Linux C++ 调试神技--如何将Linux C++ 可执行文件逆向工程到Intel格式汇编
对于许多在windows 上调试代码的人而言， Intel IA32格式的汇编代码可能并不陌生，因为种种的原因，很多软件工程师不得不去尝试理解汇编代码。Windows PE的反汇编格式默认是Intel格式的，但是由于历史的原因，在这个世界上还存在另外一种汇编，ATT格式，这也是Linux C++ 可执行代码的默认反汇编格式。
难道还真需要哥们学习两种格式的汇编么？一种学会了已经很NB了？
难道哥们在Windows上学的汇编到Linux上就白费了么?玩不转了么？
底层的处理器都是一个架构的，机器码都是一样的，这两种汇编代码一定可以互相转换，否则工具设计者智商一定低到写不出来工具。
对于这个问题且听兄弟以一个例子详细到道来。假设有如下的C++代码，我们将其在Linux上编译为一个名字为hellod的执行文件。
#include<iostream>
int main()
{
std::cout << "Enter two numbers:" << std::endl;
int v1 = , v2 = ;
std::cin >> v1 >> v2;
std::cout << "The sum of " << v1 << " and " << v2
<< " is " << v1 + v2 << std::endl;
return ;
}
如果想看看现在的默认反汇编格式是什么，可以使用如下命令，可以看到Linux默认的是ATT格式的
(gdb) show disassembly-flavor
The disassembly flavor is "att".
反汇编结果如下:
(gdb) disassemble main
Dump of assembler code for function main():
0x000000000040092d <+>: push %rbp
0x000000000040092e <+>: mov %rsp,%rbp
0x0000000000400931 <+>: push %r13
0x0000000000400933 <+>: push %r12
0x0000000000400935 <+>: push %rbx
0x0000000000400936 <+>: sub $0x18,%rsp
0x000000000040093a <+>: mov $0x400ad4,%esi
0x000000000040093f <+>: mov $0x6011a0,%edi
0x0000000000400944 <+>: callq 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x0000000000400949 <+>: mov $0x400830,%esi
0x000000000040094e <+>: mov %rax,%rdi
0x0000000000400951 <+>: callq 0x400820 <_ZNSolsEPFRSoS_E@plt>
0x0000000000400956 <+>: movl $0x0,-0x28(%rbp)
0x000000000040095d <+>: movl $0x0,-0x24(%rbp)
0x0000000000400964 <+>: lea -0x28(%rbp),%rax
0x0000000000400968 <+>: mov %rax,%rsi
0x000000000040096b <+>: mov $0x601080,%edi
0x0000000000400970 <+>: callq 0x400810 <_ZNSirsERi@plt>
0x0000000000400975 <+>: lea -0x24(%rbp),%rdx
0x0000000000400979 <+>: mov %rdx,%rsi
0x000000000040097c <+>: mov %rax,%rdi
---Type <return> to continue, or q <return> to quit---
0x000000000040097f <+>: callq 0x400810 <_ZNSirsERi@plt>
0x0000000000400984 <+>: mov -0x28(%rbp),%edx
0x0000000000400987 <+>: mov -0x24(%rbp),%eax
0x000000000040098a <+>: lea (%rdx,%rax,),%r13d
0x000000000040098e <+>: mov -0x24(%rbp),%ebx
0x0000000000400991 <+>: mov -0x28(%rbp),%r12d
0x0000000000400995 <+>: mov $0x400ae7,%esi
0x000000000040099a <+>: mov $0x6011a0,%edi
0x000000000040099f <+>: callq 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009a4 <+>: mov %r12d,%esi
0x00000000004009a7 <+>: mov %rax,%rdi
0x00000000004009aa <+>: callq 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009af <+>: mov $0x400af3,%esi
0x00000000004009b4 <+>: mov %rax,%rdi
0x00000000004009b7 <+>: callq 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009bc <+>: mov %ebx,%esi
0x00000000004009be <+>: mov %rax,%rdi
0x00000000004009c1 <+>: callq 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009c6 <+>: mov $0x400af9,%esi
0x00000000004009cb <+>: mov %rax,%rdi
0x00000000004009ce <+>: callq 0x400800 <_ZStlsISt11char_traitsIcEERSt13---Type <return> to continue, or q <return> to quit---
basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009d3 <+>: mov %r13d,%esi
0x00000000004009d6 <+>: mov %rax,%rdi
0x00000000004009d9 <+>: callq 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009de <+>: mov $0x400830,%esi
0x00000000004009e3 <+>: mov %rax,%rdi
0x00000000004009e6 <+>: callq 0x400820 <_ZNSolsEPFRSoS_E@plt>
0x00000000004009eb <+>: mov $0x0,%eax
0x00000000004009f0 <+>: add $0x18,%rsp
0x00000000004009f4 <+>: pop %rbx
0x00000000004009f5 <+>: pop %r12
0x00000000004009f7 <+>: pop %r13
0x00000000004009f9 <+>: pop %rbp
0x00000000004009fa <+>: retq
End of assembler dump.
