HCMC University Of Technology

Faculty of Computer Science & Engineering

Lab 2
Process & Multithreaded Process
Course: Operating Systems

November 14, 2022

Goal: This lab helps student to review the data segment of a process and distinguish
the differences between thread and process.

Content In detail, this lab requires student identify the memory regions of process’s
data segment, practice with examples such as creating a multithread program, showing
the memory region of threads:

• view process memory regions: Data segment, BSS segment, Stack and Heap.

• Write a multithread program.

• Solve an example which can run in parallel using thread.

• Show the differences between process and thread in term of memory region.

Result After doing this lab, students can understand the mechanism of distributing
memory region to allocate the data segment for specific processes. In addition, they
understand how to write a multithreaded program.

Requirement Student need to review the theory of process memory and thread.

1
 Contents
1. Introduction 2
1.1. Process ’s memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2. Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3. Interprocess Communication . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4. Introduction to thread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2. Practice 7
2.1. Looking inside a process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2. How to transfer data between processes? . . . . . . . . . . . . . . . . . . . 8
2.2.1. Shared Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.2. Pipe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3. How to create multiple threads? . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.1. Thread libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2. Multithread programming . . . . . . . . . . . . . . . . . . . . . . . 15

3. Exercise (Required) 17
3.1. Problem 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2. Problem 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3. Problem 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

A. Memory-related data structures in the kernel 20

1. Introduction
1.1. Process ’s memory
Traditionally, a Unix process is divided into segments. The standard segments are code
segment, data segment, BSS (block started by symbol), and stack segment.
The code segment contains the binary code of the program which is running as the pro-
cess (a “process” is a program in execution). The data segment contains the initialized
global variables and data structures. The BSS segment contains the uninitialized global
data structures and finally, the stack segment contains the local variables, return ad-
dresses, etc. for the particular process.
Under Linux, a process can execute in two modes - user mode and kernel mode. A
process usually executes in user mode, but can switch to kernel mode by making sys-
tem calls. When a process makes a system call, the kernel takes control and does the
requested service on behalf of the process. The process is said to be running in kernel
mode during this time. When a process is running in user mode, it is said to be “in
userland” and when it is running in kernel mode it is said to be “in kernel space”. We
will first have a look at how the process segments are dealt with in userland and then
take a look at the book keeping on process segments done in kernel space.
In Figure 1.1, blue regions represent virtual addresses that are mapped to physical

2
memory, whereas white regions are unmapped. The distinct bands in the address space
correspond to memory segments like the heap, stack, and so on.

Figure 1.1: Layout of memory segments with process.

Userland’s view of the segments

• The Code segment consists of the code - the actual executable program. The code
of all the functions we write in the program resides in this segment. The addresses
of the functions will give us an idea where the code segment is. If we have a
function func() and let p be the address of func() (p = &func;). We know that p
will point within the code segment.
• The Data segment consists of the initialized global variables of a program. The
Operating system needs to know what values are used to initialize the global vari-
ables. The initialized variables are kept in the data segment. To get the address of
the data segment we declare a global variable and then print out its address. This
address must be inside the data segment.
• The BSS consists of the uninitialized global variables of a process. To get an
address which occurs inside the BSS, we declare an uninitialized global variable,
then print its address.

3
 • The automatic variables (or local variables) will be allocated on the stack, so
printing out the addresses of local variables will provide us with the addresses
within the stack segment.

• A process may also include a heap,which is memory that is dynamically allocated

during process run time.

1.2. Stack
Stack is one of the most important memory region of a process. It is used to store tem-
porary data used by the process (or thread). The name “stack” is used to described the
way data put and retrieved from this region which is identical to stack data structure:
The last item pushed to the stack is the first one to be removed (popped).

Stack organization makes it suitable from handling function calls. Each time a function
is called, it gets a new stack frame. This is an area of memory which usually contains, at
a minimum, the address to return when it complete, the input arguments to the function
and space for local variables.

