Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 1

Buffer-Overflow Vulnerabilities and Attacks

1 Memory
In the PC architecture there are four basic read-write memory regions in a program: Stack, Data, BSS
(Block Started by Symbol), and Heap. The data, BSS, and heap areas are collectively referred to as the
”data segment”. In the tutorial titled “Memory Layout And The Stack” [1], Peter Jay Salzman described
memory layout in a great detail.

• Stack: Stack typically located in the higher parts of memory. It usually ”grows down”: from high
address to low address. Stack is used whenever a function call is made.

• Data Segment

– Data area: contains global variables used by the program that are not initialized to zero. For
instance the string “hello world” defined by char s[] = "hello world" in C would
exist in the data part.
– BSS segment: starts at the end of the data segment and contains all global variables that are
initialized to zero. For instance a variable declared static int i would be contained in the
BSS segment.
– Heap area: begins at the end of the BSS segment and grows to larger addresses from there. The
Heap area is managed by malloc, realloc, and free. The Heap area is shared by all shared
libraries and dynamic load modules in a process.

2 Stack Buffer Overflow

2.1 Background about Stack
• Stack Layout: the following figure shows the stack layout after the execution has entered the function
func().

High Address
void func (char *str) {
char buffer[12];
int variable_a;
strcpy (buffer, str); str (a pointer to a string)
Current Frame

} Return Address

Int main() { Previous Frame Pointer (FP)

Current FP
char *str = “I am greater than 12 bytes”; buffer[0] … buffer[11]
func (str);
} variable_a

Low Address

(a) A code example (b) Active Stack Frame in func()

• Stack Direction: Stack grows from high address to low address (while buffer grows from low address
to high address)
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 2

• Return Address: address to be executed after the function returns.

– Before entering a function, the program needs to remember where to return to after return from
the function. Namely the return address has to be remembered.
– The return address is the address of the instruction right after the function call.
– The return address will be stored on the stack. In Intel 80x86, the instruction call func will
pushes the address of the next instruction that immediately follows the call statement into the
stack (i.e. in the return address region), and then jumps to the code of function func().

• Frame Pointer (FP): is used to reference the local variables and the function parameters. This pointer
is stored in a register (e.g. in Intel 80x86, it is the ebp register). In the following, we use $FP to
represent the value of the FP register.

– variable_a will be referred to as ($FP-16).

– buffer will be referred to as ($FP-12).
– str will be referred to as ($FP+8).

• Buffer-Overflow Problem: The above program has a buffer-overflow problem.

– The function strcpy(buffer, str) copies the contents from str to buffer[].
– The string pointed by str has more than 12 chars, while the size of buffer[] is only 12.
– The function strcpy() does not check whether the boundary of buffer[] has reached. It
only stops when seeing the end-of-string character ’\0’.
– Therefore, contents in the memory above buffer[] will be overwritten by the characters at
the end of str.

2.2 A Vulnerable Program

Now, let us look at a more complicated program. Unlike the previous program, the string that is used to
overflow the return address is not a static string; it is actually provided by users. In other words, users can
decide what should be included in this string.

/* stack.c */

/* This program has a buffer overflow vulnerability. */

/* Our task is to exploit this vulnerability */
#include <stdlib.h>
#include <stdio.h>
#include <string.h>

int func (char *str)

{
char buffer[12];

/* The following statement has a buffer overflow problem */

strcpy(buffer, str);
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 3

return 1;
}

int main(int argc, char **argv)

{
char str[517];
FILE *badfile;

badfile = fopen("badfile", "r");

fread(str, sizeof(char), 517, badfile);
func (str);
printf("Returned Properly\n");
return 1;
}

It is not so difficult to see that the above program has a buffer overflow problem. The program first reads
an input from a file called “badfile”, and then passes this input to another buffer in the function bof(). The
original input can have a maximum length of 517 bytes, but the buffer in bof() has only 12 bytes long.
Because strcpy() does not check boundaries, buffer overflow will occur. If this program is running as
a set-root-uid program, a normal user can exploit this buffer overflow vulnerability and take over the root
privileges.

2.3 Exploit the Buffer-Overflow Vulnerability

To fully exploit a stack buffer-overflow vulnerability, we need to solve several challenging problems.

