The SeQ Streams Primer The Santa Cruz Operation, Inc.

The SeQ Streams Primer The Santa Cruz Operation, Inc.
The SeQ Streams
Primer
The Santa Cruz Operation, Inc.
Information in this document is subject to change without notice and does not
represent a commitment on the part of The Santa Cruz Operation, Inc. nor Microsoft
Corporation. The software described in this document is furnished under a license
agreement or nondisclosure agreement. The software may be used or copied only in
accordance with the terms of the agreement. It is against the law to copy this software
on magnetic tape, disk, or any other medium for any purpose other than the
purchaser's personal usc.
ALL USE, DUPLICATION, OR DISCLOSURE WHATSOEVER BY THE
GOVERNMENT SHALL BE EXPRESSLY SUBJECT TO RESTRICTIONS AS
SET FORTH IN SUBDIVISION (b) (3) (ii) FOR RESTRICTED RIGHTS IN
COMPUTER SOFTWARE AND SUBDIVISION (b) (2) FOR LIMITED RIGHTS
INTECHNICALDATA,BOTH AS SET FORTH IN FAR52.227-7013.
Portions© 1987 AT&T.
All rights reserved.
Portions © 1980, 1981, 1982, 1983, 1984, 1985, 1986, 1987, 1988 Microsoft
Corporation
All rights reservcd.
Portions © 1983, 1984, 1985, 1986, 1987,1988 The Santa Cruz Operation, Inc.
All rights rcservcd.
DocumcntNumber: XG-11-1-88-1.0A
Processed Date: Fri Dec 214:26:32 PST 1988
XENIX is a registered trademark of Microsoft Corporation.
UNIX is a registered tradcmark of AT&T.
Contents
1
1.1
1.2
1.3
1.4
2
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3
3.1
3.2
3.3
4
4.1
4.2
4.3
4.4
4.5
4.6
4.7
5
5.1
5.2
5.3
5.4
5.5
5.6
Introduction
Introduction 1-1
How This Document Is Organized
Other Documents 1-3
Terms You Should Know 1-4
1-3
Overview
A Basic View of a Stream 2-1
System Calls 2-2
Benefits of STREAMS 2-2
Creating Service Interfaces 2-2
Manipulating Modules 2-3
An Advanced View of a Stream 2-6
Stream Head 2-7
Modules 2-8
Stream End 2-9
Building a Stream
Building a Stream 3-1
Expanded Streams 3-2
Pushable Modules 3-3
User Level Functions
STREAMS System Calls 4-1
An Asynchronous Protocol Stream Example
Initializing the Stream 4-2
Message Types 4-3
Sending and Receiving Messages 4-4
Using Messages in the Example 4-5
Other User Functions 4-8
Kernel Level Functions
Introduction 5-1
Messages 5-1
Message Allocation 5-3
Put and Service Procedures 5-4
Kernel Processing 5-6
Read Side Processing 5-6
-i-
4-1
5.7
5.8
6
6.1
6.2
6.3
6.4
6.5
6.6
7
7.1
7.2
7.3
7.4
Write Side Processing
Analysis 5-9
5-8
Other Facilities
Introduction 6-1
Message Queue Priority
Flow Control 6-2
Multiplexing 6-5
Monitoring 6-8
Error and Trace Logging
6-1
6-9
Driver Design Comparisons
Introduction 7-1
Environment 7-1
Drivers 7-1
Modules 7-2
- ii-
Chapter 1
Introduction
1.1
Introduction
1-1
1.2
How This Document Is Organized
1.3
Other Documents
1-3
1.4 Terms You Should Know
1-4
1-3
Introduction
1.1 Introduction
This Primer describes STREAMS, a major building block of support for
networking services in UNIX. The Primer provides a high-level, technical overview of STREAMS; it is intended for managers and developers
who have prior knowledge of the UNIX System and of networking or
other data communication facilities. For a more detailed description of
STREAMS, see the STREAMS Programmer's Guide.
The UNIX System was originally designed as a general-purpose, multiuser, interactive operating system for minicomputers. Initially
developed in the 1970's, the system's communications environment
included slow to medium speed, asynchronous terminal devices. The original design, the communications environment, and hardware state of the
art influenced the character input/output (I/O) mechanism, but the character I/O area did not require the same emphasis on modularity and performance as other areas of the system.
Support for a broader range of devices, speeds, modes, and protocols has
since been incorporated into the system, but the original character I/O
mechanism, which processes one character at a time, made such development difficult.
The current generation of networking protocols is exemplified by Open
Systems Interconnection (OSI), Systems Network Architecture (SNA),
Transmission Control Protocol/lnternet Protocol (TCPIIP), X.25, and
Xerox Network Systems (XNS). These protocols provide diverse functionality, layered organization, and various feature options. When
developing these protocol suites, developers faced additional problems,
because there were no relevant standard interfaces in the UNIX System.
Attempts to compensate for the above problems have led to diverse, adhoc implementations; for example, protocol drivers are often intertwined
with the hardware configuration in which they were developed. As a
result, functionally equivalent protocol software often cannot interface
with alternate implementations of adjacent protocol layers. Portability,
adaptability, and reuse of software have been hindered.
STREAMS is a general, flexible facility and a set of tools for development
of UNIX System communication services. With STREAMS, developers
can provide services ranging from complete networking protocol suites to
individual device drivers.
STREAMS defines standard interfaces for character I/O within the UNIX
System kernel, and between the kernel and the rest of the UNIX System.
The associated mechanism is simple and open-ended. It consists of a set
1-1
Streams Primer
of system calls, kernel resources, and kernel utility routines. The standard interface and open-ended mechanism enable modular, portable
development and easy integration of higher-performance network services and their components. STREAMS does not impose any specific network architecture. Instead, it provides a powerful framework with a consistent user interface that is compatible with the existing character I/O
interface still available in UNIX.
The modularity and design of STREAMS reflect the "layers and options"
characteristics of contemporary networking architectures. The basic
components in a STREAMS implementation are referred to as modules.
Modules reside in the kernel and offer a set of processing functions and
associated service interfaces. Modules can be dynamically selected from
the user level and then interconnected to provide any rational processing
sequence. Kernel programming, assembly, and link editing are not
required to create the interconnection. Modules can also be dynamically
"plugged into" existing connections from user level. STREAMS' modularity makes the following features possible:
•
User-level programs that are independent of underlying protocols
and physical communication media
•
Network architectures and higher-level protocols that are independent of underlying protocols, drivers, and physical communication
media
•
Higher-level services that can be created by selecting and connecting lower-level services and protocols
•
Enhanced
portability
of protocol modules,
resulting
from
STREAMS' well-defined structure and interface standards
In addition to modularity, STREAMS provides developers with integral
functions, a library of utility routines, and facilities that expedite software
design and implementation. The principal facilities are as follows:
•
Buffer management -To maintain STREAMS' own independent
buffer pool
•
Flow control-To conserve STREAMS' memory and processing
resources
•
Scheduling-To incorporate STREAMS' own scheduling mechanism
•
Multiplexing-For processing interleaved data streams, such as
1-2
Introduction
occur
in SNA, X.25, and windows
•
Asynchronous operation of STREAMS and user processes-Allows
STREAMS-related operations to be performed efficiently from user
level
•
Error and trace loggers-For debugging and administrative functions
1.2 How This Document Is Organized
The remainder of this manual is organized as follows:
•
Chapter 2 provides an overview of the applications and benefits of
STREAMS and the STREAMS mechanism.
•
Chapter 3 describes how to set up a Stream from user .level and
how this initialization affects the kernel. This and following
chapters are aimed at developers.
•
Chapter 4 contains a detailed example and discusses it from user
level.
•
Chapter 5 describes kernel operations associated with the Chapter
4 example, as well as basic STREAMS kernel facilities.
•
Chapter 6 includes kernel and user facilities not otherwise
described.
•
Chapter 7 compares certain design features of character I/O device
drivers with STREAMS modules and drivers.
1.3 Other Documents
The STREAMS Programmer's Guide contains more detailed STREAMS
information for programmers. This guide discusses how programmers can
develop networking applications with STREAMS user-level facilities and
how system programmers can use STREAMS kernel-level facilities to
build modules and drivers.
1-3
Streams Primer
The STREAMS reference materials are divided among several locations.
STREAMS system calls are specified in Section S of the Programmer's
Reference Manual. STREAMS utilities are specified in Section ADM of
the System Administration Guide. Appendix C of the STREAMS
Programmer's Reference contains the reference for STREAMS kernel utilities. STREAMS-specific ioetl calls are specified in streamio(STR). The
modules and drivers available are also described in Section STR. This
section is in Appendix F of the STREAMS Programmer's Reference.
1.4 Terms You Should Know
To understand this guide, you need to be familiar with the following
terms.
DOWNSTREAM
The direction from Stream head to driver.
