On 04/14/2014 04:32 AM, Claudio Fontana wrote: > the problem is not in the two nibbles you show, but in the third nibble: > 31 30 29 28 27 26 25 24 23 22 21 20 > size 1 1 1 v 0 0 opc 0 x > > the third nibble in your mask is 'E' and the expected result is 0, which > forces opc to be 0 for writes. > However, for 128bit SIMD STR, the opc is 2 (0b10).
Ah, I see what you mean. Unfortunately, this means a simple mask won't work for this case; it'll take two masks to decode that properly. r~
