More x86 fixes for floating point operations.

- Fixed encoding for cvtss2sd and cvtsd2ss (they were swapped).

- Implemented neg-float and neg-double in assembly.

- Fixed float-to-int and double-to-int to return 0 if given NaN.

- Created helpers for rem-float and rem-double.

Change-Id: I456fcc1252a63dbb96c8280562e601f4efa3a5df
5 files changed
tree: 40417b04b3599c5bdef8fbbc7387f371d2b1601d
  1. build/
  2. jdwpspy/
  3. src/
  4. test/
  5. tools/
  6. .gitignore
  7. Android.mk