CSI3131 - Module 2:

Processes
Reading: Chapter 3 (Silberchatz)

Objective: To understand the nature of processes

including the following concepts:
 Process states, the PCB (process control block),
and process switching.
 Process scheduling
 Operations on processes
 Mechanisms for supporting cooperation
between processes
1
 Co
nc
g ep
in t
u l
ed
Sch Process
O n:
pe t io
ra r a
t io p e C
ns oo IP
C

2
 Program
Execution
Process
Co State
nc
ep
t

g
lin
du
he
Sc Process

n:
O

C io
pe

IP rat
ra

pe
tio

oo
ns

3
Process Concept
 Process executes a
program
 Text – code to execute
 Data
 Heap (dynamic memory)
 Stack (local variables)
 An operating system
executes a variety of
programs:
 Batch system – jobs
 Time-shared systems –
user programs or tasks
 We will use the terms
job, task, and process
almost interchangeably
4
Process States
 Which properties of process states should OS
distinguish?
 running vs not running
 blocked/waiting vs ready
 Main process states:
 new: The process is being created
 running: Instructions are being executed
 waiting: The process is waiting for some event to occur
 ready: The process is waiting to be assigned to a process
 terminated: The process has finished execution
 In real OSes, there could be additional states
 i.e. has the process been swapped out?

5
 Simple Diagram of Process States

6
Process state
transitions
 New  Ready
 Represents the creation of the process.
• In the case of batch systems – a list of waiting in new
state processes (i.e. jobs) is common.
 Ready  Running
 When the CPU scheduler selects a process for
execution.
 Running  Ready
 Happens when an interruption caused by an event
independent of the process
• Must deal with the interruption – which may take away
the CPU from the process
• Important case: the process has used up its time with the
7 CPU.
Process state
transitions
 Running  Waiting
 When a process requests a service from the OS and
the OS cannot satisfy immediately (software
interruption due to a system call)
• An access to a resource not yet available
• Starts an I/O: must wait for the result
• Needs a response from another process
 Waiting  Ready
 When the expected event has occurred.
 Running  Terminated
 The process has reached the end of the program (or
an error occurred).
8
 PCB
Program
Execution
Process
Process
Co State
nc Switching
g ep
in t
u l
ed
Sch Process
O n:
pe t io
ra r a
t io p e C
ns oo IP
C

9
Saving process information
 In a multitasking system, a process executes in an
intermittent fashion on the CPU.
 Each time a process is assigned the CPU
(transitions from ready to running), execution
must continue in the same situation the process
left the CPU last (i.e. same content in the CPU
registers, etc.)
 Thus when a process leaves the running state, it
is necessary to save important process
information that will be retrieved when it comes
back to the running state.

10
Process Control Block (PCB)

11
Process Control Block (PCB)
PCB Contains all information the OS needs to remember about
a process
So, what is needed?
 To resume process execution?
• Program counter
• CPU registers
 To know how to schedule the process?
• CPU scheduling information (i.e. priority)
 To know which memory addresses the process can
access?
• Memory-management information (base and limit pointers,
…)
 What else?
• Process state
• I/O status information (e.g. open files)
• Owner, parent process, children
12 • Accounting information
Process switching
Also called context switching
 When the CPU moves from executing of one
process, say process 0, to another process, say
process 1
 The context switching occurs when the CPU is
interrupted during the execution of process 0 (can be
either hardware or software interrupt).
 Updates and saves the PCB of process 0
 Finds and accesses the PCB of process 1, that was
previously saved.
 Updates the CPU registers, program counter, etc. to
restore the context as defined in PCB 1
 Returns from the interrupt to allow the CPU to
continue the execution of process 1.
13
CPU Switch From Process to Process

It is possible that much time passes before returning

14 to Process 0, and that many other processes executes
during this time
 PCB
Program
Execution
Process
Scheduling Process
Co State
Queues nc Switching
g ep
n t
u li
ed
c h
S Process
O n:
pe t io
ra r a
t io p e C
ns oo IP
C

