arm64: lift scvtf

[yrp604]

Partially fixes #5417

From the attached example:

  41 @ 100003ea0  s1 = [x29 - 8 {var_18}].d
  42 @ 100003ea4  s0 = s1
  43 @ 100003ea8  unimplemented  {sxtl v0.2d, v0.2s}
  44 @ 100003eac  d1 = int.q(d0) ; <- this is a lifted scvtf
  45 @ 100003eb0  d0 = 10.0
  46 @ 100003eb4  d0 = d0 f+ d1
  47 @ 100003eb8  s0 = fconvert.s(d0)
  48 @ 100003ebc  x8 = sx.q([x29 - 8 {var_18}].d)
  49 @ 100003ec0  x9 = 0x2c
  50 @ 100003ec4  x9 = x8 * x9
  51 @ 100003ec8  x8 = &main::x
  52 @ 100003ecc  x8 = x8
  53 @ 100003ed0  x8 = x8 + x9
  54 @ 100003ed4  [x8 + 0x28].d = s0
  55 @ 100003ed8  [x29 - 0xc {var_1c_1}].d = 0
  56 @ 100003edc  goto 57 @ 0x100003ee0

NOTE: the neon code gen script is broken for me:

C:\Users\x\OneDrive\Documents\binja\binaryninja-api\arch\arm64\misc\neon_intrins.py:59: SyntaxWarning: invalid escape sequence '\d'
  m = re.match('^.* - (Vt(\d)\..*)}', reg0)
Traceback (most recent call last):
  File "C:\Users\x\OneDrive\Documents\binja\binaryninja-api\arch\arm64\misc\neon_intrins.py", line 124, in <module>
    (fsig, asig) = l.split('; // ')
    ^^^^^^^^^^^^
ValueError: not enough values to unpack (expected 2, got 1)

So I deleted the relevant scvtf code from the source and sink and called it Good Enough.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[yrp604]

Also, we might want to do UCVTF at the same time...

[galenbwill]

Superseded by 12d47835