DRIVER
The end of the Stream closest to a device. The principal functions of the
driver are to handle any associated device and to transform data and
information between the device and Stream. It can also be a pseudodriver, not directly associated with a device, which performs functions
internal to a Stream, such as a multiplexer or log driver.
MESSAGE
One or more linked blocks of data or information, with associated
STREAMS control structures containing a message type. Messages are
the only means of transferring data and communicating within a Stream.
MESSAGE QUEUE
A linked list of messages connected to a QUEUE.
MESSAGE TYPE
A defined set of values identifying the contents of a message.
1-4
Introduction
MODULE
Software that performs functions on
Stream head and driver. A module is
commands in a Shell pipeline, except
functions which allow independent
upstream) data flow and processing.
messages as they flow between
the STREAMS counterpart to the
that a module contains a pair of
bidirectional (downstream and
MULTIPLEXER
A mechanism for connecting multiple Streams to a multiplexing driver.
The mechanism supports the processing of interleaved data Streams and
the processing of internetworking protocols. The multiplexing driver
routes messages among the connected Streams. The other end of a
Stream connected to a multiplexing driver is typically connected to a device driver.
PUSHABLE MODULE
A module between the Stream head and driver. A driver is a nonpushable module and a Stream head includes a non-pushable module.
QUEUE
The set of structures that forms a module. A module is composed of two
QUEUEs, a read (upstream) QUEUE and a write (downstream) QUEUE.
READ QUEUE
The message queue in a module or driver containing messages moving
upstream. It is associated with input data coming from a driver.
STREAM
The kernel aggregate created· by connecting STREAMS components,
resulting from an application of the STREAMS mechanism. The primary
components are a Stream head, a driver and zero or more push able
modules between the Stream head and driver. A Stream forms a fullduplex processing and data-transfer path in the kernel, between a user
process and a driver. A Stream is analogous to a Shell pipeline, except
that data flow and processing are bidirectional.
1-5
Streams Primer
STREAM HEAD
The end of the Stream closest to the user process. The Stream head provides the interface between the Stream and the user process. The principal functions of the Stream head are processing STREAMS-related system
calls and bidirectional transfer of data and information between a user
process and messages in STREAMS' kernel space.
STREAMS
A kernel mechanism that supports development of network services and
data communication drivers. It defines interface standards for character
input/output within the kernel, and between the kernel and user level.
The STREAMS mechanism comprises integral functions, utility routines,
kernel facilities and a set of structures.
UPSTREAM
The direction from driver to Stream head.
WRITE QUEUE
The message queue in a module or driver containing messages moving
downstream. It is associated with output from a user process.
1-6
Chapter 2
Overview
2.1
A Basic View of a Stream
2-1
2.2
System Calls
2.3
Benefits of STREAMS
2.4
Creating Service Interfaces
2.5
Manipulating Modules 2-3
2.5.1 Protocol Portability 2-4
2.5.2 Protocol Substitution 2-4
2.5.3 Protocol Migration 2-5
2.5.4 Module Reusability 2-5
2.6
An Advanced View of a Stream
2.7
Stream Head
2.8
Modules
2.9
Stream End
2-2
2-7
2-8
2-9
2-2
2-2
2-6
Overview
2.1 A Basic View of a Stream
"STREAMS" is a collection of system calls, kernel resources, and kernel
utility routines that can create, use, and dismantle a "Stream". A Stream
is a full-duplex processing and data transfer path between a driver in kernel space and a process in user space (see Figure 2-1).
User Space
-Kernel space -
Stream
Head
ldownstream
(optional)
Module
t upstream
Driver
External
Interface
Figure 2-1 Basic Stream
A Stream has three parts: a Stream head, zero or more modules, and a
driver (also referred to as the Stream end). The Stream head provides the
interface between the Stream and user processes. Its principal function is
to process STREAMS-related user system calls. A module processes data
that travel between the Stream head and driver. A STREAMS driver can
be a device driver, providing the services of an external I/O device, or it
can be an internal software driver, commonly called a pseudo-device
driver.
Using a combination of system calls, kernel routines, and kernel utilities,
STREAMS passes data between a driver and the Stream head in the form
of messages. Messages that are passed from the Stream head toward the
driver are said to travel downstream, and messages passed in the other
direction travel upstream.
2-1
Streams Primer
Data sent to a driver from a user process are packaged into STREAMS
messages and passed downstream. STREAMS can insert one or more
modules into a Stream between the Stream head and driver to perform
intermediate processing of data passing between the Stream head and
driver.
2.2 System Calls
Applications programmers can use the STREAMS facilities by means of a
set of system calls. This system call interface is upwardly compatible
with the existing character I/O facilities. The open [see open(S)] system
call recognizes a STREAMS file and creates a Stream to the specified
driver. A user process can send and receive data using read [see read(S)]
and write [see write(S)] in the same manner as character files and devices. The ioctI [see ioctI(S)] system call enables application programs to
perform functions specific to a particular device. In addition, a set of generic STREAMS ioetl commands [see streamio(STR)] support a variety of
functions for accessing and controlling Streams. A close [see close(S)]
system call dismantles a Stream.
The open, close, read, write, and ioetI system calls support the basic set
of operations on Streams. In addition, new system calls support advanced
STREAMS facilities. The poll [see poU(S)] system call enables an application program to poll multiple Streams for various events. When used
with the STREAMS CSETSIG ioetI command, poll allows an application
to process I/O in an asynchronous manner. The putmsg [see putmsg(S)]
and getmsg [see getmsg(S)] system calls enable application programs to
interact with STREAMS modules and drivers through a service interface.
These calls are discussed in this document and in the STREAMS
Programmer's Guide.
2.3 Benefits of STREAMS
STREAMS offers two major benefits for applications programmers: easy
creation of modules that offer standard data communications services and
the ability to manipulate those modules on a Stream.
2.4 Creating Service Interfaces
One benefit of STREAMS is that it simplifies the creation of modules that
present a service interface to any neighboring application program,
module, or device driver. A service interface is defined at the boundary
2-2
Overview
of two neighbors. In STREAMS, a service interface is a specified set of
messages and the rules for allowable sequences of these messages across
the boundary. A module that implements a service interface receives a
message from a neighbor and responds with an appropriate action (for
example, sending back a request to retransmit) based on the specific message received and the preceding sequence of messages.
STREAMS provides features that make it easy to design various applica-
tion processes and modules that comply with common service interfaces.
If these modules are written to comply with industry-standard service
interfaces, they are called protocol modules.
In general, any two modules can be connected anywhere in a Stream.
However, rational sequences are generally constructed by connecting
modules with compatible protocol service interfaces. For example, a
module that implements an X.25 protocol layer, as shown in Figure 2-2,
presents a protocol service interface at its input and output sides. In this
case, other modules should be connected to the input and output side only
if they have the compatible X.25 service interface.
2.5 Manipulating Modules
STREAMS provides the capabilities to manipulate modules from user
level, interchange modules with common service interfaces and to present
a service interface to a Stream user process. As stated in Chapter 1, the
benefits these capabilities yield when implementing networking services
and protocols include
•
User-level programs that can be independent of underlying protocols and physical communication media
•
Network architectures and higher-level protocols that can be
independent of underlying protocols, drivers, and physical communication media
•
Higher-level services that can be created by selecting and connecting lower-level services and protocols
The following examples illustrate how STREAMS can benefit developers
who are creating service interfaces and manipulating modules.
2-3
Streams Primer
Note
All protocol modules used below were selected for illustrative purposes. Their use does not imply that SCQ offers such modules as
products.
2.5.1 Protocol Portability
Figure 2-2 shows how the same X.25 protocol module can be used with
different drivers on different machines by implementing compatible service interfaces. The X.25 protocol module interfaces are the
Connection-Oriented Network Service (CONS) and the Link Access Protocol - Balanced (LAPB) driver.
MACHINE A
MACHINE B
-----r----
CONS
INTERFACE
X.25
Protocol Layer
Module
SAME
MODULE
------
LAPB
INTERFACE
------
LAPB
Driver
Machine A
------
-l------i
I
DIFFERENT
DRIVER
X.25
Protocol Layer
Module
------
(
------
LAPB
Driver
Machine B
I
Figure 2-2 Protocol Module Portability
2.5.2 Protocol Substitution
Alternative protocol modules (and device drivers) can be interchanged on
the same machine if they are implemented for an equivalent service
interface(s).
2-4
Overview
2.5.3 Protocol Migration
Figure 2-3 illustrates how STREAMS can migrate functions between kernel software and front-end finnware. A common downstream service
interface allows the transport protocol module to be independent of the
number or type of modules below. The same transport module connects
without modification to either an X.25 module or an X.25 driver that has
the same service interface.
By shifting functions between software and finnware, developers can
produce cost effective, functionally equivalent systems over a wide range
of configurations. They can rapidly incorporate technological advances.
