ad majorem Dei gloriam

CPS 356: Operating Systems

Department of Computer Science, University of Dayton

Fall 2017

Table of Contents (TOC)

Course Description

CPS 356 (3 hours) is a course that introduces the theoretical and practical issues underlying an operating system's structure and operation. Topics include process and thread creation and management, scheduling, concurrent, multi-threaded programming and synchronization, deadlock, memory management, virtual memory, and computer security. Concepts are demonstrated using the C, Go, and Elixir programming languages in a Linux environment. This course assumes no prior experience with Linux, C, Go, Lua, or Elixir.

Course Details

CPS 250 (Introduction to Computer Organization) or ECE 314 (Fundamentals of Computer Architecture) & CPS 350 (Data Structures & Algorithms).
Meeting times:
T Th 5:05pm-6:20pm, MH 203.
Dr. Saverio Perugini, e-mail id: sperugini1, tel: 229-4079, office: AN 145
OHs: T Th 6:30pm-7:30pm; Th 3:30pm-4:30pm (only when classes are in session); & by appointment.
Student helper:
Matthew Weiler (e-mail id: weilerm3) and Branden McNally (e-mail id: mcnallyb2); AN 135
OHs: M W 10:30am-1:00pm & 5:10pm-6:00pm (only when class is in session) & by appointment.


Required Textbooks

    [LP] Linux Programming by S. Perugini. 2017. Draft (Available as a Resource in Isidore).

    [PLCI] Programming Languages: Concepts and Implementation by S. Perugini. 2017. Draft (Available as a Resource in Isidore; only Chapter 14: Concurrency and Synchronization needed for this course.).

    [SCMSW] Seven Concurrency Models in Seven Weeks: When Threads Unravel by P. Butcher. Pragmatic Programmers, Dallas, TX, 2014. ISBN-13:978-1-937785-65-9 (textbook webpage contains links to the source code of all programs in the text). An eBook of [SCMSW] is available free to all UD students in the library's eContent collection. To access it conduct a search for the title in the library's catalog at

Steps to access [SCMSW] off-campus:
  1. Go to
  2. Login using your standard UD credentials.
  3. Click on the link labeled: EBSCO Ebook Collection (formerly NetLibrary).
  4. This link will will take you to a page were you can perform a search. Search for "Seven concurrency models in seven weeks: When threads unravel" Click on "eBook Full Text".

    Note that there is a limit on how many people can view the book and once someone checks it out, no one else can use it. Also, if you incorrectly break your session (closing the page), the book remains checked out for a period of time and you cannot use the book in until the session ends.
    [SMLSW] Seven More Languages in Seven Weeks: Languages that are Shaping the Future by B.A. Tate, F. Daoud, I. Dees, & J. Moffit. Pragmatic Programmers, Dallas, TX, 2014. ISBN-13:978-1-941222-15-7 (textbook webpage contains links to the source code of all programs in the text). An eBook of [SMLSW] is available free to all UD students in the library's eContent collection. To access it conduct a search for the title in the library's catalog at Steps to access [SMLSW] off-campus: follow same steps above for accessing [SCMSW] (with the title of this book).


Course Wordle

CPS 356: Operating Systems/Fall 2017


Course Objectives


Course Outline

Course outline, required reading assignments, lecture notes, & homeworks & projects:
  1. Introduction to operating systems & the UNIX/Linux & C programming environment ([OSCJ8] Ch 1-2; [LP] Ch 1-5; [USP] Ch 1-2, 4)
    1. introduction to operating systems (review of computer organization, C exercises): Aug 24 29 31
    2. the UNIX philosophy (class UNIX/Linux page, Linux quick reference sheet, vi quick reference sheet, vi editor, UW vi reference): Sep 7
    3. files & directories (manipulation & management): self-study
    4. system libraries & I/O (C quick reference sheet): Aug 24 29 31 Sep 5 7
    5. compiling C in UNIX (static vs. dynamic linking, macros, conditional compilation, error handling, & debugging; RMS's gdb tutorial, valgrind): Aug 29 31 Sep 5 19

