Day01_Notes.
md 2023-11-27
Operating System Concepts
OS Concepts
Linux commands
Shell scripts
Linux System call Programming
Learning OS
step 1: End user
Linux commands
step 2: Administrator
Install OS (Linux)
Configuration - Users, Networking, Storage, ...
Shell scripts
step 3: Programmer
Linux System call programming
step 4: Designer/Internals
UNIX & Linux internals
What is OS?
Interface between end user and computer hardware.
Interface between Programs and computer hardware.
Control program that controls execution of all other programs.
Resource manager/allocator that manage all hardware resources.
Bootable CD/DVD = Core OS + Applications + Utilities
Core OS = Kernel -- Performs all basic functions of OS.
OS Functions
CPU scheduling
Process Management
Memory Management
File & IO Management
Hardware abstraction
User interfacing
Security & Protection
Networking
Process Management
Program
Set of instructions given to the computer --> Executable file.
Program --> Sectioned binary --> "objdump" & "readelf".
1/8
Day01_Notes.md 2023-11-27
Exe header --> Magic number, Address of entry-point function, Information about all sections.
(objdump -h [Link])
Text --> Machine level code (objdump -S [Link])
Data --> Global and Static variables (Initialized)
BSS --> Global and Static variables (Uninitialized)
RoData --> String constants
Symbol Table --> Information about the symbols (Name, Size, section, Flags, Address) (objdump
-t [Link])
Program (Executable File) Format
Windows -- PE
Linux -- ELF
Program are stored on disk (storage).
Process
Progam under execution
Process execute in RAM.
Process control block contains information about the process (required for the execution of process).
Process id
Exit status
0 - Indicate successful execution
Non-zero - Indicate failure
Scheduling information (State, Priority, Sched algorithm, Time, ...)
Memory information (Base & Limit, Segment table, or Page table)
File information (Open files, Current directory, ...)
IPC information (Signals, ...)
Execution context (Values of CPU registers)
Kernel stack
PCB is also called as process descriptor (PD), uarea (UNIX), or task_struct (Linux).
In Linux size of task_struct is approx 4KB
Process
Process is program in execution.
Process has multiple sections i.e. text, data, rodata, heap, stack. ... into user space and its metadata is
stored into kernel space in form of PCB struct.
PCB contains
id, exit status,
scheduling info (state, priority, time left, scheduling policy, ...),
files info (current directory, root directory, open file descriptor table, ...),
memory information (base & limit, segment table, or page table),
ipc information (signals, ...),
execution context, kernel stack, ...
OS Data Structures:
Job queue / Process table: PCBs of all processes in the system are maintained here.
Ready queue: PCBs of all processes ready for the CPU execution and kept here.
2/8
Day01_Notes.md 2023-11-27
Waiting queue: Each IO device is associated with its waiting queue and processes waiting for that IO
device will be kept in that queue
Process Life Cycle
Process States
New
New process PCB is created and added into job queue. PCB is initialized and process get ready
for execution.
Ready
The ready process is added into the ready queue. Scheduler pick a process for scheduling from
ready queue and dispatch it on CPU.
Running
The process runs on CPU. If process keeps running on CPU, the timer interrupt is used to forcibly
put it into ready state and allocate CPU time to other process.
Waiting
If running process request for IO device, the process waits for completion of the IO. The waiting
state is also called as sleeping or blocked state.
Terminated
If running process exits, it is terminated.
Linux: TASK_RUNNING (R), TASK_INTERRUPTIBLE (S), TASK_UNINTERRUPTIBLE (D), TASK_STOPPED(T),
TASK_ZOMBIE (Z), TASK_DEAD (X)
Classification of OS
OS can be categorized based on the target system (computers).
Mainframe systems
Desktop systems
Multi-processor (Parallel) systems
Distributed systems
Hand-held systems
Real-time systems
Mainframe systems
Resident Monitor
Early (oldest) OS resides in memory and monitor execution of the programs. If it fails, error is reported.
