Esempi di File System: UNIXbotta/didattica/EsempiDiFileSystems.pdf · Esempi di File System: UNIX The UNIX file system supports two main objects: files and directories. Directories

1

Silberschatz, Galvin and Gagne 2002A.1Operating System Concepts

Esempi di File System: UNIX

� The UNIX file system supports two main objects: files anddirectories.

� Directories are just files with a special format, so therepresentation of a file is the basic UNIX concept.


Blocks and Fragments

� Most of the file system is taken up by data blocks.� 4.2BSD uses two block sized for files which have no

indirect blocks:✦ All the blocks of a file are of a large block size (such as 8K),

except the last.✦ The last block is an appropriate multiple of a smaller

fragment size (i.e., 1024) to fill out the file.✦ Thus, a file of size 18,000 bytes would have two 8K blocks

and one 2K fragment (which would not be filled completely).


Blocks and Fragments (Cont.)

� The block and fragment sizes are set during file-systemcreation according to the intended use of the file system:✦ If many small files are expected, the fragment size should

be small.✦ If repeated transfers of large files are expected, the basic

block size should be large.� The maximum block-to-fragment ratio is 8 : 1; the

minimum block size is 4K (typical choices are 4096 : 512and 8192 : 1024).

2


Inodes

� A file is represented by an inode — a record that storesinformation about a specific file on the disk.

� The inode also contains 15 pointer to the disk blockscontaining the file’s data contents.✦ First 12 point to direct blocks.✦ Next three point to indirect blocks

✔ First indirect block pointer is the address of a singleindirect block — an index block containing theaddresses of blocks that do contain data.

✔ Second is a double-indirect-block pointer, the address ofa block that contains the addresses of blocks thatcontain pointer to the actual data blocks.

✔ A triple indirect pointer is not needed; files with as manyas 232 bytes will use only double indirection.


Struttura di un I-node


Directories

� The inode type field distinguishes between plain files anddirectories.

� Directory entries are of variable length; each entrycontains first the length of the entry, then the file nameand the inode number.

� The user refers to a file by a path name,whereas the filesystem uses the inode as its definition of a file.✦ The kernel has to map the supplied user path name to an

inode✦ Directories are used for this mapping.

3


Directories (Cont.)

� First determine the starting directory:✦ If the first character is “/”, the starting directory is the root

directory.✦ For any other starting character, the starting directory is the

current directory.� The search process continues until the end of the path

name is reached and the desired inode is returned.� Once the inode is found, a file structure is allocated to

point to the inode.� 4.3BSD improved file system performance by adding a

directory name cache to hold recent directory-to-inodetranslations.


Mapping of a File Descriptor to an Inode

� System calls that refer to open files indicate the file ispassing a file descriptor as an argument.

� The file descriptor is used by the kernel to index a table ofopen files for the current process.

� Each entry of the table contains a pointer to a filestructure.

� This file structure in turn points to the inode.� Since the open file table has a fixed length which is only

setable at boot time, there is a fixed limit on the numberof concurrently open files in a system.


File-System Control Blocks

4


Disk Structures

� The one file system that a user ordinarily sees mayactually consist of several physical file systems, each ona different device.

� Partitioning a physical device into multiple file systemshas several benefits.✦ Different file systems can support different uses.✦ Reliability is improved✦ Can improve efficiency by varying file-system parameters.✦ Prevents one program form using all available space for a

large file.✦ Speeds up searches on backup tapes and restoring

partitions from tape.


Disk Structures (Cont.)

� The root file system is always available on a drive.

� Other file systems may be mounted — i.e., integrated intothe directory hierarchy of the root file system.

� The following figure illustrates how a directory structure ispartitioned into file systems, which are mapped ontological devices, which are partitions of physical devices.


Mapping File System to Physical Devices

5


Implementations

� The user interface to the file system is simple and welldefined, allowing the implementation of the file system itselfto be changed without significant effect on the user.

� For Version 7, the size of inodes doubled, the maximum fileand file system sized increased, and the details of free-listhandling and superblock information changed.

� In 4.0BSD, the size of blocks used in the file system wasincreased form 512 bytes to 1024 bytes — increasedinternal fragmentation, but doubled throughput.

� 4.2BSD added the Berkeley Fast File System, whichincreased speed, and included new features.✦ New directory system calls✦ truncate calls✦ Fast File System found in most implementations of UNIX.


Layout and Allocation Policy

� The kernel uses a <logical device number, inodenumber> pair to identify a file.✦ The logical device number defines the file system involved.✦ The inodes in the file system are numbered in sequence.

� 4.3BSD introduced the cylinder group — allowslocalization of the blocks in a file.✦ Each cylinder group occupies one or more consecutive

cylinders of the disk, so that disk accesses within thecylinder group require minimal disk head movement.

✦ Every cylinder group has a superblock, a cylinder block, anarray of inodes, and some data blocks.


4.3BSD Cylinder Group

6


File Systems

� To the user, Linux’s file system appears as a hierarchicaldirectory tree obeying UNIX semantics.

� Internally, the kernel hides implementation details andmanages the multiple different file systems via anabstraction layer, that is, the virtual file system (VFS).

� The Linux VFS is designed around object-orientedprinciples and is composed of two components:✦ A set of definitions that define what a file object is allowed to

look like✔ The inode-object and the file-object structures represent

individual files✔ the file system object represents an entire file system

✦ A layer of software to manipulate those objects.