  2. Processes & threads ([OSCJ8] Ch 3-4; [LP] Ch 6-7; [USP] Ch 2-6)
    1. processes (identification; getpid, creation; fork, & termination): Aug 24 29 31 Sep 7
    2. memory allocation/deallocation: Aug 31
    3. (shell) job control (through signals) & terminals: Sep 12
    4. storage & linkage classes in C: Sep 19
    5. threads & thread-safe functions (strtok): Sep 12 19 21
    6. process manipulation (wait): Sep 14
    7. process manipulation (exec & building an argument vector): Sep 21
    8. the Linux shell (Learning the Shell) & process environment (variables, configuration, customization; Advanced Linux quick reference sheet): Sep 21
    9. low-level I/O (open & close, & read & write): Sep 26
    10. context switching: Sep 26
    11. files & directories (data structures, inodes, & hard & symbolic links): value added

    12. Exam I (closed book, closed notes) [practice problems]: Oct 3

  3. Scheduling ([OSCJ8] Ch 5)
    1. types, evaluation criteria, & algorithms (FCFS, SJF, SRTF, RR): Oct 3 10 12
    2. multi-level queues & multi-level feedback queues: Oct 10 12

  4. Concurrency and Synchronization ([PG] Ch 7, [SCMSW] Ch 5, & [SMLSW] Ch4)
    1. mutual exclusion & the critical section problem (§§ 6.1-6.4): Oct
    2. classical problems of synchronization (§ 6.6): Oct
    3. introduction to Go programming (Go by example, a tour of Go, command-line programming with Go, effective Go): Oct 24
    4. communicating sequential processes (CSP) (Go/CSP quick reference sheet) in Go ([PG] Ch7): Oct 24 26 31 Nov 2

    5. Exam II (closed book, closed notes) [practice problems]: Nov 7

    6. the Actor model of concurrency in Elixir (Elixir quick reference sheet; [SCMSW] Ch5 & [SMLSW] Ch4): Nov 14 16 20

      Exam III (closed book, closed notes) [practice problems]: Dec 5

  5. Deadlock ([OSCJ8] Ch 7)
    1. deadlock prevention (§§ 7.1-7.4):
    2. deadlock avoidance & Banker's algorithm (§ 7.5) (Banker's examples):
    3. deadlock detection & recovery (§§ 7.6-7.8):

  6. Memory Management ([OSCJ8] Ch 8)
    1. fundamentals (overview of hardware, logical vs. physical address space, dynamic loading & linking) (§§ 8.1-8.2): Nov
    2. contiguous allocation (§ 8.3): Nov
    3. paging (§§ 8.4-8.5): Nov
      paging (§§ 8.4-8.5): Nov
    4. segmentation (§§ 8.6-8.8): Nov
    5. segmentation faults & buffer overflows (§§ 8.6-8.8): Nov

  7. Virtual Memory ([OSCJ8] Ch 9)
    1. demand paging (§§ 9.1-9.3):
    2. page replacement algorithms (FIFO, optimal, LRU; § 9.4):
    3. allocation (§ 9.5) & thrashing (§ 9.6):
    4. course reflection (terms): Dec

  8. Appendices ([LP] Appendices)
    1. Vim quick reference sheet
    2. Linux quick reference sheet
    3. C quick reference sheet
    4. Advanced Linux quick reference sheet
    5. Go quick reference sheet
    6. Elixir quick reference sheet

