IO Systems

I/O Systems
Amir H. Payberah
amir@sics.se
Amirkabir University of Technology
(Tehran Polytechnic)
Amir H. Payberah (Tehran Polytechnic) I/O Systems 1393/9/15 1 / 57

Motivation

Overview
I/O management is a major component of OS 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

Overview
Ports, busses, device controllers connect to various devices.

Overview
Ports, busses, device controllers connect to various devices.
Device drivers encapsulate device details.
• Present uniform device-access interface to I/O subsystem.

I/O Hardware

I/O Hardware
Incredible variety of I/O devices
• Storage, e.g., disks, tapes
• Transmission, e.g., network connections, bluetooth
• Human-interface, e.g., screen, keyboard, mouse, audio in and out

I/O Hardware
Incredible variety of I/O devices
• Storage, e.g., disks, tapes
• Transmission, e.g., network connections, bluetooth
• Human-interface, e.g., screen, keyboard, mouse, audio in and out
We only need to understand how the devices are attached and how
the software can control the hardware.

Common Concepts in I/O Hardware
Common concepts: signals from I/O devices interface with com-
puter.

puter.
Port: connection point for device

puter.
Bus: set of wires and a protocol that speciﬁes a set of messages
that can be sent on the wires.

puter.
Bus: set of wires and a protocol that speciﬁes a set of messages
that can be sent on the wires.
Controller: a collection of electronics that can operate a port, a bus,
or a device.
• Sometimes integrated and sometimes separate circuit board (host
adapter)
• Contains processor, microcode, private memory, bus controller, etc

Bus
PCI bus: connects the processor-memory subsystem to fast devices.
Expansion bus: connects relatively slow devices.

Processor/Controller Interaction
How can the processor give commands and data to a controller to
accomplish an I/O transfer?

Devices usually have registers where device driver places commands,
addresses, and data.

Processor/controller interaction:

• Direct I/O instructions: triggers bus lines to select the proper
device and to move bits into or out of a device register.

• Direct I/O instructions: triggers bus lines to select the proper
device and to move bits into or out of a device register.
• Memory-mapped I/O: device data and command registers mapped
to processor address space.

Device I/O Port Locations on PCs

I/O Port Registers
The data-in register: read by the host to get input.

I/O Port Registers
The data-out register: written by the host to send output.

I/O Port Registers
The status register: read by the host, and indicates states.
• whether the current command has completed.
• whether a byte is available to be read from the data-in register.
• whether a device error has occurred.

I/O Port Registers
The status register: read by the host, and indicates states.
• whether the current command has completed.
• whether a byte is available to be read from the data-in register.
• whether a device error has occurred.
The control register: written by the host to start a command or to
change the mode of a device.

Host Device Interaction
Polling
Interrupt
Direct memory access (DMA)

Polling (1/2)
A handshake between the host and a controller.

Polling (1/2)
Assume 2 bits for coordination. For each byte of I/O:

Polling (1/2)
1 Host reads the busy bit from the status register until 0.

Polling (1/2)
2 Host sets the read or write bit and if write copies data into the
data-out register.

Polling (1/2)
data-out register.
3 Host sets the command-ready bit.

Polling (1/2)
data-out register.
4 Controller sets the busy bit, executes transfer.

Polling (1/2)
data-out register.
4 Controller sets the busy bit, executes transfer.
5 Controller clears the busy bit, error bit, and command-ready bit
when transfer done.

Polling (2/2)
Step 1 is busy-wait cycle (polling) 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.

Interrupts
Polling can happen in 3 instruction cycles.
• read status, logical-and to extract status bit, and branch if not
zero.
• Ineﬃcient, but more eﬃcient way?

Interrupts
zero.
CPU interrupt-request line is triggered by I/O device.
• Checked by processor after each instruction.
• Saves state and jumps to interrupt-handler routine at a ﬁxed
address in memory.

Interrupts
zero.
CPU interrupt-request line is triggered by I/O device.
• Checked by processor after each instruction.
• Saves state and jumps to interrupt-handler routine at a ﬁxed
address in memory.
Two interrupt request lines:
• Nonmaskable: reserved for events such as unrecoverable memory
errors.
• Maskable: it can be turned oﬀ by the CPU.