• Injecting the malicious code: We need to be able to inject the malicious code into the memory of the
target process. This can be done if we can control the contents of the buffer in the targeted program.
For example, in the above example, the program gets the input from a file. We can store the malicious
code in that file, and it will be read into the memory of the targeted program.

• Jumping to the malicious code: With the malicious code already in the memory, if the targeted
program can jump to the starting point of the malicious code, the attacker will be in control.

• Writing malicious code: Writing a malicious code is not trivial. We will show how a special type of
malicious code, shellcode, can be written.

2.4 Injecting Malicious Code

With the buffer overflow vulnerability in the program, we can easily inject malicious code into the memory
of the running program. Let us assume that the malicious code is already written (we will discuss how to
write malicious code later).
In the above vulnerable program, the program reads the contents from the file "badfile", and copy
the contents to buffer. Therefore, we can simply store the malicious code (in binary form) in the
"badfile", the vulnerable program will copy the malicious code to the buffer on the stack (it will
overflow the buffer).
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 4

2.5 Jumping to the Malicious Code

Malicious Code Malicious Code

Stack’s growing direction

NOP
NOP
…… (many NOP’s)
NOP

str str

Return Address Return Address

Previous FP Previous FP

buffer [0] …... buffer [11] buffer [0] …... buffer [11]

(a) Jump to the malicious code (b) Improve the chance

• To jump to the malicious code that we have injected into the target program’s stack, we need to know
the absolute address of the code. If we know the address before hand, when overflowing the buffer,
we can use this address to overwrite the memory that holds the return address. Therefore, when the
function returns, it will return to our malicious code.

• The challenge is to find where the malicious code starts.

• If the target program is a Set-UID program, you can make a copy of this program, and run it with
your own privilege; this way you can debug the program (note that you cannot debug a Set-UID
program). In the debugger, you can figure out the address of buffer[], and thus calculate the
starting point of the malicious code. The address of buffer[] may be slightly different when you
run the Set-UID copy, instead of of your copy, but you should be quite close. You can try several
values.

• If the target program is running remotely, and you may not be able to rely on the debugger to find
out the address. However, you can always guess. The following facts make guessing a quite feasible
approach:

– Stack usually starts at the same address.

– Stack is usually not very deep: most programs do not push more than a few hundred or a few
thousand bytes into the stack at any one time.
– Therefore the range of addresses that we need to guess is actually quite small.

• Improving the chance: To improve the chance of success, we can add many NOP operations to the
beginning of the malicious code. NOP is a special instruction that does nothing other than advancing
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 5

to the next instruction. Therefore, as long as the guessed address points to one of the NOPs, the attack
will be successful. With NOPs, the chance of guessing the correct entry point to the malicious code
is significantly improved.

2.6 Malicious Code: Shellcode

In the previous discussion, we assume that the malicious code is already available. In this subsection, we
discuss how to write such malicious code.
If we can ask the privileged program to run our code, what code do we want it to run? The most powerful
code that we want it to run is to invoke a shell, so we can run any command we want in that shell. A program
whose only goal is to launch a shell is called a shellcode. To learn how to write a shellcode, let us see the
following C program:

#include <stdio.h>

int main( ) {
char *name[2];

name[0] = ‘‘/bin/sh’’;
name[1] = NULL;
execve(name[0], name, NULL);
}

After we compile the above program into binary code, can we directly use the binary code as our shell-
code in the buffer-overflow attack? Things are not that easy. There are several problems if we directly use
the above code:

• First, to invoke the system call execve(), we need to know the address of the string “/bin/sh”.
Where to store this string and how to derive the location of this string are not trivial problems.

• Second, there are several NULL (i.e., 0) in the code. This will cause strcpy to stop. If the vulnera-
bility is caused by strcpy, we will have a problem.