  9. Final Exam (comprehensive, closed book, closed notes) [practice problems]: T Dec 12, 4:30pm-6:20pm, MH 203.


Quick Reference Sheets



  1. Exam I (closed book, closed notes) [practice problems]: Oct 3

  2. Exam II (closed book, closed notes) [practice problems]: Nov 7

  3. Exam III (closed book, closed notes) [practice problems]: Dec 5

  4. Final Exam (comprehensive, closed book, closed notes) [practice problems]: T Dec 12, 4:30pm-6:20pm, MH 203



Homework Assigned Due Total points
Homework #1 Aug 24 Sep 7 40
Homework #2 Sep 7 Sep 14 30
Homework #3 Sep 14 Sep 21 40
Homework #4 Sep 21 Sep 28 30
Homework #5 Oct 10 Oct 12 20
Homework #6 Oct 26 Nov 2 26
Homework #7 Nov 6 Nov 8 18
Homework #8 Nov 14 Nov 16 13
Total Homework Points: 245


Homework #1

Assigned: August 24
Due: September 7, 5:05pm

Total points: 20 points

Style guide | Academic Integrity | Evaluation Criteria

[LP] Programming Exercise 4.31.31. (Use filename /home/<logname>/homeworks/hw1/countsubsstdin.c)


Homework #2

Assigned: September 7
Due: September 14, 5:05pm

Total points: 30 points

Style guide | Academic Integrity | Evaluation Criteria

  1. (20 points) [LP] Programming Exercise 4.31.34. (Use filename /home/<logname>/homeworks/hw2/removesubsargs.c)

  2. (10 points) [LP] Programming Exercise 7.3.19. (Use filename /home/<logname>/homeworks/hw2/alternate.c)


Homework #3

Assigned: September 14
Due: September 21, 5:05pm

Total points: 40 points

Style guide | Academic Integrity | Evaluation Criteria

  1. (20 points) [LP] Programming Exercise 4.31.36. (Use filename /home/<logname>/homeworks/hw3/allsubsargs.c)

  2. (20 points) [LP] Programming Exercise 4.31.38. (Use filename /home/<logname>/homeworks/hw3/parsestring.c)


Homework #4

Assigned: September 21
Due: September 28, 5:05pm

Total points: 30 points

Style guide | Academic Integrity | Evaluation Criteria

  1. (15 points) [LP] Programming Exercise 7.7.5 as /home/<logname>/homeworks/hw4/mathfib.c

  2. (15 points) [LP] Programming Exercise 7.7.7 as /home/<logname>/homeworks/hw4/interface.c


Homework #5

Assigned: October 10
Due: October 12, 5:05pm

Total points: 20 points

Style guide | Academic Integrity | Evaluation Criteria

(5+6+6+3=20 points) [OSCJ8] Exercise 5.12 (parts a, b, c, and d) on pp. 235-236, or exercise 5.4 (parts a, b, c, and d) on p. 205 of the 7th ed.


Homework #6

Assigned: October 26
Due: November 2, 5:05pm

Total points: 26 points

Style guide | Academic Integrity | Evaluation Criteria

  1. (13 points) [PLCI] Programming Exercise 14.3.19 as /home/<logname>/homeworks/hw6/PingPongFan.go

  2. (13 points) [PLCI] Programming Exercise 14.3.21 as /home/<logname>/homeworks/hw6/PingPongChain.go


Homework #7

Assigned: November 6
Due: November 9, 5:05pm

Total points: 18 points

Style guide | Academic Integrity | Evaluation Criteria

  1. (5 points) [PLCI] Programming Exercise 14.3.23 as /home/<logname>/homeworks/hw7/HardwareData.go.

  2. (13 points) [PLCI] Programming Exercise 14.3.22 as /home/<logname>/homeworks/hw7/FairDiningPhilosophers.go.


Homework #8

Assigned: November 14
Due: November 16, 5:05pm

Total points: 13 points

Style guide | Academic Integrity | Evaluation Criteria

[PLCI] Programming Exercise 14.4.16 as /home/<logname>/homeworks/hw8/SleepingBarber.ex. You must use the model for messages and the skeleton provided in the problem specification.


Midterm Project

Assigned: October 12
+20 points Extra Credit Due: October 19, 11:59pm
Due: October 26, 5:05pm

Total points: 80 points

Isidore | Style guide | Academic Integrity | Evaluation Criteria


Design and implement a program (in any language) that simulates some of the job and CPU scheduling of a time-shared operating system.

