Lecture 2:

IO Management

Presented by Dr. Huy Nguyen

Operating Systems
 Lecture 2A:
Mass Storage Systems

Presented by Dr. Huy Nguyen

Operating Systems
 Lecture 2A: Mass Storage Systems

 Overview of Mass Storage Structure

 Disk Structure
 Disk Attachment
 Disk Scheduling
 Disk Management
 Swap Space Management
 RAID Structure
 Stable Storage Implementation

Operating Systems 2.3

 Objectives

 To describe the physical structure of secondary storage devices and its

effects on the uses of the devices
 To explain the performance characteristics of mass storage devices
 To evaluate disk scheduling algorithms
 To discuss operating system services provided for mass storage,
including RAID

Operating Systems 2.4

 Overview of Mass Storage Structure

 Magnetic disks provide bulk of secondary storage of modern

computers
 Drives rotate at 60 to 250 times per second
 Transfer rate is rate at which data flow between drive and
computer
 Positioning time (random access time) is time to move disk arm
to desired cylinder (seek time) and time for desired sector to
rotate under the disk head (rotational latency)
 Head crash results from disk head making contact with the disk
surface - That’s bad
 Disks can be removable
 Drive attached to computer via I/O bus
 Busses vary, including EIDE, ATA, SATA, USB, Fibre Channel,
SCSI, SAS, Firewire
 Host controller in computer uses bus to talk to disk controller
built into drive or storage array

Operating Systems 2.5

 Moving Head Disk Mechanism

Operating Systems 2.6

 Hard Disks

 Platters range from 0.85” to 14” (historically)

 Commonly 3.5”, 2.5”, and 1.8”
 Range from 30GB to 3TB per drive
 Performance
 Transfer Rate - theoretical - 6 Gb/sec
 Effective Transfer Rate - real - 1Gb/sec
 Seek time from 3ms to 12ms - 9ms
common for desktop drives
 Average seek time measured or
calculated based on 1/3 of tracks
 Latency based on spindle speed
(From Wikipedia)
 1 / (RPM / 60) = 60 / RPM
 Average latency = ½ latency

Operating Systems 2.7

 Hard Disk Performance

 Access Latency = Average access time = average seek time +

average latency
 For fastest disk 3ms + 2ms = 5ms
 For slow disk 9ms + 5.56ms = 14.56ms
 Average I/O time = average access time + (amount to transfer /
transfer rate) + controller overhead
 For example to transfer a 4KB block on a 7200 RPM disk with a 5ms
average seek time, 1Gb/sec transfer rate with a 0.1ms controller
overhead =
 5ms + 4.17ms + 0.1ms + transfer time =
 Transfer time = 4KB / 1Gb/s * 8Gb / GB * 1GB / 10242KB = 32 /
(10242) = 0.031 ms
 Average I/O time for 4KB block = 9.27ms + 0.031ms = 9.301ms

Operating Systems 2.8

 First Commercial Disk Drive

In 1956, IBM RAMDAC

computer included the IBM
Model 350 disk storage
system
5M (7 bit) characters
50 x 24” platters
Access time  1 second

Operating Systems 2.9

 Solid State Disks

 Nonvolatile memory used like a hard drive

 Many technology variations
 Can be more reliable than HDDs
 More expensive per MB
 Maybe have shorter life span
 Less capacity
 But much faster
 Busses can be too slow  connect directly to PCI for example
 No moving parts, so no seek time or rotational latency

Operating Systems 2.10

 Magnetic Tape

 Was early secondary storage medium

 Evolved from open spools to cartridges
 Relatively permanent and holds large quantities of data
 Access time slow
 Random access ~1000 times slower than disk
 Mainly used for backup, storage of infrequently used data, transfer
medium between systems
 Kept in spool and wound or rewound past read-write head
 Once data under head, transfer rates comparable to disk
 140 MB/sec and greater
 200 GB to 1.5 TB typical storage
 Common technologies are LTO-{3,4,5} and T10000

Operating Systems 2.11

 Disk Structure

 Disk drives are addressed as large 1-dimensional arrays of logical

blocks, where the logical block is the smallest unit of transfer
 Low level formatting creates logical blocks on physical media
 The 1-dimensional array of logical blocks is mapped into the sectors of
the disk sequentially
 Sector 0 is the first sector of the first track on the outermost
cylinder
 Mapping proceeds in order through that track, then the rest of the
tracks in that cylinder, and then through the rest of the cylinders
from outermost to innermost
 Logical to physical address should be easy
 Except for bad sectors
 Non constant # of sectors per track via constant angular velocity

Operating Systems 2.12

 Disk Attachment

 Host attached storage accessed through I/O ports talking to I/O

busses
 SCSI itself is a bus, up to 16 devices on one cable, SCSI initiator
requests operation and SCSI targets perform tasks
 Each target can have up to 8 logical units (disks attached to
device controller)
 FC is high speed serial architecture
 Can be switched fabric with 24 bit address space - the basis of
storage area networks (SANs) in which many hosts attach to
many storage units
 I/O directed to bus ID, device ID, logical unit (LUN)

Operating Systems 2.13

 Storage Array

 Can just attach disks, or arrays of disks

 Storage Array has controller(s), provides features to attached host(s)
 Ports to connect hosts to array
 Memory, controlling software (sometimes NVRAM, etc)
 A few to thousands of disks
 RAID, hot spares, hot swap (discussed later)
 Shared storage  more efficiency
 Features found in some file systems
 Snaphots, clones, thin provisioning, replication, deduplication,
etc

Operating Systems 2.14

 Storage Area Network

 Common in large storage environments

 Multiple hosts attached to multiple storage arrays - flexible

Operating Systems 2.15

 Storage Area Network (cont.)

 SAN is one or more storage arrays

 Connected to one or more Fibre Channel switches
 Hosts also attach to the switches
 Storage made available via LUN Masking from specific arrays to
specific servers
 Easy to add or remove storage, add new host and allocate it storage
 Over low latency Fibre Channel fabric
 Why have separate storage networks and communications networks?
 Consider iSCSI, FCOE

Operating Systems 2.16

 Network Attached Storage

 Network attached storage (NAS) is storage made available over a

network rather than over a local connection (such as a bus)
 Remotely attaching to file systems
 NFS and CIFS are common protocols
 Implemented via remote procedure calls (RPCs) between host and
storage over typically TCP or UDP on IP network
 iSCSI protocol uses IP network to carry the SCSI protocol
 Remotely attaching to devices (blocks)

Operating Systems 2.17

 Disk Scheduling

 The operating system is responsible for using hardware efficiently - for

the disk drives, this means having a fast access time and disk
bandwidth
 Minimize seek time
 Seek time  seek distance
 Disk bandwidth is the total number of bytes transferred, divided by the
total time between the first request for service and the completion of
the last transfer

Operating Systems 2.18

 Disk Scheduling (cont.)