The Linux Ext2fs File System

� Ext2fs uses a mechanism similar to that of BSD FastFile System (ffs) for locating data blocks belonging to aspecific file.

� The main differences between ext2fs and ffs concerntheir disk allocation policies.✦ In ffs, the disk is allocated to files in blocks of 8Kb, with

blocks being subdivided into fragments of 1Kb to storesmall files or partially filled blocks at the end of a file.

✦ Ext2fs does not use fragments; it performs its allocationsin smaller units. The default block size on ext2fs is 1Kb,although 2Kb and 4Kb blocks are also supported.

✦ Ext2fs uses allocation policies designed to place logicallyadjacent blocks of a file into physically adjacent blocks ondisk, so that it can submit an I/O request for several diskblocks as a single operation.


Ext2fs Block-Allocation Policies

7


File System di Windows 2000

� La struttura fondamentale del file system di Windows2000 (NTFS) è il volume.✦ Creato dal programma di utilità disk administrator.✦ Si basa su una partizione logica del disco.✦ Può occupare una porzione di un disco, un intero disco, o

spaziare su più dischi.� Tutti i metadata, cioè le informazioni sul volume, sono

memorizzate in un normale file.� NTFS usa clusters (gruppi) come unità di allocazione del

disco.✦ Un cluster è un numero di settori del disco che è una

potenza di 2.✦ Siccome la dimensione del cluster è più piccola di quella del

file system FAT a 16 bit, la frammentazione interna èridotta.


File System — Struttura interna

� NTFS usa logical cluster numbers (LCNs) come indirizzidi disco.

� Un file in NTFS non è una semplice sequenza di byte,come in MS-DOS o UNIX, ma è un oggetto strutturatoche consiste di attributi.

� Ogni file in NTFS è descritto da uno o più record in unvettore memorizzato in un file speciale chiamato MasterFile Table (MFT).

� Ogni file in un volume NTFS ha un ID univoco chiamatofile reference.✦ Valore a 64-bit suddiviso in 48-bit per numero file e 16-bit

per numero di sequenza.✦ Può essere usato per eseguire controlli di consistenza.

� Lo spazio dei nomi in NTFS è organizzato come unagerarchia di directory.


File System — Ripristino

� Tutte le modifiche alle strutture del file system sonoeseguite come transazioni che vengono tracciate.✦ Prima di alterare una struttura dati, la transazione scrive un

record di log che contiene informazioni sull’operazione dafare.

✦ Dopo che la struttura dati è stata modificata, viene scritto unrecord di commit per dire che la transazione ha avutosuccesso.

✦ Dopo un crash del sistema, le strutture dati del file systempossono essere ripristinate ad uno stato consistenteripetendo le transazioni nel log.

8


File System — Ripristino (Cont.)

� Questo schema non garantisce che tutti i dati dell’utentesiano ripristinati dopo un crash, ma solo le strutture delfile system sono riportate ad uno stato consistente cheriflette una situazione stabile prima del crash.

� Il file di log è memorizato nel terzo file di metadatiall’inizio del volume.

� La funzionalità di log è fornita dal servizio di Windows2000 log file service.


File System — Sicurezza

� La sicurezza su un volume NTFS è derivata dal modelload oggetti di Windows 2000.

� Ogni oggetto di tipo file ha un attributo di sicurezzamemorizzato nella Master File Table.

� Questo attributo contiene le informazioni di accesso delproprietario del fule, e una lista di accesso che elenca iprivilegi di accessi che sono permessi a ciascun utenteche accede al file.


Gestione dei Volumi e Tolleranza ai guasti

FtDisk, è il driver di disco tollerante ai guasti di 2000, efornisce diversi modi per combinare i dischi in un unicovolume logico.

� Un volume set è una concatenzaione logica di più dischia formare un grande volume logico.

� Uno stripe set organizza più partizioni fisiche in manieracircolare (anche chiamato RAID di livello 0, o “diskstriping”).✦ Variazione: stripe set with parity, o RAID livello 5.

� Disk mirroring, o RAID livello 1, è uno schema robustoche usa un mirror set : due partizioni di egual misura sidue dischi diversi con identici contenuti.

� Per gestire malfunzionamenti, FtDisk, usa una tecnicahardware chiamata sector sparing e NTFS usa unattecnica software chiamata cluster remapping.

9


Volume Set On Two Drives


Stripe Set on Two Drives


Stripe Set With Parity on Three Drives

10


Mirror Set on Two Drives


File System — Compression

� To compress a file, NTFS divides the file’s data intocompression units, which are blocks of 16 contiguousclusters.

� For sparse files, NTFS uses another technique to savespace.✦ Clusters that contain all zeros are not actually allocated or

stored on disk.✦ Instead, gaps are left in the sequence of virtual cluster

numbers stored in the MFT entry for the file.✦ When reading a file, if a gap in the virtual cluster numbers is

found, NTFS just zero-fills that portion of the caller’s buffer.


File System — Reparse Points

� A reparse point returns an error code when accessed.The reparse data tells the I/O manager what to do next.

� Reparse points can be used to provide the functionality ofUNIX mounts

� Reparse points can also be used to access files that havebeen moved to offline storage.

Documents

Esempi di File System: UNIXbotta/didattica/EsempiDiFileSystems.pdf · Esempi di File System: UNIX The UNIX file system supports two main objects: files and directories. Directories