To solve the first problem, we can push the string “/bin/sh” onto stack, and then use the stack pointer
esp to get the location of the string. To solve the second problem, we can convert the instructions that
contain 0 into another instructions that do not contain 0. For example, to store 0 to a register, we can use
XOR operation, instead of directly assigning 0 to that register. The following is an example of shellcode in
assembly code:

Line 1: xorl %eax,%eax

Line 2: pushl %eax # push 0 into stack (end of string)
Line 3: pushl $0x68732f2f # push "//sh" into stack
Line 4: pushl $0x6e69622f # push "/bin" into stack
Line 5: movl %esp,%ebx # %ebx = name[0]
Line 6: pushl %eax # name[1]
Line 7: pushl %ebx # name[0]
Line 8: movl %esp,%ecx # %ecx = name
Line 9: cdq # %edx = 0
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 6

Line 10: movb $0x0b,%al

Line 11: int $0x80 # invoke execve(name[0], name, 0)

A few places in this shellcode are worth mentioning:

• First, the third instruction pushes “//sh”, rather than “/sh” into the stack. This is because we need a
32-bit number here, and “/sh” has only 24 bits. Fortunately, “//” is equivalent to “/”, so we can get
away with a double slash symbol.
• Second, before calling the execve() system call, we need to store name[0] (the address of the
string), name (the address of the array), and NULL to the %ebx, %ecx, and %edx registers, respec-
tively.

– Line 5 stores name[0] to %ebx;

– Line 8 stores name to %ecx;
– Line 9 sets %edx to zero. There are other ways to set %edx to zero (e.g., xorl %edx,
%edx); the one (cdq) used here is simply a shorter instruction: it copies the sign (bit 31) of the
value in the EAX register (which is 0 at this point) into every bit position in the EDX register,
basically setting %edx to 0.

• Third, the system call execve() is called when we set %al to 11, and execute “int $0x80”.

If we convert the above shellcode into binary code, and store it in an array, we can call it from a C
program:

#include <stdlib.h>
#include <stdio.h>

const char code[] =

"\x31\xc0" /* Line 1: xorl %eax,%eax */
"\x50" /* Line 2: pushl %eax */
"\x68""//sh" /* Line 3: pushl $0x68732f2f */
"\x68""/bin" /* Line 4: pushl $0x6e69622f */
"\x89\xe3" /* Line 5: movl %esp,%ebx */
"\x50" /* Line 6: pushl %eax */
"\x53" /* Line 7: pushl %ebx */
"\x89\xe1" /* Line 8: movl %esp,%ecx */
"\x99" /* Line 9: cdq */
"\xb0\x0b" /* Line 10: movb $0x0b,%al */
"\xcd\x80" /* Line 11: int $0x80 */
;

int main(int argc, char **argv)

{
char buf[sizeof(code)];
strcpy(buf, code);
((void(*)( ))buf)( );
}
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 7

The statement ((void(*)( ))buf)() in the above main function will invoke a shell, because the
shellcode is executed.

3 Countermeasures
3.1 Apply Secure Engineering Principles
• Use strong type language, e.g. java, C#, etc. With these languages, buffer overflows will be detected.

• Use safe library functions.

– Functions that could have buffer overflow problem: gets, strcpy, strcat, sprintf,
scanf, etc.
– These functions are safer: fgets, strncpy, strncat, and snprintf.

3.2 Systmetic Code Modification

• StackShield: seperate control (return address) from data.

– It is a GNU C compiler extension that protects the return address.

– When a function is called, StackShield copies away the return address to a non-overflowable
area.
– Upon returning from a function, the return address is stored. Therefore, even if the return address
on the stack is altered, it has no effect since the original return address will be copied back before
the returned address is used to jump back.

• StackGuard: mark the boundary of buffer

– Observation: one needs to overwrite the memory before the return address in order to overwrite
the return address. In other words, it is difficult for attackers to only modify the return address
without overwriting the stack memory in front of the return address.
– A canary word can be placed next to the return address whenever a function is called.
– If the canary word has been altered when the function returns, then some attempt has been made
on the overflow buffers.
– StackGuard is also built into GNU C compiler.
– We can understand how StackGuard work through the following proram (we emulate the com-
piler, and manually add the protection code to the function). For the sake of simplicity, we only
use an integer for the canary word in the following example; this is not strong enough. We can
use several integers for the canary word.
/* This program has a buffer overflow vulnerability. */
/* However, it is protected by StackGuard */
#include <stdlib.h>
#include <stdio.h>
#include <string.h>

int func (char *str)

{
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 8

int canaryWord = secret;

char buffer[12];

/* The following statement has a buffer overflow problem */

strcpy(buffer, str);

if (canaryWord == secret) // Return address is not modified

return 1;
else // Return address is potentially modified
{ ... error handling ... }
}

static int secret; // a global variable

int main(int argc, char **argv)