 There are many sources of disk I/O request

 OS
 System processes
 Users processes
 I/O request includes input or output mode, disk address, memory
address, number of sectors to transfer
 OS maintains queue of requests, per disk or device
 Idle disk can immediately work on I/O request, busy disk means work
must queue
 Optimization algorithms only make sense when a queue exists

Operating Systems 2.19

 Disk Scheduling (cont.)

 Note that drive controllers have small buffers and can manage a queue
of I/O requests (of varying “depth”)
 Several algorithms exist to schedule the servicing of disk I/O requests
 The analysis is true for one or many platters
 We illustrate scheduling algorithms with a request queue (0 - 199)
98, 183, 37, 122, 14, 124, 65, 67
Head pointer 53

Operating Systems 2.20

 FCFS

 Illustration shows total head movement of 640 cylinders

Operating Systems 2.21

 SSTF

 Shortest Seek Time First selects the request with the minimum seek
time from the current head position
 SSTF scheduling is a form of SJF scheduling; may cause starvation of
some requests
 Illustration shows total head movement of 236 cylinders

Operating Systems 2.22

 SCAN

 The disk arm starts at one end of the disk, and moves toward the
other end, servicing requests until it gets to the other end of the disk,
where the head movement is reversed and servicing continues
 SCAN algorithm Sometimes called the elevator algorithm
 Illustration shows total head movement of 208 cylinders
 But note that if requests are uniformly dense, largest density at other
end of disk and those wait the longest

Operating Systems 2.23

 SCAN (cont.)

Operating Systems 2.24

 C-SCAN

 Provides a more uniform wait time than SCAN

 The head moves from one end of the disk to the other, servicing
requests as it goes
 When it reaches the other end, however, it immediately returns to
the beginning of the disk, without servicing any requests on the
return trip
 Treats the cylinders as a circular list that wraps around from the last
cylinder to the first one
 Total number of cylinders?

Operating Systems 2.25

 C-SCAN (cont.)

Operating Systems 2.26

 C-LOOK

 LOOK a version of SCAN, C-LOOK a version of C-SCAN

 Arm only goes as far as the last request in each direction, then
reverses direction immediately, without first going all the way to the
end of the disk
 Total number of cylinders?

Operating Systems 2.27

 C-LOOK (cont.)

Operating Systems 2.28

 Selecting a Disk Scheduling Algorithm

 SSTF is common and has a natural appeal

 SCAN and C-SCAN perform better for systems that place a heavy
load on the disk
 Less starvation
 Performance depends on the number and types of requests
 Requests for disk service can be influenced by the file allocation
method
 And metadata layout
 The disk scheduling algorithm should be written as a separate module
of the operating system, allowing it to be replaced with a different
algorithm if necessary
 Either SSTF or LOOK is a reasonable choice for the default algorithm
 What about rotational latency?
 Difficult for OS to calculate
 How does disk based queueing effect OS queue ordering efforts?

Operating Systems 2.29

 Disk Management

 Low level formatting, or physical formatting - dividing a disk into

sectors that the disk controller can read and write
 Each sector can hold header information, plus data, plus error
correction code (ECC)
 Usually 512 bytes of data but can be selectable
 To use a disk to hold files, the operating system still needs to record its
own data structures on the disk
 Partition the disk into one or more groups of cylinders, each
treated as a logical disk
 Logical formatting or “making a file system”
 To increase efficiency most file systems group blocks into clusters
 Disk I/O done in blocks
 File I/O done in clusters

Operating Systems 2.30

 Disk Management (cont.)

 Raw disk access for apps that want to do their own block management,
keep OS out of the way (databases for example)
 Boot block initializes system
 The bootstrap is stored in ROM
 Bootstrap loader program stored in boot blocks of boot partition
 Methods such as sector sparing used to handle bad blocks

Operating Systems 2.31

 Booting from a Disk in Windows

Operating Systems 2.32

 Swap Space Management

 Swap space - virtual memory uses disk space as an extension of main memory
 Less common now due to memory capacity increases
 Swap space can be carved out of the normal file system, or, more commonly, it
can be in a separate disk partition (raw)
 Swap space management
 4.3BSD allocates swap space when process starts; holds text segment (the
program) and data segment
 Kernel uses swap maps to track swap space use
 Solaris 2 allocates swap space only when a dirty page is forced out of
physical memory, not when the virtual memory page is first created
 File data written to swap space until write to file system requested
 Other dirty pages go to swap space due to no other home
 Text segment pages thrown out and reread from the file system as needed
 What if a system runs out of swap space?
 Some systems allow multiple swap spaces

Operating Systems 2.33

 Data Structures for Swapping on Linux Systems

Operating Systems 2.34

 RAID Structure

 RAID - redundant array of inexpensive disks

 multiple disk drives provides reliability via redundancy
 Increases the mean time to failure
 Mean time to repair - exposure time when another failure could cause
data loss
 Mean time to data loss based on above factors
 If mirrored disks fail independently, consider disk with 100,000 hours
mean time to failure and 10 hours mean time to repair
 Mean time to data loss is 100,0002 / (2 ∗ 10) = 500 ∗ 106 hours, or
57,000 years!
 Frequently combined with NVRAM to improve write performance
 Several improvements in disk use techniques involve the use of
multiple disks working cooperatively

Operating Systems 2.35

 RAID Structure (cont.)

 Disk striping uses a group of disks as one storage unit

 RAID is arranged into six different levels
 RAID schemes improve performance and improve the reliability of the
storage system by storing redundant data
 Mirroring or shadowing (RAID 1) keeps duplicate of each disk
 Striped mirrors (RAID 1+0) or mirrored stripes (RAID 0+1)
provides high performance and high reliability
 Block interleaved parity (RAID 4, 5, 6) uses much less
redundancy
 RAID within a storage array can still fail if the array fails, so automatic
replication of the data between arrays is common
 Frequently, a small number of hot spare disks are left unallocated,
automatically replacing a failed disk and having data rebuilt onto them

Operating Systems 2.36

 RAID Levels

Operating Systems 2.37

 RAID (0 + 1) and (1 + 0)

Operating Systems 2.38

 Other Features

 Regardless of where RAID implemented, other useful features can be

added
 Snapshot is a view of file system before a set of changes take place
(i.e. at a point in time)
 More in Lecture 2C
 Replication is automatic duplication of writes between separate sites
 For redundancy and disaster recovery
 Can be synchronous or asynchronous
 Hot spare disk is unused, automatically used by RAID production if a
disk fails to replace the failed disk and rebuild the RAID set if possible
 Decreases mean time to repair

Operating Systems 2.39

 Extensions

 RAID alone does not prevent or detect data corruption or other errors,
just disk failures
 Solaris ZFS adds checksums of all data and metadata
 Checksums kept with pointer to object, to detect if object is the right
one and whether it changed
 Can detect and correct data and metadata corruption
 ZFS also removes volumes, partitions
 Disks allocated in pools
 Filesystems with a pool share that pool, use and release space like
malloc() and free() memory allocate / release calls

Operating Systems 2.40

 ZFS Checksums All Metadata & Data

Operating Systems 2.41

 Traditional & Pooled Storage

Operating Systems 2.42

 Stable Storage Implementation

 Write ahead log scheme requires stable storage

 Stable storage means data is never lost (due to failure, etc)
 To implement stable storage
 Replicate information on more than one nonvolatile storage media
with independent failure modes
 Update information in a controlled manner to ensure that we can
recover the stable data after any failure during data transfer or
recovery
 Disk write has 1 of 3 outcomes
1. Successful completion - the data were written correctly on disk
2. Partial failure - a failure occurred in the midst of transfer, so only
some of the sectors were written with the new data, and the
sector being written during the failure may have been corrupted
3. Total failure - the failure occurred before the disk write started,
so the previous data values on the disk remain intact

Operating Systems 2.43

 Stable Storage Implementation (cont.)

