Linux architecture

1. LINUX ARCHITECTURE

2. Outline
● Linux history
● Linux system architecture
● Linux kernel
● Kernel Architecture overview
● Kernel Functional overview
– File system
– Process management
– Memory management
– Device driver
– Networking

3. History of Linux
● First truly free UNIX like OS
● 1983 , GNU developed by Richard Stallman but kernel was missing
● Personal project by Finnish student Linus Torvalds to create a new
free operating system kernel in year 1991.
● Linus Torvalds began a project that later became the Linux kernel.
● First version of kernel 0.01
● 1991 from few c files to 18 million lines in 2015

4. History of Linux

5. Characteristics of linux OS
● Multiuser Capability
● Multitasking
● Portability – Works on different hardware
● Security
– Authentication
– Authorization
– Encryption

6. Linux system architecture

7. Linux system architecture
● Hardware layer – Consist of all peripherical devices like RAM, CPU, HD
● Kernel
– Core of OS
– Interacts with hardware
– Perform low level services
Shell
– Interface to kernel. Hides complexity of kernel from users.
– Take commands and executes kernel functions.
● Utilities

8. Linux system architecture

9. Linux system architecture

10. Linux kernel
● Monolithic kernel
● Supports
– Preemptive mutitasking
– Virtual memory
– Shared libraries
– Demand loading
– Memory mangement
– Threading
– Loadable kernel modules

11. Kernel Functional overview

12. Linux system architecture
● Virtual address space – 0x0000000 to 0xFFFFFFFF for 32bit linux
● User space - normal programs run
– Contains data, stack
● Kernel space - where kernel run
– Contains kernel code, pagetable..

13. File systems
● In UNIX everything is a file
● Filesystem is a hierarchical storage of data adhering to a specific
structure
● Tree like structure
● Folders named as directories
● Root - '/ '

14. File systems
● Sort of files : directories (d), special files(s), links(l), sockets (s),
named pipes(p)
● Major partitions on a Linux system:
– data partition: system data, including the root partition
containing all the data to start up and run the system
– swap partition: expansion of the computer's physical memory,
extra memory on hard disk(twice amount of RAM)
● Mount point
– Partitions attached through mount point

15. File systems
● Virtual file system

16. File system
● Concepts in VFS
– Super block – filesystem metadata
– Inode – File metadata(permissions, owner, creation time)
● VFS is designed to work with filesystems that understand and
implement such concepts

17. File systems - VFS
● Four object types of VFS
– Superblock object – specific mounted filesystem.
– inode object - a specific file.
– dentry object - directory entry, which is a single component of
a path.
– file object - an open file as associated with a process.

18. Process management and Scheduling
● Illusion of multitasking
● Issues kernel must resolve
– Applications must not interfere.
– CPU time shared fairly.
● Process priorities
– Hard real time process
– Soft real time process
● Allocation of CPU time

19. Process management and Scheduling
● Fork and exec system calls
● Process life cycle
– Running, Waiting, Sleeping, Zombie

20. Process management

21. Process management and Scheduling
● User and kernel mode
● Interrupts and kernel threads
● Pre emptive multitasking
– Scheduler can take global decisions to start and stop a process

22. Process management and scheduling
● I/O bound and processor bound process
– I/O bound : much time on waiting request from I/O
– Process bound : much time on executing code
● Nice values – range from -20 to 19 (default 0)
– Larger nice value low priority
● Time slice (default 20 ms)
– Numeric value showing
how long a task can run

23. Process management and scheduling
● Scheduling policy with example
– Consider two runnable task text editor and video encoder
– Text editor – I/O bound, video encoder – processor bound
– Text editor : Interactive process hence larger timeslice and high
priority

24. Memory management
● Virtual memory – makes system appear to have more memory than it
actually has
– Virtual memory greater than physical memory
– Each process have its own virtual address space
– Sharing of physical memory
● Zones
– ZONE_DMA - This zone contains pages that can undergo DMA.
– ZONE_DMA32 - these pages are accessible only by 32-bit devices
– ZONE_NORMAL - contains normal, regularly mapped, pages.
– ZONE_HIGHMEM - contains “high memory,” which are pages not
permanently mapped into the kernel’s address space.

25. Memory management
● Page table - Data structures used to map virtual to physical address
● Pages – memory divided into chunks of equal size. (Intel x86 4Kbytes)

26. Memory management
● Linux page tables

27. Memory management
● Demand paging
– Creating page faults and bringing image/file/data from disk to
physical memory
– Process should wait in queue until it is fetched
● Swapping
– No free space in RAM swapping done and done by kswapd
– The page selected for swapping has modified (dirty page)it is saved
as swap file.
– Least recently used(LRU) used for swapping.
● Shared virtual memory
– Two process sharing same physical memory

28. Memory management
● Page allocation and deallocation
– mem_map_t : decribes single physical page
– free_area : to find free pages
– Buddy algorithm to allocate and deallocate pages
– Page allocation results in page fragmentation
– Reallocation combiles fragmented pages into large blocks
● Malloc(), vmalloc() and kmalloc()

29. Memory management

30. Devices drivers
● Communicate I/O devices through device files.
● Character and Block devices
– Character - continous stream of data, eg:modem
– Block – BLock of data, eg: hard disc

31. Device drivers
● Dynamic creation of device files

32. Thank you