{
// getRandomNumber will return a random number
secret = getRandomNumber();

char str[517];
FILE *badfile;

badfile = fopen("badfile", "r");

fread(str, sizeof(char), 517, badfile);
func (str);
printf("Returned Properly\n");
return 1;
}

3.3 Operating System Approach

• Address Space Randomization: Guessing the addresses of the malicious code is one of the critical
steps of buffer-overflow attacks. If we can make the address of the malicious code difficult to pre-
dict, the attack can be more difficult. Several Linux distributions have already used address space
randomization to randomize the starting address of heap and stack. This makes guessing the exact
addresses difficult. The following commands (can only run by root) enable or disable the address
space randomization.

# sysctl -w kernel.randomize_va_space=2 // Enable Randomization

# sysctl -w kernel.randomize_va_space=0 // Disable Randomization

Unfortunately, in 32-bit machines, even if the the addresses are randomized, the entropy is not large
enough against random guesses. In practice, if you try many times, your chance of success is quite
high. Our experience has shown that a few minutes of tries are enough to succeed in a Intel 2GHz
machine.
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 9

• Non-executable stack: From the attack, we can observe that the attackers put the malicious code in
the stack, and jump to it. Since the stack is a place for data, not for code, we can configure the stack
to be non-executable, and thus preventing the malicious code from being executed.
This protection scheme is called ExecShield. Several Linux distributions have already implemented
this protection mechanism. ExecShield essentially disallows executing any code that is stored in the
stack. The following commands (can only run by root) enable or disable ExecShield.

# sysctl -w kernel.exec-shield=1 // Enable ExecShield

# sysctl -w kernel.exec-shield=0 // Disable ExecShield

In the next section, we can see that such a protection scheme does not solve the buffer-overflow
problem, because another type of attack, called Return-to-libc attack does not need the stack to be
executable.

4 Non-Executable Stack and Return-to-libc Attack

To exploit the stack-based buffer overflow vulnerability, adversaries need to inject a piece of code into
the user stack, and then execute the code from the stack. If we can make the memory segement used for
stack non-executable, even if the code is injected into the stack, the code will not be able to execute. This
way, we can prevent the buffer-overflow attacks. Technically this is easy to achieve because modern CPU
architectures (such as Intel 386) do allow operating systems to turn a block of memory into non-executable
memory. However, things are not so easy: many operating systems, such as Linux, do save code into stacks,
and thus need the stack to be executable. For example, in Linux, to handle signals, a small sequence of
code is put on the user stack; this sequence will be executed when handling signals.
Newer versions of Linux have since made the stack for data only. Therefore, stacks can be configured to
be non-executable. In Fedora Linux, we can run the following commands to make stacks non-executable:

# /sbin/sysctl -w kernel.exec-shield=1

Unfortunately, making stacks non-executable cannot totally defeat the buffer-overflow attacks. It makes
running malicious code from the stack infeasible, but there are other ways to exploit buffer-overflow vulner-
abilities, without running any code from the stack. Return-to-libc attack is such an attack.
To understand this new attack, let us recall the main purpose of running the malicious shellcode from
the stack. We know it is to invoke a shell. The question is whether we can invoke a shell without using
injected code. This is actually doable: we can use the code in the operating system itself to invoke a shell.
More specifically, we can use the library functions of operating systems to achieve our goal. In Unix-like
operating systems, the shared library called libc provides the C runtime on UNIX style systems. This
library is essencial to most C programs, because it defines the “system calls” and other basic facilities such
as open, malloc, printf, system, etc. The code of libc is already in the memory as a shared runtime
library, and it can be accessed by all applications.
Function system is one of the functions in libc. If we can call this function with the argument
“/bin/sh”, we can invoke a shell. This is the basic idea of the Return-to-libc attack. The first part of
Return-to-libc attack is similar to the attack using shellcode, i.e., it overflows the buffer, and modify the
return address on the stack. The second part is different. Unlike the shellcode approach, the return address
is not pointed to any injected code; it points to the entry point of the function system in libc. If we do
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 10

it correctly, we can force the target program to run system("/bin/sh"), which basically launches a
shell.
Challenges. To succeed in the Return-to-libc attack, we need to overcome the following challenges:

• How to find the location of the function system?

• How to find the address of the string "/bin/sh"?