The same transport protocol module can be used on a lower-capacity
machine, where economics may preclude the use of front-end hardware,
and also on a larger scale system where a front-end is economically
justified.
Class 1
Transport
Protocol
SAME
MODULES
t
CONS
---------
---------Li~fface--
Class 1
Transport
Protocol
-------
-------
X.25
Packet Layer
Protocol
X.25
KERNEL I Packet Layer
-1
Driver
HARDWARE
LAPB
Driver
\
r
Figure 2-3 Protocol Migration
2.5.4 Module Reusability
Figure 2-4 shows the same canonical module (for example, one that provides delete and kill processing on character strings) reused in two
different Streams. This module is typically implemented as a filter, with
no downstream service interface. In both cases, a TrY interface is
presented to the Stream's user process, since the module is nearest the
Stream head.
2-5
Streams Primer
User
Process
SAME
INTERFACE
User
Process
Canonical
Module
SAME
MODULE
Canonical
Module
Raw
TTY
Driver
Figure 2-4 Module Reusability
2.6 An Advanced View of a Stream
The STREAMS mechanism constructs a Stream by serially connecting
kernel-resident STREAMS components, each constructed from a specific
set of structures. As described earlier, and shown in Figure 2-5, the primary STREAMS components are the Stream head, optional module(s), and
Stream end.
2-6
Overview
___ !1 ~e:. SY,:c: ___ _
Kernel Space
Module
B
Module
A
§
M'~:d~~e
.
.
..
..
. . . . . . .. . .................. u~~~~~
Stream
Module
End
External
Interface
Figure 2-5 Stream in More Detail
2.7 Stream Head
The Stream head provides the interface between the Stream and an application program. The Stream head processes STREAMS-related system
calls from the application and performs the bidirectional transfer of data
and information between the application (in user space) and messages (in
STREAMS' kernel space).
Messages are the only means of transferring data and communicating
within a Stream. A STREAMS message contains data, status/control
information, or a combination of the two. Each message includes a
specified message type indicator that identifies the contents.
2-7
Streams Primer
2.8 Modules
A module perfonns intennediate transfonnations on messages passing
between Stream head and driver. There can be zero or more modules in a
Stream. (There are zero when the driver perfonns all the required character and device processing.)
Each module is constructed from a pair of QUEUE structures (see Au/Ad
and Bu/Bd in Figure 2-5). A pair is required to implement the bidirectional and symmetrical attributes of a Stream. One QUEUE perfonns
functions on messages passing upstream through the module (Au and Bu
in Figure 2-5). The other QUEUE (Ad and Bd) performs another set of
functions on downstream messages. (A QUEUE, which is part of a
module, is different from a message queue, which is described later.)
The two QUEUEs in a module generally have distinct functions, that is,
unrelated processing procedures and data. The QUEUEs operate independently, and so Au does not know if a message passes through Ad unless
Ad is programmed to infonn Au. Messages and data can be shared only if
the developer specifically programs the module functions to perform the
sharing.
Each QUEUE can directly access the adjacent QUEUE in the direction of
message flow (for example, Au to Bu or Stream head to Bd). In addition,
within a module, a QUEUE can readily locate its mate and access its messages (for example, for echoing) and data.
Each QUEUE in a module can contain or point to messages, processing
procedures, or data:
•
Messages-These are dynamically attached to the QUEUE on a
linked list as they pass through the module. (This linked list is a
"message queue"; see Au and Bd in Figure 2-5.)
•
Processing procedures-A put procedure to process messages must
be incorporated into each QUEUE. An optional service procedure
to share the message processing with the put procedure can also be
incorporated. According to their function, the procedures can send
messages upstream and/or downstream, and they can also modify
the private data in their module.
•
Data-Developers can provide private data (for example, state
information and translation tables) if the QUEUE requires it to perform message processing.
2-8
Overview
In general, each of the two QUEUEs in a module has a distinct set of all of
these elements. Additional module elements are described later.
Although depicted as distinct from modules, the Stream head and Stream
end also contain a pair of QUEUEs.
2.9 Stream End
A Stream end is a module in which the module's processing procedures
are the driver routines. The procedures in the Stream end are different
from those in other modules, because they are accessible from an external
device and because the STREAMS mechanism allows multiple Streams to
be connected to the same driver.
The driver can be a device driver, providing an interface between kernel
space and an external communications device, or it can be an internal
pseudo-device driver. A pseudo-device driver is not directly related to
any external device, and it performs functions internal to the kernel. The
multiplexing driver discussed in Chapter 6 is a pseudo-device driver.
Device drivers must transform all data and status/control information
between STREAMS message formats and their external representation.
Differences between STREAMS and character device drivers are discussed in Chapter 7.
2-9
Chapter 3
Building a Stream
3.1
Building a Stream
3-1
3.2 Expanded Streams
3-2
3.3
3-3
Pushable Modules
Building a Stream
3.1 Building a Stream
A Stream is created on the first open system call to a character special
file corresponding to a STREAMS driver. A STREAMS device is distinguished from other character devices by a field contained in the associated cdevsw device table entry.
A Stream is usually built in two steps. Step one creates a minimal Stream
consisting of just the Stream head and device driver. Step one has three
parts:
1.
The head and driver structures are allocated and initialized
2.
The modules in the head and end are linked to each other to form a
Stream
3.
The driver open routine is called
Step two adds modules to produce an expanded Stream (see Figure 3-1).
Minimal
STREAM
Expanded
STREAM
STREAM
HE~D
QUEQEpair
raw TTY
device driver
Figure 3-1 Setting Up a Stream
If the driver performs all of the required character and device processing,
no modules need be added to a Stream. Examples of STREAMS drivers
include a raw tty driver (one that passes along input characters without
change) and a driver with multiple Streams open to it (corresponding to
multiple minor devices opened to a character device driver).
3-1
Streams Primer
When the driver receives characters from the device, it places them into
messages. The messages are then transferred to the next Stream component, the Stream head, which extracts the contents of the message and
copies them to user space. Similar processing occurs for downstream
character output; the Stream head copies data from user space into messages and sends them to the driver.
3.2 Expanded Streams
As the second step in building a Stream, modules can be added to the
Stream. In the right-hand Stream in Figure 3-1, the CANONPROC module
was added to provide additional processing of the characters sent between
head and driver.
Modules are added and removed from a Stream in last-in-first-out (LIFO)
order. They are inserted and deleted at the Stream head by means of the
ioetl system call. In the Stream on the left of Figure 2-4, the X.25 module
was the first added to the Stream, followed by the Class 1 Transport and
Canonical modules. To replace the Class 1 module with a Class 0
module, the Canonical module is removed first, followed by the Class 1
module. Then a Class 0 module is added and the Canonical module put
back.
Because adding and removing modules resembles stack operations, the
add is called a push, and the remove, a pop. Push and pop are two of the
ioetl functions included in the STREAMS subset of ioetl system calls.
These commands perfonn various manipulations and operations on
Streams. The modules manipulated in this manner are called pushable
modules, in contrast to the modules contained in the Stream head and end.
This stack tenninology applies only to the setup, modification, and breakdown of a Stream. Subsequent use of the word "module" refers to the
push able modules between Stream head and end.
The Stream head processes the ioctl and executes the push, which is
analogous to opening the Stream driver. Modules are referenced by a
unique symbolic name, contained in the STREAMS fmodsw module table
(similar to the edevsw table associated with a device file). The module
table and module name are internal to STREAMS and are accessible from
user space only through STREAMS ioctl system calls. The fmodsw table
points to the module template in the kernel. When a module is pushed,
the template is located, the module structures for both QUEUES are allocated, and the tempi ave values are copied into the structures.
Each module contains pointers to an open routine and a close routine.
The open is called when the module is pushed, and the close is called
3-2
Building a Stream
when the module is popped. Module open and close procedures are similar to a driver open and close. These procedures are in addition to the
module elements described in the section "A Basic View of a Stream" in
Chapter 2.
As with other files, a STREAMS file is closed when the last process open
to it closes the file by a close system call. This system call causes the
Stream to be dismantled (any modules are popped and the driver close is
executed).
3.3 Pushable Modules
Modules are pushed onto a Stream to provide special functions and/or
additional protocol layers. In Figure 3-1, the Stream on the left is opened
in a minimal configuration with a raw tty driver and no other module
added. The driver receives one character at a time from the device,
places the character in a message, and sends the message upstream. The
Stream heaf receives the message, extracts the single character, and
copies it into the reading process buffer to send to the user process in
response to a read system call. When the user process wants to send
characters back to the driver, it issues a write system call, and the characters are sent to the Stream head. The head copies the characters into one
or more multicharacter messages and sends them downstream. An application program requiring no further kernel character processing can use
this minimal Stream.