In Linux, stack starts at a high address in memory and grows down to increase its size.
Each time a new function is called, the process will create a new stack frame for this
function. This frame will be place right after that of its caller. When the function re-
turns, this frame is clean from memory by shrinking the stack (stack pointer goes up).
The following program illustrates to identify the relative location of stack frames create
by nested function calls.

Example: Trace function calls

1 #include <s t d i o . h>
2 void func (unsigned long number) {
3 unsigned long local_f = number ;
4 p r i n t f ( "%2lu ␣−−>␣%p\n" , local_f , &local_f ) ;
5 local_f −−;
6 i f ( local_f > 0) {
7 func ( local_f ) ;
8 }
9 }
10 int main ( ) {
11 func (10) ;
12 }

Similar to heap, stack has a pointer name stack pointer (as heap has program break)
which indicate the top of the stack. To change stack size, we must modify the value of
this pointer. Usually, the value of stack pointer is hold by stack pointer register inside
the processor. Stack space is limited, we cannot extend the stack exceed a given size.

4
 If we do so, stack overflow will occurs and crash our program. To identify the default
stack size, use the following command
1 ulimit −s

Different from heap, data of stack are automatically allocated and cleaned through
procedure invocation termination, Therefore, in C programming, we do not need to
allocate and free local variables. In Linux, a process is permitted to have multiple stack
regions. Each regions belongs to a thread.

1.3. Interprocess Communication

Processes in OS are independent but we also need them cooperating. Cooperating
process can affect or be affected by other processes, including sharing data. Some reasons
for the need of cooperating processes are as follows:

• Information sharing Since several applications may be interested in the same

piece of information (for instance, copying and pasting), we must provide an envi-
ronment to allow concurrent access to such information.

• Computation speedup . If we want a particular task to run faster, we must

break it into subtasks, each of which will be executing in parallel with the others.
Notice that such a speedup can be achieved only if the computer has multiple
processing cores.

• Modularity We may want to construct the system in a modular fashion, dividing

the system functions into separate processes or threads.

Therefore, we need communication methods to transfer data among processes, which

are also known as inter-process communication (IPC) protocols. For example, a Web
browser may request a Web page from a Web server, which then sends HTML data. This
transfer of data usually uses sockets in a telephone-like connection. The shell creates an
ls process and a separate lpr process, connecting the two with a pipe, represented by the
"|" symbol. A pipe permits one-way communication between two related processes. The
ls process writes data into the pipe, and the lpr process reads data from the pipe. There
are several IPC methods, in this lab, we will practice with two of them, namely shared
memory and pipe. In the shared-memory model, a region of memory that is shared by the
cooperating processes is established. Processes can then exchange information by reading
and writing data to the shared region. In the message-passing model, communication
takes place by means of messages exchanged between the cooperating processes. The
two communications models are contrasted in Figure 1.2. In this lab, we will focus
on the shared memory method.

1.4. Introduction to thread

A thread is a basic unit of CPU utilization; it comprises a thread ID, a program counter,
a register set, and a stack. It shares with other threads belonging to the same process

5
 Figure 1.2: Shared memory vs message passing model.

its code section, data section, and other operating-system resources, such as open files
and signals. A traditional (or heavyweight) process has a single thread of control. If a
process has multiple threads of control, it can perform more than one task at a time.

Figure 1.3: Single-threaded and multithreaded processes.

Figure 1.3 illustrates the difference between a traditional single-threaded process and a
multithreaded process. The benefits of multithreaded programming can be broken down
into four major categories:

• Responsiveness

• Resource sharing

• Economy

6
 • Scalability

Question: What resources are used when a thread is created? How do they differ from
those used when a process is created?

Multicore programming Earlier in the history of computer design, in response to

the need for more computing performance, single-CPU systems evolved into multi-CPU
systems. A more recent, similar trend in system design is to place multiple computing
cores on a single chip. Each core appears as a separate processor to the operating system.
Whether the cores appear across CPU chips or within CPU chips, we call these systems
multicore or multiprocessor systems. Multithreaded programming provides a mechanism
for more efficient use of these multiple computing cores and improved concurrency.

Figure 1.4: Parallel execution on a multicore system.

