, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) 		msg = vld1q_u8(buf) ^ s;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) 		vst1q_u8(buf, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) 		memcpy(dst, buf, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) 		st = aegis128_update_neon(st, msg);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) 	aegis128_save_state_neon(st, state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) void crypto_aegis128_final_neon(void *state, void *tag_xor, uint64_t assoclen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) 				uint64_t cryptlen)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) 	struct aegis128_state st = aegis128_load_state_neon(state);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) 	uint8x16_t v;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) 	int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) 	preload_sbox();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) 	v = st.v[3] ^ (uint8x16_t)vcombine_u64(vmov_n_u64(8 * assoclen),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) 					       vmov_n_u64(8 * cryptlen));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) 	for (i = 0; i < 7; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) 		st = aegis128_update_neon(st, v);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) 	v = vld1q_u8(tag_xor);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) 	v ^= st.v[0] ^ st.v[1] ^ st.v[2] ^ st.v[3] ^ st.v[4];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) 	vst1q_u8(tag_xor, v);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) }