Closed egyjuzer closed 10 years ago
How stable is "msvcrt" going to be? New application installs can upgrade this, even across the same OS/SP combination, let alone language pack.
Hi. From irb
:
1.9.3-p547 :001 > [ 0x77c34fbf, 0x200f0704 ].pack("V")
=> "\xBFO\xC3w"
1.9.3-p547 :002 > "" << [ 0x77c34fbf ].pack("V")
=> "\xBFO\xC3w"
1.9.3-p547 :003 > _ << [ 0x200f0704 ].pack("V")
=> "\xBFO\xC3w\x04\a\x0F "
1.9.3-p547 :004 > [ 0x77c34fbf, 0x200f0704 ].pack("VV")
=> "\xBFO\xC3w\x04\a\x0F "
1.9.3-p547 :005 > [ 0x77c34fbf, 0x200f0704 ].pack("V*")
=> "\xBFO\xC3w\x04\a\x0F "
1.9.3-p547 :006 >
So, it looks like the pack
should be either VV
or V*
. Thanks.
the problem as i saw saw not with msvcrt, but the source code itself. i'm mainly a asm/c developer, so beyond python i don't know any other widely used scripting language in detail, but i noticed that the way it was originally, the actual generated file did not contain the stack address 0x200f0704. i dont know ruby that well but it was as if the " , 0x200f0704 ] " part was ignored
so it was like:
[generated junk] [generated junk] [generated junk] [generated junk] [0x77c34fbf]# pop esp # ret # from msvcrt [generated junk] [generated junk] [generated junk] [generated junk]
instead of
[generated junk] [generated junk] [generated junk] [generated junk] [0x77c34fbf]# pop esp # ret # from msvcrt [0x200f0704] [generated junk] [generated junk] [generated junk]
you may be right, but what bothers me is that this particular installation was a freshly installed xp, sp3, with absolutely no updates, i think it was installed from an few yeard old cd.
i just wanted to share this information because i thought that it could help others to know where to look if any problem arises when trying to run the exploit
On Tue, Aug 26, 2014 at 6:59 PM, HD Moore notifications@github.com wrote:
How stable is "msvcrt" going to be? New application installs can upgrade this, even across the same OS/SP combination, let alone language pack.
— Reply to this email directly or view it on GitHub https://github.com/rapid7/metasploit-framework/issues/3706#issuecomment-53444539 .
@egyjuzer, @hmoore-r7: Fix is up at #3710. Closing this. We can resume discussion there.
Tried to reproduce (http://www.rapid7.com/db/modules/exploit/windows/fileformat/mswin_tiff_overflow) on a Windows XP SP3, with Office 2010 Professional. The guide says the exploit works on Office 2010 Standard SP1.
Another user discovered the same issue with the module: (http://www.crowdstrike.com/blog/analysis-cve-2013-3906-exploit/).
I have found a semi-solution to the problem, that at least allowed the rop-chain to reach and successfully execute VirtualProtect.
Here is the fix:
change to: