Chapter 4: File System Management & I/O System

4.5. I/O Systems

GV: Nguyễn Thị Thanh Vân

Operating System Concepts – 10th Edition Silberschatz, Galvin and Gagne ©2018

Chapter 12: 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 System Concepts – 10th Edition 12.2 Silberschatz, Galvin and Gagne ©2018

1
 Objectives
 Explore the structure of an operating system’s I/O subsystem

 Discuss the principles and complexities of I/O hardware

 Explain the performance aspects of I/O hardware and software

Operating System Concepts – 10th Edition 12.3 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.4 Silberschatz, Galvin and Gagne ©2018

2
 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
 Serial-attached SCSI (SAS) common disk interface
• 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 System Concepts – 10th Edition 12.5 Silberschatz, Galvin and Gagne ©2018

A Typical PC Bus Structure

Operating System Concepts – 10th Edition 12.6 Silberschatz, Galvin and Gagne ©2018

3
 I/O Hardware (Cont.)
 Fibre channel (FC) is complex controller, usually separate circuit board
(host-bus adapter, HBA) plugging into bus
 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 System Concepts – 10th Edition 12.7 Silberschatz, Galvin and Gagne ©2018

Device I/O Port Locations on PCs (partial)

Operating System Concepts – 10th Edition 12.8 Silberschatz, Galvin and Gagne ©2018

4
 Polling
 The complete protocol for interaction between the host and a controller can
be used by handshaking. For each byte of I/O
1. The host repeatedly reads the busy bit until that bit becomes clear.
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, and executes transfer
5. Controller clears busy bit, error bit, command-ready bit when transfer done
 In step 1, host is busy-wait or polling: 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 System Concepts – 10th Edition 12.9 Silberschatz, Galvin and Gagne ©2018

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 non-zero 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 System Concepts – 10th Edition 12.10 Silberschatz, Galvin and Gagne ©2018

5
 Interrupt-Driven I/O Cycle

Operating System Concepts – 10th Edition 12.11 Silberschatz, Galvin and Gagne ©2018

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
 Multi-CPU systems can process interrupts concurrently
• If operating system designed to handle it
 Used for time-sensitive processing, frequent, must be fast

Operating System Concepts – 10th Edition 12.12 Silberschatz, Galvin and Gagne ©2018

6
 Latency
 Stressing interrupt management because even single-user systems
manage hundreds of interrupts per second and servers hundreds of
thousands per second
 For example, a quiet macOS desktop generated 23,000 interrupts
over 10 seconds

Operating System Concepts – 10th Edition 12.13 Silberschatz, Galvin and Gagne ©2018

Intel Pentium Processor Event-Vector Table

 Interrupt vector: contains the memory addresses of specialized interrupt
handlers. => to reduce the need for a single interrupthandler to search all
possible sources of interrupts to determine which one needs service.
 Ex: Interrupt vector for the Intel Pentium processor

Operating System Concepts – 10th Edition 12.14 Silberschatz, Galvin and Gagne ©2018

7
 Direct Memory Access
 DMA is used to avoid programmed I/O (one byte at a time) for large
data movement
 Requires DMA controller: Proceeds to operate the memory bus directly
• transfer data directly between I/O device and memory (without CPU)
 Handshaking between the DMA controller and the device controller
• is performed via a pair of wires called DMA-request and DMA- acknowledge
 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 - direct virtual memory access

Operating System Concepts – 10th Edition 12.15 Silberschatz, Galvin and Gagne ©2018

Six Step Process to Perform DMA Transfer

Operating System Concepts – 10th Edition 12.16 Silberschatz, Galvin and Gagne ©2018

8
 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: transfers bytes one by one
• Sequential or random-access: transfers data in a fixed order determined
• Synchronous or asynchronous (or both)
• Sharable or dedicated
• Speed of operation
• read-write, read only, or write only

Operating System Concepts – 10th Edition 12.17 Silberschatz, Galvin and Gagne ©2018

A Kernel I/O Structure

Operating System Concepts – 10th Edition 12.18 Silberschatz, Galvin and Gagne ©2018

9
 Characteristics of I/O Devices
Devices vary on many dimensions

Operating System Concepts – 10th Edition 12.19 Silberschatz, Galvin and Gagne ©2018

Characteristics of I/O Devices (Cont.)