Detailed Description and Requirements

When jobs initially arrive in the system, they are put on the job scheduling queue which is maintained in FIFO order. The job scheduling algorithm is run when a job arrives or terminates. Job scheduling allows as many jobs to enter the ready state as possible given the following restriction: a job cannot enter the ready state if there is not enough free memory to accommodate that job's memory requirement. Do not start a job unless it is the first job on the job scheduling queue. When a job terminates, its memory is released, which may allow one or more waiting jobs to enter the ready state.

A job can only run if it requires less than or equal to the system's main memory capacity. The system has a total of 512 blocks of usable memory. If a new job arrives needing more than 512 blocks, it is rejected by the system with an appropriate error message. Rejected jobs do not factor into the final statistics (described below).

Note that all jobs in the ready state must fit into available main memory.

Process scheduling is managed as a multilevel feedback queue. The queue has two levels, each queue is organized as a FIFO, and both use a round robin scheduling technique. New jobs are put on the first level when arriving in the ready state. When a job from the first level is given access to the CPU, it is allowed a quantum of 100 time units. If it exceeds that time quantum, it is preempted and moves to the second level.

The jobs on the second level may only be allocated the CPU if there are no jobs on the first level. When a job on the second level is given access to the CPU, it is allowed a quantum of 300 time units. If it exceeds that, it is preempted and put back on the second level of the ready queue.

Process scheduling decisions are made whenever any process leaves the CPU for any reason (e.g., expiration of a quantum or job termination). When a job terminates, do job scheduling first, then process scheduling. Also, give preference to first level jobs (i.e., if a job from the second level of the ready queue is running, and a new job enters the first level, the running job is preempted to the second level in favor of the first level job).

While executing on the CPU, a job may require I/O, which preempts it to the I/O wait queue for the duration of its I/O burst.

When a job completes, put it on a finished list for later processing.

The simulator is driven by the events read from standard input. Examples of possible events are given below. The first field will be the first character of the line, and subsequent fields will be separated by one of more spaces or tabs. The header of each field in the following examples does not appear in the input stream.

A new job arrives:

Event Time Job Memory Run Time
A     140  12  24     2720

Interpretation: job 12 arrives at time 140, requires 24 blocks of memory and uses the CPU for a total of 2720 time units.

A job needs to perform I/O:

Event Time I/O Burst Time
I     214  85

Interpretation: the job currently running on the CPU will not finish its quantum because at time 214 it needs to perform I/O for a duration of 85 time units.

Display the status of the simulator:

Event Time
D     214

Interpretation: display the status of the simulator at time 214.

You may assume that events appear on the input stream in ascending time order and no two events happen at the same time. However, realize that the events given in the input stream are not only events which your simulator must handle. For instance, a time quantum expiration is not an event given in the input stream, but it is an event which your simulator must handle. Furthermore, an internal event, such as a time quantum expiration, not in the input stream, may occur at the same time as an event in the input stream (e.g., a new job arrival). Events in the input stream are external events.

The following is a list of internal events (i.e., not given on the input stream) which your simulator must handle:

Assume that context switching and displays take no simulator time (an unrealistic assumption in a real operating system).

When a display is requested, print the contents of all queues as well as the job currently running on the CPU to standard output using only the format used in the sample output given below.

After processing all jobs, write the following to standard output (in this order, as shown on the sample output given below):

  1. completion time for each job (in order of completion),
  2. average turnaround time (where turnaround time is defined as completion time minus arrival time), and
  3. average job scheduling wait time (where wait time is defined as the number of time units spent in the job scheduling queue).