Interrupt-Driven I/O Cycle

Interrupt Vector
The interrupt mechanism accepts an address: a number that selects
a speciﬁc interrupt-handling routine.

Interrupt Vector
This address is an oﬀset in a table called the interrupt vector.

Interrupt Vector
The interrupt vector contains the memory addresses of specialized
interrupt handlers.

Interrupt Vector
interrupt handlers.
Computers have more devices than they have address elements in
the interrupt vector.

Interrupt Vector
interrupt handlers.
Computers have more devices than they have address elements in
the interrupt vector.
• Use interrupt chaining: each element in the interrupt vector points
to the head of a list of interrupt handlers.

Intel Pentium Processor Event-Vector Table

Other Interrupt Usages
Interrupt mechanism also used for exceptions.
• Divide by zero, terminate process, 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.

Multi-CPU systems can process interrupts concurrently.
Used for time-sensitive processing, frequent, must be fast.

Direct Memory Access (1/2)
Used to avoid programmed I/O (one byte at a time) for large data
movement.
Requires Direct Memory Access (DMA) controller
Bypasses CPU to transfer data directly between I/O device and
memory.
Version that is aware of virtual addresses can be even more eﬃcient.

Direct Memory Access (2/2)
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

Six Step Process to Perform DMA Transfer

Application I/O Interface

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

Device-driver layer hides diﬀerences among I/O controllers from ker-
nel.

nel.
New devices talking already-implemented protocols need no extra
work.

nel.
New devices talking already-implemented protocols need no extra
work.
Each OS has its own I/O subsystem structures and device driver
frameworks.

A Kernel I/O Structure

Characteristics of I/O Devices (1/2)
Devices vary in many dimensions
• Data-transfer mode: character or block
• Access method: sequential or random-access
• Transfer schedule: synchronous or asynchronous (or both)
• Sharing: sharable or dedicated
• Device speed: speed of operation
• I/O direction: read-write, read only, or write only

Characteristics of I/O Devices (2/2)

Character Devices
Character devices include keyboards, mouse, serial ports.

Character Devices
A character device transfers bytes one by one.

Character Devices
A character device transfers bytes one by one.
Commands include get() and put().

Block Devices
Block devices include disk drives.

Block Devices
Commands include read() and write() and seek() for random-
access devices.

Block Devices
access devices.
Raw I/O: access a block device as a simple linear array of blocks.

Block Devices
access devices.
Direct I/O: disable buﬀering and locking.

Block Devices
access devices.
Direct I/O: disable buffering and locking.
Memory-mapped file access: file mapped to virtual memory and
clusters brought via demand paging.

Network Devices
Varying enough from block and character to have own interface.

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.

Clocks and Timers
Provide current time, elapsed time, and timer (trigger operation X
at time T)

Clocks and Timers
at time T)
Programmable interval timer, the hardware used for timings, and
periodic interrupts.

Clocks and Timers
at time T)
Normal resolution about 1/60 second.

Clocks and Timers
at time T)
Normal resolution about 1/60 second.
Some systems provide higher-resolution timers.

Blocking, Nonblocking and Asynchronous I/O
Blocking: process suspended until I/O completed
• Insuﬃcient for some needs

Nonblocking: I/O call returns as much as available
• User interface, data copy (buﬀered I/O)

• Implemented via multi-threading

• select() to ﬁnd if data ready then read() or write() to
transfers.

• select() to ﬁnd if data ready then read() or write() to
transfers.
Asynchronous: process runs while I/O executes
• I/O subsystem signals process when I/O completed.

Synchronous vs. Asynchronous I/O Methods

Vectored I/O
Vectored I/O allows one system call to perform multiple I/O oper-
ations.

Vectored I/O
ations.
The scatter-gather method is better than multiple individual I/O
calls.

Vectored I/O
ations.
calls.
• Decreases context switching and system call overhead.

Vectored I/O
ations.
calls.
• Some versions provide atomicity: avoid for example worry about
multiple threads changing data as reads/writes occurring.