 If failure occurs during block write, recovery procedure restores block

to consistent state
 System maintains 2 physical blocks per logical block and does the
following
1. Write to 1st physical
2. When successful, write to 2nd physical
3. Declare complete only after second write completes
successfully
 Systems frequently use NVRAM as one physical to accelerate

Operating Systems 2.44

 End of Lecture 2A

Operating Systems
 Lecture 2B:
File System Interface

Presented by Dr. Huy Nguyen

Operating Systems
 Lecture 2B: File System Interface

 File Concept
 Access Methods
 Disk and Directory Structure
 File System Mounting
 File Sharing
 Protection

Operating Systems 2.47

 Objectives

 To explain the function of file systems

 To describe the interfaces to file systems
 To discuss file system design tradeoffs, including access methods, file
sharing, file locking, and directory structures
 To explore file system protection

Operating Systems 2.48

 File Concept

 Contiguous logical address space

 Types
 Data
 numeric
 character
 binary
 Program
 Contents defined by file’s creator
 Many types
 Consider text file, source file, executable file

Operating Systems 2.49

 File Attributes

 Name - only information kept in human readable form

 Identifier - unique tag (number) identifies file within file system
 Type - needed for systems that support different types
 Location - pointer to file location on device
 Size - current file size
 Protection - controls who can do reading, writing, executing
 Time, date, and user identification - data for protection, security, and
usage monitoring
 Information about files are kept in the directory structure, which is
maintained on the disk
 Many variations, including extended file attributes such as file
checksum
 Information kept in the directory structure

Operating Systems 2.50

 File Info Window on Mac OS X

Operating Systems 2.51

 File Operations

 File is an abstract data type

 Create
 Write - at write pointer location
 Read - at read pointer location
 Reposition within file - seek
 Delete
 Truncate
 Open(Fi) - search the directory structure on disk for entry Fi, and move
the content of entry to memory
 Close (Fi) - move the content of entry Fi in memory to directory
structure on disk

Operating Systems 2.52

 Open Files

 Several pieces of data are needed to manage open files

 Open file table: tracks open files
 File pointer: pointer to last read/write location, per process that has
the file open
 File open count: counter of number of times a file is open - to
allow removal of data from open file table when last processes
closes it
 Disk location of the file: cache of data access information
 Access rights: per process access mode information

Operating Systems 2.53

 Open File Locking

 Provided by some operating systems and file systems

 Similar to reader-writer locks
 Shared lock similar to reader lock - several processes can acquire
concurrently
 Exclusive lock similar to writer lock
 Mediates access to a file
 Mandatory or advisory
 Mandatory - access is denied depending on locks held and
requested
 Advisory - processes can find status of locks and decide what to
do

Operating Systems 2.54

 File Locking Example - Java API
import java.io.*;
import java.nio.channels.*;
public class LockingExample {
public static final boolean EXCLUSIVE = false;
public static final boolean SHARED = true;
public static void main(String arsg[]) throws IOException {
FileLock sharedLock = null;
FileLock exclusiveLock = null;
try {
RandomAccessFile raf = new RandomAccessFile("file.txt", "rw");
// get the channel for the file
FileChannel ch = raf.getChannel();
// this locks the first half of the file - exclusive
exclusiveLock = ch.lock(0, raf.length()/2, EXCLUSIVE);
/** Now modify the data . . . */
// release the lock
exclusiveLock.release();

Operating Systems 2.55

 File Locking Example - Java API (cont.)

// this locks the second half of the file - shared

sharedLock = ch.lock(raf.length()/2+1, raf.length(), SHARED);
/** Now read the data . . . */
// release the lock
sharedLock.release();
} catch (java.io.IOException ioe) {
System.err.println(ioe);
} finally {
if (exclusiveLock != null)
exclusiveLock.release();
if (sharedLock != null)
sharedLock.release();
}
}
}

Operating Systems 2.56

 File Types - Name, Extension

Operating Systems 2.57

 File Structure

 None sequence of words, bytes

 Simple record structure
 Lines
 Fixed length
 Variable length
 Complex Structures
 Formatted document
 Relocatable load file
 Can simulate last two with first method by inserting appropriate control
characters
 Who decides
 Operating system
 Program

Operating Systems 2.58

 Sequential Access File

Operating Systems 2.59

 Access Methods
 Sequential Access
read next
write next
reset
no read after last write
(rewrite)
 Direct Access - file is fixed length logical records
read n
write n
position to n
read next
write next
rewrite n
n = relative block number
 Relative block numbers allow OS to decide where file should be placed
 See allocation problem in Lecture 2C
Operating Systems 2.60
 Simulation of Sequential Access on Direct Access File

Operating Systems 2.61

 Other Access Methods

 Can be built on top of base methods

 General involve creation of an index for the file
 Keep index in memory for fast determination of location of data to be
operated on (consider UPC code plus record of data about that item)
 If too large, index (in memory) of the index (on disk)
 IBM indexed sequential access method (ISAM)
 Small master index, points to disk blocks of secondary index
 File kept sorted on a defined key
 All done by the OS
 VMS operating system provides index and relative files as another
example (see next slide)

Operating Systems 2.62

 Example of Index & Relative Files

Operating Systems 2.63

 Directory Structure

 A collection of nodes containing information about all files

Directory

Files
F1 F2 F4
F3
Fn

 Both the directory structure and the files reside on disk

Operating Systems 2.64

 Disk Structure

 Disk can be subdivided into partitions

 Disks or partitions can be RAID protected against failure
 Disk or partition can be used raw - without a file system, or formatted
with a file system
 Partitions also known as minidisks, slices
 Entity containing file system known as a volume
 Each volume containing file system also tracks that file system’s info in
device directory or volume table of contents
 As well as general purpose file systems there are many special
purpose file systems, frequently all within the same operating system
or computer

Operating Systems 2.65

 Typical File System Organization

Operating Systems 2.66

 Types of File Systems

 We mostly talk of general purpose file systems

 But systems frequently have may file systems, some general and some
special purpose
 Consider Solaris has
 tmpfs - memory based volatile FS for fast, temporary I/O
 objfs - interface into kernel memory to get kernel symbols for
debugging
 ctfs - contract file system for managing daemons
 lofs - loopback file system allows one FS to be accessed in place of
another
 procfs - kernel interface to process structures
 ufs, zfs - general purpose file systems

Operating Systems 2.67

 Operations Performed on Directory

 Search for a file

 Create a file
 Delete a file
 List a directory
 Rename a file
 Traverse the file system

Operating Systems 2.68

 Directory Organization

The directory is organized logically to obtain

 Efficiency - locating a file quickly
 Naming - convenient to users
 Two users can have same name for different files
 The same file can have several different names
 Grouping - logical grouping of files by properties, (e.g., all Java
programs, all games…)

Operating Systems 2.69

 Single Level Directory

 A single directory for all users

 Naming problem
 Grouping problem

Operating Systems 2.70

 Two Level Directory

 Separate directory for each user

 Path name
 Can have the same file name for different user
 Efficient searching
 No grouping capability

Operating Systems 2.71

 Tree Structured Directories

Operating Systems 2.72

 Tree Structured Directories (cont.)