Event Collisons

Often more than one event happen at the same time. Use the following rules to determine which events to process first:

Architectural View of the Simulator

Additional Requirements

  1. Your system must be developed and run on the cpssuse systems. Mutliple programming languages are available on these systems, including C (/usr/bin/gcc), C++ (/usr/bin/g++), Java (/usr/bin/javac, /usr/bin/java), Go (/usr/local/go/bin/go), Perl (/usr/bin/perl), and Python (/usr/bin/python). If the language you want to develop your simulator is absent from the system, let us know and we will try to get it installed.
  2. Your implementation must be distributed across more than one source code file, in some sensible manner which reflects the logical purpose of the various components of your design, to encourage problem decomposition and modular design.
  3. Include a README file in your submission which describes (i) the language you used to develop your program and (ii) the name of the compiler or interpreter which you used to compile or interpret your program.

Hints and notes

Test data: sample input and output streams

  1. (only events A & D, & E & T) p2stdin_a and p2stdout_a [image]
  2. (all events A, I, & D, & E, C, & T) p2stdin_b and p2stdout_b [image]

This test data is available at /home/perugini_cps356/share/project/. At first, simply try to get only one job through your system; see the test file /home/perugini_cps356/share/project/testdata/one.txt. Once you are confident that your system processes only one job properly, try to get two jobs through the system; see the test file /home/perugini_cps356/share/project/testdata/two.txt

While developing your simulator, you are encouraged to get it to work on the simple test input first (p2d.dat) and progressively enhance and refine your system to the point where it works on the most complex test input (p2a.dat).

Use the Linux diff utility to compare your output to the correct output. For full credit, the output produced by your program must have zero differences, as defined by diff, with the output posted here.

There is also a reference executable of a solution for this project available at /home/perugini_cps356/share/project/OSsim


How to submit

Note: All directory and filenames below are case-sensitive. You must use the directory and filenames exactly as shown below, (i.e., all lower case).

Prepare your submission file as /home/<logname>/project/project.tar. This archive must contain only the most minimal set of files necessary to build your simulator from scratch. Only the file /home/<logname>/project/project.tar will be electronically collected from your account on the deadline.

Failure to follow these submission requirements will result in a 10% penalty.


  1. How do a create my tar file?
      The requirements require you to make a .tar and place it in /home//project/project.tar. A tar file is a single file that contains a packaged up files and/or directories. The command to make a tar file on UNIX systems (Mac OS X and Linux) is as follows:
      tar cvf project.tar <directory_with_code>
      tar cvf project.tar <list of source code files>
      Be very careful to ensure the file name after the cvf is the name of the tar file. If you run the command as follows, you will overwrite one of your source files!
      tar cvf
      The command to untar a tar file is as follows:
      tar xvf project.tar 
      This will create the directory structure with all the files indicated by project.tar.

  2. I developed my project on Windows using and IDE. What should I do?

      Some of you may work locally on your personal computer in some IDE. To upload your code from your computer to the suse server you can either copy and paste the contents of a file into a file on the server or you can use a secure copy program.

      To transfer files to and from your account, use a secure file transfer program such as FileZilla. You can also use PSCP or PSFTP available for free download from the PuTTY Download Page. You can use pscp.exe or psfpt.exe.

      For Windows users you will need to download pscp or psfpt.exe. You will need to have the pscp.exe or psftp.exe in your CMD path (system environment variable PATH) or in your current working directory for this command to work or you will have to specify the path to pscp in the command (C:\putty\pscp.exe for example). Below is an example of the command:

      pscp.exe -r <path_to_local_directory_with_code> <logname><logname>/project/

      For Mac OS X and Linux users, there is a command built into your terminal called scp (for secure copy. The format of this command is:

      scp -r <path_to_local_directory_with_code> 

      Lastly, it is strongly recommended that you make multiple copies of your project. This will allow you to revert changes if something accidentally gets overwritten.


Ninety percent of your score will come from correctness and 10% of your score will come from following our programming style guide. Applicable submission penalties will then be applied.