15
Process Scheduling
What?
 Choosing which process to run next
Why?
 Multiprocessing – to achieve good utilization
 Time sharing – to achieve low response time
How?
 We will talk in detail about that in Module 4 (Chapter 5).
 Now we just introduce the basic concepts of CPU
scheduling.
 We have already seen the first one
• Context switch from process to process

16
Process Scheduling Queues
What?
 data structures supporting scheduling
algorithms
Job queue
 set of all processes in the system
Ready queue
 set of all processes residing in main memory,
ready and waiting to execute
Device queues
 for I/O devices
 set of processes waiting for an I/O on that
device

Processes (i.e. their PCB’s) migrate between the

17
various queues as they change state
Ready Queue And Various I/O Device Queues

18
 The PCBs are not moved around in memory to
be put into different queues: it is pointers in the
PCBs that are used to create the queues.

term. unit 0 ready

. . . PCB2 PCB3 PCB4 PCB5 PCB6 PCB7 . . . PCB14

disk unit 0

19
Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
g ep
in t
Context d ul
e
switch
S ch Process
O n:
pe t io
ra r a
t io p e C
ns oo IP
C

20
Schedulers
 Long-term scheduler (or job scheduler)
 selects which new processes should be brought into
the memory; new  ready (and into the ready queue
from a job spool queue) (used in batch systems)
 Short-term scheduler (or CPU scheduler)
 selects which ready process should be executed
next; ready  running

 Which of these schedulers must execute really fast,

and which one can be slow? Why?
 Short-term scheduler is invoked very frequently
(milliseconds)  (must be fast)
 Long-term scheduler is invoked very infrequently
(seconds, minutes)  (may be slow)
21
Schedulers (Cont.)
 Processes differ in their resource utilization:
 I/O-bound process – spends more time doing
I/O than computations, many short CPU bursts
 CPU-bound process – spends more time doing
computations; few very long CPU bursts
 The long-term scheduler controls the degree
of multiprogramming
 the goal is to efficiently use the computer
resources
 ideally, chooses a mix of I/O bound and CPU-
bound processes
• but difficult to know beforehand
22
Representation of Process Scheduling

23
Medium Term Scheduling
 Due to memory shortage, the OS might decide
to swap-out a process to disk.
 Later, it might decide to swap it back into
memory when resources become available
 Medium-term scheduler – selects which
process should be swapped out/in

24
Context Switch
What?
 Assigning the CPU from one process to another
How?
 We have already seen that
 Save the CPU state to the PCB of the current
process
 Load the CPU with the information from the PCB of
the new process
 Set the program counter to the PC of the new
process
 Additional work as well (accounting, …)
Comments:
 Can be quite a lot of work - pure overhead, as no
process is executing at that time
 Context switch time is dependent on hardware
support (and OS smarts)
25
 Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
g ep
in t
Context d ul
e
switch ch
crea
te Proc
ess
fork/
S Process
exec
O n:
s pe t io
es n r a r a
r
c
o atio t io p e C
P re ns oo IP
C C

26
Process Creation
So, where do all processes come from?
 Parent process create children processes, which,
in turn create other processes, forming a tree of
processes
 Usually, several properties can be specified at
child creation time:
• How do the parent and child share resources?
• Share all resources
• Share subset of parent’s resources
• No sharing
• Does the parent run concurrently with the child?
• Yes, execute concurrently
• No, parent waits until the child terminates
• Address space
• Child duplicate of parent
27
• Child has a program loaded into it
Process Creation (Cont.)
UNIX example:
 fork() system call creates new process with
the duplicate address space of the parent
• no shared memory, but a copy
• copy-on-write used to avoid excessive cost
• returns child’s pid to the parent, 0 to the new
child process
• the parent may call wait() to wait until the child
terminates
 exec(…) system call used after a fork() to
replace the process’ memory space with a
new program

28
UNIX: fork(), exec(), exit() & wait()

29
C Program Forking Separate Process
#include <stdio.h>
#include <unistd.h>
int main(int argc, char *argv[]) {
int pid;
pid = fork(); /* fork another process */
if (pid < 0) { /* error occurred */
fprintf(stderr, "Fork Failed");
exit(-1);
} else if (pid == 0) { /* child process */
execlp("/bin/ls","ls",NULL);
} else { /* parent process, will wait for the
child to complete */
wait(NULL);
printf("Child Complete");
exit(0);
}
30 }
Fork example
int pid, a = 2, b=4;
pid = fork(); /* fork another process */
if (pid < 0) exit(-1); /* fork failed */
else if (pid == 0) { /* child process */
a = 3; printf(“%d\n”, a+b);
} else {
wait();
b = 1;
printf(“%d\n”, a+b);
}

What would be the output printed?