 Efficient searching
 Grouping Capability
 Current directory (working directory)
 cd /spell/mail/prog
 type list

Operating Systems 2.73

 Tree Structured Directories (cont.)

 Absolute or relative path name

 Creating a new file is done in current directory
 Delete a file
rm <file-name>
 Creating a new subdirectory is done in current directory
mkdir <dir-name>
 Example: if in current directory /mail
mkdir count

Deleting “mail”  deleting the entire subtree rooted by “mail”

Operating Systems 2.74

 Acyclic Graph Directories

 Have shared subdirectories and files

Operating Systems 2.75

 Acyclic Graph Directories (cont.)

 Two different names (aliasing)

 If dict deletes list  dangling pointer
Solutions
 Backpointers, so we can delete all pointers
 Variable size records a problem
 Backpointers using a daisy chain organization
 Entry hold count solution
 New directory entry type
 Link - another name (pointer) to an existing file
 Resolve the link - follow pointer to locate the file

Operating Systems 2.76

 General Graph Directory

Operating Systems 2.77

 General Graph Directory (cont.)

 How do we guarantee no cycles?

 Allow only links to file not subdirectories
 Garbage collection
 Every time a new link is added use a cycle detection algorithm to
determine whether it is OK

Operating Systems 2.78

 File System Mounting

 A file system must be mounted before it can be accessed

 A unmounted file system (i.e., Fig. (b)) is mounted at a mount point

Operating Systems 2.79

 Mount Point

Operating Systems 2.80

 File Sharing

 Sharing of files on multiuser systems is desirable

 Sharing may be done through a protection scheme
 On distributed systems, files may be shared across a network
 Network File System (NFS) is a common distributed file sharing
method
 If multiuser system
 User IDs identify users, allowing permissions and protections to be
per user
Group IDs allow users to be in groups, permitting group access
rights
 Owner of a file / directory
 Group of a file / directory

Operating Systems 2.81

 File Sharing - Remote File Systems

 Uses networking to allow file system access between systems

 Manually via programs like FTP
 Automatically, seamlessly using distributed file systems
 Semi automatically via the world wide web
 Client - server model allows clients to mount remote file systems from
servers
 Server can serve multiple clients
 Client and user on client identification is insecure or complicated
 NFS is standard UNIX client - server file sharing protocol
 CIFS is standard Windows protocol
 Standard operating system file calls are translated into remote calls
 Distributed Information Systems (distributed naming services) such
as LDAP, DNS, NIS, Active Directory implement unified access to
information needed for remote computing

Operating Systems 2.82

 File Sharing - Failure Modes

 All file systems have failure modes

 For example corruption of directory structures or other non user
data, called metadata
 Remote file systems add new failure modes, due to network failure,
server failure
 Recovery from failure can involve state information about status of
each remote request
 Stateless protocols such as NFS v3 include all information in each
request, allowing easy recovery but less security

Operating Systems 2.83

 File Sharing - Consistency Semantics

 Specify how multiple users are to access a shared file simultaneously

 Similar to Lecture 3C process synchronization algorithms
 Tend to be less complex due to disk I/O and network latency
(for remote file systems
 Andrew File System (AFS) implemented complex remote file
sharing semantics
 Unix file system (UFS) implements
 Writes to an open file visible immediately to other users of the
same open file
 Sharing file pointer to allow multiple users to read and write
concurrently
 AFS has session semantics
 Writes only visible to sessions starting after the file is closed

Operating Systems 2.84

 Protection

 File owner/creator should be able to control

 What can be done
 By whom
 Types of access
 Read
 Write
 Execute
 Append
 Delete
 List

Operating Systems 2.85

 Access Lists & Groups

 Mode of access: read, write, execute

 Three classes of users on Unix / Linux
RWX
a) owner access 7  111
RWX
b) group access 6  110
RWX
c) public access 1  001
 Ask manager to create a group (unique name), say G, and add some users
to the group
 For a particular file (say game) or subdirectory, define an appropriate
access
 Attach a group to a file
chgrp G game

Operating Systems 2.86

 Windows 7 Access Control List Management

Operating Systems 2.87

 Sample UNIX Directory Listing

Operating Systems 2.88

 End of Lecture 2B

Operating Systems
 Lecture 2C:
File System Implementation

Presented by Dr. Huy Nguyen

Operating Systems
 Lecture 2C: File System Implementation

 File System Structure

 File System Implementation
 Directory Implementation
 Allocation Methods
 Free Space Management
 Efficiency and Performance
 Recovery
 NFS
 Example: WAFL File System

Operating Systems 2.91

 Objectives

 To describe the details of implementing local file systems and directory

structures
 To describe the implementation of remote file systems
 To discuss block allocation and free block algorithms and trade offs

Operating Systems 2.92

 File System Structure

 File structure
 Logical storage unit
 Collection of related information
 File system resides on secondary storage (disks)
 Provided user interface to storage, mapping logical to physical
 Provides efficient and convenient access to disk by allowing data
to be stored, located retrieved easily
 Disk provides in place rewrite and random access
 I/O transfers performed in blocks of sectors (usually 512 bytes)
 File control block - storage structure consisting of information about
a file
 Device driver controls the physical device
 File system organized into layers

Operating Systems 2.93

 Layered File System

Operating Systems 2.94

 File System Layers

 Device drivers manage I/O devices at the I/O control layer

 Given commands like “read drive1, cylinder 72, track 2, sector 10,
into memory location 1060” outputs low level hardware specific
commands to hardware controller
 Basic file system given command like “retrieve block 123” translates
to device driver
 Also manages memory buffers and caches (allocation, freeing,
replacement)
 Buffers hold data in transit
 Caches hold frequently used data
 File organization module understands files, logical address, and
physical blocks
 Translates logical block # to physical block #
 Manages free space, disk allocation

Operating Systems 2.95

 File System Layers (cont.)

 Logical file system manages metadata information

 Translates file name into file number, file handle, location by
maintaining file control blocks (inodes in UNIX)
 Directory management
 Protection
 Layering useful for reducing complexity and redundancy, but adds
overhead and can decrease performanceTranslates file name into file
number, file handle, location by maintaining file control blocks (inodes
in UNIX)
 Logical layers can be implemented by any coding method
according to OS designer

Operating Systems 2.96

 File System Layers (cont.)

