Exploiting a buffer overflow in a jmp_buf struct

Question

I need help to exploit a buffer overflow on a jmp_buf structure.

I have the following values on the stack (seen by gdb):

0xbffffc40:     0xb7fd8ff4      0x080485a0      0x080483f0      0xbffffcf8
0xbffffc50:     0xebf06081      0x1d0a15ee

The second one is the next IP (after the setjmp) but if i modify it i do not change the behavior of my program.

I noticed that I can change the flow (and obtain a Segmentation) only if i modify the value of the last one.

But, I can not understand the meaning of the last word and how to change it to obtain the desired behavior.

Answer 1

I am trying to do the same at the moment - i do not know what os you are using, but i found the following docu (@ http://freebsd.active-venture.com/FreeBSD-srctree/newsrc/dev/vinum/vinummemory.c.html) for freebsd which simpy discribes the jmpbuf struct:

struct JmpBuf {
    int jb_ebx; //in your case --> 0xb7fd8ff4      
    int jb_esp; //in your case --> 0x080485a0
    int jb_ebp; //in your case --> 0x080483f0      
    int jb_esi; //in your case --> 0xbffffcf8
    int jb_edi; //in your case --> 0xebf06081      
    int jb_eip; //in your case --> 0x1d0a15ee
    };

Therefore 0x080485a0 would be the stored ESP and not EIP which would explain the segfault in your case...