In an effort to award partial credit to students who are unable to complete certains parts of this project, students earn up to two different portions of the 70 possible points for correctness:

Note: Depending on how you order the jobs on your I/O wait queue, your dump of the jobs waiting for I/O may not match our output exactly, and for just that queue, that is acceptable. For instance, you might organize your I/O wait queue as a priority queue where the job which comes off first is at the head, or you might maintain jobs on the I/O wait queue in the order in which they are put on and then search for the job to take off when the I/O is complete.


Evaluation Criteria

(point values below are approximate)
Component Quantity Points per Total points
Homeworks and pop quizzes ~10 varies (33 EC) 245
Midterm Project 1 80 80
Exams 3 125 375
Final exam (comprehensive) 1 300 300

Homeworks involves analytical, theoretical, and programming exercises. The programming requires a fair amount of critical thought and design, and approximately 500-1000 lines of code. To prepare students for the realities of computer science problems in industry and graduate school (and beyond) this course encourages (and rewards) self-reliance and independent, self-directed work. Handwritten assignments are not accepted. Assignments are due at 5:05pm in class. Late assignments are not accepted. No exceptions. All exams are in-class, closed-book, and closed-notes. Attendance is mandatory at all examinations; make-ups are not given. Any missed examination will result in a zero. Make no assumptions about anything; always consult the instructor first. Final letter grades of A, A-, B+, B, B-, C+, C, C-, and D start approximately at 93, 90, 87, 83, 80, 77, 73, 70, and 60 percent, respectively.



CPS 356 is a challenging course and moves at a very fast pace. Spending a minimum of 9 hours outside of class each week reading, studying, and programming is required. I advise you to see me to discuss any problems you may have before you are evaluated. Having said this, CPS 356 is exciting, fun, and essential. The advent of multi-core processors on the desktop makes mastery of core operating system concepts and concurrent programming more necessary than ever for the modern computer scientist.


Classroom & Course Policies

Students are expected to conduct themselves with respect, integrity, and virtue. Keep phones and similar devices in a silent mode and put away during class (i.e., out of sight). The use of laptop computers and similar devices is not permitted in class. Audio or video recording of any kind in class is strictly prohibited.


Academic Integrity

To achieve the course objectives, homework assignments must be a sole result of your work, not be shared with other students, and prepared in accordance with the University Honor Pledge (see below). Moreover, you may not plagiarize code from any, cited or uncited, textbooks, online resources, or other authors. There is no team, group-work assignments in this class. Discussions among classmates must never include pending assignments. No exemptions. All questions and comments about a pending assignment must only be directed to the instructor and teaching assistants. Evidence indicating a violation of this policy will be handled according to the University Academic Honor Code and result in a doubly-weighted zero which will not be dropped (e.g., if the assignment is worth 100 points, you receive a 0/200) or a zero on the next exam. Make no assumptions about this policy; always consult the instructor first. No student should ever feel that they must resort to academic dishonesty. You are encouraged to consult the instructor if you are struggling with the course or an assignment. No grade is worth your integrity. Honesty in your academic work will develop into professional integrity. The faculty and students of the University of Dayton will not tolerate any form of academic dishonesty.

The Honor Pledge as listed in the Academic Honor Code section of the Undergraduate Catalog applies in full to this course.

Honor code FAQ


Other Helpful ...

Programming style guide

Quick Reference Sheets:
  • Vim quick reference sheet
  • Linux quick reference sheet
  • C quick reference sheet
  • Advanced Linux quick reference sheet
  • Go quick reference sheet
  • Elixir quick reference sheet

  • Practice problems:
    see conceptual and programming exercises in [LP] and [PLCI] (Chapter 14); (outdated, but still relevant) practice problems.

    available in Isidore

    Computer accounts:
    Linux account access | UDit | A beginner's guide to effective e-mail

    Helpful links:
    academic calendar (PDF) | student handbook (PDF) | UDit FERPA policies

    Dr. Perugini welcomes any feedback you may have on the course motif and approach, style of the lectures, the concepts presented in class, the course webpage, homeworks, projects, deadlines, exams, course and grading policies, or your general experience in the course.


    Last modified:
    ad majorem Dei gloriam