7
3
31
Fork example
int pid, a = 2, b=4;
pid = fork(); /* fork another process */
if (pid < 0) exit(-1); /* fork failed */
else if (pid == 0) { /* child process */
a = 3; printf(“%d\n”, a+b);
} else {
b = 1;
wait();
printf(“%d\n”, a+b);
}

What would be the output printed?

7
3
32
Fork example
int pid, a = 2, b=4;
pid = fork(); /* fork another process */
if (pid < 0) exit(-1); /* fork failed */
else if (pid == 0) { /* child process */
a = 3; printf(“%d\n”, a+b);
} else {
b = 1;
printf(“%d\n”, a+b);
wait();
}

What would be the output printed?

7 or 3
3 7
33
Understanding fork()
Before fork After fork After child executes
a=3
Parent process Parent process Parent process
page table page table page table
* *
*
a: 2 a: 2 a: 2
b: 4 b: 4 * b: 4 *
* *
Child process Child process
page table page table
* *
*
* a: 3 *
b: 4
34
* *
Process Creation (Cont.)
Windows:
 CreateProcess(…)
 Similar to fork() immediately followed by
exec()

Discussion
 fork() very convenient for passing
data/parameters from the parent to the child
 all code can be conveniently at one place
 direct process creation more efficient when
you really just want to launch a new
program
35
 Example Process Tree (Solaris)

36
 Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
e pt

g
lin
Context

du
he
c re a
tePro switch
cess
fork/
Sc Process
exec

:
on
O

C i
at ss

pe

IP rat
term
inate
r e ce
n

Proc
ra

pe
io

ess
o

t
Pr

io

oo
exit
ns
C

C
in s s
n
io
rm ce

kill
at
Te P r o

37
Process Termination
How do processes terminate?
 Process executes last statement and asks the
operating system to delete it (by making exit()
system call)
 Abnormal termination
 Division by zero, memory access violation, …
 Another process asks the OS to terminate it
 Usually only a parent might terminate its children
• To prevent user’s terminating each other’s processes
 Windows: TerminateProcess(…)
 UNIX: kill(processID, signal)

38
Process Termination
What should the OS do?
 Release resources held by the process
 When a process has terminated, but not all of its
resources has been released, it is in state terminated
(zombie)
 Process’ exit state might be sent to its parent
 The parent indicates interest by executing wait()
system call

What to do when a process having children is exiting?

 Some OSs (VMS) do not allow children to continue
 All children terminated - cascading termination
 Other find a parent process for the orphan processes

39
Time for questions
Process states
 Can a process move from waiting state to running
state?
 From ready state to terminated state?
PCB
 Does PCB contain program’s global variables?
 How is the PCB used in context switches?
CPU scheduling
 What is the difference between long term and medium
term scheduler?
 A process (in Unix) has executed wait() system call. In
which queue it is located?
 What happens if there is no process in the ready
queue?
40
Other questions!
Process creation
 Understand fork(), exec(), wait()
 Do you think you understand them?
• So, how many processes are created in this
code fragment?
for(i=0; i<3; i++)
fork();

Process termination
 How/when.
 What should the OS do?

