49 #define roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, t0, t1, t2, t3, t4, t5, t6, \  argument
64 	vpshufb t4, x2, x2; \
76 	filter_8bit(x2, t0, t1, t7, t6); \
91 	vaesenclast t4, x2, x2; \
107 	filter_8bit(x2, t2, t3, t7, t6); \
135 	vpxor x7, x2, x2; \
138 	vpxor x2, x4, x4; \
145 	vpxor x5, x2, x2; \
151 	vpxor x2, x7, x7; /* note: high and low parts swapped */ \
179 	vpxor t5, x2, x2; \
180 	vpxor 6 * 16(mem_cd), x2, x2; \
211 #define two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
222 	vmovdqu x2, 6 * 16(mem_cd); \
228 	store_ab(x0, x1, x2, x3, x4, x5, x6, x7, mem_ab);
230 #define dummy_store(x0, x1, x2, x3, x4, x5, x6, x7, mem_ab) /* do nothing */  argument
232 #define store_ab_state(x0, x1, x2, x3, x4, x5, x6, x7, mem_ab) \  argument
236 	vmovdqu x2, 2 * 16(mem_ab); \
243 #define enc_rounds16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
245 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
247 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
249 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
252 #define dec_rounds16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
254 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
256 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
258 	two_roundsm16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \
418 #define transpose_4x4(x0, x1, x2, x3, t1, t2) \  argument
422 	vpunpckldq x3, x2, t1; \
423 	vpunpckhdq x3, x2, x2; \
428 	vpunpckhqdq x2, t2, x3; \
429 	vpunpcklqdq x2, t2, x2;
481 #define inpack16_pre(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
499 	vpxor 13 * 16(rio), x0, x2; \
504 #define inpack16_post(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
506 	byteslice_16x16b(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, \
511 	vmovdqu x2, 2 * 16(mem_ab); \
527 #define outunpack16(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, \  argument
529 	byteslice_16x16b(y0, y4, x0, x4, y1, y5, x1, x5, y2, y6, x2, x6, y3, \
550 	vpxor x0, x2, x2; \
554 #define write_output(x0, x1, x2, x3, x4, x5, x6, x7, y0, y1, y2, y3, y4, y5, \  argument
558 	vmovdqu x2, 2 * 16(rio); \