On a system with multiple cores, however, concurrency means that the threads can run
in parallel, because the system can assign a separate thread to each core, as Figure 1.4
shown.
Question: Is it possible to have concurrency but not parallelism? Explain.

2. Practice
2.1. Looking inside a process
Looking at the following C program with basic statements:
1 #include <s t d i o . h>
2 #include <s t d l i b . h>
3 #include <sys /types . h>
4 #include <unistd . h>
5
6 int glo_init_data = 99;
7 int glo_noninit_data ;
8
9 void print_func ( ) {
10 int local_data = 9 ;

7
11 p r i n t f ( "Process␣ID␣=␣%d\n" , getpid ( ) ) ;
12 p r i n t f ( "Addresses␣ of ␣the␣ process : \ n" ) ;
13 p r i n t f ( " 1 . ␣glo_init_data␣=␣%p\n" , &glo_init_data ) ;
14 p r i n t f ( " 2 . ␣glo_noninit_data␣=␣%p\n" , &glo_noninit_data ) ;
15 p r i n t f ( " 3 . ␣print_func ( ) ␣=␣%p\n" , &print_func ) ;
16 p r i n t f ( " 4 . ␣local_data␣=␣%p\n" , &local_data ) ;
17 }
18
19 int main( int argc , char ∗∗argv ) {
20 print_func ( ) ;
21 return 0 ;
22 }

Let’s run this program many times and give the discussion about the segments of a
process. Where is data segment/BSS segment/stack/code segment?

2.2. How to transfer data between processes?

2.2.1. Shared Memory
A process allocates a shared memory segment using shmget ("SHared Memory GET").
int shmget (key_t key , size_t s i z e , int shmflg ) ;

• Its first parameter is an integer key that specifies which segment to create and
unrelated processes can access the same shared segment by specifying the same
key value. Moreover, other processes may have also chosen the same fixed key,
which could lead to conflict. So that you should be careful when generating keys
for shared memory regions. A solution is that you can use the special constant
IPC_PRIVATE as the key value guarantees that a brand new memory segment is
created.

• Its second parameter specifies the number of bytes in the segment. Because seg-
ments are allocated using pages, the number of actually allocated bytes is rounded
up to an integral multiple of the page size.

• The third parameter is the bitwise or of flag values that specify options to shmget.
The flag values include these:
– IPC_CREAT: This flag indicates that a new segment should be created. This
permits creating a new segment while specifying a key value.
– IPC_EXCL: This flag, which is always used with IPC_CREAT, causes shmget
to fail if a segment key is specified that already exists. If this flag is not given
and the key of an existing segment is used, shmget returns the existing seg-
ment instead of creating a new one.

8
 – Mode flags: This value is made of 9 bits indicating permissions granted to
owner, group, and world to control access to the segment.

To make the shared memory segment available, a process must attach it by calling
shmat().
void ∗shmat( int shmid , const void ∗shmaddr , int shmflg ) ;

This function take three arguments as follows:

• the shared memory segment identifier SHMID returned by shmget().

• a pointer that specifies where in your process’s address space you want to map the
shared memory; if you specify NULL, Linux will choose an available address.

• The third argument is a flag. You can read more details about this argument
in Linux manual page. https://man7.org/linux/man-pages/man3/shmat.3p.
html.

When you’re finished with a shared memory segment, the segment should be detached
using shmdt. Pass it the address returned by shmat. If the segment has been deallocated
and this was the last process using it, it is removed. Examples: Run the two following
processes in two terminals. At the writer process, you can type an input string and
observe returns from the reader process.

