Open
Description
Add supports for floating pointer registers, basic instructions and inline assembly.
Foundation
- Subtarget feature. 68k doesn't have builtin floating point support until 68040. Before that, 68020 and '030 use M68881 and M68882 coprocessor for FP computations, respectively. Thus we need a subtarget feature for this. ([M68k] Add basic floating point infrastructures #61802)
- Register file definitions. Data registers (FP0 - FP7), FPCR, FPSR, and FPIAR. ([M68k] Add basic floating point infrastructures #61802)
Basic instruction definitions
Instruction definitions and MC supports for common FP instructions, for instance:
- FMOVE ([M68k] Add basic floating point infrastructures #61802)
- FADD, FSUB, FMUL, FDIV, FABS ([M68k] Add definitions for floating point arithmetic instructions #61803)
- FTST, FCMP, FBcc, FScc ([M68k] Add definitions for basic floating point compare and control flow instructions #61804)
Advanced instruction supports
- Inline assembly supports. ([M68k] Inline assembly support for floating point constructions #61806)
- Codegen supports for the aforementioned instructions. ([M68k] Add codegen support for basic floating point instructions #61805)