• How to pass the address of the string "/bin/sh" to the system function?

4.1 Finding the location of the system function.

In most Unix operating systems, the libc library is always loaded into a fixed memory address. To find
out the address of any libc function, we can use the following gdb commands (let a.out is an arbitrary
program):

$ gdb a.out
(gdb) b main
(gdb) r
(gdb) p system
$1 = {<text variable, no debug info>} 0x9b4550 <system>
(gdb) p exit
$2 = {<text variable, no debug info>} 0x9a9b70 <exit>

From the above gdb commands, we can find out that the address for the system() function is
0x9b4550, and the address for the exit() function is 0x9a9b70. The actual addresses in your sys-
tem might be different from these numbers.
We call also use functions dlopen and dlsym to find out the address location of a libc function:

#include <dlfcn.h>

#define LIBCPATH "/lib/libc.so.6" /* on Fedora */

void *libh, *sys;

if ((libh = dlopen(LIBCPATH, RTLD_NOW)) == NULL){

// report error
}

if (( sys = dlsym (libh, "system")) == NULL){

// report error
}
printf("system @ %p\n", sys);

4.2 Finding the address of “/bin/sh”.

There are many ways to find the address of such a string:
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 11

• Insert the string directly into the stack using the buffer overflow problem, and then guess its address.

• Before running the vulnerable program, create an environment variable with value “/bin/sh”. When
a C program is executed from a shell, it inherits all the environment variables from the shell. In the
following, we define a new shell variable MYSHELL and let its value be /bin/sh:

$ export MYSHELL=/bin/sh

We will use the address of this variable as an argument to system() call. The location of this
variable in the memory can be found out easily using the following program:

void main()
{ char* shell = getenv("MYSHELL");
if (shell)
printf("%x\n", shell);
}

If the stack address is not randomized, we will find out that the same address is printed out. However,
when we run another program, the address of the environment varable might not be exactly the same
as the one that you get by running the above program; such an address can even change when you
change the name of your program (the number of characters in the file name makes difference). The
good news is, the address of the shell will be quite close to what you print out using the above program.
Therefore, you might need to try a few times to succeed.

• We also know that the function system uses "/bin/sh" in its own code. Therefore, this string
must exist in libc. If we can find out the location of the string, we can use directly use this string.
You can search the libc library file (/lib/libc.so.6) for the string "rodata":

$ readelf -S /lib/lib.so.6 | egrep ’rodata’

[15] .rodata PROGBITS 009320e0 124030 ......

The result of the above command indicates that the ".rodata" section starts from 0x009320e0.
The ".rodata" section is used to store constant data, and the constant string "/bin/sh" should
be stored in this section. You can write a program to search for the string in the memory starting from
0x00932030).

4.3 Passing the address of “/bin/sh” to system.

In order to let system run the command "/bin/sh", we need to pass the address of this command string
as an argument to system. Just like invoking any function, we need to pass the argument via the stack.
Therefore, we need to put the argument in a correct place on the stack. To do this correctly, we need to
clearly understand how the stack frame for a function is constructed when invoking a function. We use a
small C program to understand the effects of a function invocation on the stack.

/* foobar.c */
#include<stdio.h>
void foo(int x)
{
printf("Hello world: %d\n", x);
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 12

int main()
{
foo(1);
return 0;
}

We can use "gcc -S foobar.c" to compile this program to the assembly code. The resulting file
foobar.s will look like the following:

......
8 foo:
9 pushl %ebp
10 movl %esp, %ebp
11 subl $8, %esp
12 movl 8(%ebp), %eax
13 movl %eax, 4(%esp)
14 movl $.LC0, (%esp) : string "Hello world: %d\n"
15 call printf
16 leave
17 ret
......
21 main:
22 leal 4(%esp), %ecx
23 andl $-16, %esp
24 pushl -4(%ecx)
25 pushl %ebp
26 movl %esp, %ebp
27 pushl %ecx
28 subl $4, %esp
29 movl $1, (%esp)
30 call foo
31 movl $0, %eax
32 addl $4, %esp
33 popl %ecx
34 popl %ebp
35 leal -4(%ecx), %esp
36 ret

Calling and Entering foo(). Let us concentrate on the stack while calling foo(). We can ignore the
stack before that. Please note that line numbers instead of instruction addresses are used in this explanation.