• writer.c
#include <sys /types . h>
#include <sys / ipc . h>
#include <sys /shm . h>
#include <s t d i o . h>
#include <unistd . h>
#include <s t d l i b . h>
#define SHM_KEY 0x123
int main( int argc , char ∗argv [ ] ) {
int shmid ;
char ∗shm ;

shmid = shmget (SHM_KEY, 1000 , 0644 | IPC_CREAT) ;

i f ( shmid < 0) {
perror ( "Shared␣memory" ) ;
return 1 ;
}else {
p r i n t f ( "shared␣memory : ␣%d\n" , shmid ) ;
}

shm = (char ∗)shmat(shmid , 0 , 0) ;

9
 i f (shm == (char ∗)−1) {
perror ( "shmat" ) ;
e x i t (1) ;
}
s p r i n t f (shm, " h e l l o ␣world\n" ) ;
p r i n t f ( "shared␣memory␣content : ␣%s \n" , shm) ;
s l e e p (10) ;
// detach from the shared memory
i f (shmdt(shm) == −1) {
perror ( "shmdt" ) ;
return 1 ;
}
// Mark the shared segment to be destroyed .
i f ( shmctl (shmid , IPC_RMID, 0) == −1) {
perror ( "shmctl" ) ;
return 1 ;
}
return 0 ;
}

• reader.c
#include <sys /types . h>
#include <sys / ipc . h>
#include <sys /shm . h>
#include <s t d i o . h>
#include <s t d l i b . h>

#define SHM_KEY 0x123

int main( int argc , char ∗argv [ ] ) {

int shmid ;
char ∗shm ;
/∗

∗/
shmid = shmget (SHM_KEY, 1000 , 0644|IPC_CREAT) ;
i f ( shmid < 0) {
perror ( "shmget" ) ;
return 1 ;
}
else {
p r i n t f ( "shared␣memory : ␣%d\n" , shmid ) ;
}

10
 shm = (char ∗)shmat(shmid , 0 ,0) ;
i f (shm == (char ∗)−1) {
perror ( "shmat" ) ;
e x i t (1) ;
}
p r i n t f ( "shared␣memory : ␣%p\n" , shm) ;
i f (shm != 0) {
p r i n t f ( "shared␣memory␣content : ␣%s \n" , shm) ;
}

s l e e p (10) ;
i f (shmdt(shm) == −1) {
perror ( "shmdt" ) ;
return 1 ;
}
return 0 ;
}

2.2.2. Pipe
Pipe actually is very common method to transfer data between processes. For example,
the "pipe" operator ’|’ can be used to transfer the output from a command to another
command as in the following example:
# the output from " h i s t o r y " w i l l be input to the grep command.
history | grep "a"

In terms of C programming, the standard library named "unistd.h" defined the following
function to create a pipe. This function creates a pipe, a unidirectional data channel
that can be used for interprocess communication. The array pipefd is used to return two
file descriptors referring to the ends of the pipe. pipefd[0] refers to the read end of the
pipe. pipefd[1] refers to the write end of the pipe. Data written to the write end of the
pipe is buffered by the kernel until it is read from the read end of the pipe.
int pipe ( int pipefd [ 2 ] ) ;

Pipe can be used for one-way communication as follows:

#include<s t d i o . h>
#include<unistd . h>
int main ( ) {
int pipefds [ 2 ] ;
int returnstatus ;
int pid ;
char writemessages [20]= " Hello " ;

11
 char readmessage [ 2 0 ] ;
returnstatus = pipe ( pipefds ) ;
i f ( returnstatus == −1) {
p r i n t f ( "Unable␣to␣ create ␣pipe\n" ) ;
return 1 ;
}
pid = fork ( ) ;

// Child process
i f ( pid == 0) {
read ( pipefds [ 0 ] , readmessage , sizeof ( readmessage ) ) ;
p r i n t f ( "Child␣Process : ␣Reading , ␣message␣ i s ␣%s \n" , readmessage ) ;
return 0 ;
}
//Parent process
p r i n t f ( "Parent␣Process : ␣Writing , ␣message␣ i s ␣%s \n" , writemessages ) ;
write ( pipefds [ 1 ] , writemessages , sizeof ( writemessages ) ) ;
return 0 ;
}

In the above program, firstly the parent process will create a pipline and call fork() to
create a child process. Then, the parent process will write a message to the pipeline.
At the same time, the child process will read data from the pipeline. Noticeably, both
write() and read() need to know the size of the message.