Vectored I/O
ations.
calls.
• Some versions provide atomicity: avoid for example worry about
multiple threads changing data as reads/writes occurring.
For example, Unix readve() accepts a vector of multiple buﬀers to
read into or write from.

Kernel I/O Subsystem

Kernel I/O Subsystem
Kernels provide many services related to I/O:
• Scheduling
• Buﬀering
• Caching
• Spooling
• Device reservation
• Error handling

Scheduling (1/2)
Determine a good order in which to execute I/O requests.
Some I/O request ordering via per-device queue.
Some OSs try fairness.

Scheduling (2/2)
In asynchronous I/O the kernel must be able to keep track of many
I/O requests at the same time.

Scheduling (2/2)
• The OS attaches the wait queue to a device-status table.

Scheduling (2/2)
• The table contains an entry for each I/O device.

Scheduling (2/2)
• The table contains an entry for each I/O device.
• If the device is busy with a request, the type of request and other
parameters will be stored in the table entry for that device.

Buﬀering and Caching
Buﬀering: stores data in memory while transferring between devices.

• To cope with device speed mismatch.

• To cope with device transfer size mismatch, e.g., fragmentation and
reassembly of messages

• To maintain copy semantics, e.g., copy semantics

• To maintain copy semantics, e.g., copy semantics
Caching: faster device holding copy of data.
• Always just a copy
• Key to performance
• Sometimes combined with buﬀering

Spooling and Device Reservation
Spooling: a buﬀer that holds output for a device.
• If device can serve only one request at a time, i.e., printing

Spooling and Device Reservation
Spooling: a buﬀer that 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

Error Handling
OS can recover from disk read, device unavailable, transient write
failures

Error Handling
failures
• Retry a read or write.

Error Handling
failures
• Track error frequencies, stop using device with increasing frequency
of retry-able errors.

Error Handling
failures
Most return an error number or code when I/O request fails.

Error Handling
failures
Most return an error number or code when I/O request fails.
System error logs hold problem reports.

I/O Protection
User process may accidentally or purposefully attempt to disrupt
normal operation via illegal I/O instructions.

I/O Protection
All I/O instructions deﬁned to be privileged.

I/O Protection
I/O must be performed via system calls.

I/O Protection
I/O must be performed via system calls.
Memory-mapped and I/O port memory locations must be protected
too.

Use of a System Call to Perform I/O

Kernel Data Structures
Kernel keeps state info for I/O components, including open ﬁle ta-
bles, network connections, character device state

Many complex data structures to track buﬀers, memory allocation,
dirty blocks.

dirty blocks.
Some use object-oriented methods and message passing to imple-
ment I/O

dirty blocks.
Some use object-oriented methods and message passing to imple-
ment I/O
E.g., 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.

UNIX I/O Kernel Structure

Transforming I/O Requests
to Hardware Operations

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

Life Cycle of An I/O Request

STREAMS

STREAMS (1/2)
STREAM: a full-duplex communication channel between a user-level
process and a device in Unix System V and beyond.

STREAMS (1/2)
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.

STREAMS (2/2)
Each module contains a read queue and a write queue.
Message passing is used to communicate between queues.
Asynchronous internally, synchronous where user process communi-
cates with stream head.

Performance

Performance
I/O is 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

Inter-computer Communications

Improving Performance
Reduce number of context switches

Reduce data copying

Reduce data copying
Reduce interrupts by using large transfers, smart controllers, polling

Reduce data copying
Use DMA

Reduce data copying
Use DMA
Use smarter hardware devices

Reduce data copying
Use DMA
Balance CPU, memory, bus, and I/O performance for highest
throughput

Reduce data copying
Use DMA
Balance CPU, memory, bus, and I/O performance for highest
throughput
Move user-mode processes to kernel

Summary

Summary
I/O hardware: port, bus, controller

Summary
I/O port registers: data-in, data-out, status, control

Summary
Host-device interaction: polling, interrupt, DMA

Summary
Devices: char, block, network

Summary
Kernel I/O: schedulling, buﬀering, caching, spooling, device reser-
vation, error handling

Summary
STREAMS

Summary
STREAMS
Performance

Questions?
Acknowledgements
Some slides were derived from Avi Silberschatz slides.

IO Systems

More Related Content

IO Systems