A user requiring a more terminal-like interface needs to insert a module
to perform functions such as echoing, character-erase, and line-kill.
Assuming that the CANONPROC module in Figure 3-1 fulfills this need,
the application program first opens a raw tty Stream. Then the
CANONPROC module is pushed above the driver to create a Stream of the
form shown on the right of the figure. The driver is not aware that a
module has been placed above it and therefore continues to send singlecharacter messages upstream. The module receives single-character messages from the driver, processes the characters, and accumulates them
into line strings. Each line is placed into a message and sent to the
Stream head. The head now finds more than one character in the messages it receives from downstream.
3-3
Streams Primer
Stream head implementation accommodates this change in format
automatically and transfers the multicharacter data into user space. The
Stream head also keeps track of messages partially transferred into user
space, as occurs when the current user read buffer can only hold part of
the current message. Downstream operation is not affected: multiple
character messages are sent by the head and received by the driver.
Note that the Stream head provides the interface between the Stream and
the user process. Modules and drivers do not have to implement user
interface functions other than open and close.
3-4
Chapter 4
User Level Functions
4.1
STREAMS System Calls
4-1
4.2 An Asynchronous Protocol Stream Example 4-1
4.3 Initializing the Stream 4-2
4.4 Message Types 4-3
4.5 Sending and Receiving Messages 4-4
4.6 Using Messages in the Example 4-5
4.7 Other User Functions 4-8
User Level Functions
4.1 STREAMS System Calls
After a Stream has been opened, STREAMS-related system calls allow a
user process to insert and delete (push and pop) modules. That process
can then control the operation of the Stream head, modules, and drivers,
and can send and receive messages containing data and control infonnation. This chapter presents an example of some of the basic functions
available to STREAMS-based applications via the system calls. Additional functions are described at the end of this chapter and in Chapter 6.
The full set of STREAMS-related system calls follows:
open
Open a Stream (described in Chapter 3)
close
Close a Stream (described in Chapter 3)
read
Read data from a Stream
write
Write data to a Stream
ioctl
Control a Stream
getmsg
Receive the message at Stream head
putmsg
Send a message downstream
poll
Notify the application program when selected
events occur on a Stream
The following two-part example describes a Stream that controls the data
communication characteristics of a connection between an asynchronous
tenninal and a tty port. It illustrates basic user level STREAMS features
and then shows how messages can be used. Chapter 5 discusses the kernel level Stream operations corresponding to the user-level operations
described in this chapter. See the STREAMS Programmer's Guide for
more detailed examples of STREAMS applications, modules, and drivers.
4.2 An Asynchronous Protocol Stream Example
In this example our computer runs the UNIX System and supports
different kinds of asynchronous tenninals, each logging in on its own port.
The port hardware is limited in function; for example, it detects and
reports line and modem status, but does not check parity.
Communications software support for these tenninals is provided by
4-1
Streams Primer
means of a STREAMS-implemented asynchronous protocol. The protocol
includes a variety of options that are set when a terminal operator dials in
to log on. The options are determined by a getty-type STREAMS user
process, getstrm, which analyzes data sent to it through a series of dialogs
(prompts and responses) between the process and terminal operator.
The process sets the terminal options for the duration of the connection by
pushing modules onto the Stream or by sending control messages to cause
changes in modules, or in the device driver, already on the Stream. The
options supported include the following:
•
ASCII or EBCDIC character codes
•
Parity (odd, even, or none) for ASCII code
•
Echo or not echo input characters
•
Canonical input and output processing or transparent (raw) character handling
These options are set with the following modules:
CHARPROC
Provides input character processing functions,
including dynamically settable character echo and
parity checking. (This is done via control messages
passed to the module.) The module's default settings are to echo characters and not check character
parity.
CANONPROC
Performs canonical processing on ASCII characters
upstream and downstream. Note that this performs
some processing in. a different manner from the
standard UNIX System character I/O tty subsystem.
ASCEBC
Translates EBCDIC code to ASCII upstream and
ASCII to EBCDIC downstream.
4.3 Initializing the Stream
At system initialization a user process, getstrm, is created for each tty
port. getstrm opens a Stream to its port and pushes the CHARPROC
module onto the Stream, by use of an ioctl CPUSH command. Then the
process issues a getmsg system call to the Stream and sleeps until a message reaches the Stream head. The Stream is now in its idle state.
4-2
User Level Functions
The initial idle Stream, shown in Figure 4-1, contains only one pushable
module, CHARPROC. The device driver is a limited-function, raw tty
driver connected to a limited-function communication port. The driver
and port transparently transmit and receive one unbuffered character at a
time.
I
ge1bm I
-----
--------------
STREAM
HE~D
Figure 4-1 Idle Stream Configuration for Example
Upon receipt of initial input from a tty port, getstrm establishes a connection with the terminal, analyzes the option requests, verifies them, and
issues STREAMS system calls to set the options. After setting up the
options, getstrm creates a user application process. Later, when the user
terminates that application, getstrm restores the Stream to its idle state by
use of system calls.
The next step is to analyze in more detail how the Stream sets up the
communications options. Before doing so, let's examine how messages
are handled in STREAMS.
4.4 Message Types
All STREAMS messages are assigned message types to indicate their
intended use by modules and drivers and to determine their handling by
the Stream head. A driver or module can assign most types to a message
it generates, and a module can modify a message's type during processing. The Stream head converts certain system calls to specified message
types and sends them downstream, and it responds to other calls by
4-3
Streams Primer
copying the contents of certain message types that were sent upstream.
Messages exist only in the kernel, and so a user process can only send and
receive buffers. The process is not explicitly aware of the message type,
but it may be aware of message boundaries, depending on the system call
used (see the distinction between getmsg and read in the next section).
Most message types are internal to STREAMS and can only be passed
from one STREAMS module to another. A few message types, including
M_DATA, M_PROTO, and M_PCPROTO, can also be passed between a
Stream and a user process. M_DATA messages carry data within a Stream
and between a Stream and a user process. M_PROTO and M_PCPROTO
messages carry both data and control information. However, the distinction between control information and data is generally determined by the
developer when implementing a particular Stream. Control information
includes service interface information, which is carried between two
Stream entities that present service interfaces, and condition or status
infonnation, which can be sent between any two Stream entities regardless of their interface. An M_PCPROTO message has the same general
use as an M_PROTO, but the former moves faster through a Stream (see
"Message Queue Priority" in Chapter 6).
4.5 Sending and Receiving Messages
putmsg is a STREAMS-related system call that sends messages; it is similar to write. putmsg provides a data buffer which is converted into an
M_DATA message, and it can also provide a separate control buffer to be
placed into an M_PROTO or M_PCPROTO block. write provides bytestream data to be converted into M_DATA messages.
getmsg is a STREAMS-related system call that accepts messages; It IS
similar to read. One difference between the two calls is that read accepts
only data (messages sent upstream to the Stream head as message type
M_DATA), such as the characters entered from the terminal. getmsg can
simultaneously accept both data and control information (messages sent
upstream as types M_PROTO or M_PCPROTO). getmsg also differs from
read in that it preserves message boundaries so that the same boundaries
exist above and below the Stream head (that is, between a user process
and a Stream). read generally ignores message boundaries, processing
data as a byte stream.
Certain STREAMS ioctl commands, such as CSTR, also cause messages
to be sent or received on the Stream. CSTR provides the general ioctl
capability of the character I/O subsystem. A user process above the
Stream head can issue putmsg, getmsg, the CSTR ioctl command, and
certain other STREAMS-related system calls. Other STREAMS ioctls
4-4
User Level Functions
perfonn functions that include changing the state of the Stream head,
pushing and popping modules, or returning special infonnation. ioctl
commands are described in more detail in the STREAMS Programmer's
Guide.
In addition to message types that explicitly transfer data to a process,
. some messages sent upstream result in infonnation transfer. When these
messages reach the Stream head, they are transfonned into various fonns
and sent to the user process. These fonns include signals, error codes, and
call return values.
4.6 Using Messages in the Example
Returning to the asynchronous protocol example, the Stream was in its
idle configuration (see Figure 4-1). getstrm had issued a getmsg and was
sleeping until the arrival of a message from the Stream head. Such a message results when the driver detects activity on the associated tty port.
An incoming call arrives at port one and causes a ring-detect signal in the
modem. The driver receives the ring signal, answers the call, and sends
upstream an M_PROTO message containing infonnation indicating an
incoming call. getstrm is notified of all incoming calls, although it can
choose to refuse the call because of system limits. In this idle state,
getstrm also accepts M_PROTO messages, such as those indicating error
conditions (for example, detection of line or modem problems on the idle
line).