41
 Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
e pt

g
lin
Context

du
he
c re a
tePro switch
cess
fork/
Sc Process
exec

:
on
O

C i
at ss

pe

IP rat
term
inate
r e ce
n

Proc
ra

pe
io

ess
o

t
Pr

io

oo
exit
ns
C

C
Ba
in s s
n

m o si c
io
rm ce

kill
at

Memory de
Te P r o

sharing ls
Message passing
42
Cooperating Processes
 Independent processes cannot affect or be affected by
the execution of other processes
 Cooperating processes can affect or be affected by the
execution of other processes
 Advantages of process cooperation
 Information sharing
 Computation speed-up (parallel processing)
 Modularity
 Nature of the problem may request it
 Convenience

So, we really want Inter-Process Communication (IPC)

43
Interprocess Communication (IPC)
 Mechanisms for processes to
communicate and to synchronize their
actions
 So, how can processes communicate?
 Fundamental models of IPC
• Through shared memory
• Using message passing
 Examples of IPC mechanisms
• signals
• pipes & sockets
44
• semaphores …
Shared Memory
 By default, address spaces of processes are disjoint
 To protect from unwanted interference
 The OS has to allow one process to access (a
precisely specified) part of the address space of
another process
 One process makes a system call to create a shared
memory region
 Other process makes system call to map this shared
memory region into its address space
 Special precaution is needed in accessing the data in
the shared memory, it is way too easy to get into
inconsistent state
 We will talk about this extensively in Chapter 6

45
Shared Memory - Server:
…
#define SHMSZ 27
main() {
char c;
int shmid;
key_t key;
char *shm, *s;

key = 5678;
if ((shmid = shmget(key, SHMSZ, IPC_CREAT | 0666)) < 0) {
perror("shmget"); exit(1); }

if ((shm = shmat(shmid, NULL, 0)) == (char *) -1) {

perror("shmat"); exit(1); }

s = shm;
for (c = 'a'; c <= 'z'; c++) *s++ = c;
*s = ‘\0’;

while (*shm != '*') sleep(1);

exit(0);
}
46
Shared Memory - Client
…
#define SHMSZ 27
main(){
int shmid;
key_t key;
char *shm, *s;

key = 5678;
if ((shmid = shmget(key, SHMSZ, 0666)) < 0) {
perror("shmget");
exit(1);
}

if ((shm = shmat(shmid, NULL, 0)) == (char *) -1) {

perror("shmat"); exit(1); }

for (s = shm; *s != ‘\0’; s++) putchar(*s);

putchar('\n');

*shm = '*';

exit(0);
}47
Message Passing
 Two basic operations:
 send(destination, message)
 receive(source, message)
 If processes wish to communicate, they need
to establish a communication link between
them
 figure out what destination and source to use
 There are many variants of how send() and
receive() behave
 Direct or indirect communication
 Synchronous or asynchronous communication
 Automatic or explicit buffering
48
Direct Communication
 Processes must name each other explicitly:
 send (P, message) – send a message to
process P
 receive(Q, message) – receive a message from
process Q
 Properties of the communication link
 Links are established automatically, exactly
one link for each pair of communicating
processes
 The link may be unidirectional, but is usually
bi-directional

49
Indirect Communication
 Messages are sent and received to/from mailboxes
(also referred to as ports)
 Each mailbox has unique id
 Processes can communicate only if they share a
mailbox
 Basic primitives:
send(A, message) – send a message to mailbox A
receive(A, message) – receive a message from
mailbox A
 Operations
 create a new mailbox
 send and receive messages through mailbox
 destroy a mailbox
50
Indirect Communication

51
Indirect Communication
 Mailbox sharing
 P1, P2, and P3 share mailbox A
 P1, sends; P2 and P3 receive
 Who gets the message?
 Solutions
 Allow a link to be associated with at most two
processes
 Allow only one process at a time to execute a
receive operation
 Allow the system to select arbitrarily the
receiver. Sender is notified who the receiver
was.
52
Blocking Message Passing

Also called synchronous message passing

 sender waits until the receiver receives the message
 receiver waits until the sender sends the message
 advantages:
 inherently synchronizes the sender with the receiver
 single copying sufficient
 disadvantages:
 possible deadlock problem

53
Synchronous Message Passing
send() before corresponding receive()

request
send()
sender
suspended ack recv()
data

receive() before corresponding send()

recv()
receiver
send() suspended

54
Non-Blocking Message Passing

Also called asynchronous message

passing
 Non-blocking send: the sender
continues before the delivery of the
message
 Non-blocking receive: check whether
there is a message available, return
immediately

