Scribd
Upload
191 views

Case Study of Linux - Linux Kernel Version 2.6

Linux is the most popular open-source, freely distributed, fullfeatured operating system. It is available to the public for examination and modification and is free to download and install. The name Linux is derived from "linus" and "UNIX"

Uploaded by

Vijay Kumar Reddy
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
191 views

Case Study of Linux - Linux Kernel Version 2.6

Linux is the most popular open-source, freely distributed, fullfeatured operating system. It is available to the public for examination and modification and is free to download and install. The name Linux is derived from "linus" and "UNIX"

Uploaded by

Vijay Kumar Reddy
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 23

Case Study of Linux Introduction Linux kernel version 2.

6
Core of the most popular open-source, freely distributed, fullfeatured operating system Linux source code is available to the public for examination and modification and is free to download and install

Popular in high-end servers, desktop computers and embedded systems Supports many advanced features
Symmetric multiprocessing (SMP), Access to multiple file systems Support for broad spectrum of hardware architectures

History Created in 1991 by Linus Torvalds a student at the University of Helsinki, Finland
The name Linux is derived from Linus and UNIX

The Minix source code served as a starting point Torvalds sought advice from the community Developers continued to support the concept of a new, freely available operating system

History Distribution
Enables users unfamiliar with Linux details to install and use Linux Includes software such as
The Linux kernel System applications (e.g., user account management, network management and security tools) User applications (e.g., GUIs, Web browsers, text editors, e-mail applications, databases, and games) Tools to simplify the installation process

Linux Overview Linux systems include user interfaces and applications in addition to the kernel Borrows from the UNIX layered system approach System contains kernel threads to perform services
Implemented as daemons, which sleep until awakened by a kernel component

Multiuser system
Restricts access to important operations to users with superuser (also called root) privileges

Development and Community Torvalds controls all modifications to the kernel Relies on a group of about 20 lieutenants to manage kernel enhancements As a development kernel nears completion:
Feature freeze: no new features are added to the kernel Code freeze: only code that fixes important bugs are accepted

Many corporations support Linux development Linux is distributed under the GNU Public License (GPL) Linux is free, copyrighted software

Distributions Over 300 distributions available Typically organized into packages, each containing a single service or application Popular distributions include:
Debian Mandrake Red Hat SuSE Slackware

User Interface Can be accessed via the command-line via shells such as bash, csh and esh Most Linux GUIs are layered
X Window System
Lowest level Provides to higher GUI layers mechanisms to create and manipulate graphical components

Window manager
Builds on mechanisms in the X Window System interface to control the placement, appearance, size and other window attributes

Desktop environment (e.g., KDE, GNOME)


Provide user applications and services

Kernel Architecture Monolithic kernel


Contains modular components, however

UNIX-like or UNIX-based operating system Six primary subsystems:


Process management Interprocess communication Memory management File system management
VFS: provides a single interface to multiple file systems

I/O management Networking

Kernel Architecture
Linux architecture.

Process Management Process and Thread Organization Tasks


Processes and thread are internally represented as tasks using the task_struct structure

Task states:
Running Sleeping Zombie Dead Stopped Active and expired (not stored by state)

Process Scheduling
Preemptive scheduler
Each task runs until its quantum, or time slice, expires, a higher priority process becomes runnable or the process blocks Tasks placed in run queues (similar to multilevel feedback queues) Priority array maintains pointer to each level of the run queue
Task of priority i is placed in the ith entry of the priority array for a run queue

Scheduler dispatches the task at the front of the list in the highest level of the priority array
If more than one task exists in a level of the priority array, tasks are dispatched from the priority array round-robin

Process Scheduling
Scheduler priority array.

Memory Management Memory Organization Linux uses paging exclusively


Often implemented using a single page size On 32-bit systems, kernel can address 4GB of data
On 64-bit systems, the kernel supports up to 2 petabytes of data

Linux uses demand-paged system Linux uses Buddy Algorithm for Memory Allocation.

Buddy Algorithm Buddy Algorithm is a memory management algorithm used to speed up the combining of adjacent holes when a process is terminated. In buddy say allocation, is always in terms of 2n. Consider a memory size of 1MB(1024 KB). Now process A request for 70K then 128 K is allocated to the process because this is the smallest power of 2k which is greater than 70K. The memory is split as follows:
128K Process A Allocates (70K) 128K 256K 512K

Process B request for 35K


128K (A) 64K(B) 64K 256K 512K

Process C request for 80K


128K (A) 64K(B) 64K 128K(C) 128K 512K

Process A completes execution


128K 64K(B) 64K 128K(C) 128K 512K

Process D request for 60K


128K (D) 64K(B) 64K 128K(C) 128K 512K

Page Replacement(LFU)
LFU assumes that the page which is used less number of times in the past may not be used in the near future and hence marks that page for replacement. The O.S. must maintain a counter for each page present in the core. This counter is updated every time the page is used. When the page fault occurs the page with smallest counter is deleted. The idea is that if a page is not heavily used so far, then we assume that it behaves in the same fashion even in the future also and hence delete that page.

Ex: Let us assume that pages of a program are used in the following fashion:
7 7 7 0 1 1 2 0 0 3 0 0 3 3 4 Let us assume that there are 3 frames in the memory
7 7 0 7 0 1 7 1 2 7 1 0 7 0 3 7 0 4

1 Page fault = 7

File Systems Second Extended File System (ext2fs) Ext2 characteristics


Goal: high-performance, robust file system with support for advanced features Typical block sizes are 1,024, 2,048, 4,096 or 8,192 bytes

Second Extended File System (ext2fs) ext2 inode


Represents files and directories in an ext2 file system Stores information relevant to a single file or directory, such as time stamps, permissions, the identity of the files owner and pointers to data blocks
First 12 pointers directly locate the first 12 data blocks 13th pointer is an indirect pointer locates a block that contains pointers to data blocks 14th pointer is a doubly indirect pointer locates a block of indirect pointers. 15th pointer is a triply indirect pointer locates a block of doubly indirect pointers

Provides fast access to small files, while supporting larger files

Second Extended File System (ext2fs)


Ext2 inode contents.

Input/Output Management
Kernel provides common interface for I/O system calls Devices are grouped into classes
Members of each device class perform similar functions Allows the kernel to address the performance needs of certain devices (or classes of devices) individually

Device driver: software interface between system calls and a hardware device
Most have been written by independent developers Typically implemented as loadable kernel modules

Device special files


Most devices are represented by device special files Entries in the /dev directory that provide access to devices List of devices in the system can be obtained by reading the contents of
/proc/devices

Device Drivers
/proc/devices file contents.

Disk Scheduling Disk scheduling algorithms Default is the LOOK algorithm


Can lead to synchronous read request starvation during concurrent writes

Deadline scheduler
Eliminates read request starvation by ensuring all read operations are performed by a certain deadline

Authentication Default authentication


User enters username and password via login Passwords are hashed (using MD5 or DES)
Encryption cannot be reversed

Stored in /etc/passwd or /etc/shadow

Pluggable authentication modules (PAMs)


Can reconfigure the system at run time to include enhanced authentication techniques For example:
Disallow terms found in a dictionary and require users to choose new passwords regularly Supports smart cards, Kerberos and voice authentication

You might also like