那如果我看不懂，我想使用Intel格式的汇编怎么办呢？下面的命令就可以做到，牛X吧？哈哈
(gdb) set disassembly-flavor intel
(gdb) show disassembly-flavor
The disassembly flavor is "intel".
再来看看这下我们反出来的汇编代码， 已经变成Intel 格式的了
(gdb) disassemble main
Dump of assembler code for function main():
0x000000000040092d <+>: push rbp
0x000000000040092e <+>: mov rbp,rsp
0x0000000000400931 <+>: push r13
0x0000000000400933 <+>: push r12
0x0000000000400935 <+>: push rbx
0x0000000000400936 <+>: sub rsp,0x18
0x000000000040093a <+>: mov esi,0x400ad4
0x000000000040093f <+>: mov edi,0x6011a0
0x0000000000400944 <+>: call 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x0000000000400949 <+>: mov esi,0x400830
0x000000000040094e <+>: mov rdi,rax
0x0000000000400951 <+>: call 0x400820 <_ZNSolsEPFRSoS_E@plt>
0x0000000000400956 <+>: mov DWORD PTR [rbp-0x28],0x0
0x000000000040095d <+>: mov DWORD PTR [rbp-0x24],0x0
0x0000000000400964 <+>: lea rax,[rbp-0x28]
0x0000000000400968 <+>: mov rsi,rax
0x000000000040096b <+>: mov edi,0x601080
0x0000000000400970 <+>: call 0x400810 <_ZNSirsERi@plt>
0x0000000000400975 <+>: lea rdx,[rbp-0x24]
0x0000000000400979 <+>: mov rsi,rdx
0x000000000040097c <+>: mov rdi,rax
---Type <return> to continue, or q <return> to quit---
0x000000000040097f <+>: call 0x400810 <_ZNSirsERi@plt>
0x0000000000400984 <+>: mov edx,DWORD PTR [rbp-0x28]
0x0000000000400987 <+>: mov eax,DWORD PTR [rbp-0x24]
0x000000000040098a <+>: lea r13d,[rdx+rax*]
0x000000000040098e <+>: mov ebx,DWORD PTR [rbp-0x24]
0x0000000000400991 <+>: mov r12d,DWORD PTR [rbp-0x28]
0x0000000000400995 <+>: mov esi,0x400ae7
0x000000000040099a <+>: mov edi,0x6011a0
0x000000000040099f <+>: call 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009a4 <+>: mov esi,r12d
0x00000000004009a7 <+>: mov rdi,rax
0x00000000004009aa <+>: call 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009af <+>: mov esi,0x400af3
0x00000000004009b4 <+>: mov rdi,rax
0x00000000004009b7 <+>: call 0x400800 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009bc <+>: mov esi,ebx
0x00000000004009be <+>: mov rdi,rax
0x00000000004009c1 <+>: call 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009c6 <+>: mov esi,0x400af9
0x00000000004009cb <+>: mov rdi,rax
0x00000000004009ce <+>: call 0x400800 <_ZStlsISt11char_traitsIcEERSt13---Type <return> to continue, or q <return> to quit---
basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009d3 <+>: mov esi,r13d
0x00000000004009d6 <+>: mov rdi,rax
0x00000000004009d9 <+>: call 0x4007a0 <_ZNSolsEi@plt>
0x00000000004009de <+>: mov esi,0x400830
0x00000000004009e3 <+>: mov rdi,rax
0x00000000004009e6 <+>: call 0x400820 <_ZNSolsEPFRSoS_E@plt>
0x00000000004009eb <+>: mov eax,0x0
0x00000000004009f0 <+>: add rsp,0x18
0x00000000004009f4 <+>: pop rbx
0x00000000004009f5 <+>: pop r12
0x00000000004009f7 <+>: pop r13
0x00000000004009f9 <+>: pop rbp
0x00000000004009fa <+>: ret
End of assembler dump.
总结
对于很多计算机工程领域的技术问题，理解原理是最重要的，软件工程师很忙，忙着学东西，但是有的东西你一旦知道了道理，靠分析已经能节省很多的时间，根本不用去再学一遍，就拿本例来说，如果不去分析，再去学一遍另一个版本的，可能也是事倍功半，站的高方能望的远，和大家共勉。