OS provides hardware interfacing that can be reused by all the programs.
Batch Systems
3/8
Day01_Notes.md 2023-11-27
The batch/group of similar programs is loaded in the computer, from which OS loads one program in
the memory and execute it. The programs are executed one after another.
In this case, if any process is performing IO, CPU will wait for that process and hence not utilized
efficiently.
Multi-Programming
In multi-programming systems, multiple program can be loaded in the memory.
The number of program that can be loaded in the memory at the same time, is called as "degree of
multi-programming".
In these systems, if one of the process is performing IO, CPU can continue execution of another
program. This will increase CPU utilization.
Each process will spend some time for CPU computation (CPU burst) and some time for IO (IO burst).
If CPU burst > IO burst, then process is called as "CPU bound".
If IO burst > CPU burst, then process is called as "IO bound".
To efficiently utilize CPU, a good mix of CPU bound and IO bound processes should be loaded into
memory. This task is performed by an unit of OS called as "Job scheduler" OR "Long term scheduler".
If multiple programs are loaded into the RAM by job scheduler, then one of process need to be
executed (dispatched) on the CPU. This selection is done by another unit of OS called as "CPU
scheduler" OR "Short term scheduler".
Multi-tasking OR time-sharing
CPU time is shared amoung multiple processes in the main memory is called as "multi-tasking".
In such system, a small amount of CPU time is given to each process repeatedly, so that response time
for any process < 1 sec.
With this mechanism, multiple tasks (ready for execution) can execute concurrently.
There are two types of multi-tasking:
Process based multitasking: Multiple independent processes are executing concurrently.
Processes running on multiple processors called as "multi-processing".
Thread based multi-tasking OR multi-threading: Multiple parts/functions in a process are
executing concurrently.
Multi-user
Multiple users can execute multiple tasks concurrently on the same systems. e.g. IBM 360, UNIX,
Windows Servers, etc.
Each user can access system via different terminal.
There are many UNIX commands to track users and terminals.
tty, who, who am i, whoami, w
Desktop systems
Personal computers -- desktop and laptops
User convinence and Responsiveness
Examples: Windows, Mac, Linux, few UNIX, ...
Multiprocessor systems
4/8
Day01_Notes.md 2023-11-27
The systems in which multiple processors are connected in a close circuit is called as "multiprocessor
computer".
The programs/OS take advantage of multiple processors in the computer are called as "Multiprocssing"
programs/OS.
Windows Vista: First Windows OS designed for multi-processing.
Linux 2.5+: Linux started supporting multi-processing.
Modern PC architectures are multi-core arch i.e. multiple CPUs on single chip.
Since multiple tasks can be exeuted on these processors simultaneously, such systems are also called as
"parallel systems".
Parallel systems have more throughput (Number of tasks done in unit time).
There are two types of multiprocessor systems:
Asymmetric Multi-processing
Symmetric Multi-processing
Asymmetric Multi-processing
OS treats one of the processor as master processor and schedule task for it. The task is in turn divided
into smaller tasks and get them done from other processors.
Symmetric Multi-processing
OS considers all processors at same level and schedule tasks on each processor individually.
All modern desktop systems are SMP.
Distributed systems
Multiple computers connected together in a close network is called as "distributed system".
Its advantages are high availability (24x7), high scalability (many clients, huge data), fault tolerance (any
computer may fail).
The requests are redirected to the computer having less load using "load balancing" techniques.
The set of computers connected together for a certain task is called as "cluster". Examples: Linux.
Handheld systems
OS installed on handheld devices like mobiles, PDAs, iPODs, etc.
Challenges:
Small screen size
Low end processors
Less RAM size
Battery powered
Examples: Symbian, iOS, Linux, PalmOS, WindowsCE, etc.
Realtime systems
The OS in which accuracy of results depends on accuracy of the computation as well as time duration in
which results are produced, is called as "RTOS".
If results are not produced within certain time (deadline), catastrophic effects may occur.
These OS ensure that tasks will be completed in a definite time duration.