2.3. How to create multiple threads?

2.3.1. Thread libraries
A thread library provides the programmer with an API for creating and managing
threads. There are two primary ways of implementing a thread library. Three main
thread libraries are in use today: POSIX Pthreads, Windows, and Java. In this lab, we
use POSIX Pthread on Linux and Mac OS to practice with multithreading programming.

Creating threads
pthread_create ( thread , attr , start_routine , arg )

Initially, your main() program comprises a single, default thread. All other threads
must be explicitly created by the programmer.

• thread: An opaque, unique identifier for the new thread returned by the subrou-
tine.

• attr: An opaque attribute object that may be used to set thread attributes. You
can specify a thread attributes object, or NULL for the default values.

12
 • start: the C routine that the thread will execute once it is created.

• arg: A single argument that may be passed to textttstart_routine. It must be

passed by reference as a pointer cast of type void. NULL may be used if no argument
is to be passed.

Example: Pthread Creation and Termination

1 #include <pthread . h>
2 #include <s t d i o . h>
3 #define NUM_THREADS 5
4
5 void ∗ PrintHello (void ∗ threadid )
6 {
7 long t i d ;
8 t i d = (long) threadid ;
9 p r i n t f ( " Hello ␣World ! ␣ I t ’ s ␣me, ␣thread␣#%ld ! \ n" , t i d ) ;
10 pthread_exit (NULL) ;
11 }
12
13 int main ( int argc , char ∗argv [ ] )
14 {
15 pthread_t threads [NUM_THREADS] ;
16 int rc ;
17 long t ;
18 for ( t =0; t<NUM_THREADS; t++){
19 p r i n t f ( "In␣main : ␣ creating ␣thread␣%ld \n" , t ) ;
20 rc = pthread_create(&threads [ t ] ,NULL, PrintHello , (void ∗) t ) ;
21 i f ( rc ) {
22 p r i n t f ( "ERROR; ␣ return ␣from␣pthread_create ( ) ␣ i s ␣%d\n" , rc ) ;
23 e x i t (−1) ;
24 }
25 }
26
27 /∗ Last thing t h a t main() should do ∗/
28 pthread_exit (NULL) ;
29 }

Passing argument to Thread We can pass a structure to each thread such as the
example below. Using the previous example to implement this example:
1 struct thread_data{
2 int thread_id ;
3 int sum ;
4 char ∗message ;
5 };

13
 6
7 struct thread_data thread_data_array [NUM_THREADS] ;
8
9 void ∗ PrintHello (void ∗thread_arg )
10 {
11 struct thread_data ∗my_data ;
12 ...
13 my_data = ( struct thread_data ∗) thread_arg ;
14 taskid = my_data−>thread_id ;
15 sum = my_data−>sum ;
16 hello_msg = my_data−>message ;
17 ...
18 }
19
20 int main ( int argc , char ∗argv [ ] )
21 {
22 ...
23 thread_data_array [ t ] . thread_id = t ;
24 thread_data_array [ t ] . sum = sum ;
25 thread_data_array [ t ] . message = messages [ t ] ;
26 rc = pthread_create(&threads [ t ] , NULL, PrintHello ,
27 (void ∗) &thread_data_array [ t ] ) ;
28 ...
29 }

Joining and Detaching Threads “Joining” is one way to accomplish synchronization

between threads, For example:

Figure 2.1: Joining threads.

• The pthread_join() subroutine blocks the calling thread until the specified threa-
did thread terminates.

• The programmer is able to obtain the target thread’s termination return status if
it was specified in the target thread’s call to pthread_exit().

14
 • A joining thread can match one pthread_join() call. It is a logical error to
attempt multiple joins on the same thread.

