movsx disassembly

movsx.py

import subprocess
import io
import re

def main():
    #         64     32      16      8
    regs = [('rax', 'eax' , 'ax'  , 'al'), #r0
            ('rdi', 'edi' , 'di'  , 'dl'),
            ('rsi', 'esi' , 'si'  , 'sil'),
            ('rdx', 'edx' , 'dx'  , 'dl'),
            ('rcx', 'ecx' , 'cx'  , 'cl'),
            ('r8' , 'r8d' , 'r8w' , 'r8b'),
            ('rbx', 'ebx' , 'bx'  , 'bl'),
            ('r13', 'r13d', 'r13w', 'r13b'),
            ('r14', 'r14d', 'r14w', 'r14b'),
            ('r15', 'r15d', 'r15w', 'r15b'), #r9
        ]
    B64, B32, B16, B8 = 0, 1, 2, 3
    BIT_WIDTH = { B64: 64, B32: 32, B16: 16, B8: 8 }

    def ereg(s):
        return not not re.match(r'r[0-9+]', s)

    def gen_asm(dst_size, src_size, edst, esrc):
        dst_kind = '<ereg>' if edst else '<non-ereg>'
        src_kind = '<ereg>' if esrc else '<non-ereg>'
        dst_width = BIT_WIDTH[dst_size]
        src_width = BIT_WIDTH[src_size]
        hdr=f'# {dst_width} dst, {src_width} src, movsx {dst_kind} {src_kind}'
        with io.StringIO() as f:
            for _dst in regs:
                if ereg(_dst[0]) != edst:
                    continue
                dst = _dst[dst_size]
                for _src in regs:
                    if ereg(_src[0]) != esrc:
                        continue
                    src = _src[src_size]
                    print(f'movsx {dst}, {src}', file=f)
            return hdr, f.getvalue()

    def print_asm(asm):
        shell = '''
        as -o /dev/null -aln -msyntax=intel -mnaked-reg
        '''
        p = subprocess.run(['bash', '-c', shell],
                           input=asm, capture_output=True, text=True)
        if p.stderr:
            print(p.stderr)
        print()
        for line in p.stdout.split('\n'):
            line = line.strip().split()
            if len(line) < 2:
                continue
            hex_bytes = re.sub(r'..', lambda m: f'{m[0]} ', line[2]).split()
            if len(hex_bytes) != 4:
                hex_bytes.insert(0, '  ')
            hex_encoding = ' '.join(hex_bytes)
            disasm = ' '.join(line[3:])
            print(f'{hex_encoding}   {disasm}')
        print()

    for dst in [B64, B32]:
        for src in [B32, B16, B8]:
            if dst == src:
                continue
            for edst in [False, True]:
                for esrc in [False, True]:
                    hdr, asm = gen_asm(dst, src, edst, esrc)
                    print(hdr)
                    print_asm(asm)

if __name__ == '__main__':
    main()