 Subtleties of devices handled by device drivers
 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
 UNIX and Linux use tuple of “major” and “minor” device numbers to identify
type and instance of devices (here major 8 and minors 0-4)
% ls –l /dev/sda*

Operating System Concepts – 10th Edition 12.20 Silberschatz, Galvin and Gagne ©2018

10
 Block and 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 System Concepts – 10th Edition 12.21 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.22 Silberschatz, Galvin and Gagne ©2018

11
 Clocks and 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 System Concepts – 10th Edition 12.23 Silberschatz, Galvin and Gagne ©2018

Nonblocking and 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 multi-threading
• 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 System Concepts – 10th Edition 12.24 Silberschatz, Galvin and Gagne ©2018

12
 Two I/O Methods

Synchronous Asynchronous

Operating System Concepts – 10th Edition 12.25 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.26 Silberschatz, Galvin and Gagne ©2018

13
 Kernel I/O Subsystem

 Scheduling
 Buffering
 Caching
 Spooling and device Reservation
 Error Handling
 I/O Protection
 Kernel Data structure
 Power management

Operating System Concepts – 10th Edition 12.27 Silberschatz, Galvin and Gagne ©2018

Kernel I/O Subsystem

 Scheduling a set of I/O requests: determine a good order to execute
• Some I/O request ordering via per-device queue
• Some OSs try fairness among processes
• 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 System Concepts – 10th Edition 12.28 Silberschatz, Galvin and Gagne ©2018

14
 Device-status Table
Device-status table: a place that OS might attach the wait queue to keep
track of many I/O requests at the same time

Operating System Concepts – 10th Edition 12.29 Silberschatz, Galvin and Gagne ©2018

Common PC and Data-center I/O devices and Interface Speeds

Operating System Concepts – 10th Edition 12.30 Silberschatz, Galvin and Gagne ©2018

15
 Kernel I/O Subsystem
 Caching - faster holding copy of data - Access to the cached copy is
more eficient than access to the original.
• Always just a copy
• Key to performance
• Sometimes combined with buffering
 Spooling - a buffer holds 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 de-allocation
• Watch out for deadlock

Operating System Concepts – 10th Edition 12.31 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.32 Silberschatz, Galvin and Gagne ©2018

16
 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 System Concepts – 10th Edition 12.33 Silberschatz, Galvin and Gagne ©2018

Use of a System Call to Perform I/O

Operating System Concepts – 10th Edition 12.34 Silberschatz, Galvin and Gagne ©2018

17
 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 System Concepts – 10th Edition 12.35 Silberschatz, Galvin and Gagne ©2018

UNIX I/O Kernel Structure

Operating System Concepts – 10th Edition 12.36 Silberschatz, Galvin and Gagne ©2018

18
 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 System Concepts – 10th Edition 12.37 Silberschatz, Galvin and Gagne ©2018

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)
 Modern systems use advanced configuration and power interface
(ACPI) firmware providing code that runs as routines called by kernel for
device discovery, management, error and power management

Operating System Concepts – 10th Edition 12.38 Silberschatz, Galvin and Gagne ©2018

19
 Kernel I/O Subsystem Summary
 In summary, the I/O subsystem coordinates an extensive collection of
services that are available to applications and to other parts of the
kernel
• Management of the name space for files and devices
• Access control to files and devices
• Operation control (for example, a modem cannot seek())
• File-system space allocation
• Device allocation
• Buffering, caching, and spooling
• I/O scheduling
• Device-status monitoring, error handling, and failure recovery
• Device-driver configuration and initialization
• Power management of I/O devices
 The upper levels of the I/O subsystem access devices via the uniform
interface provided by the device drivers

Operating System Concepts – 10th Edition 12.39 Silberschatz, Galvin and Gagne ©2018

Transforming 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 System Concepts – 10th Edition 12.40 Silberschatz, Galvin and Gagne ©2018

20
 Life Cycle of An I/O Request

Operating System Concepts – 10th Edition 12.41 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.42 Silberschatz, Galvin and Gagne ©2018

21
 The STREAMS Structure

Operating System Concepts – 10th Edition 12.43 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.44 Silberschatz, Galvin and Gagne ©2018

22
 Intercomputer Communications

Operating System Concepts – 10th Edition 12.45 Silberschatz, Galvin and Gagne ©2018

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 System Concepts – 10th Edition 12.46 Silberschatz, Galvin and Gagne ©2018

23
 Device-Functionality Progression

Operating System Concepts – 10th Edition 12.47 Silberschatz, Galvin and Gagne ©2018

I/O Performance of Storage (and Network Latency)

Operating System Concepts – 10th Edition 12.48 Silberschatz, Galvin and Gagne ©2018

24
 End of Chapter 4.5

Operating System Concepts – 10th Edition Silberschatz, Galvin and Gagne ©2018

25