• Line 28-29:: These two statements push the value 1, i.e. the argument to the foo(), into the stack.
This operation increments %esp by four. The stack after these two statements is depicted in Fig-
ure 1(a).
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 13

bfffe764 bfffe764 bfffe764

Parameters Parameters Parameters 00000001 bfffe760
00000001 00000001 bfffe760
bfffe760
esp Return addr 080483dc bfffe75c
Return addr 080483dc bfffe75c
Old ebp bfffe768
esp bfffe758
esp ebp
(a) Line 28: subl $4, %esp
(b) Line 30: call foo (c) Line 9: push %ebp
Line 29: movl $1, (%esp)
Line 10: movl %esp, %ebp

bfffe764 bfffe764 bfffe764

Parameters 00000001
Parameters 00000001 Parameters 00000001 bfffe760
bfffe760 bfffe760
esp
Return addr 080483dc Return addr 080483dc bfffe75c
bfffe75c
Old ebp bfffe768 bfffe758 esp

Local ebp
variables
bfffe750
esp

(d) Line 11: subl $8, %esp (e) Line 16: leave (f) Line 17: ret

Figure 1: Entering and Leaving foo()

• Line 30: call foo: The statement pushes the address of the next instruction that immediately
follows the call statement into the stack (i.e the return address), and then jumps to the code of
foo(). The current stack is depicted in Figure 1(b).

• Line 9-10: The first line of the function foo() pushes %ebp into the stack, to save the previous
frame pointer. The second line lets %ebp point to the current frame. The current stack is depicted in
Figure 1(c).

• Line 11: subl $8, %esp: The stack pointer is modified to allocate space (8 bytes) for local
variables and the two arguments passed to printf. Since there is no local variable in function foo,
the 8 bytes are for arguments only. See Figure 1(d).

Leaving foo(). Now the control has passed to the function foo(). Let us see what happens to the stack
when the function returns.

• Line 16: leave: This instruction implicitly performs two instructions (it was a macro in earlier x86
releases, but was made into an instruction later):

mov %ebp, %esp

pop %ebp

The first statement release the stack space allocated for the function; the second statement recover the
previous frame pointer. The current stack is depicted in Figure 1(e).

• Line 17: ret: This instruction simply pops the return address out of the stack, and then jump to the
return address. The current stack is depicted in Figure 1(f).
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 14

• Line 32: addl $4, %esp: Further resotre the stack by releasing more memories allocated for
foo. As you can clearly see that the stack is now in exactly the same state as it was before entering
the function foo (i.e., before line 28).

Setting up the frame for system(). From Lines 9 and 10, we can see that the first thing that a function
does is to push the current %ebp value to stack, and then set the register %ebp to the top of the stack.
Although we see this from our example function foo(), other functions behave the same, including those
functions in libc. Therefore, within each function, after executing the first two instructions, %ebp points
to the the frame pointer of the previous frame, (%ebp + 4) points to the return address, and the location
above the return address should be where the arguments are stored. For function system(), (%ebp +
8) should be the address of the string passed to the function.
Therefore, if we can figure out what the stack pointer %esp points to after returnning from foo(),
we can put the the address of the string "/bin/sh" to the correct place, which is (%esp + 4). For
example, in Figure 1(d), if we want the function foo to return to system, we should put the starting
address of function system at %esp - 4 (0xbfffe75c), a return address at %esp (0xbfffe760)
and the address of the string "/bin/sh" at (%esp + 4) (0xbfffe764).
If we want the function system() to return to another function, such as exit(0), we can use the
starting address of function exit() as the return address of system, and put it in 0xbfffe760.

Note: Details of how to setup the frame for system() are intentionally left out. Students are asked to
work on a lab, in which they need to figure out all the details of the return-to-libc attack. We do not want
this lecture note to give students all the details.

