Internal representation of files ppt

1. 1
INTERNAL REPRESENTATION
OF FILES
Prepared by,
A. V. Surve

2. 2
Contents
 Inodes
 Structure of a regular file
 Directories
 Conversion of a path name to an inode
 Super block
 Inode assignment to a new file
 Allocation of Disk Blocks.
 Other File Types

3. 3
Lower Level File system Algorithms
namei
alloc free ialloc ifree
iget iput bmap
buffer allocation algorithms
getblk brelse bread breada bwrite

4. 4
Inodes
 Contains the information necessary for a process to
access a file.
 Exists in a static form on disk.
 The kernel reads them into an in-core inode to
manipulate them.
 2 Types:
 Disk Inode
 In-core Inode

5. 5
Disk Inodes (1/2)
Consists of following fields:
 File Owner identifier
 Individual owner
 “Group” owner
 Set of users who have access rights to a file
 File Type
 File

Regular, directory, character or block special
 FIFO (pipe)
 File Access permissions
 To protect by three classes(owner, group, other)
 Read, write, execute

6. 6
Disk Inodes (2/2)
 File Access times
 Last modified time, Last access time, Last modification time of Inode
 Number of links to the file
 Represents no. of names the file has in directory hierarchy.
 Table of contents for the disk address of data in a file
 Kernel saves the data in discontiguous disk blocks
 The Inodes identifies the disk blocks that contain file’s data.
 File Size
 The inode does not specify the path name(s) that access the
file.

7. 7
Disk Inodes - Sample
 Distinction between writing the contents of an inode to disk
and writing the contents of a file to disk

8. 8
in-core inode
 Contents in addition to the fields of the disk inode
 Status of the in-core inode
 the inode is locked,
 a process is waiting for the inode to become unlocked,
 the in-core representation of the inode differs from the disk copy as a
result of a change to the data in the inode,
 the in-core representation of the file differs from the disk copy as a result
of a change to the file data,
 The file is a mount point

9. 9
in-core inode
 Logical Device Number of the file system that contains file.
 Inode Number
 Pointers to other in-core inodes
 Reference Count
 Number of instances of the active file..
 The kernel links inodes on hash queues and on a free list in
the same way that it links buffers on buffer hash queues and
on the buffer free list.
 A hash queue is identified according to the inode's logical
device number and inode number.

10. 10
Difference between in-core inode
and buffer header
 In-core inode has In-core reference count.
 If count=0, Inode in inactive. (kernel can reallocate in-core
inode)
 If count = or >= 1, Inode is Active.
 Buffer header don’t have reference count.
 Buffer is locked when it is allocated.
 Buffer is unlocked; it will be in free list.

11. 11
Accessing Inodes (IGET Algo.)
(1/4)
 Kernel
 Identifies particular inodes by their file system and inode number

Allocates in-core inodes
 Using algorithm iget
 Allocates an in-core copy of an inode
 Map the device number and inode number into a hash queue
 Search the queue for the inode
 If not find the inode, allocates one from the free list and locks it
 Read the disk copy of the newly accessed inode into the in-core copy

12. 12

13. 13
Accessing Inodes (2/4)
 Computation of logical disk block
block num = ((inode number –1)/number of
inodes per block) + start block of inode list
Where,
Block num=Disk block number
Start block of inode list=Beginning of Inode List.
Inode number= Inode number to be searched

14. 14
Accessing Inodes (3/4)
 Read the block using the algorithm bread
 Computation of the byte offset of the inode in the
block
((inode number –1) modulo (number of inodes per
block)) * size of disk inode
Where,
Inode number= Inode number.
Size of disk inode = Size of disk inode

15. Example
 Find the block number and block offset of inode in
the block for following inode number:
 6895
 4258
 Assumptions:
1. Start Block of Inode List: 200
2. Disk Block size: 1024 bytes
3. Size of Disk Inode: 64 bytes
15

16. 16

17. 17
Accessing Inodes (4/4)
 Hold lock during execution of a system call for
possibly consistency
 Release it at the end of system call
 The lock is free between system calls
 To allow processes to share simultaneous access to a file
 The reference count remains set between system
calls
 To prevent the kernel from reallocating an active in-core
inode

18. 18
Release Inodes (IPUT Algo.)
 Using algorithm iput
 decrements in-core reference count
 Write the inode to disk

Reference count is 0

The in-core copy differs from the disk copy
 Add the inode on the free list of inodes

For caching

19. 19

20. 20
Structure of a Regular File
 Table of contents in an inode
 Location of a file’s data on disk
 A set of disk block #

Each block on a disk is addressable by number
 Not contiguous file allocation strategy

Why ?

When a file expand or contract…

Fragmentations occur

21. 21
Sample - Fragmentation
 File B was expanded
 Garbage collection – too high cost
File A File B File C
40 50 60 70
…. …….
File A Free File C
40 50 60 70
…. …….File B
85

22. 22
Structure of a Regular File –
UNIX System V
13 entries in the inode table of contents
 10 Direct, 1 Single Indirect, 1 Double Indirect, 1 Triple Indirect Block
 Assume
 a logical block = 1K bytes
 a block number is addressable by a 32 bit (4 bytes) integer
 a block can hold up to 256 block numbers
 Byte Capacity of a File
10 direct blocks with 1K bytes each= 10K bytes
1 indirect block with 256 direct blocks= 1K*256 = 256K bytes
1 double indirect block with 256 indirect blocks = 256K*256= 64M bytes
1 triple indirect block with 256 double indirect blocks=64M*256= 16G bytes

23. 23
Direct and Indirect Blocks in
Inode – UNIX System V
Inode Data Blocks
direct0
direct1
direct2
direct3
direct4
direct5
direct6
direct7
direct8
direct9
single indirect
double indirect
triple indirect
………..

