This patch adds half-precision floating-point conversion instructions introduced in v8.2 as part of the FEAT_FP16 feature. These are an extension of existing v8.0 floating-point conversion instructions handled in codec_v80.txt.
FCVTAS <Vd>.<T>, <Vn>.<T> FCVTNS <Vd>.<T>, <Vn>.<T>
FCVTAS <Hd>, <Hn> FCVTNS <Hd>, <Hn>
FCVTAS <Wd>, <Hn> FCVTNS <Wd>, <Hn>
FCVTAS <Xd>, <Hn> FCVTNS <Xd>, <Hn>
FCVTAU <Vd>.<T>, <Vn>.<T> FCVTPS <Vd>.<T>, <Vn>.<T>
FCVTAU <Hd>, <Hn> FCVTPS <Hd>, <Hn>
FCVTAU <Wd>, <Hn> FCVTPS <Wd>, <Hn>
FCVTAU <Xd>, <Hn> FCVTPS <Xd>, <Hn>
FCVTMS <Vd>.<T>, <Vn>.<T> FCVTPU <Vd>.<T>, <Vn>.<T>
FCVTMS <Hd>, <Hn> FCVTPU <Hd>, <Hn>
FCVTMS <Wd>, <Hn> FCVTPU <Wd>, <Hn>
FCVTMS <Xd>, <Hn> FCVTPU <Xd>, <Hn>
Issue: #2626