4.4 Protection in /bin/bash

If the "/bin/sh" is pointed to "/bin/bash", even if we can invoke a shell within a Set-UID program
that is running with the root privilege, we will not get the root privilege. This is because bash automatically
downgrades its privilege if it is executed in the Set-UID root context;
However, there are ways to get around this protection scheme. Although /bin/bash has restriction on
running Set-UID programs, it does allow the real root to run shells. Therefore, if we can turn the current
Set-UID process into a real root process, before invoking /bin/bash, we can bypass that restriction
of bash. The setuid(0) system call can help you achieve that. Therefore, we need to first invoke
setuid(0), and then invoke system("/bin/sh"); all of these can be done using the Return-to-libc
mechanism.
Basically, we need to “return to libc” twice. We first let the target program to return to the setuid
function in libc. When this function returns, it will fetch the return address from the stack, and jump to
that address. If we can let this return address point to system, we can force the function setuid to return
to the entry point of system. We have to be very careful when conducting this process, because we have
to put the appropriate arguments in the right place of the stack.

5 Heap/BSS Buffer Overflow

• Contents in Heap/BSS

– Constant strings
– Global variables
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 15

– Static variables
– Dynamic allocated memory.

• Example: Overwriting File Pointers

/* The following variables are stored in the BSS region */

static char buf[BUFSIZE], *tmpfile;

tmpfile = "/tmp/vulprog.tmp";
gets(buf); /* buffer overflow can happen here */

... Open tmpfile, and write to it ...

0x884080
tmpfile: 0x884080 “/tmp/vulprog.tmp”
overflow

buff [16]
After the buffer overflow

0x903040
“/etc/shadow”

– The (Set-UID) program’s file pointer points to /tmp/vulprog.tmp.

– The program needs to write to this file during execution using the user’s inputs.
– If we can cause the file point to point to /etc/shadow, we can cause the program to write to
/etc/shadow.
– We can use the buffer overflow to change the content of the variable tmpfile. Originally, it
points to the "/tmp/vluprog.tmp" string. Using the buffer overflow vulnerability, we can
change the content of tmpfile to 0x903040, which is the address of the string "/etc/shadow".
After that, when the program use tmpfile varialble to open the file to write, it actually opens
the shadow file.
– How to find the address of /etc/shadow?
∗ We can pass the string as the argument to the program, this way the string /etc/shadow
is stored in the memory. We now need to guess where it is.

• Example: Overwriting Function Pointers

int main(int argc, char **argv)

{ static char buf[16]; /* in BSS */
static int (*funcptr)(const char *str); /* in BSS */
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 16

funcptr = (int (*)(const char *str))goodfunc;

/* We can cause buffer overflow here */

strncpy(buf, argv[1], strlen(argv[1]));

(void)(*funcptr)(argv[2]);
return 0;
}

/* This is what funcptr would point to if we didn’t overflow it */

int goodfunc(const char *str) { ... ... }

0xC48824
funcptr: 0xC48824 goodfunc(){ … }
overflow

buff [16]
After the buffer overflow

0x804408

malicious code

– A function pointer (i.e., "int (*funcptr)(char *str)") allows a programmer to dy-

namically modify a function to be called. We can overwrite a function pointer by overwriting its
address, so that when it’s executed, it calls the function we point it to instead.
– argv[] method: store the shellcode in an argument to the program. This causes the shellcode
to be stored in the stack. Then we need to guess the address of the shellcode (just like what we
did in the stack-buffer overflow). This method requires an executable stack.
– Heap method: store the shellcode in the heap/BSS (by using the overflow). Then we need to
guess the address of the shellcode, and assign this estimated address to the function pointer.
This method requires an executable heap (which is more likely than an executable stack).

• Function Pointers

– Function pointers can be stored in heap/BSS through many different means. The do not need to
be defined by the programmer.
– If a program calls atexit(), a function pointer will be stored in the heap by atexit(), and
will be invoked when the program terminates.
– The svc/rpc registration functions (librpc, libnsl, etc.) keep callback functions stored
on the heap.

• Other Examples
Lecture Notes (Syracuse University) Buffer-Overflow Vulnerabilities and Attacks: 17

– The BSDI crontab heap-based overflow: Passing a long filename will overflow a static buffer.
Above that buffer in memory, we have a pwd structure, which stores a user name, password, uid,
gid, etc. By overwriting the uid/gid field of the pwd, we can modify the privileges that crond
will run our crontab with (as soon as it tries to run our crontab). This script could then put
out a suid root shell, because our script will be running with uid/gid 0.

References
[1] P. J. Salzman. Memory Layout And The Stack. In Book Using GNU’s GDB Debugger. URL:
http://dirac.org/linux/gdb/02a-Memory Layout And The Stack.php.