The M_PROTO message containing notification of the incoming call flows
upstream from the driver into CHARPROC. CHARPROC inspects the
message type, detennines that message processing is not required, and
passes the unmodified message upstream to the Stream head. The Stream
head copies the message into the getmsg buffers associated with getstrm
(one buffer for control infonnation, the other for data). The Stream head
then wakes up the process. getstrm sends its acceptance of the incoming
call with a putmsg system call, which results in a downstream M_PROTO
message to the driver.
Then getstrm sends a prompt to the operator with a write and issues a
getmsg to receive the response. A read could have been used to receive
the response, but the getmsg call allows concurrent monitoring for control (M_PROTO and M_PCPROTO) infonnation. getstrm now sleeps until
the response characters, or infonnation regarding possible error conditions detected by modules or driver, are sent upstream.
4-5
Streams Primer
The first response, sent upstream in a M_DATA block, indicates that the
code set is' ASCII and that canonical processing is requested. getstrm
implements these options by pushing CANONPROC onto the Stream
above CHARPROC to perform canonical processing on the input ASCII
characters.
The response to the next prompt requests even parity checking. getstrm
sends an ioctl CSTR command to CHARPROC, requesting the module to
perform even parity checking on upstream characters. When the dialog
indicates protocol option setting is complete, getstrm creates an application process. At the end of the connection, getstrm pops CANONPROC
and then sends an CSTR to CHARPROC, requesting the module to restore
the no-parity idle state (CHARPROC remains on the Stream).
As a result of the above dialogs, the terminal at port one operates in the
following configuration:
•
ASCII, even parity
•
Echo
•
Canonical processing
In similar fashion, an operator at a different type of terminal on port two
requests a different set of options, resulting in the following configuration:
•
EBCDIC
•
No Echo
•
Canonical processing
The resulting Streams for the two ports are shown in Figure 4-2. For port
one, on the left, the modules in the Stream are CANONPROC and CHARPROC.
For port two, on the right, the resulting modules are CANONPROC,
ASCEBC and CHARPROC. ASCEBC has been pushed on this Stream to
translate between the ASCII interface at the downstream side of
CANONPROC and the EBCDIC interface of the upstream output side of
CHARPROC. In addition, getstrm has sent an CSTR to the CHARPROC
module in this Stream requesting it to disable echo. The resulting
modification to CHARPROC's functions is indicated by the word
"modified" in the right Stream of Figure 4-2.
4-6
User Level Functions
PORT
1
PORT
2
Figure 4-2 Asynchronous Tenninal Streams
Since CHARPROC is now perfonning no function for port two, it might
have been popped from the Stream to be reinserted by getstrm at the end
of connection. However, the low overhead of STREAMS does not require
its removal. The module remains on the Stream, passing messages
unmodified between ASCEBC and the driver. At the end of the connection, getstrm restores this Stream to its idle configuration of Figure 4-1 by
popping the added modules and then sending an CSTR to CHARPROC to
restore the echo default.
Note that the tty driver shown in Figure 4-2 handles minor devices. Each
minor device has a distinct Stream connected from user space to the
driver. This ability to handle multiple devices is a standard STREAMS
feature, similar to the minor device mechanism in character I/O device
drivers.
4-7
Streams Primer
4.7 Other User Functions
The previous example illustrates basic STREAMS concepts. Alternative,
more efficient STREAMS calls or mechanisms could have been used in
place of those described earlier. Some of the alternatives are described in
Chapter 6, and others are addressed in the STREAMS Programmer's Guide.
For example, the initialization process that created a getstrm for each tty
port could have been implemented as a "supergetty" by use of the
STREAMS-related poll system call. As described in Chapter 6, poll
allows a single process to efficiently monitor and control multiple
Streams. The supergetty process handles all of the Stream and terminal
protocol initialization and creates application processes only for established connections.
The M_PROTO notification sent to getstrm could have been sent by the
driver as an M_SIG message that causes a specified signal to be sent to the
process. As discussed previously under "Message Types," error and
status information can also be sent upstream from a driver or module to
user processes by means of different message types. These messages are
transformed by the Stream head into a signal or error code.
Finally, an ioetl CSTR command could have been used in place of a
putmsg M_PROTO message to send information to a driver. The sending
process must receive an explicit response from an CSTR by a specified
time period or an error is returned. A response message must be sent
upstream by the destination module or driver to be translated into the user
response by the Stream head.
4-8
Chapter 5
Kernel Level Functions
5.1
Introduction
5-1
5.2
Messages
5.3
Message Allocation
5.4
Put and Service Procedures 5-4
5.4.1 Put Procedures 5-4
5.4.2 Service Procedures 5-5
5.5
Kernel Processing
5.6
Read Side Processing 5-6
5.6.1 Driver Processing 5-7
5.6.2 CHARPROC 5-7
5.6.3 CANONPROC 5-8
5.7
Write Side Processing
5.8
Analysis
5-1
5-9
5-3
5-6
5-9
Kernel Level Functions
5.1 Introduction
This chapter introduces the use of the STREAMS mechanism in the kernel
and describes some of the tools provided by STREAMS to assist in the
development of modules and drivers. In addition to the basic messagepassing mechanism and QUEUE Stream linkage described previously, the
STREAMS mechanism consists of various facilities, namely, buffer
management, the STREAMS scheduler, processing and message priority,
flow control, and multiplexing. Over 30 STREAMS utility routines and
macros are available to manipulate and use these facilities.
The key elements of a STREAMS kernel implementation are the processing routines in the module and drivers, and preparation of required data
structures. The structures are described in the STREAMS Programmer's
Guide. The following sections provide further infonnation on messages
and on the processing routines that operate on them. The example
described in Chapter 4 is continued, associating the user-level operations
described there with kernel operations.
5.2 Messages
As shown in Figure 5-1, a STREAMS message consists of one or more
linked message blocks. That is, the first message block of a message can
be attached to other message blocks that are part of the same message.
Multiple blocks in a message can occur as the result of processing that
adds header or trailer data to the data contained in the message, or when
message-buffer size limitations cause the data to span multiple blocks.
When a message is composed of multiple message blocks, the message
type of the first block detennines the type of the entire message, regardless of the types of the attached message blocks.
5-1
Streams Primer
Message
Block
(type)
1
Message
Block
1
Message
Block
I
V
Figure 5-1 A Message
STREAMS allocates a message as a single block containing a buffer of a
certain size (see the next section). If the data for a message exceed the
size of the buffer containing the data, the procedure can allocate a new
block containing a larger buffer, copy the current data to it, insert the new
data, and de-allocate the old block. Alternately, the procedure can allocate an additional
(smaller) block, place the new data in the new message block, and link it
before or after the initial message block. Both alternatives yield one new
message.
Messages can exist standalone, as shown in Figure 5-1, when the message
is being processed by a procedure, or a message can await processing on a
linked list of messages, called a message queue, in a QUEUE. In Figure
5-2, Message 1 is linked to Message 2.
5-2
Kernel Level Functions
Message
Message
2
1
I
queue
header
I
<:- - -
i
;;..
Message
Block
(type)
I
next
message
Message
Block
(type)
J
J
Message
Block
Message
Block
1
next
message
<:- - - - - - - - - - -:>
V
Message
Block
Figure 5-2 Messages on a Message Queue
When a message is on a queue, the first block of the message contains
links to preceding and succeeding messages on the same message queue,
and it can contain a link to the second block of the message (if present).
The message queue head and tail are contained in the QUEUE.
STREAMS utility routines enable developers to manipulate messages and
message queues.
5.3 Message Allocation
STREAMS maintains its own storage pool for messages. A procedure can
request the allocation of a message of a speci fied size at one of three
message-pool priorities. The allocb utility returns a message containing a
single block with a buffer of at least the size requested, provided there is a
buffer available at the priority requested. When requesting priority for
messages, developers must weigh their process's need for resources
against the needs of other processes on the same machine.
5-3
Streams Primer
Message pool priority generally has no effect on allocation until the pool
falls below internal STREAMS thresholds. When this occurs, allocb may
refuse a lower priority request for a message of size x while granting a
higher priority request for the same size message. For example, storage
for an urgent control message, such as M_HANGUP or M_PCPROTO,
could be requested at high priority. An M_DATA buffer for holding input
might be requested at medium priority, and an output buffer at lowest
priority (assuming the output data can wait in user space).
5.4 Put and Service Procedures
The procedures in the QUEUE are the software routines that process messages as they transit the QUEUE. The processing is generally performed
according to the message type and can result in a modified message, new
message(s), or no message. The resulting message is generally sent in the
same direction in which it was received by the QUEUE, but it can be sent
in either direction. A QUEUE always contains a put procedure and can
also contain an associated service procedure.
5.4.1 Put Procedures
A put procedure is the QUEUE routine which receives messages from the
preceding QUEUE in the Stream. Messages are passed between QUEUEs
when a procedure in one QUEUE calls the put procedure contained in the
following QUEUE. A call to the put procedure in the appropriate direction is generally the only way to pass messages between modules.
