66 #define roundsm32(x0, x1, x2, x3, x4, x5, x6, x7, t0, t1, t2, t3, t4, t5, t6, \  argument
93 	vextracti128 $1, x7, t1##_x; \
112 	vaesenclast t4##_x, t1##_x, t1##_x; \
113 	vinserti128 $1, t1##_x, x7, x7; \
123 	vbroadcasti128 .Lpost_tf_hi_s1, t1; \
140 	filter_8bit(x0, t0, t1, t7, t6); \
141 	filter_8bit(x7, t0, t1, t7, t6); \
142 	filter_8bit(x3, t0, t1, t7, t6); \
143 	filter_8bit(x6, t0, t1, t7, t6); \
158 	vpsrldq $1, t0, t1; \
160 	vpshufb t7, t1, t1; \
218 	vpxor t1, x6, x6; \
306 #define rol32_1_32(v0, v1, v2, v3, t0, t1, t2, zero) \  argument
311 	vpcmpgtb v1, zero, t1; \
313 	vpabsb t1, t1; \
325 	vpor t1, v2, v2; \
336 #define fls32(l, l0, l1, l2, l3, l4, l5, l6, l7, r, t0, t1, t2, t3, tt0, \  argument
349 	vpshufb tt0, t0, t1; \
354 	vpand l1, t1, t1; \
358 	rol32_1_32(t3, t2, t1, t0, tt1, tt2, tt3, tt0); \
363 	vpxor l5, t1, l5; \
380 	vpshufb tt0, t0, t1; \
385 	vpor 5 * 32(r), t1, t1; \
390 	vpxor 1 * 32(r), t1, t1; \
395 	vmovdqu t1, 1 * 32(r); \
408 	vpshufb tt0, t0, t1; \
413 	vpand 1 * 32(r), t1, t1; \
417 	rol32_1_32(t3, t2, t1, t0, tt1, tt2, tt3, tt0); \
420 	vpxor 5 * 32(r), t1, t1; \
425 	vmovdqu t1, 5 * 32(r); \
439 	vpshufb tt0, t0, t1; \
444 	vpor l5, t1, t1; \
450 	vpxor l1, t1, l1; \
457 #define transpose_4x4(x0, x1, x2, x3, t1, t2) \  argument
461 	vpunpckldq x3, x2, t1; \
464 	vpunpckhqdq t1, x0, x1; \
465 	vpunpcklqdq t1, x0, x0; \