2.3.2. Multithread programming

Problem: Constructing a multithreaded program that calculates the summation of a
non-negative integer in a separate thread.
1 #include <pthread . h>
2 #include <s t d i o . h>
3
4 int sum ; /∗ t h i s data i s shared by the thread ( s ) ∗/
5 void ∗runner (void ∗param) ; /∗ threads c a l l t h i s function ∗/
6
7 int main( int argc , char ∗argv [ ] )
8 {
9 pthread_t t i d ; /∗ the thread i d e n t i f i e r ∗/
10 pthread_attr_t a t t r ; /∗ s e t of thread a t t r i b u t e s ∗/
11
12 i f ( argc != 2) {
13 f p r i n t f ( stderr , "usage : ␣a . out␣<i n t e g e r ␣value>\n" ) ;
14 return −1;
15 }
16
17 i f ( a t o i ( argv [ 1 ] ) < 0) {
18 f p r i n t f ( stderr , "%d␣must␣be␣>=␣0\n" , a t o i ( argv [ 1 ] ) ) ;
19 return −1;
20 }
21 /∗ get the d e f a u l t a t t r i b u t e s ∗/
22 pthread_attr_init(&a t t r ) ;
23 /∗ create the thread ∗/
24 pthread_create(&tid , &attr , runner , argv [ 1 ] ) ;
25 /∗ wait f o r the thread to e x i t ∗/
26 pthread j o i n ( tid ,NULL) ;
27
28 p r i n t f ( "sum␣=␣%d\n" ,sum) ;
29 }
30
31 /∗ The thread w i l l begin control in t h i s function ∗/
32 void ∗runner (void ∗param)
33 {
34 int i , upper = a t o i (param) ;
35 sum = 0 ;
36 for ( i = 1 ; i <= upper ; i++)
37 sum += i ;

15
38 pthread e x i t (0) ;
39 }

16
 3. Exercise (Required)
3.1. Problem 1
Firstly, downloading two text files from the url: https://drive.google.com/file/
d/1fgJqOeWbJC4ghMKHkuxfIP6dh2F911-E/view?usp=sharing These file contains the
100000 ratings of 943 users for 1682 movies in the following format:
userID <tab> movieID <tab> r a t i n g <tab> timeStamp
userID <tab> movieID <tab> r a t i n g <tab> timeStamp
...
Secondly, you should write a program that spawns two child processes, and each of them
will read a file and compute the average ratings of movies in the file. You implement
the program by using shared memory method.

3.2. Problem 2
Given the following function:

sum(n) = 1 + 2 + ... + n

This is the sum of a large set including n numbers from 1 to n. If n is a large num-
ber, this will take a long time to calculate the sum(n). The solution is to divide
this large set into pieces and calculate the sum of these pieces concurrently by using
threads. Suppose the number of threads is numT hreads, so the 1st thread calculates
the sum of {1,n/numThreads}, the 2nd thread carries out the sum of {n/numThreads
+1, 2n/numThreads},...
Write two programs implementing algorithm describe above: one serial ver-
sion and one multi-thread version.
The program takes the number of threads and n from user then creates multiple threads
to calculate the sum. Put all of your code in two files named “sum_serial.c” and
“sum_multi-thread.c”. The number of threads and n are passed to your program as
an input parameter. For example, you will use the following command to run your
program for calculating the sum of 1M :
. / sum_serial 1000000
. / sum_multi−thread 10 1000000
(#numThreads=10)

Requirement: The multi-thread version may improve speed-up compared to the serial
version. There are at least 2 targets in the Makefile sum_serial and sum_multi-thread
to compile the two program.

3.3. Problem 3
Conventionally, pipe is a one-way communication method.(In the example at section 3,
you can test by add a read() call after the writer() call at the parent process, a write()