(Unless otherwise indicated in this discussion, "modules" infers
"module, driver, and Stream head"). QUEUEs in pushable modules contain a put procedure. In general, there is a separate put procedure for the
read and write QUEUEs in a module, because most Streams have "full
duplex" operation.
A put procedure is associated with immediate processing on a message.
(Deferred processing is discussed later in this chapter.) Each module
accesses the adjacent put procedure as a subroutine. For example, suppose that modA, modB, and mode are three consecutive modules in a
Stream, with mode connected to the Stream head. If modA receives a
message to be sent upstream, modA processes that message and calls
modB's put procedure, which processes it and calls mode's put procedure,
which processes it and calls the Stream head's put procedure. Thus, the
message is passed along the Stream in one continuous processing
sequence. This sequence has the benefit of completing the entire processing in a short time with low overhead (subroutine calls).
5-4
Kernel Level Functions
In addition, there are situations where the put procedure cannot immediately process the message but must hold it until processing is allowed.
The most typical examples of this are a driver that must wait until the
current output completes before sending the next message, and the Stream
head, when it must wait until a process initiates a read on the Stream.
5.4.2 Service Procedures
STREAMS allows a service procedure as well as a put procedure to be
contained in each QUEUE. A service procedure is not required in a
QUEUE and is associated with deferred processing. If a QUEUE has both
a put and service procedure, message processing is generally divided
between the procedures. The put procedure is always called first from the
preceding QUEUE. After the put procedure completes its part of the message processing, it arranges for the service procedure to be called, by
passing the message to the putq routine. putq does two things: it places
the message on the message queue of the QUEUE (see Figure 5-2), and it
links the QUEUE to the end of the STREAMS scheduling queue. When
putq returns to the put procedure, the procedure typically exits. Some
time later, the service procedure is called by the STREAMS scheduler.
The STREAMS scheduler is separate and distinct from the UNIX System
process scheduler. It is concerned only with QUEUEs linked on the
STREAMS scheduling queue. The scheduler calls the service procedure
of the scheduled QUEUE in a FIFO (first-in-first-out) manner, one at a
time.
The use of both put and service procedures in a QUEUE enables
STREAMS to provide the rapid response and the queuing required in multiuser systems. The put procedure allows rapid response to certain data
and events, such as software echoing of input characters. Put procedures
effectively have higher priority than any scheduled service procedures.
When called from the preceding STREAMS component, a put procedure
executes before the scheduled service procedures of any QUEUE are executed.
The service procedure implies message queuing. Queuing results in
deferred processing of the service procedure, following all other QUEUEs
currently on the scheduling queue. For example, terminal output and
input erase and kill processing are typically performed in a service procedure because this type of processing does not have to be as timely as
echoing. Use of a service procedure also allows processing time to be
more evenly spread among multiple Streams. As with the put procedure,
there is generally a separate service procedure for each QUEUE in a
module. The flow control mechanism (see Chapter 6) uses the service
5-5
Streams Primer
procedures.
5.5 Kernel Processing
This section continues the example of Chapter 4. It describes STREAMS
kernel operations and associates them, where relevant, with Chapter 4
user-level system calls in the example. After initializing operations and
pushing a module, the Stream for port one has the following
configuration:
write
Figure 5-3 Operational Stream for Example
As shown in Figure 5-3, the upstream QUEUE is also referred to as the
read QUEUE, reflecting the message flow in response to a read system
call. Correspondingly, downstream is referred to as the write QUEUE.
Read side processing is discussed first.
5.6 Read Side Processing
ill our example, read side processing consists of driver processing, CHARPROC processing, and CANONPROC processing.
5-6
Kernel Level Functions
5.6.1 Driver Processing
In the example the user process is blocked on the getmsg system call
while waiting for a message to reach the Stream head, and the device
driver independently waits for input of a character from· the port
hardware, or for a message from upstream. Upon receipt of an input character interrupt from the port, the driver places the associated character in
a previously allocated M_DATA message. Then the driver sends the message to the CHARPROC module by calling CHARPROC's upstream put
procedure. On return from CHARPROC, the driver calls the allocb utility
routine to get another message for the next character.
5.6.2 CHARPROC
CHARPROC has both put and service procedures on its read side. In the
example the other QUEUEs in the modules also have put and service procedures.
write
read
..........................................
··
.
CANONPRO'
Module
(service) :
··l
l·
1.. .l.
(pin) l
.
·......... ......................... ....... ..
··.........
CHARPROG.
Module ~
......................... ....... .
(p~t)
l· (se~ice)
..
(serxice) ~
(pitt)
......... ......................... ....... ..
Figure 5-4 Module Put and Service Procedures
When the driver calls CHARPROC's read QUEUE put procedure, the procedure checks private data flags in the QUEUE. In this case, the flags
indicate that echoing is to be performed. (Recall that echoing is optional
and that we are working with port hardware which cannot automatically
echo.) CHARPROC causes the echo to be transmitted back to the terminal, by first making a copy of the message with a STREAMS utility. Then
5-7
Streams Primer
CHARPROC uses another utility to obtain the address of its own write
QUEUE. Finally, the CHARPROC read put procedure calls its write put
procedure and passes it the message copy. The write procedure sends the
message to the driver to effect the echo and then returns to the read procedure.
This part of read side processing is implemented with put procedures, so
the entire processing sequence occurs as an extension of the driver input
character interrupt. The CHARPROC read and write put procedures
appear to the driver as subroutines (nested in the case of the write procedure). This manner of processing is intended to produce the character
echo in a minimal time frame.
After returning from echo processing, the CHARPROC read put procedure
checks another of its private data flags and determines that parity checking should be performed on the input character. Parity should most reasonably be checked as part of echo processing. However, for this example, parity is checked only when the characters are sent upstream. This
relaxes the timing constraints for the checking; that is, it can be deferred
along with the canonical processing. CHARPROC uses putq to schedule
the (original) message for parity check processing by its read service procedure. When the CHARPROC read service procedure is complete, it forwards the message to the read put procedure of CANONPROC. Note that
if parity checking were not required, the CHARPROC put procedure
would call the CANONPROC put procedure directly.
5.6.3 CANONPROC
CANONPROC performs canonical processing. As implemented, all read
QUEUE processing is performed in its service procedure so that
CANONPROC's put procedure simply calls putq to schedule the message
for its read service procedure and then exits. The service procedure
extracts the character from the message buffer and places it in the "line
buffer," which is contained in another M_DATA message the service procedure is constructing. Then the message which contained the single
character is returned to the buffer pooL If the character received was not
an end-of-line, CANONPROC exits. Otherwise, a complete line has been
assembled, and CANONPROC sends the message upstream to the Stream
head, which unblocks the user process from the getmsg call and passes it
the contents of the message.
5-8
Kernel Level Functions
5.7 Write Side Processing
The write side of this Stream carries two kinds of messages from the user
process: ioctl messages for CHARPROC, and M_DATA messages to be
output to the terminal.
ioctl messages are sent downstream as a result of an CSTR ioctl system
call. When CHARPROC receives an ioctl message type, it processes the
message contents to modify internal QUEUE flags and then uses a utility
to send an acknowledgment message upstream (read side) to the Stream
head. The Stream head acts on the acknowledgment message by unblocking the user from the ioctl.
For terminal output, it is assumed that· M_DATA messages sent by write
system calls contain multiple characters. In general, STREAMS returns to
the user process immediately after processing the write call so that the
process can send additional messages. Flow control eventually blocks the
sending process. (Flow control is described in Chapter 6.) The messages
can queue on the write side of the driver because of character transmission timing. When a message is received by the driver's write put procedure, the procedure uses putq to place the message on its write side service message queue if the driver is currently transmitting a previous message buffer. However, there is generally no write QUEUE service procedure in a device driver. Driver output interrupt processing takes the
place of scheduling and performs the service procedure functions, removing messages from the queue.
5.S Analysis
For reasons of efficiency, a module implementation should generally
avoid placing one character per message and using separate routines to
echo and parity check each character, as was done in this example.
Nevertheless, even this design yields potential benefits. Consider a case
where alternate, more intelligent port hardware was substituted. If the
hardware processed multiple input characters and performed the echo and
parity-checking functions of CHARPROC, then the new driver could be
implemented to present the same interface as CHARPROC. Other
modules such as CANONPROC could continue to be used without
modification.
5-9
Chapter 6
Other Facilities
6.1
Introduction
6-1
6.2 Message Queue Priority 6-1
6.3 Flow Control 6-2
6.4 Multiplexing
6-5
6.5 Monitoring 6-8
6.6 Error and Trace Logging 6-9
Other Facilities
6.1 Introduction
The previous chapters described the basic concepts of constructing a
Stream and utilizing the STREAMS mechanism. Additional STREAMS
features are provided to assist in development and to handle characteristic
problems of protocol implementation, such as flow control.