 Many file systems, sometimes many within an operating system

 Each with its own format (CD-ROM is ISO 9660; Unix has UFS,
FFS; Windows has FAT, FAT32, NTFS as well as floppy, CD,
DVD Blu-ray, Linux has more than 40 types, with extended file
system ext2 and ext3 leading; plus distributed file systems, etc.)
 New ones still arriving - ZFS, GoogleFS, Oracle ASM, FUSE

Operating Systems 2.97

 File System Implementation

 We have system calls at the API level, but how do we implement their
functions?
 On disk and in memory structures
 Boot control block contains info needed by system to boot OS from
that volume
 Needed if volume contains OS, usually first block of volume
 Volume control block (superblock, master file table) contains
volume details
 Total # of blocks, # of free blocks, block size, free block pointers
or array
 Directory structure organizes the files
 Names and inode numbers, master file table

Operating Systems 2.98

 File System Implementation (cont.)

 Per file File Control Block (FCB) contains many details about the file
 inode number, permissions, size, dates
 NFTS stores into in master file table using relational DB structures

Operating Systems 2.99

 In Memory File System Structures

 Mount table storing file system mounts, mount points, file system types
 The following figure illustrates the necessary file system structures
provided by the operating systems
 Figure (a) refers to opening a file
 Figure (b) refers to reading a file
 Plus buffers hold data blocks from secondary storage
 Open returns a file handle for subsequent use
 Data from read eventually copied to specified user process memory
address

Operating Systems 2.100

 In Memory File System Structures (cont.)

Operating Systems 2.101

 Partitions & Mounting

 Partition can be a volume containing a file system (“cooked”) or raw -

just a sequence of blocks with no file system
 Boot block can point to boot volume or boot loader set of blocks that
contain enough code to know how to load the kernel from the file
system
 Or a boot management program for multios booting
 Root partition contains the OS, other partitions can hold other Oses,
other file systems, or be raw
 Mounted at boot time
 Other partitions can mount automatically or manually
 At mount time, file system consistency checked
 Is all metadata correct?
 If not, fix it, try again
 If yes, add to mount table, allow access

Operating Systems 2.102

 Virtual File Systems

 Virtual File Systems (VFS) on Unix provide an object oriented way of

implementing file systems
 VFS allows the same system call interface (the API) to be used for
different types of file systems
 Separates file system generic operations from implementation
details
 Implementation can be one of many file systems types, or network
file system
 Implements vnodes which hold inodes or network file details
 Then dispatches operation to appropriate file system
implementation routines

Operating Systems 2.103

 Virtual File Systems (cont.)

 The API is to the VFS interface, rather than any specific type of file system

Operating Systems 2.104

 Virtual File System Implementation

 For example, Linux has four object types

 Inode, file, superblock, dentry
 VFS defines set of operations on the objects that must be
implemented
 Every object has a pointer to a function table
 Function table has addresses of routines to implement that
function on that object
 For example
• int open(. . .) - open a file
• int close(. . .) - close an already open file
• ssize t read(. . .) - read from a file
• ssize t write(. . .) - write to a file
• int mmap(. . .) - memory map a file

Operating Systems 2.105

 Directory Implementation

 Linear list of file names with pointer to the data blocks

 Simple to program
 Time consuming to execute
 Linear search time
 Could keep ordered alphabetically via linked list or use B+ tree
 Hash Table - linear list with hash data structure
 Decreases directory search time
 Collisions - situations where two file names hash to the same
location
 Only good if entries are fixed size, or use chained overflow method

Operating Systems 2.106

 Allocation Methods - Contiguous

 An allocation method refers to how disk blocks are allocated for files
 Contiguous allocation - each file occupies set of contiguous blocks
 Best performance in most cases
 Simple - only starting location (block #) and length (number of
blocks) are required
 Problems include finding space for file, knowing file size, external
fragmentation, need for compaction offline (downtime) or
online

Operating Systems 2.107

 Contiguous Allocation

 Mapping from logical to physical

LA/512

Block to be accessed = Q + starting address

Displacement into block = R

Operating Systems 2.108

 Extent Based Systems

 Many newer file systems (i.e., Veritas File System) use a modified
contiguous allocation scheme
 Extent based file systems allocate disk blocks in extents
 An extent is a contiguous block of disks
 Extents are allocated for file allocation
 A file consists of one or more extents

Operating Systems 2.109

 Allocation Methods - Linked

 Linked allocation - each file a linked list of blocks

 File ends at nil pointer
 No external fragmentation
 Each block contains pointer to next block
 No compaction, external fragmentation
 Free space management system called when new block needed
 Improve efficiency by clustering blocks into groups but increases
internal fragmentation
 Reliability can be a problem
 Locating a block can take many I/Os and disk seeks

Operating Systems 2.110

 Allocation Methods - Linked (cont.)

 FAT (File Allocation Table) variation

 Beginning of volume has table, indexed by block number
 Much like a linked list, but faster on disk and cacheable
 New block allocation simple

Operating Systems 2.111

 Linked Allocation

 Each file is a linked list of disk blocks: blocks may be scattered

anywhere on the disk

block = pointer

 Mapping

Q
LA/511
R

Block to be accessed is the Qth block in the linked chain of blocks

representing the file
Displacement into block = R + 1

Operating Systems 2.112

 Linked Allocation (cont.)

Operating Systems 2.113

 File Allocation Table

Operating Systems 2.114

 Allocation Methods - Indexed

 Indexed allocation
 Each file has its own index block(s) of pointers to its data blocks
 Logical view

index table

Operating Systems 2.115

 Example of Indexed Allocation

Operating Systems 2.116

 Indexed Allocation

 Need index table

 Random access
 Dynamic access without external fragmentation, but have overhead of
index block
 Mapping from logical to physical in a file of maximum size of 256K
bytes and block size of 512 bytes. We need only 1 block for index
table

Q
LA/512
R

Q = displacement into index table

R = displacement into block

Operating Systems 2.117

 Indexed Allocation - Mapping

 Mapping from logical to physical in a file of unbounded length (block

size of 512 words)
 Linked scheme - link blocks of index table (no limit on size)

Q1
LA / (512 x 511)
R1

Q1 = block of index table

R1 is used as follows
Q2
R1 / 512
R2

Q2 = displacement into block of index table

R2 displacement into block of file

Operating Systems 2.118

 Indexed Allocation - Mapping (cont.)

 Two level index (4K blocks could store 1,024 four byte pointers in outer
index  1,048,567 data blocks and file size of up to 4GB)

Q1
LA / (512 x 512)
R1

Q1 = displacement into outer index

R1 is used as follows
Q2
R1 / 512
R2

Q2 = displacement into block of index table

R2 displacement into block of file

Operating Systems 2.119

 Indexed Allocation - Mapping (cont.)