55
Buffering
 With blocking direct communication, no
buffering is needed – the message stays in the
sender’s buffer until it is copied into the
receiver’s buffer
 With non-blocking communication, the sender
might reuse the buffer,
 therefore the message is copied to a system
buffer,
 and from there to the receiver’s buffer
 If the system buffer becomes full, the sender
would still be suspended

56
 Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
e pt

g
lin
Context

du
Sockets

he
c re a
tePro switch
cess
fork/ Sc
Process
exec RPC/RMI

:
on
O

C i
at ss

pe

IP rat
term
inate
re ce
n

Proc
ra

pe
io

ess Pipes
o

t
Pr

io

oo
exit
ns
C

IP C Signals
C Ba E x am
in s s
n

m o si c ple
io
r m ce

kill s
at

Memory de
Te P r o

sharing ls
Message passing
57
Examples of IPC Mechanisms
 Pipes
 Signals

 Sockets
 Remote Procedure Calls
 Remote Method Invocation (Java)

 Higher-level concepts
 Implemented either using shared memory (between
processes on the same computer) or message
passing (between different computers)

58
Pipes – a few facts
Process  Each UNIX/Linux
process is created
with 3 open files:
0 1 2 ...  0: standard input
 1: standard output
Kernel
 2: standard error
 Often attached to a
terminal
 Can redirect to a file
 cmd >file

59
Pipes - communication

Process Process
 What is it?
 A unidirectional
channel
 One write end
... ... 
0 1 2 0 1 2
One read end

Kernel

 Why?
 The UNIX/Linux pipes allows linking of processes
to obtain complex functions from simple
commands
 who | sort
60
Pipes – how?
 System call: pipe(int *fd)
 Create a pipe with two file descriptors:
 fd[0] – reference to the read end
 fd[1] - reference to the write end.

Process Process
/* Array for storing 2 file
descriptors */
int fd[2];
0 1 2 3 4 ... 0 1 2 3 4 ...
/* creation of a pipe */
int ret = pipe(fd);
if (ret == -1) { /* error */
perror("pipe");
Kernel Kernel exit(1);
}
(before) (after) pipe
61
Unix Pipes
So, what did we get?
 fd[0] is a file descriptor for the read end of the pipe
 read(fd[0], dataBuf, count) would read count bytes or less
from the pipe into the dataBuf character array
 fd[1] is a file descriptor for the write end of the pipe
 write(fd[1], dataBuf, count) would write count bytes from
the dataBuf character array into the pipe

 Nice, but we want to communicate between different

processes, and pipe() gives both endpoints to the
process calling pipe()
 Too bad, we are allowed to talk only with ourselves :-(
 Wait! Can we talk at least within the family?

62
Unix Pipes
Note:
 When parent calls fork(), the child gets the parent’s
open file descriptors.
 That includes the pipe endpoints!
 If the child calls exec() to run another program, the new
open files are preserved.
So, how do we make the parent talk to its child:
 The parent creates pipe
 Calls fork() to create child
 The parent closes the read end of the pipe and writes to
the write end
 The child closes the write end of the pipe and reads
from the read end
 Can we have the communication in the opposite way?
63
Unix Pipes
int fd[2], pid, ret;
ret = pipe(fd);
if (ret == -1) return PIPE_FAILED;
pid = fork();
if (pid == -1) return FORK_FAILED;
if (pid == 0) { /* child */
close(fd[1]);
while(…) {
read(fd[0], …);
…
}
} else { /* parent */
close(fd[0]);
while(…) {
write(fd[1], …);
…
}
}
64
After Spawning New Child Process

Parent Process Child Process

0 1 2 3 4 ... 0 1 2 3 4 ...

Kernel

pipe

65
After Closing Unused Ends of Pipe

Parent Process Child Process

0 1 2 3 4 ... 0 1 2 3 4 ...

Kernel

pipe

66
Pipes – attaching to 0 and 1
 System call: dup2(fdSrc,fdDst)
 Creates a copy of fdSrc to fdDst
 If fdDst is open, it is first closed
 To attach a pipe to the standard output, dup2(fd[1],1)

 Note
Processus Processus fd 4 still refers to the
pipe
It is possible to have
multiple references
0 1 2 3 4 ... 0 1 2 3 4 ...
to the ends of the
pipe, including
from other
Noyau Noyau
processes

(avant) (après dup2)

67
Unix Pipes – Shell running who | sort
int fd[2], pid1, pid2, ret;
ret = pipe(fd); // Create the pipe
if (ret == -1) return PIPE_FAILED;
pid1 = fork(); // for who
if (pid1 == -1) return FORK_FAILED;
if (pid1 == 0) { /* child */
dup2(fd[1],1); // attach to stdout
close(fd[1]);
close(fd[0]);
execlp(“who”,”who”,NULL);
}
pid2 = fork(); // for who
if (pid2 == -1) return FORK_FAILED;
if (pid2 == 0) { /* child */
dup2(fd[0],0); // attach to stdout
close(fd[0]);
close(fd[1]);
execlp(“sort”,”sort”,NULL);
}
if(wait(NULL) == pid1) printf(“who finished”)
else printf(“sort finihed”);
if(wait(NULL)
68 == pid2) printf(“sort finished”)
else printf(“who finihed”);
Unix Pipes
Nice, nice, but we would like 2-way communication
 2 pipes can be used

Gotcha’s:
 Each pipe is implemented using a fixed-size system buffer
 If the buffer becomes full, the writer is blocked until the reader
reads enough
 If the buffer is empty, the reader is blocked until the data is
written
 What happens if both processes start by reading from pipes?
 Deadlock, of course!
 Similar with writing when the pipe becomes full
 When working with several pipes, one has to be extra
careful to avoid deadlock
 A reader will block on the read end of the pipe as long as
there is a reference to the write end – be careful about
69 cleaning up unused references.
Unix Named Pipes
But we can still communicate only between
related processes.

Solution: Named pipes

 Creating a named pipe (using mkfifo()
system call) creates the corresponding file
in the pipe file system
 Opening that file for reading returns a file
descriptor to the read end of the pipe
 Opening that file for writing returns the
write end of the pipe
70
Unix Signals
 Similar to hardware interrupts without priorities
 Each signal is represented by a numeric value. Ex:
 02, SIGINT: to interrupt a process
 09, SIGKILL: to terminate a process
 Each signal is maintained as a single bit in the process table
entry of the receiving process: the bit is set when the
corresponding signal arrives (no waiting queues)
 A signal is processed as soon as the process runs in user mode
 A default action (eg: termination) is performed unless a signal
handler function is provided for that signal (by using the signal
system call)
 Want to know more?
http://users.actcom.co.il/~choo/lupg/tutorials/signals/signals-pr
ogramming.html

71
Sockets
 A socket is defined as an endpoint for
communication
 Concatenation of IP address and port
 The socket 161.25.19.8:1625 refers to port 1625 on
host 161.25.19.8
 Communication link corresponds to a pair of
sockets

72
Remote Procedure Calls
 Wouldn’t it be nice to be able to call procedures
residing on other computers?
 Well, but the parameters and return values need to
be somehow transferred between the computers.
 The computers might have different data format
 Support for locating the server/procedure needed
 Stubs – client and server – side proxies
implementing the needed communication
 The client-side stub locates the server and
marshalls the parameters.
 The server-side stub receives this message,
unpacks the marshalled parameters, and performs
the procedure on the server.
73
Execution of RPC

74
Remote Method Invocation
 Remote Method Invocation (RMI) is a Java
mechanism similar to RPCs.
 RMI allows a Java program on one
machine to invoke a method on a remote
object.

75
Marshalling Parameters

 Local parameters: copies using « object serialization »

technique
 Distant parameters: utilizes references.
76
 Long Term PCB
Program
Short term Execution
Medium Process
Scheduling Process
Term Co State
Schedulers Queues nc Switching
e pt

g
lin
Context

du
Sockets

he
c re a
tePro switch
cess
fork/ Sc
Process
exec RPC/RMI

:
on
O

C i
at ss

pe

IP rat
term
inate
re ce
n

Proc
ra

pe
io

ess Pipes
o

t
Pr

io

oo
exit
ns
C

IP C Signals
C Ba E x am
in s s
n

m o si c ple
io
r m ce

kill s
at

Memory de
Te P r o

sharing ls
Message passing
77