There are also kernel and user-level facilities that support the implementation of advanced functions, such as multiplexers, and allow asynchronous operation of a user process and STREAMS input and output.
6.2 Message Queue Priority
As mentioned in the previous chapter, the STREAMS scheduler operates
strictly FIFO so that each QUEUE's service procedure receives control in
the order it was scheduled. When a service procedure receives control, it
can encounter multiple messages on its message queue. This buildup can
occur if there is a long interval between the time a message is queued by
a put procedure and the time that the STREAMS scheduler calls the associated service procedure. In this interval, there can be multiple calls to
the put procedure causing multiple messages. The service procedure
always processes all messages on its message queue unless prevented by
flow control (see next section). Each message must pass through all the
modules connecting its origin and destination in the Stream.
If service procedures were used in all QUEUES and there was no message
priority, then the most recently scheduled message would be processed
after all the other scheduled messages on all Streams had been processed.
In certain cases, message types containing urgent information (such as a
break or alarm conditions) must pass through the Stream quickly. To
accommodate these cases, STREAMS provides two classes of message
queuing priority, ordinary and high. STREAMS prevents high-priority
messages from being blocked by flow control and causes a service procedure to process them ahead of all ordinary priority messages on the
procedure's queue. Thus the high-priority message transits each module
with minimal delay.
6-1
Streams Primer
Message queue
QUEUE
:::~~~::--···i I I I
:'"
:
I
High
:> : '"
Prionty:
I I I I I I I 1
Ordinary
Prionty
Head
;.:
I
Tail
Figure 6-1 Streams Message Priority
The pnonty mechanism operates as shown in Figure 6-1. Message
queues are generally not present in a QUEUE unless that QUEUE contains
a service procedure. When a message is passed to putq to schedule the
message for service procedure processing, putq places the message on the
message queue in priority order. High priority messages are placed ahead
of all ordinary priority messages but behind any other high priority messages on the queue. STREAMS utilities deliver the messages to the processing service procedure FIFO within each priority class. The service
procedure is unaware of the message priority and simply receives the next
message.
Message priority is defined by the message type; once a message is
created, its priority cannot be changed. Certain message types come in
equivalent high/ordinary priority pairs (for example, M_PCPROTO and
M_PROTO) so that a module or device driver can choose between the two
priorities when sending information.
6.3 Flow Control
Even on a well-designed system, general system delays, malfunctions,
and excessive message accumulation on one or more Streams can cause
the message buffer pools to become depleted. Additionally, processing
bursts can arise when a service procedure in one module has a long message queue and processes all its messages in one pass. STREAMS provides two independent mechanisms to guard its message buffer pools from
being depleted and to minimize long processing bursts at anyone module.
6-2
Other Facilities
Note
Flow control is only applied to nonnal priority messages (see previous section) and not to high priority messages.
The first flow control mechanism is global and automatic. It is related to
the message pool priority, discussed in the "Message Storage Pool" section of Chapter 5. When the Stream head requests a message buffer in
response to a putmsg or write system call, it uses the lowest level of
priority. Since buffer availability is based on priority and buffer pool levels, the Stream head will be among the first modules refused a buffer
when the pool becomes depleted. In response, the Stream head will block
user output until the STREAMS buffer pool recovers. As a result, output
has a lower priority than input.
The second flow control mechanism is local to each Stream and advisory
(voluntary). It limits the number of characters that can be queued for processing at any QUEUE in a Stream. This mechanism limits the buffers and
related processing at anyone QUEUE and in anyone Stream, but does not
consider buffer pool levels or buffer usage in other Streams.
The advisory mechanism operates between the two nearest QUEUEs in a
Stream containing service procedures (see Figure 6-2). Messages are
generally held on a message queue only if a service procedure is present
in the associated QUEUE.
Messages accumulate at a QUEUE when its service procedure processing
does not keep pace with the message arrival rate, or when the procedure
is blocked from placing its messages on the following Stream component
by the flow control mechanism. Pushable modules contain independent
upstream and downstream limits, which are set when a developer
specifies high-water and low-water control values for the QUEUE. The
Stream head contains a preset upstream limit, which can be modified by a
special message sent from downstream, and a driver can contain a downstream limit.
Flow control operates as follows:
1.
Each time a STREAMS message-handling routine (for example,
putq) adds or removes a message from a message queue in a
QUEUE, the limits are checked. STREAMS calculates the total size
of all message blocks on the message queue.
6-3
Streams Primer
2.
The total is compared to the QUEUE high-water and low-water
values. If the total exceeds the high-water value, an internal full
indicator is set for the QUEUE. The operation of the service procedure in this QUEUE is not affected if the indicator is set, and the
service procedure continues to be scheduled.
3.
The next part of flow-control processing occurs in the nearest
preceding QUEUE that contains a service procedure. In the
diagram in Figure 6-2, if D is full and C has no service procedure,
then B is the nearest preceding QUEUE.
I
I
V
Message
Queue
V
Message
Queue
Figure 6-2 Flow Control
4.
The service procedure in B uses a STREAMS utility routine to see
if a QUEUE ahead is marked full. If messages cannot be sent, the
scheduler blocks the service procedure in B from further execution. B remains blocked until the low-water mark of the full
QUEUE, D, is reached.
5.
While B is blocked, any non-priority messages that arrive at B will
accumulate on its message queue (recall that priority messages are
not blocked). In tum, B can reach a full state and the full condition
will propagate back to the last module in the Stream.
6.
When the service procedure processing on D causes the message
block total to fall below the low-water mark, the full indicator is
turned off. Then STREAMS automatically schedules the nearest
preceding blocked QUEUE (B in this case), getting things moving
again. This automatic scheduling is known as back-enabling a
QUEUE.
Note that to use flow control, a developer need only call the utility that
tests if a full condition exists ahead, plus perform some housekeeping if it
does. Everything else is automatically handled by STREAMS. Additional
flow control features are described in the STREAMS Programmer's Guide.
6-4
Other Facilities
6.4 Multiplexing
STREAMS multiplexing supports the development of internetworking pro-
tocols such as IP and ISO CLNS, as well as the processing of interleaved
data streams such as those in SNA, X.25, and terminal window facilities.
STREAMS multiplexers (also called pseudo-device drivers) are created in
the kernel by interconnecting multiple Streams. Conceptually, there are
two kinds of multiplexers that developers can build with STREAMS,
namely, upper and lower multiplexers. Lower multiplexers have multiple
lower Streams between device drivers and the multiplexer, and upper
multiplexers have multiple upper Streams between user processes and the
multiplexer.
User
Processes
AAA
----------------------------------
~·····Up~~r·····~
Multiplexer o~
L. .. M.?~~!~ .... l
Multiplexer
Driver
Figure 6-3 Internet Multiplexing Stream
Figure 6-3 shows an example of a lower multiplexer. This configuration
typically occurs where internetworking functions are included in the system. This Stream contains two types of drivers: the Ethernet, LAPB, and
IEEE 802.2, which are hardware device drivers that terminate links to
other nodes, and the IP (Internet Protocol), which is a multiplexer.
6-5
Streams Primer
The IP multiplexer switches messages among the various nodes (lower
Streams) or sends them upstream to user processes in the system. In this
example, the multiplexer expects to see an 802.2 interface downstream;
for the Ethernet and LAPB drivers, the Net 1 and Net 2 modules provide
service interfaces between the two non-802.2 drivers and the IP multiplexer.
Figure 6-3 depicts the IP multiplexer as part of a larger Stream. The
Stream, as shown in the dotted rectangle, generally has an upper TCP
multiplexer and additional modules. Multiplexers can also be cascaded
below the IP driver if the device drivers are replaced by multiplexer
drivers.
PVC
SVC
Processes
Processes
Processes
-----------------------------~-----
........ i ...... .
: Modules
.. ~...~~..:.:~. '" ...
~.,
X.25
Packet Layer Protocol
Multiplexer Driver
r····LAPS·
:
or
·nvei"····~
:
~.~?~~~.~~~~~P.~~~~~.~
Figure 6-4 X.25 Multiplexing Stream
6-6
Other Facilities
Figure 6-4 shows an upper multiplexer. In this configuration, the driver
routes messages between the lower Stream and one of the upper Streams.
This Stream performs X.25 multiplexing to multiple independent SVC
(Switched Virtual Circuit) and PVC (Permanent Virtual Circuit) user
processes. Upper multiplexers are a specific application of standard
STREAMS facilities which support multiple minor devices in a device
driver. This figure also shows that more' complex configurations can be
built by having one or more multiplexed LAPB drivers below and multiple
modules above the X.25 Packet Layer Protocol multiplexer diver.