Operating Systems 2.120

 Combined Scheme: UNIX UFS

4K bytes per block, 32 bit addresses

More index blocks than can be addressed with 32 bit file pointer

Operating Systems 2.121

 Performance

 Best method depends on file access type

 Contiguous great for sequential and random
 Linked good for sequential, not random
 Declare access type at creation  select either contiguous or linked
 Indexed more complex
 Single block access could require 2 index block reads then data
block read
 Clustering can help improve throughput, reduce CPU overhead

Operating Systems 2.122

 Performance (cont.)

 Adding instructions to the execution path to save one disk I/O is

reasonable
 Intel Core i7 Extreme Edition 990x (2011) at 3.46Ghz = 159,000
MIPS
 http://en.wikipedia.org/wiki/Instructions_per_second
 Typical disk drive at 250 I/Os per second
 159,000 MIPS / 250 = 630 million instructions during one disk
I/O
 Fast SSD drives provide 60,000 IOPS
 159,000 MIPS / 60,000 = 2.65 millions instructions during one
disk I/O

Operating Systems 2.123

 Free Space Management

 File system maintains free space list to track available blocks/clusters

 (Using term “block” for simplicity)
 Bit vector or bit map (n blocks)

0 1 2 n-1
…

1  block[i] free
bit[i] =
0  block[i] occupied

Block number calculation

(number of bits per word) * (number of 0 value words) + offset of first 1 bit

CPUs have instructions to return offset within word of first “1” bit

Operating Systems 2.124

 Free Space Management (cont.)

 Bit map requires extra space

 Example
block size = 4KB = 212 bytes
disk size = 240 bytes (1 terabyte)
n = 240/212 = 228 bits (or 32MB)
if clusters of 4 blocks  8MB of memory
 Easy to get contiguous files

Operating Systems 2.125

 Linked Free Space List on Disk

 Linked list (free list)

 Cannot get contiguous
space easily
 No waste of space
 No need to traverse the
entire list (if # free blocks
recorded)

Operating Systems 2.126

 Free Space Management (cont.)

 Grouping
 Modify linked list to store address of next n-1 free blocks in first free
block, plus a pointer to next block that contains free block pointers
(like this one)
 Counting
 Because space is frequently contiguously used and freed, with
contiguous allocation allocation, extents, or clustering
 Keep address of first free block and count of following free
blocks
 Free space list then has entries containing addresses and
counts

Operating Systems 2.127

 Free Space Management (cont.)

 Space Maps
 Used in ZFS
 Consider metadata I/O on very large file systems
 Full data structures like bit maps couldn’t fit in memory 
thousands of I/Os
 Divides device space into metaslab units and manages metaslabs
 Given volume can contain hundreds of metaslabs
 Each metaslab has associated space map
 Uses counting algorithm
 But records to log file rather than file system
 Log of all block activity, in time order, in counting format
 Metaslab activity  load space map into memory in balanced tree
structure, indexed by offset
 Replay log into that structure
 Combine contiguous free blocks into single entry

Operating Systems 2.128

 Efficiency & Performance

 Efficiency dependent on
 Disk allocation and directory algorithms
 Types of data kept in file’s directory entry
 Preallocation or as needed allocation of metadata structures
 Fixed size or varying size data structures

Operating Systems 2.129

 Efficiency & Performance (cont.)

 Performance
 Keeping data and metadata close together
 Buffer cache - separate section of main memory for frequently
used blocks
 Synchronous writes sometimes requested by apps or needed by
OS
 No buffering / caching - writes must hit disk before
acknowledgement
 Asynchronous writes more common, bufferable, faster
 Free behind and read ahead - techniques to optimize sequential
access
 Reads frequently slower than writes

Operating Systems 2.130

 Page Cache

 A page cache caches pages rather than disk blocks using virtual
memory techniques and addresses
 Memory mapped I/O uses a page cache
 Routine I/O through the file system uses the buffer (disk) cache
 This leads to the following figure

Operating Systems 2.131

 I/O Without a Unified Buffer Cache

Operating Systems 2.132

 Unified Buffer Cache

 A unified buffer cache uses the same page cache to cache both
memory mapped pages and ordinary file system I/O to avoid double
caching
 But which caches get priority, and what replacement algorithms to
use?

Operating Systems 2.133

 I/O Using a Unified Buffer Cache

Operating Systems 2.134

 Recovery

 Consistency checking - compares data in directory structure with

data blocks on disk, and tries to fix inconsistencies
 Can be slow and sometimes fails
 Use system programs to back up data from disk to another storage
device (magnetic tape, other magnetic disk, optical)
 Recover lost file or disk by restoring data from backup

Operating Systems 2.135

 Log Structured File Systems

 Log structured (or journaling) file systems record each metadata

update to the file system as a transaction
 All transactions are written to a log
 A transaction is considered committed once it is written to the log
(sequentially)
 Sometimes to a separate device or section of disk
 However, the file system may not yet be updated
 The transactions in the log are asynchronously written to the file
system structures
 When the file system structures are modified, the transaction is
removed from the log
 If the file system crashes, all remaining transactions in the log must
still be performed
 Faster recovery from crash, removes chance of inconsistency of
metadata

Operating Systems 2.136

 Sun Network File System (NFS)

 An implementation and a specification of a software system for

accessing remote files across LANs (or WANs)
 The implementation is part of the Solaris and SunOS operating
systems running on Sun workstations using an unreliable datagram
protocol (UDP/IP protocol and Ethernet)

Operating Systems 2.137

 NFS (cont.)

 Interconnected workstations viewed as a set of independent machines

with independent file systems, which allows sharing among these file
systems in a transparent manner
 A remote directory is mounted over a local file system directory
 The mounted directory looks like an integral subtree of the local
file system, replacing the subtree descending from the local
directory
 Specification of the remote directory for the mount operation is
nontransparent; the host name of the remote directory has to be
provided
 Files in the remote directory can then be accessed in a
transparent manner
 Subject to access rights accreditation, potentially any file system
(or directory within a file system), can be mounted remotely on top
of any local directory

Operating Systems 2.138

 NFS (cont.)

 NFS is designed to operate in a heterogeneous environment of

different machines, operating systems, and network architectures; the
NFS specifications independent of these media
 This independence is achieved through the use of RPC primitives built
on top of an External Data Representation (XDR) protocol used
between two implementation independent interfaces
 The NFS specification distinguishes between the services provided by
a mount mechanism and the actual remote file access services

Operating Systems 2.139

 Three Independent File Systems

Operating Systems 2.140

 Mounting in NFS

Mounts Cascading mounts

Operating Systems 2.141

 NFS Mount Protocol

 Establishes initial logical connection between server and client

 Mount operation includes name of remote directory to be mounted and
name of server machine storing it
 Mount request is mapped to corresponding RPC and forwarded to
mount server running on server machine
 Export list - specifies local file systems that server exports for
mounting, along with names of machines that are permitted to
mount them
 Following a mount request that conforms to its export list, the server
returns a file handle - a key for further accesses
 File handle - a file system identifier, and an inode number to identify
the mounted directory within the exported file system
 The mount operation changes only the user’s view and does not affect
the server side

Operating Systems 2.142

 NFS Protocol

 Provides a set of remote procedure calls for remote file operations.