17
 call after the read() call at the child process and observe what happens?). However, we
still can have some tricks to adapt it for two-way communication by using two pipes. In
this exercise, you should implement the TODO segment in the below program.
1 #include <s t d i o . h>
2 #include <s t d l i b . h>
3 #include <unistd . h>
4 static int pipefd1 [ 2 ] , pipefd2 [ 2 ] ;
5
6 void INIT(void) {
7 i f ( pipe ( pipefd1 )<0 | | pipe ( pipefd2 )<0){
8 perror ( "pipe" ) ;
9 e x i t (EXIT_FAILURE) ;
10 }
11 }
12 void WRITE_TO_PARENT(void) {
13 /∗ send parent a message through pipe ∗/
14 // TO DO
15 p r i n t f ( "Child␣send␣message␣to␣parent ! \ n" ) ;
16 }
17 void READ_FROM_PARENT(void) {
18 /∗ read message sent by parent from pipe ∗/
19 // TO DO
20 p r i n t f ( "Child␣ r e c e i v e ␣message␣from␣parent ! \ n" ) ;
21 }
22 void WRITE_TO_CHILD(void) {
23 /∗ send c h i l d a message through pipe ∗/
24 // TO DO
25 p r i n t f ( "Parent␣send␣message␣to␣ c h i l d ! \ n" ) ;
26 }
27 void READ_FROM_CHILD(void) {
28 /∗ read the message sent by c h i l d from pipe ∗/
29 // TO DO
30 p r i n t f ( "Parent␣ r e c e i v e ␣message␣from␣ c h i l d ! \ n" ) ;
31 }
32 int main( int argc , char∗ argv [ ] ) {
33 INIT ( ) ;
34 pid_t pid ;
35 pid = fork ( ) ;
36 // s e t a timer , process w i l l end a f t e r 1 second .
37 alarm (10) ;
38 i f ( pid==0){
39 while (1) {
40 s l e e p ( rand ( )%2+1);

18
41 WRITE_TO_CHILD( ) ;
42 READ_FROM_CHILD( ) ;
43 }
44 }else{
45 while (1) {
46 s l e e p ( rand ( )%2+1);
47 READ_FROM_PARENT( ) ;
48 WRITE_TO_PARENT( ) ;
49 }
50 }
51 return 0 ;
52 }

19
 A. Memory-related data structures in the kernel
In the Linux kernel, every process has an associated struct task_struct. The definition
of this struct is in the header file include /linux/sched.h.
1 struct task_struct {
2 volatile long s t a t e ;
3 /∗ −1 unrunnable , 0 runnable , >0 stopped ∗/
4 struct thread_info ∗thread_info ;
5 atomic_t usage ;
6 ...
7 struct mm_struct ∗mm, ∗active_mm ;
8 ...
9 pid_t pid ;
10 ...
11 char comm[ 1 6 ] ;
12 ...
13 } ;

Three members of the data structure are relevant to us:

• pid contains the Process ID of the process.

• comm holds the name of the process.

• The mm_struct within the task_struct is the key to all memory management
activities related to the process.

The mm_struct is defined in include/linux/sched.h as:

1 struct mm_struct {
2 struct vm_area_struct ∗ mmap; /∗ l i s t of VMAs ∗/
3 struct rb_root mm_rb;
4 struct vm_area_struct ∗ mmap_cache; /∗ l i s t of VMAs ∗/
5 ...
6 unsigned long start_code , end_code , start_data , end_data ;
7 unsigned long start_brk , brk , start_stack ;
8 ...
9 }

20
 Here the first member of importance is the mmap. The mmap contains the pointer
to the list of VMAs (Virtual Memory Areas) related to this process. Full usage of the
process address space occurs very rarely. The sparse regions used are denoted by VMAs.
The VMAs are stored in struct vm_area_struct defined in linux/mm.h:
1 struct vm_area_struct {
2 struct mm_struct ∗ vm_mm; /∗The address space we belong to . ∗/
3 unsigned long vm_start ; /∗Our s t a r t address within vm_mm. ∗/
4 unsigned long vm_end; /∗The f i r s t byte a f t e r our end
5 address within vm_mm. ∗/
6 ....
7 /∗ l i n k e d l i s t of VM areas per task , sorted by address ∗/
8 struct vm_area_struct ∗vm_next ;
9 ....
10 }

Kernel’s view of the segments

The kernel keeps track of the segments which have been allocated to a particular process
using the above structures. For each segment, the kernel allocates a VMA. It keeps track
of these segments in the mm_struct structures. The kernel tracks the data segment
using two variables: start_data and end_data. The code segment boundaries are in the
start_code and end_code variables. The stack segment is covered by the single variable
start_stack. There is no special variable to keep track of the BSS segment - the VMA
corresponding to the BSS accounts for it.

21