Developers can choose either upper or lower multiplexing (or both) when
designing their applications. For example, a window multiplexer could
have a configuration similar to the X.25 configuration of Figure 6-4, with
a window driver replacing Packet Layer, a tty driver replacing LAPB, and
the child processes of the terminal process replacing the user processes.
Although the X.25 and window multiplexing Streams have similar
configurations, their multiplexer drivers differ significantly. The.IP multiplexer of Figure 6-2 has a different configuration than the X.25 multiplexer, and the driver would implement its own set of processing and
routing requirements.
In addition to upper and lower multiplexers, more complex configurations
can be created by connecting Streams containing multiplexers to other
multiplexer drivers. With such a diversity of needs for multiplexers, it is
not possible to provide general purpose multiplexer drivers. Rather,
STREAMS provides a general purpose multiplexing facility. The facility
allows users to set up the inter-module/driver plumbing to create multiplexer configurations of generally unlimited interconnection.
The connections are created from user space through specific STREAMS
ioctl system calls. In a lower multiplexer, multiple Streams are connected below an application-speci fic, developer-implemented multiplexing driver. The multiplexing facility only connects Streams to a driver.
The ioctl call configures a multiplexer by connecting one Stream at a
time below the opened multiplexer driver. As each Stream is connected
to the driver, the connection setup procedure identifies the Stream to the
driver. The driver generally stores this setup information in a private data
structure for later use.
6-7
Streams Primer
Subsequently, when messages flow into the driver on the various connected Streams, the identity of the associated Stream is passed to the
driver as part of the standard procedure call. The driver then has available the Stream identification, the previously stored setup infonnation for
this Stream, and any internal routing infonnation contained in the message. These data are used, according to the application implemented, to
process the incoming message and route the output to the appropriate outgoing Stream.
Additionally, new Streams can be dynamically connected to a operating
multiplexer without interfering with ongoing traffic, and existing Streams
can be disconnected with similar ease.
6.5 Monitoring
STREAMS allows user processes to monitor and control Streams so that
system resources (such as CPU cycles and process slots) can be used
effectively. Monitoring is especially useful to user-level multiplexers, in
which a user process can create multiple Streams and switch messages
among them. (This is similar to STREAMS kernel-level multiplexing,
described previously.)
User processes can efficiently monitor and control multiple Streams with
two STREAMS system calls: poll and ioctl (used with the CSETSIG command). These calls allow a user process to detect events that occur at the
Stream head on one or more Streams, including receipt of a data or protocol message on the read queue and cessation of flow control.
Synchronous monitoring is provided by use of poll alone; in this case, the
user process cannot continue processing until after the system call completes. When the calls are used together, they allow asynchronous, or
concurrent, operation of the process and STREAMS input/output. This
allows the user process to monitor the Stream while carrying on other
activities.
To monitor Streams with poll, a user process issues that system call and
specifies the Streams to be monitored, the events to look for, and the
amount of time to wait for an event. poll will block the process until the
time expires or until an event occurs. If an event occurs, poll returns the
type of event and the Stream on which the event occurred.
6-8
Other Facilities
Instead of waiting for an event to occur, a user process may want to monitor one or more Streams while processing other data. It can do so by issuing the ioctl CSETSIG command, specifying one or more Streams and
events (as with poll). Unlike poll, this ioctl does not force the user process to wait for the event but returns immediately, and it issues a signal
when an event occurs. The process must also request signal or sigset to
catch the resulting SIGPOLL signal.
If any selected event occurs on any of the selected Streams, STREAMS
causes the SIGPOLL catching function to be executed in all associated
requesting processes. However, the process(es) will not know which
event occurred, nor on what Stream the event occurred. A process that
issues the CSETSIG can get more detailed information by issuing a poll
after it detects the event.
6.6 Error and Trace Logging
STREAMS includes error and trace loggers useful for debugging and
administering modules and drivers.
Any module or driver in any Stream can call the STREAMS logging function strlog. (See log(STR) in Appendix F of the Streams Programmer's
Guide.) When called, strlog sends formatted text to the error logger
strerr(ADM), the trace logger strace(ADM), or both. The call parameters for strlog include the module/driver identification, a severity level,
and the formatted text describing the condition causing the call. The call
also identifies the process (strerr and/or strace) to receive the resulting
output message.
6-9
Streams Primer
Error
Log File
Error
Messages
Trace
Log File
Trace
Message
I mOdule~ --
--1 driver I
Figure 6-5 Error and Trace Logging
strerr is intended to operate as a daemon process initiated at system
startup. A call to strlog requesting logging of an error causes an
M_PROTO message to be sent to strerr, which fonnats the contents and
places them in a daily file. The utility strclean(ADM) is provided to
periodically purge aged, unreferenced daily log files.
A call to strlog requesting logging of trace infonnation causes a similar
M_PROTO message to be sent to strace, which places it in a user-
designated file. strace is intended to be initiated by a user. The user can
designate the modules/drivers and severity level of the messages to be
accepted for logging by strace.
By using putmsg, a user process can submit its own M_PROTO messages
to the log driver for inclusion in the logger of its choice. The messages
must be in the same fonnat required by the logging processes and are
switched to the logger(s) requested in the message.
The output to the log files is formatted, ASCII text. The files can be processed by standard system commands such as grep or ed, or by
developer-provided routines.
6-10
Chapter 7
Driver Design Comparisons
7.1
Introduction
7-1
7.2
Environment
7-1
7.3
Drivers
7.4
Modules
7-1
7-2
Driver Design Comparisons
7.1 Introduction
This chapter compares operational features of character I/O device drivers
with those of STREAMS drivers and modules. It is intended for experienced developers of UNIX system character device drivers. Details are
provided in the STREAMS Programmer's Guide.
7.2 Environment
No user environment is generally available to STREAMS module procedures and drivers. The exception is the module and driver open and
close routines, both of which have access to the u_area of the calling process and can sleep. Otherwise, a STREAMS driver, module put procedure,
and module service procedure have no user context and can neither sleep
nor access any u_area.
Multiple Streams can use a copy of the same module (that is, the same
fmodsw), each containing the same processing procedures. This means
that module code is re-entrant, and so care must be exercised when using
global data in a module. Put and service procedures are always passed
the address of the QUEUE. For example, in Figure 2-5 Au calls Bu's put
procedure with Bu as a parameter. The processing procedure establishes
its environment solely from the QUEUE contents, typically the private
data (for example, state information).
7.3 Drivers
At the interface to hardware devices, character I/O drivers have interrupt
entry points; at the system interface, those same drivers generally have
direct entry points (routines) to process open, close, read, write, and
ioct) system calls.
STREAMS device drivers have similar interrupt entry points at the
hardware device interface and have direct entry points only for open and
close system calls. These entry points are accessed via STREAMS, and
the call formats differ from those of character device drivers. The put
procedure is a driver's third entry point, but it is a message (not system)
interface. The Stream head translates write and ioct) calls into messages
and sends them downstream to be processed by the driver's write QUEUE
put procedure. read is seen directly only by the Stream head, which contains the functions required to process system calls. A driver does not
know about system interfaces other than open and close, but it can detect
absence of a read indirectly if flow control propagates from the Stream
head to the driver and affects the driver's ability to send messages
7-1
Streams Primer
upstream.
For input processing, when the driver is ready to send data or other information to a user process, it does not wake up the process. It prepares a
message and sends it to the read QUEUE of the appropriate (minor device)
Stream. The driver's open routine generally stores the QUEUE address
corresponding to this Stream.
For output processing, the driver receives messages in place of a write
call. If the message cannot be sent immediately to the hardware, it can be
stored on the driver's write message queue. Subsequent output interrupts
can remove messages from this queue.
Drivers and modules can pass signals, error codes, and return values to
processes via message types provided for that purpose.
7.4 Modules
As described above, modules have user context available only during the
execution of their open and close routines. Otherwise, the QUEUEs forming the module are not associated with the user process at the end of the
Stream, nor with any other process. Because of this, QUEUE procedures
must not sleep when they cannot proceed; instead, they must explicitly
return control to the system. The system does not save state information
for the QUEUE. The QUEUE must store this information internally if it is
to proceed from the same point on a later entry.
When a module or driver which requires private working storage is
pushed, the open routine must obtain the storage from external sources.
(Private storage may be needed for such things as state information.)
STREAMS copies the module template from fmodsw for the CPUSH, and
so only fixed data can be contained in the module template. STREAMS
has no automatic mechanism to allocate working storage to a module
when it is opened. The sources for the storage typically include the
STREAMS buffer pool or a module-specific kernel array that is installed
when the system is configured. When using an array as a module storage
pool, it is necessary to determine the maximum number of copies of the
module that can exist at anyone time. For drivers, this is typically determined from the physical devices connected, such as the number of ports
on a multiplexer. However, certain types of modules may not be associated with a particular external physical limit. For example, the CANONICAL module shown in Figure 2-4 could be used on different types of
Streams.
7-2
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement