rotl32
s1 = rotl32(x07 + x3, 4); \
s2 = rotl32(x07 + x6, 8); \
s5 = rotl32(x04 + x23 + x5, 22); \
s6 = rotl32(x17 + x56 + x0, 24); \
s3 = rotl32(x26 + x7, 13); \
s4 = rotl32(x26 + x1, 17); \
s7 = rotl32(x26 + x4, 29); \
t2 = rotl32(y05 + y3, 9); \
t1 = rotl32(y04 + y6, 5); \
t4 = rotl32(y04 + y5, 15); \
t3 = rotl32(y37 + y2, 11); \
t7 = rotl32(y37 + y0, 27); \
t5 = rotl32(y27 + y4, 20); \
t6 = rotl32(y27 + y1, 25); \