arm/crypto/blake2s-core.S

24 .macro __ldrd		a, b, src, offset
26 	ldrd		\a, \b, [\src, #\offset]
28 	ldr		\a, [\src, #\offset]
33 .macro __strd		a, b, dst, offset
35 	strd		\a, \b, [\dst, #\offset]
37 	str		\a, [\dst, #\offset]
42 .macro _le32_bswap	a, tmp
44 	rev_l		\a, \tmp
48 .macro _le32_bswap_8x	a, b, c, d, e, f, g, h,  tmp
49 	_le32_bswap	\a, \tmp
59 // Execute a quarter-round of BLAKE2s by mixing two columns or two diagonals.
73 	// a += b + m[blake2s_sigma[r][2*i + 0]];
79 	// d = ror32(d ^ a, 16);
87 	// b = ror32(b ^ c, 12);
94 	// a += b + m[blake2s_sigma[r][2*i + 1]];
95 	add		\a0, \a0, \b0, ror #12
96 	add		\a1, \a1, \b1, ror #12
100 	// d = ror32(d ^ a, 8);
109 	eor		\b0, \c0, \b0, ror#12
110 	eor		\b1, \c1, \b1, ror#12
132 	// (v[0], v[4], v[8], v[12]) and (v[1], v[5], v[9], v[13]).
133 	__ldrd		r10, r11, sp, 16	// load v[12] and v[13]
152 	// (v[0], v[5], v[10], v[15]) and (v[1], v[6], v[11], v[12]).
153 	ldr		r10, [sp, #16]		// load v[12]
194 	// to the stack to save a register during the rounds.  This also has the
216 	eor		r4, r4, r10		// v[12] = IV[4] ^ t[0]
230 	_blake2s_round	0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
231 	_blake2s_round	14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3
232 	_blake2s_round	11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4
233 	_blake2s_round	7, 9, 3, 1, 13, 12, 11, 14, 2, 6, 5, 10, 4, 0, 15, 8
234 	_blake2s_round	9, 0, 5, 7, 2, 4, 10, 15, 14, 1, 11, 12, 6, 8, 3, 13
235 	_blake2s_round	2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9
236 	_blake2s_round	12, 5, 1, 15, 14, 13, 4, 10, 0, 7, 6, 3, 9, 2, 8, 11
237 	_blake2s_round	13, 11, 7, 14, 12, 1, 3, 9, 5, 0, 15, 4, 8, 6, 2, 10
238 	_blake2s_round	6, 15, 14, 9, 11, 3, 0, 8, 12, 2, 13, 7, 1, 4, 10, 5
239 	_blake2s_round	10, 2, 8, 4, 7, 6, 1, 5, 15, 11, 9, 14, 3, 12, 13, 0
260 	pop		{r8-r11}		// load v[12..15]