24. 24
Block Layout of a Sample File
and its Inode
4096
228
45423
0
0
11111
0
101
367
0
428
9156
824
331
3333
9156
331
33330 75
367
1 disk block = 1024 bytes
byte offset 9000, byte offset 350,000
816th
808th

25. 25
Structure of a Regular File
 Processes
 access data in a file by byte offset
 view a file as a stream of bytes
 The kernel
 accesses the inode
 converts the logical file block into the appropriate disk block

26. 26

27. 27
Directories (1/2)
 A directory is a file
 Its data is a sequence of entries
 Contents of each entries

an inode number and the name of a file
 Path name is a null terminated character string divided by
slash (“/”)
 UNIX System V
 Maximum of component name : 14 characters
 Inode # : 2 bytes
 Size of a directory : 16 bytes

28. 28
Directories (2/2)
Directory
layout
for
/etc

29. 29
Conversion of a Path Name
to an Inode
 The initial access to a file is by its path name
 Open, chdir, link system calls
 The kernel works internally with inodes rather than
with path name
 Converting the path names to inodes
 Using algorithm namei

parse the path name one component at a time

convert each component into an inode

finally return the inode of the input path name

30. 30

31. 31
Sample-namei(/etc/passwd)
Encounters “/” and gets the system root inode
Current working inode = root
Permission check
Search root for a file – “etc”
Access data in the root directory block by block
Search each block one entry-”etc”
Finding
Release the inode for root(iput)
Allocate the inode for etc(iget) by inode # found
Permission check for “etc”
Search “etc” block by block for a directory struct. entry for “passwd”
Finding
Relase the the inode for “etc”
Allocate the inode for “passwd”
Return that inode

32. 32
Super Block
 Contents
 the size of the file system.
 the number of free blocks in the file system.
 a list of free blocks available on the file system.
 the index of the next free block in the free block list,
 the size of the inode list.
 the number of free inodes in the file system.
 a list of free inodes in the file system.
 the index of the next free inode in the free inode list.
 lock fields for the free block and free inode lists.
 a flag indicating that the super block has been modified.

33. 33
Inode Assignment to A New File
(1/4)
 a known inode
 Algorithm iget : to allocate
 Algorithm namei : to determine inode #
 Algorithm ialloc
 To assign a disk inode to a newly created file
 Super block contains an array
 To improve performance of searching a free inode
 To cache the numbers of free inodes

34. 34

35. 35
Inode Assignment to a New File
(2/4)
free inodes 83 48 empty
18 19 20 array1
Super Block Free Inode List
index
free inodes 83 empty
18 19 20 array2
Super Block Free Inode List
index
Assigning Free Inode from Middle of List

36. 36
Inode Assignment to a New File
(3/4)
index
Assigning Free Inode – Super Block List Empty
470 empty
array1
Super Block Free Inode List
index
0
535 free inodes 476 475 471
array2Super Block Free Inode List
0
48 49 50
remembered
inode

37. 37

38. 38
Inode Assignment to a New File
(4/4)
535 476 475 471
free inodes
remembered inode
Original Super Block List of Free Inodes
index
Free Inode 499
499 476 475 471
free inodes
remembered inode index
Free Inode 601
499 476 475 471
free inodes
remembered inode index

39. 39
Race Condition
 A Race Condition Scenario in Assigning Inodes
 three processes A, B, and C are acting in time sequence
1. The kernel, acting on behalf of process A, assigns
inode I but goes to sleep before it copies the disk
inode into the in-core copy.
2. While process A is asleep, process B attempts to
assign a new inode but free inode list is empty, and
attempts assign free inode at an inode number lower
than that of the inode that A is assigning.
3. Process C later requests an inode and happens to pick
inode I from the super block free list

40. 40
Race Condition in Assigning
Inodes (1/2)

41. 41
Race Condition in Assigning Inodes
(2/2)
time
…………………………………………….I
Empty
………………………….
free inodes
free inodes
free inodes L
J I
J I K
…………………………………………………………………...
………………………………………………….…………….
……………………………………………………...
(a)
(b)
(c)
(d)
(e)

42. 42
Allocation of Disk Blocks
 When a process writes data to a file, the kernel
must allocate disk blocks
 An array in the file system super block
 To cache the numbers of free disk block in the file
system
 Mkfs

Organize the data blocks of a file system in a linked list
 Each link is a disk block

The block contains an array of free disk block numbers

One array entry is the number of the next block of the linked
list

43. 43
Linked list of free disk block number
109 109 103 100 …………………………...
109
211 208 205 202 ………………… 112
211
310 307 304 301 ………………… 214
310
409 406 403 400 ………………… 313

44. 44

45. 45
Requesting and Freeing Disk
Blocks (1/2)
109 …………………………………………………………
211 208 205 202 …………………………….. 112
109 949 …………………………………………………..
211 208 205 202 ………………………………. 112
super block list
original configuration
109
109
After freeing block number 949

46. 46
Requesting and Freeing Disk
Blocks (2/2)
211 208 205 202 ……………………………… 112
344 341 338 335 ………………………………. 243
After assigning block number(109)
replenish (fill up again) super block free list
211
109 ………………………………………………………..
211 208 205 202 ……………………………….
112
109
After assigning block number(949)

47. 47
Other File Types
 Pipe
 fifo (first-in-first-out)
 its data is transient
 Once data is read from a pipe, it cannot be read again

The data is read in the order that it was written to the pipe, no deviation from
that order
 using only direct block
 Special File (including block device, character device)
 Specifying devices

The inode does not reference any data
 The inode contains the major and minor device number
 major number

a device type such as terminal or disk
 minor number

the unit number of the device