[PATCH 0/4] x86: follow-on to making disassembler thread-safe

[Jan Beulich]

Alan had noticed ubsan fallout from that combined with my conversion
of op_riprel to bool[]. In the course of addressing that I did notice
further things which could do with improving (or so I hope).

1: properly initialize struct instr_info instance(s)
2: shrink prefix related disassembler state fields
3: avoid string copy when swapping Vex.W controlled operands
4: replace global scratch buffer

Jan