[PATCH 0/2] x86: a little bit of further opcode table rework

[Jan Beulich]

There's still a lot of redundancy and inefficiency in there, which
I'm hoping to continue to improve over time. These are two small
steps. The changes are technically independent, but of course the
2nd heavily depends on the 1st contextually, i.e. for the patch
to actually apply.

1: infer operand count of templates
2: fold some prefix related attributes into a single one

Jan