The procedures support the following operations
 Searching for a file within a directory
 Reading a set of directory entries
 Manipulating links and directories
 Accessing file attributes
 Reading and writing files
 NFS servers are stateless; each request has to provide a full set of
arguments (NFS V4 is just coming available - very different, stateful)
 Modified data must be committed to the server’s disk before results are
returned to the client (lose advantages of caching)
 The NFS protocol does not provide concurrency control mechanisms

Operating Systems 2.143

 Three Major Layers of NFS Architecture

 UNIX file system interface (based on the open, read, write, and close
calls, and file descriptors)
 Virtual File System (VFS) layer - distinguishes local files from remote
ones, and local files are further distinguished according to their file
system types
 The VFS activates file system specific operations to handle local
requests according to their file system types
 Calls the NFS protocol procedures for remote requests
 NFS service layer - bottom layer of the architecture
 Implements the NFS protocol

Operating Systems 2.144

 Schematic View of NFS Architecture

Operating Systems 2.145

 NFS Path Name Translation

 Performed by breaking the path into component names and performing

a separate NFS lookup call for every pair of component name and
directory vnode
 To make lookup faster, a directory name lookup cache on the client’s
side holds the vnodes for remote directory names

Operating Systems 2.146

 NFS Remote Operations

 Nearly one-to-one correspondence between regular UNIX system calls

and the NFS protocol RPCs (except opening and closing files)
 NFS adheres to the remote service paradigm, but employs buffering
and caching techniques for the sake of performance
 File blocks cache - when a file is opened, the kernel checks with the
remote server whether to fetch or revalidate the cached attributes
 Cached file blocks are used only if the corresponding cached
attributes are up to date
 File attribute cache the attribute cache is updated whenever new
attributes arrive from the server
 Clients do not free delayed write blocks until the server confirms that
the data have been written to disk

Operating Systems 2.147

 Example: WAFL File System

 Used on Network Appliance “Filers” - distributed file system

appliances
 “Write anywhere file layout”
 Serves up NFS, CIFS, http, ftp
 Random I/O optimized, write optimized
 NVRAM for write caching
 Similar to Berkeley Fast File System, with extensive modifications

Operating Systems 2.148

 WAFL File Layout

Operating Systems 2.149

 Snapshots in WAFL

Operating Systems 2.150

 End of Lecture 2C

Operating Systems
 Lecture 2D:
I/O Systems

Presented by Dr. Huy Nguyen

Operating Systems
 Lecture 2D: I/O Systems

 Overview
 I/O Hardware
 Application I/O Interface
 Kernel I/O Subsystem
 Transforming I/O Requests to Hardware Operations
 STREAMS
 Performance

Operating Systems 2.153

 Objectives

 Explore the structure of an operating system’s I/O subsystem

 Discuss the principles of I/O hardware and its complexity
 Provide details of the performance aspects of I/O hardware and
software

Operating Systems 2.154

 Overview

 I/O management is a major component of operating system design and

operation
 Important aspect of computer operation
 I/O devices vary greatly
 Various methods to control them
 Performance management
 New types of devices frequent
 Ports, busses, device controllers connect to various devices
 Device drivers encapsulate device details
 Present uniform device access interface to I/O subsystem

Operating Systems 2.155

 I/O Hardware

 Incredible variety of I/O devices

 Storage
 Transmission
 Human interface
 Common concepts - signals from I/O devices interface with computer
 Port - connection point for device
 Bus - daisy chain or shared direct access
 PCI bus common in PCs and servers, PCI Express (PCIe)
 Expansion bus connects relatively slow devices
 Controller (host adapter) - electronics that operate port, bus, device
 Sometimes integrated
 Sometimes separate circuit board (host adapter)
 Contains processor, microcode, private memory, bus controller, etc
– Some talk to per device controller with bus controller,
microcode, memory, etc

Operating Systems 2.156

 Typical PC Bus Structure

Operating Systems 2.157

 I/O Hardware (cont.)

 I/O instructions control devices

 Devices usually have registers where device driver places commands,
addresses, and data to write, or read data from registers after
command execution
 Data in register, data out register, status register, control register
 Typically 1 - 4 bytes, or FIFO buffer
 Devices have addresses, used by
 Direct I/O instructions
 Memory mapped I/O
 Device data and command registers mapped to processor
address space
 Especially for large address spaces (graphics)

Operating Systems 2.158

 Device I/O Port Locations on PCs (partial)

Operating Systems 2.159

 Polling

 For each byte of I/O

1. Read busy bit from status register until 0
2. Host sets read or write bit and if write copies data into data out
register
3. Host sets command ready bit
4. Controller sets busy bit, executes transfer
5. Controller clears busy bit, error bit, command ready bit when
transfer done
 Step 1 is busy wait cycle to wait for I/O from device
 Reasonable if device is fast
 But inefficient if device slow
 CPU switches to other tasks?
 But if miss a cycle data overwritten / lost

Operating Systems 2.160

 Interrupts

 Polling can happen in 3 instruction cycles

 Read status, logical and to extract status bit, branch if not zero
 How to be more efficient if nonzero infrequently?
 CPU Interrupt request line triggered by I/O device
 Checked by processor after each instruction
 Interrupt handler receives interrupts
 Maskable to ignore or delay some interrupts
 Interrupt vector to dispatch interrupt to correct handler
 Context switch at start and end
 Based on priority
 Some nonmaskable
 Interrupt chaining if more than one device at same interrupt
number

Operating Systems 2.161

 Interrupt Driven I/O Cycle

Operating Systems 2.162

 Intel Pentium Processor Event Vector Table

Operating Systems 2.163

 Interrupts (cont.)

 Interrupt mechanism also used for exceptions

 Terminate process, crash system due to hardware error
 Page fault executes when memory access error
 System call executes via trap to trigger kernel to execute request
 MultiCPU systems can process interrupts concurrently
 If operating system designed to handle it
 Used for time sensitive processing, frequent, must be fast

Operating Systems 2.164

 Direct Memory Access

 Used to avoid programmed I/O (one byte at a time) for large data
movement
 Requires DMA controller
 Bypasses CPU to transfer data directly between I/O device and
memory
 OS writes DMA command block into memory
 Source and destination addresses
 Read or write mode
 Count of bytes
 Writes location of command block to DMA controller
 Bus mastering of DMA controller - grabs bus from CPU
 Cycle stealing from CPU but still much more efficient
 When done, interrupts to signal completion
 Version that is aware of virtual addresses can be even more efficient -
DVMA

Operating Systems 2.165

 Six Step Process to Perform DMA Transfer

Operating Systems 2.166

 Application I/O Interface

 I/O system calls encapsulate device behaviors in generic classes

 Device driver layer hides differences among I/O controllers from kernel
 New devices talking already implemented protocols need no extra
work
 Each OS has its own I/O subsystem structures and device driver
frameworks
 Devices vary in many dimensions
 Character stream or block
 Sequential or random access
 Synchronous or asynchronous (or both)
 Sharable or dedicated
 Speed of operation
 read write, read only, or write only

Operating Systems 2.167

 Kernel I/O Structure

Operating Systems 2.168

 Characteristics of I/O Devices

Operating Systems 2.169

 Characteristics of I/O Devices (cont.)