Time from the arrival of interrupt till begin handling of the interrupt is called as "Interrupt Latency".
5/8
Day01_Notes.md 2023-11-27
RTOS have very small and fixed interrupt latencies.
RTOS Examples: uC-OS, VxWorks, pSOS, RTLinux, FreeRTOS, etc.
Types of Scheduling
Non-preemptive
The current process gives up CPU voluntarily (for IO, terminate or yield).
Then CPU scheduler picks next process for the execution.
If each process yields CPU so that other process can get CPU for the execution, it is referred as "Co-
operative scheduling".
Preemptive
The current process may give up CPU voluntarily or paused forcibly (for high priority process or upon
completion of its time quantum)
Scheduling criteria's
CPU utilization: Ideal - max
* On server systems, CPU utilization should be more than 90%.
* On desktop systems, CPU utilization should around 70%.
Throughput: Ideal - max
* The amount of work done in unit time.
Waiting time: Ideal - min
* Time spent by the process in the ready queue to get scheduled on the CPU.
* If waiting time is more (not getting CPU time for execution) -- Starvation.
Turn-around time: Ideal - CPU burst + IO burst
* Time from arrival of the process till completion of the process.
* CPU burst + IO burst + (CPU) Waiting time + IO Waiting time
Response time: Ideal - min
* Time from arrival of process (in ready queue) till allocated CPU for first time.
6/8
Day01_Notes.md 2023-11-27
User interfacing
UI of OS is a program (Shell) that interface between End user and Kernel.
Shell -- Commmand interpreter
End user --> Command --> Shell --> Kernel
User interfacing (Shell)
Graphical User Interface (GUI)
Command Line Interface (CLI)
Example shells
Windows
GUI shell: [Link]
CLI shell: [Link], [Link]
DOS
CLI shell: [Link]
Unix/Linux
CLI shell: bsh, "bash", ksh, csh, zsh, ...
ls /bin/*sh
echo $SHELL
shell of current user can be changed using "chsh" command.
GUI shell/standards
GNOME: GNU Network Object Model Environment (e.g. Ubuntu, Redhat, CentOS, ...)
KDE: Kommon Desktop Environment (e.g. Kubuntu, SuSE, ...)
Linux File Structure
Linux follows “/” (root) file system.
“/” is a starting point of Linux file system.
All your data is stored in this partition.
/ contains boot, bin, sbin, etc, root, home, dev, proc, mnt, media, opt
In Linux everything is a file.
Mainly there are two types of files in Linux
File
Directory (Folder)
Linux Directories
boot - files related to booting
vmlinuz - kernel Image
grub - boot loader
config - kernel configuration
initrd/initramfs - initail root file system
bin - user commands in binary format
sbin - all admin/system commands in binary format
etc - configuration files
root - home directory of root user
home - it contains sub directories for each user with its name
7/8
Day01_Notes.md 2023-11-27
devendra -> /home/devendra
sunbeam -> /home/sunbeam
osboxes -> /home/osboxes
dev - it contains all device related files
lib - shared program libraries required by kernel
mnt - it is temporary mount point
media - it is mount point for media eg cdrom
opt - stores optional files of large softwares
proc - virtual file system - it contains information about system or processes
sys - entries of each block devices, subdirectories for each physical bus type supported, every
device class registered with the kernel, global device hierarchy of all devices
tmp - temporary files that may be lost on system shutdown
usr - read only directory that stores small programs and files accessible to all users
Path
It is a unique location of any file in the file system.
It is represented by character strings with few delimiters (“/”, “\”, “:”)
Types of path
There are two types of paths in linux
Absolute path
Path which starts with “/” is called as absolute path.
E.g. /home/devendra/MyData/Demos/[Link]
Relative path
Path with respect to current directory is called as relative path
E.g. MyData/Assignments/[Link]
Types of files
Regular file (-)
Directory file (d)
Link file (l)
Socket file (s)
Pipe file (p)
Character Special file (c)
Block Special file (b)
8/8