 Subtleties of devices handled by device drivers

 Broadly I/O devices can be grouped by the OS into
 Block I/O
 Character I/O (Stream)
 Memory mapped file access
 Network sockets
 For direct manipulation of I/O device specific characteristics, usually an
escape / back door
 Unix ioctl() call to send arbitrary bits to a device control register
and data to device data register

Operating Systems 2.170

 Block & Character Devices

 Block devices include disk drives

 Commands include read, write, seek
 Raw I/O, direct I/O, or file system access
 Memory mapped file access possible
 File mapped to virtual memory and clusters brought via
demand paging
 DMA
 Character devices include keyboards, mice, serial ports
 Commands include get(), put()
 Libraries layered on top allow line editing

Operating Systems 2.171

 Network Devices

 Varying enough from block and character to have own interface

 Linux, Unix, Windows and many others include socket interface
 Separates network protocol from network operation
 Includes select() functionality
 Approaches vary widely (pipes, FIFOs, streams, queues, mailboxes)

Operating Systems 2.172

 Clocks & Timers

 Provide current time, elapsed time, timer

 Normal resolution about 1/60 second
 Some systems provide higher resolution timers
 Programmable interval timer used for timings, periodic interrupts
 ioctl() (on UNIX) covers odd aspects of I/O such as clocks and timers

Operating Systems 2.173

 Nonblocking & Asynchronous I/O

 Blocking - process suspended until I/O completed

 Easy to use and understand
 Insufficient for some needs
 Nonblocking - I/O call returns as much as available
 User interface, data copy (buffered I/O)
 Implemented via multithreading
 Returns quickly with count of bytes read or written
 select() to find if data ready then read() or write() to transfer
 Asynchronous - process runs while I/O executes
 Difficult to use
 I/O subsystem signals process when I/O completed

Operating Systems 2.174

 Two I/O Methods

Synchronous Asynchronous

Operating Systems 2.175

 Vectored I/O

 Vectored I/O allows one system call to perform multiple I/O operations
 For example, Unix readve() accepts a vector of multiple buffers to read
into or write from
 This scatter gather method better than multiple individual I/O calls
 Decreases context switching and system call overhead
 Some versions provide atomicity
 Avoid for example worry about multiple threads changing data
as reads / writes occurring

Operating Systems 2.176

 Kernel I/O Subsystem

 Scheduling
 Some I/O request ordering via per device queue
 Some OSs try fairness
 Some implement Quality Of Service (i.e. IPQOS)
 Buffering - store data in memory while transferring between devices
 To cope with device speed mismatch
 To cope with device transfer size mismatch
 To maintain “copy semantics”
 Double buffering - two copies of the data
 Kernel and user
 Varying sizes
 Full / being processed and not full / being used
 Copy on write can be used for efficiency in some cases

Operating Systems 2.177

 Device Status Table

Operating Systems 2.178

 Sun Enterprise 6000 Device Transfer Rates

Operating Systems 2.179

 Kernel I/O Subsystem

 Caching - faster device holding copy of data

 Always just a copy
 Key to performance
 Sometimes combined with buffering
 Spooling - hold output for a device
 If device can serve only one request at a time
 I.e., printing
 Device reservation - provides exclusive access to a device
 System calls for allocation and deallocation
 Watch out for deadlock

Operating Systems 2.180

 Error Handling

 OS can recover from disk read, device unavailable, transient write

failures
 Retry a read or write, for example
 Some systems more advanced - Solaris FMA, AIX
 Track error frequencies, stop using device with increasing
frequency of retry able errors
 Most return an error number or code when I/O request fails
 System error logs hold problem reports

Operating Systems 2.181

 I/O Protection

 User process may accidentally or purposefully attempt to disrupt

normal operation via illegal I/O instructions
 All I/O instructions defined to be privileged
 I/O must be performed via system calls
 Memory mapped and I/O port memory locations must be
protected too

Operating Systems 2.182

 Use of a System Call to Perform I/O

Operating Systems 2.183

 Kernel Data Structures

 Kernel keeps state info for I/O components, including open file tables,
network connections, character device state
 Many, many complex data structures to track buffers, memory
allocation, “dirty” blocks
 Some use object oriented methods and message passing to implement
I/O
 Windows uses message passing
 Message with I/O information passed from user mode into
kernel
 Message modified as it flows through to device driver and back
to process
 Pros / cons?

Operating Systems 2.184

 UNIX I/O Kernel Structure

Operating Systems 2.185

 Power Management

 Not strictly domain of I/O, but much is I/O related

 Computers and devices use electricity, generate heat, frequently
require cooling
 OSes can help manage and improve use
 Cloud computing environments move virtual machines between
servers
 Can end up evacuating whole systems and shutting them
down
 Mobile computing has power management as first class OS aspect

Operating Systems 2.186

 Power Management (cont.)

 For example, Android implements

 Component level power management
 Understands relationship between components
 Build device tree representing physical device topology
 System bus  I/O subsystem  {flash, USB storage}
 Device driver tracks state of device, whether in use
 Unused component - turn it off
 All devices in tree branch unused - turn off branch
 Wake locks - like other locks but prevent sleep of device when lock
is held
 Power collapse - put a device into very deep sleep
 Marginal power use
 Only awake enough to respond to external stimuli (button
press, incoming call)

Operating Systems 2.187

 I/O Requests to Hardware Operations

 Consider reading a file from disk for a process

 Determine device holding file
 Translate name to device representation
 Physically read data from disk into buffer
 Make data available to requesting process
 Return control to process

Operating Systems 2.188

 Life Cycle of An I/O Request

Operating Systems 2.189

 STREAMS

 STREAM - a full duplex communication channel between a user level

process and a device in Unix System V and beyond
 A STREAM consists of
 STREAM head interfaces with the user process
 driver end interfaces with the device
 zero or more STREAM modules between them
 Each module contains a read queue and a write queue
 Message passing is used to communicate between queues
 Flow control option to indicate available or busy
 Asynchronous internally, synchronous where user process
communicates with stream head

Operating Systems 2.190

 STREAMS Structure

Operating Systems 2.191

 Performance

 I/O a major factor in system performance

 Demands CPU to execute device driver, kernel I/O code
 Context switches due to interrupts
 Data copying
 Network traffic especially stressful

Operating Systems 2.192

 Intercomputer Communications

Operating Systems 2.193

 Improving Performance

 Reduce number of context switches

 Reduce data copying
 Reduce interrupts by using large transfers, smart controllers, polling
 Use DMA
 Use smarter hardware devices
 Balance CPU, memory, bus, and I/O performance for highest
throughput
 Move user mode processes / daemons to kernel threads

Operating Systems 2.194

 Device Functionality Progression

Operating Systems 2.195

 End of Lecture 2D

Operating Systems